febril netropeni

Textbook of

Febrile Neutropenia

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page i

This book is dedicated to Gerald P Bodey, MD and Jean Klastersky, MDin recognition of their pioneering observations, pivotal contributions towards the management

of infections in cancer patients, and their enthusiastic efforts to disseminate this knowledge.

Gerald P Bodey, MD Jean Klastersky, MD

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page ii

Textbook ofFebrile Neutropenia

Edited by

Kenneth VI Rolston, MD, FACPSection of Infectious DiseasesUTMD Anderson Cancer CenterHouston, Texas, USA

Edward B Rubenstein, MD, FACPSection of Medical Supportive CareUTMD Anderson Cancer CenterHouston, Texas, USA

M A R T I N D U N I T Z

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page iii

© 2001, Martin Dunitz Ltd, a member of the Taylor & Francis Group

First published in the United Kingdom in 2001 by Martin Dunitz Ltd, The Livery House, 7–9 Pratt Street, London NW1 0AE

Tel: +44-(0)20-7482-2202Fax: +44-(0)20-7267-0159E-mail: [email protected]: http://www.dunitz.co.uk

All rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted, in any form or by any means, electronic,mechanical, photocopying, recording or otherwise, without the prior permissionof the publisher or in accordance with the provisions of the Copyright, Designsand Patents Act 1988, or under the terms of any licence permitting limited copyingissued by the Copyright Licensing Agency, 90 Tottenham Court Road, LondonW1P 0LP.

Although every effort has been made to ensure that drug doses and otherinformation are presented accurately in this publication, the ultimateresponsibility rests with the prescribing physician. Neither the publishers nor theauthors can be held responsible for errors or for any consequences arising from theuse of information contained herein. For detailed prescribing information orinstructions on the use of any product or procedure discussed herein, pleaseconsult the prescribing information or instructional material issued by themanufacturer.

A CIP catalogue record for this book is available from the British Library

ISBN 1-84184-033-5 (Print Edition)

This edition published in the Taylor & Francis e-Library, 2004.

ISBN 0-203-21527-3 Master e-book ISBN

ISBN 0-203-27170-X (Adobe eReader Format)

Contents

Preface ......................................................................................................................................................... vii

Acknowledgment ...................................................................................................................................... viii

Contributors ............................................................................................................................................... ix

1. Fever and neutropenia: An historical perspective Stephen C Schimpff ................................... 1

2. Overview of pharmacokinetic and pharmacodynamic principles of anti-infective

dosing in the neutropenic patient Russell E Lewis, Randall A Prince....................................... 27

3. Controversies and quandaries: The design of clinical trials in fever and neutropenia

Linda S Elting, Marianne Paesmans .................................................................................................. 45

4. Current epidemiology of infections in neutropenic cancer patients Winfried V Kern ......... 57

5. Infections in patients with solid tumors Kenneth VI Rolston ................................................... 91

6. Infections in patients with hematologic malignancies Ben E De Pauw .................................. 111

7. Evaluation and management of fever in the neutropenic hematopoietic stem cell

transplant patient James C Wade .................................................................................................. 125

8. Initial clinical evaluation and risk assessment of the febrile neutropenic cancer patient

James A Talcott, Edward B Rubenstein .............................................................................................. 151

9. Risk-adjusted management of the febrile neutropenic cancer patient

Edward B Rubenstein, Kenneth VI Rolston ....................................................................................... 167

10. Evaluation and management of vascular access device infections in febrile neutropenic

cancer patients Claude Afif, Issam Raad........................................................................................ 189

11. Special considerations in children with fever and neutropenia Sarah W Alexander,

Philip A Pizzo...................................................................................................................................... 201

12. Prophylaxis of infections J Peter Donnelly .................................................................................. 215

13. Cytokines and WBC transfusions Katarzyna J Finiewicz, John R Wingard .............................. 245

14. Clinical practical guidelines in patients with fever and neutropenia Natasha Leighl,

Ronald Feld.......................................................................................................................................... 291

15. Healthcare economic concepts in fever and neutropenia Thomas D Szucs............................ 301

16. Fever in the neutropenic patient: Past lessons and future prospects Philip A Pizzo ............ 317

Index ............................................................................................................................................................ 323

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page v

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page vi

Preface

Following the introduction of myelosuppres-sive combination chemotherapy, infection andbleeding were the two leading causes of deathin patients with acute leukemia. The wide-spread use of chemotherapy for patients withsolid tumors and the expanding indications forstem cell transplantation have resulted in a sub-stantial increase in the population at risk fordeveloping serious infections. The introductionof potent broad-spectrum antimicrobial agentsand the acceptance of the concept of empirictherapy has led to a substantial decrease ininfection-related mortality. Over the past fourdecades, improvements in supportive care,including: transfusion medicine, antimicrobialtherapy and prophylaxis, antineoplastic ther-apy, and the development of hematopoieticgrowth factors, have enabled clinical investiga-tors to evaluate endpoints other than responserates to antimicrobial agents, adverse events,and mortality. As a result of these advances,issues such as routes of antibiotic administra-tion, time to clinical response, site and cost ofcare and quality of life have become importantconsiderations for our investigations.

In this textbook, we have assembled a groupof international experts, many of whom haveled the way in this complex and ever-changingfield, to provide a comprehensive overview ofthe historical aspects and recent developmentsin the care of cancer patients with fever andneutropenia. In addition to providing theirunique experiences and insights regarding tra-ditional evaluation and management of suchpatients, newer concepts have been included,for example, the pharmacokinetic/pharmaco-dynamic interaction of antimicrobial agents,clinical trials methodology and design, riskassessment, and risk-based treatment strategies.We are extremely grateful to our colleagueswho have gladly contributed their time andtheir expertise towards this endeavor. For us, ithas been an immensely rewarding experience,and we consider ourselves privileged and for-tunate to have had the opportunity to workwith and be mentored by Professors Bodey andKlastersky.

Kenneth VI RolstonEdward B Rubenstein

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page vii

Acknowledgment

The Editors (KVIR and EBR) are grateful to Ms Belinda DeBose for her administrative and secretar-ial expertise, which were essential in the preparation of this book.

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page viii

Claude Afif, MDDepartment of Infectious Diseases & InfectionControlUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Sarah W Alexander, MDDepartment of Hematology/OncologyThe Dana Farber Cancer Institute44 Binney StreetBoston, MA 02115USA

Ben E De Pauw, MD, PhDSupportive Care in OncologyClinical Division of Hematology andBlood Transfusion ServiceUniversity Medical Center St RadboudGeert Grootplein Zuid 86500 HB NijmegenThe Netherlands

J Peter Donnelly, FIBMS, MIBiol, PhDStudies in Supportive CareDepartment of HematologyUniversity Medical Center St RadboudGeert Grootplein Zuid 86500 HB NijmegenThe Netherlands

Linda S Elting, DrPHDepartment of Health Services ResearchUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Ronald Feld, MDDepartment of Medical Oncology andHematologyThe Ontario Cancer InstitutePrincess Margaret Hospital610 University AvenueToronto, Ontario M5G 2M9Canada

Katarzyna J Finiewicz, MDDivision of Hematology/OncologyUniversity of Florida College of MedicineJHMHC, 1600 SW Archer RoadGainesville, FL 32610-0277USA

Winfried V Kern, MDMedizinische Universitätsklinik und Poliklinik Robert-Koch-Strasse 8D-89081 UlmGermany

Natasha Leighl, MDDepartment of Medical Oncology andHematologyThe Ontario Cancer InstitutePrincess Margaret Hospital610 University AvenueToronto, Ontario, M5G 2M9Canada

Contributors

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page ix

Russell E Lewis, PharmDUniversity of Houston College of Pharmacy andSection of Infectious DiseasesUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Marianne Paesmans, MScUnité de Biostatistique Institut Jules BordetCentre des Tumeurs de l’Université deBruxellesRue Heger-Bordet, 1B-1000 BruxellesBelgium

Philip A Pizzo, MDDeanStandford University School of Medicine300 Pasteur DriveStandford, CA 94305USA

Randall A Prince, PharmDUniversity of Houston College of Pharmacy andSection of Infectious DiseasesUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Issam Raad, MDDepartment of Infectious Diseases & InfectionControlUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Kenneth VI Rolston, MD, FACPSection of Infectious DiseasesUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Edward B Rubenstein, MD, FACPSection of Medical Supportive CareUTMD Anderson Cancer Center1515 Holcombe BlvdHouston, TX 77030USA

Stephen C Schimpff, MDChief Executive OfficerUniversity of Maryland Medical Center22 South Greene StreetBaltimore, MD 21201USA

Thomas D Szucs, MDDepartment of Medical Economics Zurich University HospitalsRämistrasse 100CH-8091 ZurichSwitzerland

James A Talcott, MD, SMCenter for Outcomes ResearchMassachusetts General Hospital55 Fruit Street, B75 230Boston, MA 02114USA

James C Wade, MD, MPH Clinical Research DivisionFred Hutchinson Cancer Research Center 1100 Fairview Avenue N, D5-280Seattle, WA 98109USA

John R Wingard, MDBone Marrow Transplant ProgramDepartment of MedicineDivision of Hematology/OncologyUniversity of Florida College of MedicineJHMHC, 1600 SW Archer RoadGainesville, FL 32610-0277USA

x FEBRILE NEUTROPENIA

516_Febrile Neutro.prelims 10/04/2001 12:45 pm Page x

1Fever and neutropenia: An historicalperspectiveStephen C Schimpff

INTRODUCTION

Many serious cancers can now be treated effec-tively. Infectious complications, however, con-tinue to be a frequent cause of morbidity, andoften a leading cause of death, despite theremarkable progress that has been made intheir recognition, prevention, and therapy. Thisdichotomy stems from the intensification ofpresent-day drug and irradiation treatment reg-imens that have, in actuality, only been possiblebecause of refinements in supportive care.Although survival has improved, the price hasbeen a continued and even increased predispo-sition to infection.

This situation is not unlike that which facedoncologists in the late 1960s and early 1970s, aschemotherapy became more effective and morecommonly utilized. Infections that wereunusual, hard to diagnose, and often rapidlyfatal had become common, yet the principles ofmanagement that are taken as standard practicetoday were still being developed.

FACTORS PREDISPOSING TO INFECTION

The following are some of the most importantfactors that predispose to infection in cancerpatients:

• neutropenia and other defects in phago-cytic defenses;

• cellular immune dysfunction;• humoral immune dysfunction;• anatomic-barrier (mucosal or integumen-

tary) damage;• obstructive phenomena;• central nervous system dysfunction;• various iatrogenic procedures.

Additional considerations are the alterations inmicrobial flora and the acquisition of neworganisms in the hospital environment.

NEUTROPENIA

Neutropenia is common in patients with acuteleukemia, following bone marrow transplanta-tion, and following intensive myelosuppressivedrug therapy for other malignancies, or as aresult of aplastic anemia. The incidence andseverity of infection is inversely proportional tothe absolute neutrophil count. Figure 1.11

graphically displays the incidence of all infec-tions among 64 consecutive patients with acutenon-lymphocytic leukemia admitted for theirinitial remission induction therapy. The inci-dence of infection began to rise as the neu-trophil count fell below 500/µl, with a very

substantial rise when the neutrophil count wasbetween 0 and 100/µl. It is obvious from thefigure that most severe infections and nearly allbacteremias occurred when the neutrophilcount was less than 100/µl. An additional fac-tor, not indicated by the figure, is the effect ofthe rate of fall of the neutrophil count; rapiddeclines were more often associated with infec-tion. These observations are directly compara-ble to those first described by Bodey et al2 in1966 in a landmark article that definitivelyrelated neutropenia to infection incidence andseverity.

Not only does the level and rapidity in

decline of the neutrophil count correlate withinfection – so too does the duration of the aplas-tic phase. The current approach to remissioninduction therapy of acute myelocyticleukemia, for example, is such that patients willbecome and remain neutropenic for 20–40 days,with about one-half of that time spent with anabsolute level of circulating neutrophils of lessthan 100/µl. Likewise, patients who receiveallogeneic bone marrow transplants will have aperiod of approximately 3 weeks with essen-tially no circulating neutrophils, and are at anexceedingly high risk of infection during thattime.

2 TEXTBOOK OF FEBRILE NEUTROPENIA

6

Severe infections

5

4

3

2

1

0

Infe

ctio

ns p

er 1

00 d

ays

0–99 100–499 500–999 �1,000

Granulocyte count (cells/µl)

All infections

Bacteremias

Figure 1.1 Incidence ofinfection in acute non-lymphocytic leukemia duringinduction therapy. Reprintedfrom Joshi JH, Schimpff SC,Infections in thecompromised host. In:Principles and Practice ofInfectious Diseases, 2ndedn (Mandel GL, DouglasRG Jr, Bennett JE, eds),Copyright © 1985, JohnWiley & Sons, Inc.Reprinted by permission ofJohn Wiley & Sons, Inc.

Although neutropenia clearly predisposes toinfection, the occurrence of infection in the set-ting of neutropenia is dependent upon the pres-ence or absence of some other associatedpredisposing factors, which act in concert withthe absence of neutrophils. When cancerchemotherapy damages mucosal membranes,the opportunity for development of pharyngitisor esophagitis, typhilitis, or perianal lesions isaccentuated. Damage to the integument byvenipuncture, indwelling vascular catheters, oraxillary shaving may lead to infection. Damageto the mucosa of the trachea and bronchi, alongwith damage to ciliary function due to cancerchemotherapy, may offer the opportunity for

pneumonia to develop. Any form of obstructivephenomenon can interact with neutropenia toencourage infection, such as the developmentof a urinary tract infection in a patient withtumor infiltration of the prostate, otitis mediafollowing an enlargement of adenoid tissue inpatients with lymphocytic leukemia, or thedevelopment of axillary lesions in patients whouse occlusive antiperspirants.

SITES OF INFECTION

The most common sites (Figure 1.2) of infectionin neutropenic patients are the oropharynx, the

FEVER AND NEUTROPENIA: AN HISTORICAL PERSPECTIVE 3

Sinuses

Periodontium Pharynx

Lungs

Distalesophagus

Bone marrowaspiration site

Colon

Anus

Veniputure

fingerstick

Vascular accesscatheter

Figure 1.2 Sites ofinfection amonggranulocytopenic cancerpatients. Reprinted withpermission from SchimpffSC, Infection in patientswith cancer: overview andepidemiology. In:Comprehensive Textbook ofOncology, 2nd edn (MoossaAR, Schimpff SC, RobsonMC, eds). CopyrightWilliams & Wilkins, 1991.

lung, the perianal area, and the skin, especiallyat sites of damage/invasion. As a general rule,the organisms that cause infection (Table 1.1) atany given site are usually organisms that havecolonized (not being just transiently present in)that area or a nearby area.3 In the presence of adamaged mucosal barrier, ciliary dysfunction,or obstruction, and in the absence of normalnumbers of granulocytes, it becomes possiblefor such an organism of otherwise low patho-genicity to cause infection. Thus, pneumoniasare usually caused by organisms that have beencolonizing the patient’s oronasopharynx, andperianal lesions are caused by one or more ofthe organisms colonizing the lower intestinaltract.3,4 In addition, there are bacteremias ofunknown origin, some of which are presumedto relate to bacterial translocation along the


Table 1.1 Infections in neutropenic patientsa,3

Sites Pathogens

Alimentary canal: Gram-negative bacilli:Periodontitis Escherichia coliPharyngitis Pseudomonas Esophagitis aeruginosaColitis Klebsiella pneumoniaePerianal lesions

Gram-positive cocci:Respiratory tract: Streptococcus spp.

Sinusitis Staphylococcus aureusPneumonitis Staphylococcus

epidermidisSkin:

Local trauma Yeasts/fungi:Vascular access Candida spp.

Aspergillus fumigatus/flavus

a These patients can become infected at any site and by anypotential pathogen, but the sites and pathogens listed hererepresent more than 85% of acute infections.

intestinal wall. It is important to recognize that,although most infections are caused by organ-isms already colonizing the patient, these maywell have been acquired by the patient subse-quent to admission to the hospital. Theseacquired organisms may prove to be more viru-lent or more resistant to commonly utilizedantibiotics, or both.5

These predominating infection sites are read-ily explainable: acute periodontitis occurs as aresult of acute exacerbation of previouslyunrecognized chronic periodontal disease.Esophagitis occurs in the distal esophagusbecause of mucosal damage due to chemothera-peutic agents exacerbated by acid reflux fromthe stomach, which is secondary tochemotherapy-induced vomiting. A not uncom-mon progression is infection first with herpessimplex, followed by a mixed bacterial infec-tion, followed by invasive infection by Candida.Perianal lesions occur particularly in patientswith acute monocytic or myelomonocyticleukemia, and can reach an incidence of 33%.Patients with a history of hemorrhoids are mostfrequently affected because of the developmentof small mucosal tears at the base of the hemor-rhoid at the anal opening.6 The high pressuresdeveloped in the process of defecation exacer-bate this process. Sinusitis seems to develop inpatients with a previous history of sinus infec-tions, perhaps suggesting a tendency towardobstruction to the ostia. Pneumonia resultsfrom damaged ciliary function, with reducedtracheobronchial clearance of mucus. Thesealterations in normal clearance mechanismsallow the organisms normally aspirated duringsleep to establish local infection, which is thenunchecked by either neutrophils or pulmonarymacrophages. The axillae are common sites ofinfection because of the warm, moist environ-ment that allows for the growth of organisms inan area that has been damaged by shaving or inan area where hair follicles have been occludedby antiperspirants. Infection at areas of directdamage to the skin, such as bone marrow aspi-ration sites and fingersticks, occurs becausehealing is slow after chemotherapy and because

the number of organisms necessary to induceinfection in the individual who is neutropenic issubstantially less than in the normal host.

‘Bacterial translocation’ is a term used todefine the movement of bacteria across theintact intestinal epithelium into the mesentericlymph nodes and possibly beyond to cause sys-temic infection. This process is well recognizedwith Salmonella typhi in the production oftyphoid fever. However, animal experimenta-tion shows that certain aerobic Gram-negativeorganisms, principally Escherichia coli, Klebsiellapneumoniae, and Pseudomonas aeruginosa, canalso translocate across the normal alimentarycanal mucosa under conditions of suppressionof the anaerobic flora of the intestinal tract orsuppression of cellular immune function. It istherefore possible that many episodes of so-called bacteremia of unknown origin have theirorigin in the intestinal tract as a result of bacter-ial translocation in the absence of specificmucosal epithelial damage.

PATHOGENS CAUSING INFECTION

The most common (i.e. approximately 85%)causes of bacterial infection in the neutropenicpatient are the aerobic Gram-positive cocciStaphylococcus epidermidis (i.e. coagulase-negative staphylococci), � (viridans) Strepto-coccus spp. and Staphylococcus aureus, and theaerobic Gram-negative rods, especially E. coli,K. pneumoniae, and Ps. aeruginosa. Despite colo-nization with other aerobic Gram-positive andGram-negative organisms, patients who areneutropenic generally do not develop infectionor bacteremia other than with those notedabove. Bacteroides fragilis, an anaerobic Gram-negative rod that is known to cause infection inmany other settings, and other anaerobes arealso uncommon causes of infection during neu-tropenia. The principal yeasts and fungi tocause infection during neutropenia are Candidaspp. (especially C. albicans and C. tropicalis), andAspergillus spp. (especially A. flavus and A.fumigatus).7

ALTERATIONS IN MICROBIALFLORA/ACQUISITION OF NEW ORGANISMS

Various exogenous influences can affect thehost’s normal microbial flora. The generaldebilitation that occurs as a consequence of anysevere or chronic illness will perturb indigen-ous flora. Shifts of the normal oropharyngealflora toward a predominance of Gram-negativebacilli occur with acute illness, and the preva-lence of colonization by Gram-negative bacillicorrelates directly with the severity of illness.8

Most bacterial pneumonias result from aspira-tion of oropharyngeal contents. Colonization ofthe oropharynx by potential Gram-negativepathogens in the compromised host withdiminished pulmonary defense mechanismscan, therefore, lead to aspiration and pneumo-nia. Once an infection has become establishedin the neutropenic patient, it can rapidlyprogress and easily disseminate.

Antimicrobial agents have the most dramaticeffect on indigenous flora, and cause both rapidand radical changes. Broad-spectrum antibi-otics can suppress the non-invasive and poten-tially beneficial normal flora that may provide adegree of protection against colonization orinfection, or both, by more pathogenic microor-ganisms. Suppression of alimentary canalanaerobes may destroy a means of endogenousmicrobial protection termed ‘colonization resis-tance’.9 In neutropenic hosts, the loss of this, yetanother normal host defense barrier to infectioncan be substantially detrimental and increasethe high infection risk. The occurrence of resis-tant organisms and infections in those receivingbroad-spectrum antimicrobial therapy isanother serious liability. The role of antibioticsin predisposing to fungal infections is clear.10

Table 1.2, from the first of the EuropeanOrganisation for Research and Treatment ofCancer (EORTC) studies, demonstrates theincreasing incidence of further infections asantibiotic therapy is continued over time.11

The cancer patient may spend substantialtime in the hospital or clinic, and, as a result, isgiven an opportunity to acquire potential


pathogens from this environment.3,10 The organ-isms colonizing a patient at the time of infectionmay have been acquired only subsequent topatient admission. This is of considerableimportance, because the organisms that apatient is likely to acquire in the hospital aremore likely to be resistant to various antibiotics.Approximately one-half of all infections arecaused by organisms that have been acquiredby the patient during hospitalization in the set-ting of neutropenia.3

INFLAMMATORY RESPONSE

The absence or near absence of neutrophils sub-stantially limits the inflammatory response,which in turn affects both diagnosis and prog-nosis. There are very few early signs and symp-toms except for fever.12 It is this ability of anotherwise minor-appearing and localized infec-

tion to progress rapidly to a systemic bac-teremia that makes the need for early diagnosisand prompt empiric therapy critical. Thepatient with a Gram-negative bacteremia whois not treated promptly will usually die within24–48 hours unless antimicrobial therapy is ini-tiated within the first few hours.

Despite the presence of few of the classicmanifestations of localized infection, the vastmajority of these febrile episodes occurringduring the period of neutropenia are due toinfection.12 Approximately 20% of febrileepisodes have an associated bacteremia,another 20% have a microbiologically docu-mented infection without bacteremia, andanother 20% have clear-cut evidence of the siteof infection but an etiologic agent cannot bedefined. This leaves 20% with fever caused by anon-infectious etiology (infection doubted) andthe remaining 20% in whom infection is highlysuspected but is never proved. Overall then, at


Table 1.2 Relation between incidence of further infection and neutrophil count and duration ofantibiotic therapy for infection in neutropenic patients with cancer (EORTC Trial I)a

Patients with further infection/ Patients with further infection/patients with stable neutrophil countb patients with increased neutrophil countb

Duration of antibiotictherapy (days) Number Percentage Number Percentage

<5 017/87 20c 001/52 02d

6–10 017/96 17 08/115 0711–15 016/39 41 006/34 18>15 008/27 30c 005/20 25d

Total 58/249 23e 20/221 09e

a Modified with permission from EORTC International Antimicrobial Therapy Cooperative Group, J Infect Dis 1978; 137:14–29.11

b Patients with a stable neutrophil count had persistent neutropenia, and the neutrophil count of patients with an increaserose by 100/µl or more during therapy.c p � 0.20 for the difference between these two values (not significant); d p � 0.005 for the difference between these twovalues; e p � 0.001 for the difference between these two values.

least 60% of new febrile episodes are associatedwith infection.11

GRAM-POSITIVE INFECTION

Viridans and �-hemolytic streptococci havebecome frequent pathogens in febrile neu-tropenic patients. These streptococcal infectionsmay be severe and present with septic shock oracute respiratory distress syndrome. Since viri-dans streptococci are normal inhabitants of themouth and pharynx, it has been hypothesizedthat these infections arise from the oral cavity.These streptococcal infections may be sec-ondary to the development of severe mucositisfollowing radiation therapy or chemotherapy,particularly in patients treated with high-dosecytosine arabinoside, but may also be sec-ondary to oral ulcerations due to herpesvirusinfections. Another factor predisposing patientsto these streptococcal infections may be the useof quinolone antibiotics for the prevention ofbacterial infection.13–15

GRAM-NEGATIVE INFECTIONS

Gram-negative bacteremia is essentially a dis-ease of modern times, with fewer than 100reported cases prior to 1920. Over the last 40years, a number of studies have looked at theincidence of Gram-negative bacteremia andhave noted a continuing increase.16 Althoughmortality rates in various reports range widely,a case fatality rate of about 50% is common. Thefatality rate depends fairly dramatically on hostfactors, along with the approach to treatment,the occurrence of complications, and, to somedegree, the specific pathogen. McCabe andJackson17 were the first to emphasize the impor-tance of the host’s underlying disease by divid-ing patients into those with rapidly fatal disease(i.e. those expected to die within the course ofthe next year), those with ultimately fatal dis-ease (i.e. those who would die within about 4years), and those with non-fatal underlying dis-

ease. They found the fatality rates to be 91%,66%, and 11%, respectively.

EMPIRIC THERAPY

At the Baltimore Cancer Research Center of theNational Cancer Institute in 1969, a review ofthe microbiology records from the previousyear indicated that there were 22 episodes of Ps.aeruginosa bacteremia. Of these 22 patients, 11died within 72 hours from the time the firstpositive blood culture was drawn; all but onepatient eventually died of the infection.18 Thesite of infection was never identified in the vastmajority. Antibiotic therapy usually included acombination of cephalothin plus kanamycin(neither drug being active against Ps.aeruginosa), and was usually not started untilafter a report of a positive blood culture for aGram-negative rod had returned from the labo-ratory or until the patient developed signs ofseptic shock. Polymixin B or E tended to beadded to the regimen only when laboratoryidentification had been achieved.

In the 1960s and early 1970s, the acceptedapproach was not to institute antibiotic therapyuntil there was some definitive proof of infection– fever alone was not considered enough. But itwas clear from review of the 22 patients withPseudomonas bacteremia that one could not waitfor laboratory results to return, because half ofthe patients had died within 72 hours. Further, itseemed that, apart from fever, documentation ofinfection was uncommon except for the culturereport, which generally returned too late to be ofvalue. Therefore, it only seemed appropriate totreat all neutropenic patients who developednew fever with an antibiotic regimen empiri-cally. Carbenicillin and gentamicin (both stillinvestigational in 1969) seemed to be a logicalregimen because of the broad Gram-negativeand Gram-positive activity of gentamicin,including against Ps. aeruginosa, and the anti-pseudomonal activity of carbenicillin. Further,there were laboratory data which indicated thatcarbenicillin and gentamicin were synergistic in


vitro against Ps. aeruginosa, and there were somedata to suggest that resistance might developless rapidly when a combination was utilized.We then treated 75 febrile neutropenic patientswith this regimen: 48 had a microbiologicallydocumented infection, another 12 had a clini-cally documented infection, 12 had a possibleinfection, and 3, in retrospect, were felt not tohave been infected. Thus, the empiric approachto therapy was appropriate in all but either 3 or15 of 75 patients. Among the 48 patients with amicrobiologically documented infection were 13with a Pseudomonas bacteremia, of whom 8improved and 3 improved temporarily, plus anadditional 8 patients with a non-bacteremic Ps.aeruginosa infection, of whom 6 improved. Thiswas a fairly striking difference from the resultsof the prior year, noted above; however, it couldnot be ascertained whether the critical factorhere was the early empiric institution of antibi-otics, the effectiveness of the new investigationalcombination of agents, or both. It is noteworthy,however, that during this same time frame, anadditional 8 patients with neutropenia and fever,who proved to have a Ps. aeruginosa bacteremia,were treated with gentamicin alone shortly afterfever developed. Each patient had a strain sus-ceptible to gentamicin, yet 7 patients had persis-tently positive blood cultures while receivingthis drug, whereas 1 patient rapidly improvedwith this single-agent therapy. This suggestedthat gentamicin alone was not adequate therapyfor Pseudomonas bacteremia in neutropenicpatients.19

EARLY INITIATION OF THERAPY

The studies by Greisman et al20 are relevant tothe observation that prompt therapy is impor-tant. They studied non-neutropenic mice, eachof which received a lethal intraperitoneal doseof a Gram-negative rod such as E. coli or K.pneumoniae. They then utilized an antibiotic towhich the organism was susceptible, in a doseand schedule that would assure a blood levelabove the minimal inhibitory concentration on


a continuing basis. The only variable in theexperiment was the time of initiation of the firstdose of antibiotic. It was found that if the firstdose of the antibiotic was administered concur-rently with the intraperitoneal injection, none ofthe mice died. However, the longer the firstdose was delayed, the greater was the likeli-hood that death would occur. Even though theanimal might survive for a few days, death wasinevitable if the first dose was more than just afew hours delayed from the onset of infection(Table 1.3). Thus, it was demonstrated thatthere is a ‘window of opportunity’ withinwhich therapy must begin if death is not toensue.

This was demonstrated nicely in two studiesby Bodey et al from the University of TexasMD Anderson Cancer Center. In reviewingPseudomonas bacteremia, they found that about15% of neutropenic patients died if the first doseof therapy was given within 12 hours of theonset of fever, yet 55–75% died if the first dose

Table 1.3 Mouse mortality after Gram-negative infection20

Time (hours) sinceantibiotic begun % mortality

0 0001 0151.5 0452 0703 0954 100

Mice were injected intraperitoneally with a lethal dose(1 � 108) of E. coli 018, then treated with a bactericidalantibiotic at a dose and schedule to maintain an effectiveserum concentration, i.e. constantly above the minimumbactericidal concentration of the challenge organism. Notethe ‘window of opportunity’, i.e. mice treated quickly withfirst dose all survive, but die if treated for first time only afew hours after innoculum of bacteria injectedintraperitoneally.

of antibiotic was delayed.21 For E. coli bac-teremia, including neutropenic and non-neutropenic patients, about 12% of patients died(Figure 1.3) if antibiotics were started within thefirst 12 hours, compared with 18% when theywere started between 12 hours and 24 hours,and 30% when they were started between 24hours and 48 hours after the collection of thefirst blood culture that proved to be positive.The mortality rate continued to rise to 80% ifappropriate therapy had not been institutedpromptly.22 A similar observation was madewith Ps. aeruginosa bacteremia (Figure 1.4).21

Many published reports indicate that the siteof Gram-negative bacteremia in the neutropeniccancer patient is frequently never identified.This has not been my experience when eachpatient has been studiously examined on adaily basis.3 Perhaps what is most important isthe recognition that the signs and symptoms of

inflammation are markedly diminished andtherefore evidence of infection may be subtle. Itis particularly helpful to have seen the patienton a regular basis prior to the onset of infection,so that one can compare the prefebrile examina-tion with any changes that may have occurred.If one knows the common sites of origin ofinfection in these patients, then it is possible togive particular attention to those sites and lookfor subtle changes. For example, infection aris-ing from pharyngitis may be represented onlyby complaints of an intense sore throat andsome erythema, but little other physical evid-ence. Bacteremia due to a perianal lesion maybe detected by the patient noting pain withdefecation, and an examination that shows onlyminimal erythema but intensive tendernessover the site of a minor-appearing fissure, oftenat the base of a hemmorrhoid, which serves asthe nidus and origin for the bacteremia.


0

Cum

ulat

ive

% o

f pa

tient

s

0

Days from first positive blood culture to start of therapy

20

40

60

80

100

1 2 3 4 5 6 7

Figure 1.3 E. coli bacteremia. Mortality is related to time of onset of bacteremia after the first positive bloodculture and institution of appropriate therapy. Reprinted from Am J Med, Vol 81 (Suppl 1A), Bodey GP, Elting L,Kassameli H, Lim BP, Escherichia coli bacteremia in cancer patients, pp 85–95. Copyright 1986, withpermission from Excerpta Medica Inc.

Some organisms, notably Ps. aeruginosa, areexceptionally invasive during profound neu-tropenia (i.e., if colonized, the patient fre-quently becomes infected). Colonization withother organisms such as E. coli and K. pneumo-niae occasionally leads to bacteremia, and colo-nization with still other organisms such asnon-aeruginosa Pseudomonas spp. very rarelyproceed to infection, even during profoundneutropenia. Thus, there is a clear difference ininvasive potential among Gram-negative bacilliin this highly vulnerable population of patients.

CHOICE OF DRUGS AND LEVEL OFNEUTROPENIA: IMPLICATIONS FORSURVIVAL

Bodey et al have reviewed the common causesof Gram-negative bacteremia in cancer patients,such as E. coli,22 Ps. aeruginosa,21 and some of the

less common organisms, such as Serratiamarcescens and Enterobacter spp. Consistently,the rate per 1000 admissions is higher forpatients with acute leukemia than for patientswith other hematologic malignancies, andmuch higher than for patients with solidtumors. Bodey and his colleagues have notedthat for patients who become bacteremic, theblood culture may be positive 50% or more ofthe time when fever is first documented (Figure1.5). Another consistent finding has been theimportance of the neutrophil count with regardto the ultimate response; when the initial neu-trophil count was below 100/µl and remainedunchanged, the response rate for E. coli bac-teremia was 48%, whereas if the neutrophilcount increased, the response rate was 83%.When the initial neutrophil count was above100/µl yet less than 1000/µl and remainedunchanged, the response rate was 47%, but ifthe neutrophil count increased further, the


0

% d

ying

0

Days from first positive blood culture to start of therapy

20

40

60

80

100

1 2 3 4 5 6 7

Figure 1.4 Ps. aeruginosa bacteremia. Mortality is related to time of onset of therapy after the first positiveblood culture. Reprinted with permission from Bodey GP, Jadeja L, Elting L, Pseudomonas bacteremia:retrospective analysis of 410 episodes. Arch Intern Med 1985; 145: 1621–9. Copyright 1985, AmericanMedical Association.

response rate was 95%. In addition, and notsurprisingly, there was a marked difference insurvival depending upon whether the patientreceived appropriate therapy (i.e. antibiotics towhich the organism was susceptible) or inap-propriate therapy. For E. coli bacteremia, thesurvival rates were about 75% for those whoreceived appropriate therapy and 38% for thosewho received inappropriate therapy initially(Figure 1.6).

THERAPY OF FEBRILE NEUTROPENIA

Given all of these factors, empiric therapy forthe febrile, neutropenic patient must be:

• prompt;• empiric;• bactericidal;• broad-spectrum.

The need for prompt institution of therapy isdue to the rapid and high mortality rate ofpatients with Gram-negative bacteremia and,occasionally, the bacteremias caused byStreptococcus spp. Recall from above that therisk of dying was closely related to the intervalbetween the onset of bacteremia and the institu-tion of appropriate antibiotic therapy (Table 1.3and Figures 1.3 and 1.4).21,22 The need forprompt therapy makes the use of empiricantibacterial regimens obvious. It should beemphasized that an antibiotic regimen must bechosen that is most appropriate for the specificpatient in a specific institution. It is necessary toknow what organisms are, or are likely to be,colonizing the patient and what the likely sus-ceptibility patterns will be. It is critical to havecontinually updated information on the suscep-tibility patterns of organisms frequently recov-ered from the hospital and the area of the


0

Cum

ulat

ive

% o

f pa

tient

s

0

Days

20

40

60

80

100

1 2 3 4 5 6 7 8

Figure 1.5 E. coli bacteremia. Time of first positive blood culture relative to onset of fever. Nearly 50% ofpatients were bacteremic when fever first developed. Reprinted from Am J Med, Vol 81 (Suppl 1A), Bodey GP,Elting L, Kassameli H, Lim BP, Escherichia coli bacteremia in cancer patients, pp 85–95. Copyright 1986, withpermission from Excerpta Medica Inc.

hospital where the patient is being treated.Bactericidal rather than bacteriostatic antibi-otics are essential, since, in the absence of neu-trophils, this is a battle of ‘bugs versus drugs’.The agents should have a broad spectrum so asto ‘cover’ the great majority of the relativelylimited number of potential pathogens.

Combination empiric therapy

The options for choices of antibiotics are wide(Table 1.4). For many years, the most commonapproach was the use of �-lactam plus anaminoglycoside. These combinations havewithstood the test of time, they have beenfound to be broadly effective, the newer �-lac-tams offer ‘coverage’ for most of the Gram-negative and Gram-positive bacteria thatinvade these patients, there is synergistic activ-ity against many Gram-negative bacilli, and

there are data to suggest that the developmentof resistance to the �-lactam is less likely withthe added aminoglycoside. Among the peni-cillins, those with the broadest spectruminclude piperacillin (especially when combinedwith tazobactam) and ticarcillin (especiallywhen they are combined with the �-lactamaseinhibitor clavulanic acid). The antipseudomonalcephalosporins such as ceftazidime andcefepime likewise offer broad Gram-negativecoverage, as do the carbapenems imipenem andmeropenem. The monobactam azetreonam hasexcellent Gram-negative activity, but incorpo-rates no Gram-positive activity. Thecephalosporins and carbapenems have activityagainst S. aureus and the streptococci. None ofthese agents are effective for S. epidermidis (evenif susceptibility results suggest activity). Thepenicillins, some of the cephalosporins, andimipenem have activity against anaerobes(which rarely cause bacteremia but which are


0

% o

f pa

tient

s

0

Days

20

40

60

80

100

2 4 6 8 10 12 14

Appropriate therapy

Inappropriate therapy

No therapy

Figure 1.6 E. coli bacteremia. Mortality is related to whether or not initial therapy was appropriate, i.e.whether the organism was susceptible by in vitro testing. Reprinted from Am J Med, Vol 81 (Suppl 1A), BodeyGP, Elting L, Kassameli H, Lim BP, Escherichia coli bacteremia in cancer patients, pp 85–95. Copyright 1986,with permission from Excerpta Medica Inc.

important along the alimentary canal to pre-serve colonization resistance), whereas cef-tazidime has no activity against anaerobes. Thechoice of the �-lactam agent should be basedlargely on institutional antimicrobial suscepti-bility patterns and, preferably, knowledge ofsusceptibility patterns for recent infectionswithin the oncology unit.

The commonly utilized aminoglycosidesinclude gentamicin, tobramycin, and amikacin.For susceptible organisms, each of these amino-glycosides probably has equivalent efficacy.Aminoglycosides are both ototoxic and nephro-toxic, and it is generally advisable to measureserum levels to be sure that one is within thetherapeutic, yet below the toxic, range. If oneapproaches the aminoglycoside primarily for itsvalue in adding synergy, then it is probably notnecessary to push toward the higher, more


Table 1.4 Available antibiotics for initialempiric therapy of the febrile neutropenicpatient

�-lactams Aminoglycosides

Penicillins GentamicinPiperacillin � tazobactam TobramycinTicarcillin � clavulanic acid Amikacin

CephalosporinsCeftazidimeCefepimeCeftriaxoneMonobactamAztreonam

CarbapenemsImipenemMeropenem

toxic side of the accepted therapeutic range,and thereby lessen the opportunity for undesir-able side-effects. There are no data to demon-strate that higher peak and trough levels ofaminoglycosides, when given in combinationwith a �-lactam for this type of patient, aremore efficacious than a somewhat lower dose.There is sufficient data to demonstrate thataminoglycosides alone are not adequate in thesetting of profound neutropenia (i.e. <100/µl).

Monotherapy

With the advent of the very broad-spectrum�-lactams, such as ceftazidime, imipenem, andticarcillin plus clavulanic acid, and given theinherent toxicities of the aminoglycosides, itseemed reasonable to attempt to use single-agent therapy for initial empiric treatment. Theclassic study was completed by Pizzo et al,23

who randomly allocated patients to a combina-tion of carbenicillin, cephalothin plus gentam-icin or to ceftazidime at the onset of feverduring neutropenia. Ceftazidime alone was asefficacious as the combination. A large numberof patients required alteration in therapy, suchas the addition of vancomycin, an antifungal, oran antiviral agent to each of the two initial regi-mens. However, the ultimate responses wereequivalent, and the mortality rate wasextremely small in both groups. Ceftazidime,imipenem, meropenem, and cefepime havebeen used effectively for many patients asinitial empiric therapy for fever during neu-tropenia.

I do not believe that there is adequate data toassess whether monotherapy with one of theseagents is sufficient for the patient who has pro-found, persistent neutropenia and a Gram-negative rod bacteremia. Only about 10% ofpatients in most large studies have proven tohave a Gram-negative rod bacteremia, and onlyabout half of these patients tend to fall into thecategory of those with profound, persistentneutropenia. For example, in the study by Pizzoet al,23 there were only 13 Gram-negative rod

bacteremias out of more than 500 patiententries. In an EORTC trial,24 there were only 129Gram-negative bacteremias out of 1074 patientsentered (Table 1.5), and of the 129, only 53 hadprofound, persistent neutropenia. It is this lattergroup of patients, however, that concern me. Instudy after study, they tend to do poorlyregardless of the agent(s) used, but do less wellwith single-agent therapy (see below). In thisEORTC trial, for example, the response rates forpatients with Gram-negative rod bacteremiaand a neutrophil count of less than 100/µlthroughout therapy were 6% with ceftazidimeand short-course amikacin, compared with 50%with ceftazidime and long-course amikacin(p � 0.03) (Table 1.6 and Figure 1.7). Thiswas further evidence to suggest the value ofthe aminoglycoside in combination withthe �-lactam in this particular subgroup ofpatients.

Synergy

At our institution, de Jongh et al25 looked at aseries of 75 consecutive Gram-negative rod bac-teremias that occurred among patients who hada neutrophil count of less than 100/µl; eachreceived prompt empiric antibiotic therapyusing a combination. The critical observationswere as follows. First, there was a dramatic dif-ference in response rate between patients whoremained profoundly neutropenic and thosewhose neutrophil count began to rise duringthe next few days (Figure 1.8). Indeed, theresponse rate for those who were profoundlyneutropenic was substantially and disturbinglypoorer. Dissecting further, the patients withpersisting, profound neutropenia had aresponse rate that was significantly better ifthey had received two drugs to which theGram-negative bacillus proved to be suscepti-


Table 1.5 EORTC IV trial24

Entries 1074

Exclusions 0202Protocol violation 0052Doubted infection 0135Viral/fungal infection 0015

Evaluable episodes 0872Possible infection 0342Clinically documented 0225Microbiologically documented 0305

Without bacteremia 053With bacteremia 252

Polymicrobial 033Single-organism 219

Gram-positive 090Gram-negative 129

Persistent profound neutropenia 53


Table 1.6 Response to treatment in the presence and absence of persistent profound neutropenia24

Condition of patients Patients with response/patients with bacteremia

Azlocillin Ceftazidime Ceftazidime� amikacin � short amikacin � long amikacin Total

Persistent profoundneutropenia:

Presenta 5/25 (20%) 1/16 (6%) 6/12 (50%) 12/53 (23%)b

Absentc 11/15 (73%) 19/26 (73%) 32/35 (91%) 62/76 (82%)b

Reprinted with permission from EORTC International Antimicrobial Therapy Cooperative Group, N Engl J Med 1987; 317:1692–8. Copyright © 1987 Massachusetts Medical Society. All rights reserved.a The p values for comparison of treatment regimens were as follows: global, 0.02; azlocillin � amikacin versusceftazidime � short amikacin, 0.45; azlocillin � amikacin versus ceftazidime � long amikacin, 0.14; and ceftazidime � shortamikacin versus ceftazidime � long amikacin, 0.03.b p < 0.001.c Global p value � 0.12.

0

Prop

ortio

n w

ithou

t fa

ilure

0

Time (days)

0.2

0.4

0.6

0.8

1.0

2 4 6 8 10 12 14

C� long A

C�short A

AZ�A

Figure 1.7 Time to treatment failure according to treatment regimen (C, ceftazidime; A, amikacin; AZ,azlocillin). Continued use of the aminoglycoside with ceftazidime led to better response rates for Gram-negativebacteria. Reprinted with permission from EORTC International Antimicrobial Therapy Cooperative Group,Ceftazidime combined with a short or long course of amikacin for empirical therapy of Gram-negative bacteremiain cancer patients with granulocytopenia. N Engl J Med 1987; 317: 1692–8. Copyright © 1987 MassachusettsMedical Society. All rights reserved.

ble; the response rate when the organism wassusceptible to only one drug was particularlypoor. Further analysis demonstrated that if thecombination of two drugs was synergistic invitro against the invading organism, then thesepatients did better than if the combination wasnot synergistic.

The study demonstrates that among persis-tently neutropenic patients who are treated withtwo effective bactericidal antibiotics, combina-tion with in vitro synergism is associated with amore favorable clinical outcome than are similarcombinations that are not synergistic in vitroeven when both agents are quite active againstthe pathogen. Such synergism is of no prognosticimportance among patients with rising neu-trophil counts; only those patients with Gram-negative bacteremia who are profoundly andpersistently neutropenic benefit from the pres-ence of the two-drug synergistic combination.

Serum bactericidal activity

Klastersky and colleagues26,27 and Anderson etal28 demonstrated that synergistic combinationsof agents were more effective for Gram-negative


Gram-negative-rod bacteria and neutrophils�100/µlTreated with two-drug empiric regimen

Neutrophils�100/µl30%

Neutrophils rose85% Survival rate

Pathogen susceptible to: Both40%

One5%

Combination synergistic: Yes65%

No0%

Not evaluable

44%

Figure 1.8 Treatment of infection in cancer patientswith granulocytopenia.25 Reprinted from Schimpff SC,Gram-negative bacteremia. Support Care Cancer1993; 1: 5–18. Copyright 1993 Springer-Verlag.

bacteremia than single agents (Table 1.7).Klastersky29 reported that synergistic combina-tions also elicited a serum bactericidal activitythat was significantly greater (1 : 16 versus 1 : 4at peak and 1 : 8 versus 1 : 2 at trough) (Table1.8). However, with the advent of newer �-lac-tams such as ceftazidime and imipenem, goodbactericidal activity could be obtained with thesingle agent. For example, Standiford et al30

gave ticarcillin and amikacin or ceftazidime tovolunteers and measured serum bactericidalactivity at 1 hour and 6 hours. Ceftazidime hada notably better serum cidal profile for Ps. aerug-inosa, E. coli, and K. pneumoniae than did thecombination (Table 1.9 and Figure 1.9).30

In a later study, imipenem was compared

Table 1.7 Summary of the results of 12controlled clinical trials of therapy with singleversus multiple antibiotics and withsynergistic versus non-synergisticcombinations of antibiotics in neutropenicpatients infected with Gram-negative bacillia

Type of therapy Number of patientswith a favorableclinical response

Single antibiotic 119 (61%)(195 patients)

Multiple antibiotics 138 (81%)(170 patients)

Nonsynergistic 77 (43%)combinations (179 patients)

Synergistic 158 (76%)combinations (208 patients)

a Reproduced with permission from Klastersky J, Empirictreatment of infections in neutropenic patients with cancer.Rev Infect Dis 1983; 5(Suppl): S21–31.

with ticarcillin plus amikacin.31 At 1 hour, thegeometric mean bactericidal titers were 13 and12, respectively, while at 5�� hours, they were 3and 2, respectively. An animal model dem-onstrated that severely neutropenic rats given alethal intraperitoneal challenge of Ps. aeruginosaresponded as well to imipenem alone as to the

combination of moxalactam and amikacin.However, the rat survival was substantiallybetter still when amikacin was added to theimipenem (Figure 1.10).32

Alternatively, it may be that two antibioticscan effectively eliminate a Gram-negativebacillus during profound neutropenia only if


Table 1.8 Clinical responses and serum bactericidal activity in patients with cancer and Gram-negative bacillary infections who received synergistic or non-synergistic combinations of antibiotics(the studies were performed at the Institut Jules Bordet, Brussels, Belgium)29

Median titer of serum bactericidal activity

Type of combination Number of patients Maximum Minimumwith a favorable clinical response

Synergistic (100 patients) 80 (80%)a 1 : 16 1 : 8Non-synergistic (105 patients) 52 (50%)a 1 : 4 1 : 2

Reproduced with permission from Klastersky J, Eur J Cancer 1979; 15: 3–13.29

a p < 0.01

Table 1.9 Reciprocal geometric mean bactericidal titers generated at 1 and 6 hours by eachregimen30

Titer obtained with

Ceftazidime Ticarcillin–amikacin

Test organism 1 h 6 h 1 h 6 h

Ps. aeruginosa (31 strains) 40.7 4.7 12.2 2.1S. aureus (7 strains) 3.6 NAa 24.3 3.0E. coli (7 strains) 256.0 128.0 125.5 8.2K. pneumoniae (7 strains) 236.5 97.0 86.1 8.0

Reproduced from Standiford HC et al, Antimicrob Agents Chemother 1984; 26: 339–42.30

a No activity assayable.

they attack by different mechanisms. Using anin vitro system that exposed Ps. aeruginosa tofluctuating levels of gentamicin, Gerber et al33

noted the development of small colonies ofgentamicin-resistant variants. Although thesevariants were less pathogenic in normal andmoderately neutropenic mice than were sus-

ceptible colonies, they invariably killed severelyneutropenic mice challenged intraperitoneally.The development of these gentamicin-resistantvariants could be prevented by the additionof ticarcillin.34,35 Thus, two agents killing by sep-arate mechanisms and preventing the emer-gence of resistant organisms may be important.


Cum

ulat

ive

perc

enta

ge o

f de

term

inat

ions

bac

teric

idal

at in

dica

ted

dilu

tion

100

80

60

40

20

01:256 1:128 1:64 1:32 1:16 1:8 1:4 1:2

K. pneumoniae

*

*p�0.0067

100

80

60

40

20

01:256 1:128 1:64 1:32 1:16 1:8 1:4 1:2

E. coli

*

*p�0.003

1:256 1:128 1:64 1:32 1:16 1:8 1:4 1:2

Ps. aeruginosa

*

*p�0.0248

1:256 1:128 1:64 1:32 1:16 1:8 1:4 1:2

S. aureus

*

*p�0.003

Figure 1.9 Cumulative percentage of determinations bactericidal against pathogens commonly bacteremic inneutropenic cancer patients. Serum was obtained from volunteers 1 hour after the end of the infusion. The pvalue represents the level of significance between geometric mean bactericidal titers produced by the tworegimens. Symbols: �, ceftazidime; �, ticarcillin plus amikacin. Reprinted with permission of the AmericanSociety for Microbiology from Standiford HC, Drusano GL, Fitzgerald B et al, Bactericidal activity of ceftazidine inserum compared with that of ticarcillin combined with amikacin. Antimicrob Agents Chemother 1984; 26:339–42.

Therapy of persistent fever and persistentneutropenia

A major concern for the clinician dealing withthe febrile neutropenic patient is what to dowith the patient who has persistence of feverfollowing the administration of empiric anti-biotic therapy. The questions relate to whetherthe initial antibiotic should be continued or dis-continued, whether an additional antibacterialantibiotic should be added, and whether anantifungal agent such as amphotericin B or anantiviral agent such as acyclovir should beadded. There is no single correct answer. Thefirst step should be to carefully repeat the his-tory, physical examination, and chest X-ray,and to review the results of the original cul-

tures. More often than not, such a review willreveal an infection site, if one exists. However,other causes of fever must be considered: bloodproduct transfusions, a history of fever with theunderlying tumor, and drug fever from com-pounds such as cytosine arabinoside or fromthe empiric antibiotics themselves. Pizzo etal36,37 showed that continued therapy preventednew/recurrent bacterial infection for patientswith persistent neutropenia; however, fungalinfections became common yet could be pre-vented/treated by instituting amphotericin Bon day 7 of continued fever.

The EORTC found that the addition ofempiric amphotericin B after four days ofbroad-spectrum antibiotics and persistent feverand neutropenia had some benefit. Of the


% s

urvi

vors

Hours post bacterial challenge

20

40

60

80

100

10 20 30 40 50 60 70

2 8 14 22 30 38 46 54 62

Antibiotic administration: hours post challenge

Saline

Imipenem 30mg/kgplus

amikacin 20mg/kg

Imipenem 30mg/kg

Moxalactam 150mg/kgplus

amikacin 20mg/kg

Moxalactam 150mg/kgAmikacin 20mg/kg

Figure 1.10 Antibiotic therapy in neutropenic rats challenged with 250 LD50 (2.2 � 109) of Ps. aeruginosastrain number 228. Reprinted with permission from Johnson DE, Calia FM, Snyder MJ et al, Imipenem therapyof Pseudomonas aeruginosa bacteremia in neutropenic rats. J Antimicrob Chemother 1983; 12: 89–96.Copyright 1983 Oxford University Press.

patients with added amphotericin B, 69% hadresolution of fever, compared with 53% in thecontrol group. There was one fungemia (1 of 68patients) compared with six (6 of 64 patients),and there was one death due to fungal infectioncompared with four. Of note, those who hadfungal infections usually had some clinicalevidence to suggest that it might be present, i.e.this was not entirely ‘empiric’ therapy.38

Therapy of persistent neutropenia withfebrile response

There are some patients with persistent neu-tropenia and no specific evidence of infectionon repeated history and physical examinationwho have a ‘febrile response’, i.e. the feverabates promptly after institution of antibiotics.This raises the question as to whether thepatient was infected and whether the anti-biotic(s) should be continued? The critical stepis to repeat the history and physical examina-tion, review all cultural data, and repeat thechest X-ray. If no specific evidence of infectioncan be determined, yet it appears that thepatient has had a febrile response secondary toantibiotic therapy, then it would seem reason-able to continue the antibiotics for a total ofabout 10 days. If at that time the neutrophilcount remains very low, one would have todecide whether to continue antibiotic therapyfor a longer period. I usually discontinue antibi-otics at this time, but there is evidence to sug-gest that continuing antibiotic therapy isappropriate.36 However, continuation must bebalanced against the potential risk of predispos-ing toward fungal infection, which may requirethe addition of amphotericin B or other antifun-gal agents.38

SHIFT FROM GRAM-NEGATIVE TOGRAM-POSITIVE PREDOMINANCE

In the 1960s and 1970s, the predominantpathogens of febrile neutropenic episodes were

Gram-negative bacilli along with some S.aureus. Over the past 25 years, a change towardfewer Gram-negatives and a predominance ofcertain Gram-positives, especially coagulase-negative staphylococci and viridans strepto-cocci, has occurred (Figure 1.11).39

The reasons for this shift are not totally clear,but some hypotheses are as follows. Preventiontechniques such as attention to handwashingmay have reduced S. aureus transmission.Attention to water sources (e.g. faucet aeratorsand tamperproof ice machines) and the use oflower-microbial-content foods (e.g. avoidingsalads and uncooked tomatoes, and usingfreshly ground pepper) may have reduced theacquisition of Gram-negative bacilli.40 The useof alimentary canal microbial suppression (e.g.oral quinolones) may have reduced Gram-negative bacillary invasion of the damagedmucosa.41

Concurrently, the commonplace use ofindwelling vascular access catheters hasincreased the opportunity for S. epidermidisinfections (entry-site infections, tunnel infec-tions, and especially internal colonization of thecatheter, with bloodstream seeding).42

Streptococcal infections may be related to inten-sive oral-mucosal-damaging chemotherapy,specific agents such as high-dose cytosine arabi-noside, and the use of quinolones as prophy-laxis.

This shift in pathogen frequency has led toconsideration of changes in the choice ofempiric antibiotics.43 Some have suggested thata combination including vancomycin should beused in the initial regimen. Others have indi-cated that S. epidermidis, unlike Gram-negativebacilli, tends to cause a more indolent infectionand hence there is time to substitute or addvancomycin once culture reports are avail-able44–46 (Figure 1.12). This avoids the inherentnephrotoxicity of vancomycin, especially whenit is used in conjunction with an aminoglyco-side, amphotericin B, or both. The increasingproblem of vancomycin-resistant enterococci(VRE) and hence the need to restrict van-comycin to situations where it is truly needed


strengthens the case for withholding van-comycin as part of the initial empiric regimen.

The streptococci are generally quite suscepti-ble to the various �-lactams in use, althoughresistance to penicillins is definitely on therise.47 Prompt initiation of therapy is key,because these organisms are capable of causingvery serious infection with shock over a shorttime frame.

ORAL THERAPY INSTEAD OF INTRAVENOUSTHERAPY

The initial concepts in treating the febrile neu-tropenic patient empirically included:

• use of broad-spectrum antibiotics to assurecoverage for most of the commonpathogens;

• intravenous therapy to assure rapidachievement of adequate serum levels;

• close monitoring because of the concern forprogression to septic shock with Gram-negative bacteremia or possibly the devel-opment of respiratory failure withpneumonia.

Today, Gram-negative bacteremia is much lesscommon, and it is possible to establish thepatient’s relative risk for an adverse outcomeof febrile neutropenia. Hence, it would seemlogical to utilize oral agents that have an


140

120

100

80

60

40

20

0

1973–1978

1978–1980

1980–1983

1983–1986

1986–1988

1989–1991

1992–1994

Coagulase-negativestaphylococci

S. aureus

S. pneumoniae

Otherstreptococci(viridans)

MiscellaneousGram-positiveorganisms

Num

ber

of p

atho

gens

Figure 1.11 Changing pattern of infectious organisms in febrile neutropenia. Reprinted from Maschmeyer G,Noskin GA, Ribaud P, Sepkowitz KA, Changing patterns of infections and antimicrobial sensitivities. Oncology2000; 14(Suppl 6): 9–16. With permission of ONCOLOGY, Melville, NY.


1.0

0.8

0.6

0.4

0.2

0

Prop

ortio

n of

pat

ient

s on

prot

ocol

ant

ibio

tic t

hera

py

p�0.0001

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Days of treatment

CAV

CA

1.0

0.8

0.6

0.4

0.2

0

Prop

ortio

n of

feb

rile

patie

nts

p�0.85

0 2 4 6 8 10 12 14 16

Days of treatment

CAV

CA

1 3 5 7 9 11 13 15

Figure 1.12 Gram-positive bacteremia: proportion of febrile patients at each treatment day in the twotreatment regimens. No discrepancy was noted between the two groups in the proportion of febrile patients(p � 0.85).44 CA (�), ceftazidime and amikacin; CAV (�), ceftazidime, amikacin and vancomycin. Reprinted withpermission from EORTC International Antimicrobial Therapy Cooperative Group and the National Cancer Instituteof Canada – Clinical Trials Group, Vancomycin added to empirical combination antibiotic therapy for fever ingranulocytopenic cancer patients. J Infect Dis 1991; 163: 951–8. Copyright 1991 University of Chicago Press.

adequate spectrum and are well absorbed,especially in low-risk patients. Data supportingthis approach have recently been published.48–51

See Chapter 9.

SUMMARY

Much progress has been made over the past30–35 years regarding the treatment of neu-tropenic patients who develop fever.

• It has become accepted that empiric ther-apy is appropriate when fever develops.

• It has been recognized that the inflamma-tory response is muted, so that signs andsymptoms are limited, making site identifi-cation difficult. Repeated examinationsfocusing on the common sites will oftendefine the site over the course of a fewdays.

• Some pathogens can cause sepsis and deathquickly – notably Gram-negative bacilli andstreptococci – necessitating prompt initia-tion of therapy with a regimen designed to‘cover’ the most likely organisms.

• Selection of a regimen should take intoaccount the current antibiotic susceptibilitypattern at the hospital/oncology center.

• Monotherapy with a variety of �-lactams iseffective for most patients.

• Subgroups of patients with lower risk canbe identified, thus allowing considerationof outpatient/home therapy and/or oraltherapy.

• Oral therapy can be effective, but somepatients will be intolerant owing to nausea,vomiting, or possibly diarrhea; closefollow-up is required to be certain that theoral therapy is being ingested adequately.

• The spectrum of pathogens has shiftedincreasingly toward Gram-positive cocci,especially coagulase-negative staphylococciand viridans streptococci. In general, van-comycin is required for S. epidermidis infec-

tions, but can be withheld until culturereports confirm staphylococcal presence.

• The most important factor in response,other than the immediate initiation of theproper regimen, is return of circulatingneutrophils. Neutrophil transfusions foraplastic patients can be useful, but it is diffi-cult to obtain adequate numbers, and thoseat greatest need are often alloimmunized.Colony-stimulating factors (e.g. granulo-cyte or granulocyte–macrophage colony-stimulating factors: G-CSF and GM-CSF)may be helpful in assisting a more rapidreturn of bone marrow function.

These improvements mean that mostpatients will have rapid resolution of fever andinfection. Unfortunately, one subgroup con-tinues to have a dismal prognosis – namely,those patients with an aplastic marrow and aGram-negative bacteremia. Response rates tohighly active antibiotic(s) are poor at best. Inlarge studies, these patients represent about 5%of the total that develop fever during neutrope-nia. The absolute numbers of patients are low –but so is their survival.

CURRENT RECOMMENDATIONS

Monotherapy with drugs such as ceftazidime,cefepime, and imipenem is probably adequatefor most patients who develop fever yet haveonly moderate degrees of neutropenia. Thesepatients rarely have Gram-negative-rod bac-teremia, and their prognosis is generally good.For those patients who have profound neu-tropenia (<100/µl) and an aplastic bone mar-row, one might consider a combination ofdrugs such as a �-lactam along with an amino-glycoside. Neutrophil transfusions are rarelyutilized today. G-CSF or GM-CSF is an appro-priate adjunct for those patients with profound,persistent neutropenia in whom Gram-negativebacteremia is either proven or highly suspected.


REFERENCES

1. Joshi JH, Schimpff SC, Infections in the compro-mised host. In: Principles and Practice of InfectiousDiseases, 2nd edn (Mandel GL, Douglas RG Jr,Bennett JE, eds). New York: Wiley, 1985: 1644–9.

2. Bodey GP, Buckley M, Sathe YS et al,Quantitative relationships between circulatingleukocytes and infection in patients with acuteleukemia. Ann Intern Med 1966; 64: 328–40.

3. Schimpff SC, Young VM, Greene WH et al, Originof infection in acute nonlymphocytic leukemia:significance of hospital acquisition of potentialpathogens. Ann Intern Med 1972; 77: 707–14.

4. Williams DM, Krick JA, Remington JS,Pulmonary infection in the compromised host.Am Rev Respir Dis 1976; 114: 359.

5. Schimpff SC, Infection in patients with cancer:overview and epidemiology. In: ComprehensiveTextbook of Oncology, 2nd edn (Moossa AR,Schimpff SC, Robson MC, eds). Baltimore:Williams & Wilkins, 1991: 1720–32.

6. Schimpff SC, Wiernik PH, Block JB, Rectalabscesses in cancer patients. Lancet 1972; ii:844–7.

7. Young LS, Nosocomial infections in theimmunocompromised adult. Am J Med 1981; 70:394.

8. Johanson WG, Pierce AK, Sanford JP, Changingpharyngeal bacterial flora of hospitalizedpatients: emergence of Gram-negative bacilli. NEngl J Med 1969; 281: 1137.

9. van der Waaij D, Berghuis-de Vries JM,Lekkerkerk JEC, Colonization resistance of thedigestive tract of mice during systemic antibiotictreatment. J Hyg 1972; 70: 605.

10. Aisner J, Murillo J, Schimpff SC, Steere AC,Invasive aspergillosis in acute leukemia: correla-tion with nose cultures and antibiotic usage. AnnIntern Med 1979; 90: 4–9.

11. EORTC International Antimicrobial TherapyCooperative Group, Three antibiotic regimens inthe treatment of infection in febrile granulocy-topenic patients with cancer. J Infect Dis 1978;137: 14–29.

12. Sickles EA, Greene WH, Wiernik PH, Clinicalpresentation in granulocytopenic patients. ArchIntern Med 1975; 135: 715–19.

13. Pizzo PA, Ladisch S, Witebsky FG, �-hemolyticstreptococci: clinical significance in the cancerpatient. Med Pediatr Oncol 1978; 4: 367–70.

14. Kern W, Kurrle E, Vanek E, High risk of strepto-coccal septicemia after high dose cytosine arabi-noside treatment for acute myelogenousleukemia. Klin Wochenschr 1987; 65: 773–80.

15. Sotiropoulos SV, Jackson MA, Woods GM et al,�-streptococcal septicemia in leukemic childrentreated with continuous or large dosage inter-mittent cytosine arabinoside. Pediatr Infect Dis1989; 8: 755–8.

16. DuPont HL, Spink WW, Infections due to Gram-negative organisms: an analysis of 860 patientswith bacteremia at the University of MinnesotaMedical Center, 1958–1966. Medicine 1969; 48:307–32.

17. McCabe WR, Jackson GG, Gram-negative bac-teremia. I. Etiology and ecology. Arch Intern Med1962; 110: 847–55.

18. Schimpff SC, Greene WH, Young VM, WiernikPH, Significance of Pseudomonas aeruginosa in thepatient with leukemia or lymphoma. J Infect Dis1974; 130(Suppl): S24–31.

19. Schimpff SC, Satterlee W, Young VM, Serpick A,Empiric therapy with carbenicillin and gentam-icin for febrile patients with cancer and granulo-cytopenia. N Engl J Med 1971; 284: 1061–5.

20. Greisman SE, DuBuy JB, Woodward CL,Experimental Gram-negative bacterial sepsis:prevention of mortality not preventable byantibiotics alone. Infect Immun 1979; 25: 538–57.

21. Bodey GP, Jadeja L, Elting L, Pseudomonas bac-teremia: retrospective analysis of 410 episodes.Arch Intern Med 1985; 145: 1621–9.

22. Bodey GP, Elting L, Kassamali H, Lim BP,Escherichia coli bacteremia in cancer patients. AmJ Med 1986; 81 (Suppl 1A): 85–95.

23. Pizzo P, Hathorn J, Hiemenz J et al, A random-ized trial comparing ceftazidime alone withcombination antibiotic therapy in cancer patientswith fever and neutropenia. N Engl J Med 1986;315: 552–8.

24 EORTC International Antimicrobial TherapyCooperative Group, Ceftazidime combined witha short or long course of amikacin for empiricaltherapy of Gram-negative bacteremia in cancerpatients with granulocytopenia. N Engl J Med1987; 317: 1692–8.

25. de Jongh CA, Joshi JH, Newman KA et al,Antibiotic synergism and response in Gram-negative bacteremia in granulocytopenic cancerpatients. Am J Med 1986; 80: 96–100.

26. Sculier JP, Klastersky J, Significance of serum


bactericidal activity in Gram-negative bacillarybacteremia in patients with and without granu-locytopenia. Am J Med 1984; 76: 429–35.

27. Klastersky J, Daneau D, Swings G, Weerts D,Antibacterial activity in serum and urine as atherapeutic guide in bacterial infections. J InfectDis 1974; 129: 187–93.

28. Anderson ET, Young LS, Hewitt WL,Antimicrobial synergism in the therapy of Gram-negative rod bacteremia. Chemotherapy 1978; 24:45.

29. Klastersky J, Combinations of antibiotics fortherapy of severe infections in cancer patients.Eur J Cancer 1979; 15: 3–13.

30. Standiford HC, Drusano GL, Fitzgerald B et al,Bactericidal activity of ceftazidime in serumcompared with that of ticarcillin combined withamikacin. Antimicrob Agents Chemother 1984; 26:339–42.

31. Wade JC, Standiford HC, Drusano GL et al,Potential of imipenem as single-agent empiricantibiotic therapy of febrile neutropenic patientswith cancer. Am J Med 1985; 78(Suppl A): 54–64.

32. Johnson DE, Calia FM, Snyder MJ et al,Imipenem therapy of Pseudomonas aeruginosabacteremia in neutropenic rats. J AntimicrobChemother 1983; 12: 89–96.

33. Gerber AU, Wiprachtiger P, Stettler-Spichiger U,Lebek G, Constant infusions versus intermittentdoses of gentamicin against Pseudomonas aerugi-nosa in vitro. J Infect Dis 1982; 145: 554–60.

34. Gerber AU, Vastola AP, Brandel J, Craig WA,Selection of aminoglycoside-resistant variants ofPseudomonas aeruginosa in an in vivo model. JInfect Dis 1982; 146: 691–7.

35. Gerber AU, Craig WA, Aminoglycoside-selectedsubpopulations of Pseudomonas aeruginosa. J LabClin Med 1982; 100: 671–81.

36. Pizzo PA, Robichaud KJ, Wesley R et al, Fever inthe pediatric and young adult patient withcancer. A prospective study of 1001 episodes.Medicine 1982; 61: 153.

37. Pizzo PA, Robichaud KJ, Gill FA et al, Durationof empiric antibiotic therapy in granulocytopenicpatients with cancer. Am J Med 1979; 67: 194–200.

38. EORTC International Antimicrobial TherapyCooperative Group, Empiric antifungal therapyin febrile granulocytopenic patients. Am J Med1989; 86: 668–72.

39. Maschmeyer G, Noskin GA, Ribaud P,Sepkowitz KA, Changing patterns of infections

and antimicrobial susceptibilities. Oncology 2000;14(Suppl 6): 9–16.

40. Remington JS, Schimpff SC, Please don’t eat thesalads. N Engl J Med 1981; 304: 433–5.

41. Schimpff SC, Prevention of infection in cancerpatients. In: Supportive Care in Cancer. AHandbook for Oncologists, 2nd edn (Klastersky J,Schimpff SC, Senn HJ, eds). New York: MarcelDekker, 1999: 129–54.

42. Tenney JH, Moody MR, Newman KA et al,Adherent microorganisms on lumenal surfacesof long-term intravenous catheters: importanceof Staphylococcus epidermidis in patients withcancer. Arch Intern Med 1986; 146: 1949–54.

43. Shenep JL, Hughes WT, Roberson PK et al,Vancomycin, ticarcillin, and amikacin comparedwith ticarcillin–clavulanate and amikacin inthe empirical treatment of febrile, neutropenicchildren with cancer. N Engl J Med 1988; 319:1053–8.

44. European Organization for Research andTreatment of Cancer (EORTC) InternationalAntimicrobial Therapy Cooperative Group andthe National Cancer Institute of Canada –Clinical Trials Group, Vancomycin added toempirical combination antibiotic therapy forfever in granulocytopenic cancer patients. J InfectDis 1991; 163: 951–8.

45. Ramphal R, Bolger M, Oblon DJ et al,Vancomycin is not an essential component of theinitial empiric treatment regimen for febrile neu-tropenic patients receiving ceftazidime: a ran-domized prospective study. Antimicrob AgentsChemother 1992; 36: 1062–7.

46. Dompeling EC, Donnelly JP, Deresinksi SC et al,Early identification of neutropenic patients athigh risk of Gram-positive bacteraemia and theimpact of empirical administration of van-comycin. Eur J Cancer 1996; 32A: 1332–9.

47. Carratalá J, Alcaide F, Fernandez-Sevilla A et al,Bacteremia due to viridans streptococci that arehighly resistant to penicillin: increase amongneutropenic patients with cancer. Clin Infect Dis1995; 20: 1169–73.

48. Rubenstein EB, Rolston K, Benjamin RS et al,Outpatient treatment of febrile episodes in low-risk neutropenic patients with cancer. Cancer1993; 71: 3640–6.

49. Kern WV, Cometta A, de Bock R et al, Oral ver-sus intravenous empirical antimicrobial therapyfor fever in patients with granulocytopenia who


are receiving cancer chemotherapy. N Engl J Med1999; 341: 312–18.

50. Lucas KG, Brown AE, Armstrong D et al, Theidentification of febrile, neutropenic childrenwith neoplastic disease at low risk for bac-teremia and complications of sepsis. Cancer 1996;77: 791–8.

51. Freifeld A, Marchigiani D, Walsh T et al, Adouble-blind comparison of empirical oral andintravenous antibiotic therapy for low-riskfebrile patients with neutropenia during cancerchemotherapy. N Engl J Med 1999; 341: 305–11.


2Overview of pharmacokinetic andpharmacodynamic principles of anti-infective dosing in the neutropenic patientRussell E Lewis, Randall A Prince

INTRODUCTION

Effective antimicrobial therapy is dependentupon a number of factors, many of which arebeyond the direct control of the clinician (seeFigure 2.1). Antimicrobial selection and dosing,however, are two variables of drug therapy thatcan be controlled. Since anti-infective therapyplays such a critical role in successful outcomesfor the neutropenic patient, optimization ofdrug regimen design is essential. This chapterwill focus on pharmacokinetic and pharmaco-dynamic principles of antimicrobial dosing inthe neutropenic host. Special attention will bedevoted to describing differences between vari-ous classes of antimicrobials, as well as specialpharmacokinetic issues in the care of neu-tropenic patients.

RATIONALE OF ANTIMICROBIAL DOSING

Three components of antimicrobial pharmacol-ogy are of special interest in developing effect-ive dosage regimens:

(i) the potency of the antimicrobial against thepathogen(s) in question;

Host factorsUnderlying disease

ImmunosuppressionMucositis

Graft-versus-host diseaseInvasive devices

Pathogen factorsVirulence

ResistanceToxin production

Host tissue damage

Drug factorsPotency (MIC, MBC)PharmacodynamicsPharmacokinetics

Figure 2.1 Interrelationship of host, pathogen, anddrug factors that influence outcome in anti-infectivetherapy.

(ii) the concentration achieved by the antimi-crobial in the serum and at the site of infec-tion (pharmacokinetics);

(iii) the relationship of drug concentrations tothe rate and extent of pathogen killing(pharmacodynamics).

Antimicrobial potency is typically defined bysusceptibility-testing endpoints such as theminimum inhibitory concentration (MIC) andthe minimum bactericidal concentration (MBC).Despite its many limitations, MIC testing pro-vides a reasonable measurement of drug activ-ity that can be easily related to the drugconcentrations achieved in the body. Historic-ally, the goal of most antimicrobial dosing strat-egies has been to maintain drug concentrationsabove the MICs of common pathogens inserum/tissues for the extent of the dosing inter-val. Although this dosing strategy may produceacceptable antibacterial efficacy with somecompounds, it does not take into account fun-damental pharmacodynamic differences betweenvarious antimicrobial classes.

The study of antimicrobial pharmacodynam-ics has provided new insight into the relation-ship of drug concentrations to bacterial orfungal killing.1 Antimicrobial concentration–killing relationships generally follow one of twopatterns (see Figure 2.2). The first pattern ischaracterized by concentration-dependentkilling over a broad range of clinically achiev-able concentrations (drug B). That is, the higherthe drug concentration, the greater the rate andextent of bacterial or fungal killing. The secondpattern, however, is characterized by minimalconcentration-dependent killing over the rangeof clinically achievable levels (drug A).Generally, drug concentrations greater thanfour times the MIC do not enhance the rate orextent of activity.2–4 Since the extent of killingnoted with this second pattern is largely depen-dent on how long drug concentrations remainnear the MIC, it is also termed time-dependentkilling pharmacodynamics.

Some antimicrobial agents may produce per-sistent or prolonged inhibitory effects even asdrug concentrations fall below the MIC.5 Thephenomena related to the sub-MIC concentra-tions, including the post-antibiotic effect (PAE),sub-MIC effect (SME), and the post-antibioticleukocyte enhancement (PALE), are increas-ingly incorporated into the development of dos-ing regimen strategies. Some investigators have


100

80

60

40

20

0

% o

bser

ved

effe

ct

Drug ADrug B

Range of clinically achievedconcentrations

log10 (multiple of MIC)0.1 1 10 100

Emax

drug AEmax

drug B

Figure 2.2 Comparison of concentration-dependentand concentration-independent pharmacodynamics.

even recommended that the dosing interval ofan antibiotic should be equal to the time forwhich drug concentrations remain above theMIC plus the duration of the PAE.6 Of the threeaforementioned mechanisms, the PAE has beenstudied the most. The PAE is thought to be dueto a lag time in the disassociation of the antimi-crobial from the cellular receptors in the organ-ism or the recovery of the organism fromcellular injury.1 The significance of these phe-nomena in the neutropenic population,however, remains to be determined. With cer-tain antimicrobials, the PAE is often more pro-longed the higher the concentration ofantibiotic exposure or the greater the durationof exposure.5 This has led some investigators topropose a third pattern of pharmacodynamicactivity where concentration-independentkilling predominates initially, but persistenteffects or the PAE are concentration-dependent.1,6,7 This ‘combination’ pharmacody-namic picture may be seen with newermacrolides such as azithromycin.8

Defining the killing characteristics of an anti-biotic or antifungal is essential for optimizingdosing. For agents that exhibit concentration-dependent pharmacodynamics, dosing regi-

mens should maximize peak concentrations(Cmax) or overall exposure to the drug (areaunder the curve, AUC). In some cases, largerinfrequently administered doses may be neces-sary to achieve sufficient peak concentrations.For example, in extended-interval dosing ofaminoglycosides, high peak concentrations(>8–10 times the MIC) have been shown toresult in more rapid and extensive bacterialkilling, and may reduce the probability of resis-tance.9–11 In neutropenic patients, the benefits ofincreasing antibiotic dosages for concentration-dependent agents generally outweigh theincreased risk of adverse drug effects. Forantimicrobials with concentration-independentkilling pharmacodynamics, dosing regimensshould optimize the time for which concentra-tions remain above the MIC. Escalating antimi-crobial dosages for these antibiotics per se doesnot significantly improve antimicrobial killing.

Increasingly, pharmacokinetic/pharmacody-

PHARMACOKINETIC AND PHARMACODYNAMIC PRINCIPLES OF ANTI-INFECTIVE DOSING 29

Table 2.1 Pharmacokinetic and pharmacodynamic parameters correlating with efficacy ofantimicrobial therapy for various anti-infective classes

Pharmacokinetic : pharmacodynamic parameters Refs

Concentration-dependent killing agentsAminoglycosides Peak : MIC, AUC0–24 : MIC 9–11, 13Fluoroquinolones Peak : MIC, AUC0–24 : MIC 14–16Metronidazole Peak : MIC 56, 57Amphotericin B Peak : MIC 58–60

Time-dependent killing agents�-lactams Time > MIC 17–21Macrolides Time > MIC, AUC0-24 : MIC 1, 7, 8Vancomycin AUC0–24 : MIC 1Lincosamides Time > MIC 1Tetracyclines AUC0–24 : MIC 1Azoles Time > MIC, AUG0–24 : MIC 58–61Oxazolidinones AUC0–24 : MIC 62–64Streptogramins AUC0–24 : MIC 8, 65, 66

AUC:MIC

MIC

Antib

iotic

conc

entr

atio

n

Time (hours)0

% timeMIC

Peak:MIC

Figure 2.3 Pharmacokinetic : Pharmacodynamicparameters of interest in antimicrobial therapy.

namic relationships are being used to comparethe activity of antimicrobial agents (see Table2.1 and Figure 2.3). This approach has severalinherent advantages over comparing drugs onthe basis of MIC data alone. First, by dividing

the serum pharmacokinetic parameter value bythe MIC, drugs with dissimilar potency andpharmacokinetics can be directly compared. Forexample, if a new fluoroquinolone that wasfourfold more potent against Pseudomonasaeruginosa than an older and established fluoro-quinolone were introduced on the market, itwould appear by comparing MICs alone to be asuperior drug for Pseudomonas infections.However, if the drug concentrations achievedwith the new agent were one-sixth thatachieved with ciprofloxacin, it may actually bea less effective agent. By creating a ratio of thepharmacokinetic and pharmacodynamic para-meters, disparities in potency and pharmacoki-netics are normalized, thus allowing a moredirect comparison of the agents.12

Pharmacokinetic : pharmacodynamic (PK : PD)ratios are very useful as markers or ‘break-points’ of drug activity.12 It has been shownfor aminoglycosides, for example, that aCmax : MIC > 8–10 is associated with maximalclinical efficacy against Gram-negative organ-isms.9,13 For quinolones, such as ciprofloxacin, a serum Cmax : MIC > 12 or a serumAUC0–24 : MIC > 125 (put another way, averag-ing concentrations 4–6 times the MIC over thedosing interval) have been associated withmaximal bacteriological and clinical outcomesfor Gram-negative infections.14–16 For drugswith more concentration-independent killingcharacteristics such as �-lactams, maintenanceof drug concentrations above the MIC for atleast 50% of the dosing interval has been associ-ated with bacteriological efficacy.17–21

Once PK : PD breakpoints have beendescribed for an antimicrobial class, it is oftenpossible to compare agents with disparatepotency, pharmacokinetics, and pharmacody-namics. Table 2.2 shows PK : PD breakpointsachieved with various agents that often consti-tute empiric therapy to treat Pseudomonas spp.For most anti-pseudomonal �-lactams pre-sented in the table, the critical breakpoint ofsurpassing the MIC for greater than 50% of thedosing interval is easily achieved at standarddosages. However, if agents are compared on

the basis of the more conservative measure-ment of the MIC (MIC90), one sees that many ofthe drugs fall on the borderline of meeting thePK : PD threshold of 50%. For some �-lactams(e.g. cefepime and piperacillin/tazobactam),activity can be improved by using higherdosages and shorter dosing intervals.

Similarly, those agents with concentration-dependent killing activity can be compared inthe same fashion. Quinolones such asciproflaxacin and levofloxacin both achieve aserum Cmax : MIC > 12 if MIC50 data are con-sidered. If the MIC90 data are assessed,however, both agents fall well below theCmax : MIC threshold of 12 and AUC : MIC of125, despite ciprofloxacin appearing initially tobe the more effective agent. This same strategyof comparing antibiotics on the basis of PK : PDparameters can be individualized using institu-tional MIC data and dosing practices todevelop specific PK : PD antibiograms.Clinicians can then identify both antimicrobialagents and dosing strategies that would bemore effective for empiric therapy.

It is important to recognize, however, thatPK : PD ratios serve only as general markers ofantimicrobial activity and have many inherentlimitations. The vast majority of PK : PD dataderived for antibacterial and antifungal agentshave come from in vitro pharmacodynamicmodels and animal studies – not from humantrials. Although these PK : PD breakpoints aregenerally conserved among animal species (e.g.a quinolone Cmax : MIC ratio that results in max-imal bacteriologic efficacy is generally the samein mice and humans), PK : PD data from animalmodels may not be completely applicable tohumans. Also, it must be emphasized that thevast majority of PK : PD experiments are basedon serum/plasma data, which may not alwaysreflect the conditions at the site of infection. It isknown, for example, that tissue concentrationsachieved with fluoroquinolones are oftenhigher than concurrent serum concentrations.Therefore, the PK : PD data in Table 2.2 mayunderestimate activity at the site of infection.Most studies to date, however, have noted that


Tabl

e 2.2

Com

pariso

n of

ant

i-pse

udom

onal

PK

:PD

dat

a fo

r va

riou

s an

tim

icro

bial

s us

ed in

em

piric

ther

apya

Dru

g an

d do

seM

IC5

0:M

IC9

0Ser

umSer

umSer

um(µ

g/m

l)%

tim

e>

MIC

50/M

IC90

Cm

ax:M

IC50/M

IC90

AU

C0–2

4:M

IC50/

MIC

90

Imip

enem

2/>

8500

mg

q6h

90/

50

NA

NA

1000

mg

q6h

90/

60

Mer

open

em0.2

5/2

1000

mg

q8h

100/<

60

NA

NA

Cef

epim

e2/1

61

gq1

2h

73/3

5N

AN

A2

gq8

h100/7

5

Cef

tazi

dim

e2/1

61

gq8

h100/3

5N

AN

A2

gq8

h100/3

5

Pipe

raci

llin/

tazo

bact

am4/6

43.3

75

gq6

h75/<

25

NA

NA

4.5

gq4

h100/4

0

Cip

roflo

xaci

n0.1

2/2

.0400

mg

q12h

NA

38/2

.4106/6

.35

400

mg

q8h

38/2

.4275/1

6.5

Levo

floxa

cin

0.5

/>4.0

500

mg

q24h

NA

12.4

/1.5

110/3

aAl

l PK

dat

a ha

ve b

een

extr

acte

d fr

om r

epre

sent

ed p

acka

ge in

sets

; M

IC d

ata

have

bee

n ge

nera

ted

from

ref

eren

ce 6

8. N

A, n

ot a

pplic

able

.

serum/plasma drug concentrations (includingquinolone data) correlate best with clinicalresponse/efficacy. Finally, it is important toremember there is a paucity of data that havevalidated PK : PD ‘breakpoints’ with clinicaloutcome in the neutropenic patient.

SPECIAL PHARMACOKINETICCONSIDERATIONS IN THE NEUTROPENICHOST

Tissue penetration

Besides the spectrum and potency of an antimi-crobial agent, the penetration of an agent intoinfected tissue is perhaps the most importantdeterminant of antimicrobial efficacy.22–24

Numerous factors can affect the distribution ofthe drug from the bloodstream to the tissue,including the ionic charge of the drug molecule,lipophilicity, plasma protein binding, tissuebinding, and permeability barriers (e.g. centralnervous system (CNS) and aqueous humor).The elimination rate of the drug from the bodycan also affect the distribution/penetration ofthe drug into infected tissues. For most drugs,however, distribution occurs more rapidly thanelimination. One method used to estimate dis-tribution of different drugs is to calculate theapparent volume of distribution (Vd) of theagent. As can be seen in Table 2.3, virtually allantimicrobials have ‘volume’ values that sug-gest a distribution outside of plasma (volumeapproximately 3 l or 0.04 l/kg) and into tissues.In fact, many agents have a volume of distribu-tion value similar or greater than the total bodywater (0.65 l/kg). One must be cautious,however, in trying to utilize the Vd term to pre-dict and/or relate to specific anatomic sites andsites of drug accumulation/penetration in thebody.

For antimicrobials that are commonly uti-lized as empiric regimens for patients withfebrile neutropenia, general categorizations ofdrug penetration can be made (see Tables 2.3and 2.4). Aminoglycosides have poor-to-

moderate penetration in tissues, including thelung and CNS. These agents are likely to bemost effective for infections in the bloodstreamand urinary tract, and generally should not beemployed as monotherapy, particularly in theneutropenic patient.25,26 Anti-pseudomonalpenicillins, carbapenems, and cephalosporinsachieve moderate-to-good concentrations in thelung, tissues, and CNS (high dose), with theexception of the so-called first-generationcephalosporins (poor penetration in CNS). Inmost cases, the use of a �-lactam provides goodbaseline coverage into tissues throughout thebody. Vancomycin achieves moderate-to-goodconcentrations in the lung and tissues;however, higher dosages may be necessary toachieve even moderate CNS penetration. Ingeneral, fluoroquinolones and trimethoprim/sulfamethoxazole reach high concentrations in tissues, and both agents exhibit good pene-tration into the CNS. Amphotericin B and itslipid formulations exhibit moderate-to-goodtissue penetration, particularly in the lungs,spleen, kidney, and liver. However, CNSpenetration of amphotericin B is poor.27,28

Flucytosine, which possesses excellent CNSpenetration, should be used in combinationwith amphotericin B during initial inductiontherapy for cryptococcal meningitis and otherCNS fungal infections.29 Interestingly, liposo-mal amphotericin B (AmBisome) may exhibithigher CNS concentrations than conventionalamphotericin B.30–32 The azole, fluconazole, dis-tributes widely to most tissues and the CNS.33

Similarly, itraconazole distributes to the lungand other tissues, but does not penetrate theCNS as well as fluconazole.27,34 Lastly, mostantivirals distribute widely throughout thebody; however, only acyclovir and ganciclovirachieve clinically useful concentrations in theCNS.35–37

Renal and liver dysfunction

The kidney serves as the primary route of elimi-nation for the majority of antimicrobial agents


Tabl

e 2.3

Sel

ecte

d ph

arm

acok

inet

ic fea

ture

s of

ant

i-inf

ecti

ve a

gent

sa

Pha

rmac

okin

etic

pro

pert

ies

Mai

nten

ance

reg

imen

in r

enal

fai

lure

b

Glo

mer

ular

filt

rati

on r

ate

Dru

gO

ral d

ose

Bio

avai

labi

lity

Intr

aven

ous

Maj

orV

dH

alf-l

ife

(%)

dose

excr

etio

n(l

/kg

)(h

)50–8

0m

l/10–5

0m

l/<10

ml/

path

way

min

min

min

Ant

ibac

terial

s

Am

inog

lyco

side

s

Amik

acin

0.2

5–0

.5g

q8h

Ren

al0

.34

2TD

MTD

MTD

M

Gen

tam

icin

1.5

mg/

kgq8

hR

enal

0.2

42

TDM

TDM

TDM

Tobr

amyc

in1

.5m

g/kg

q8h

Ren

al0

.25

2.5

TDM

TDM

TDM

inha

latio

n (T

OB

I)0.3

gq1

2h

unne

cess

ary

unne

cess

ary

unne

cess

ary

�-la

ctam

s

Amox

icill

in/c

lavu

lana

te0.2

5–0

.5g

q8h

75

Ren

al0.2

6–0

.31

1.2

Usu

alq1

2–2

4h

q12–2

4h

Ampi

cilli

n/su

lbac

tam

1–2

gq6

hR

enal

0.2

6–0

.31

1U

sual

Usu

alU

sual

1–2

gq8

h1

–2g

q12

h

Aztr

eona

m1–2

gq6

hR

enal

0.1

8–0

.21.7

–21–2

gq8

–12h

1–2

gq1

2–1

8h

1–2

gq2

4h

Cef

epim

e0.5

–2g

q8–1

2h

Ren

al0.1

82.0

0.5

–2.0

gq1

2h

0.5

–1g

q24h

0.2

5–0

.5g

q24h

Cef

otax

ime

1–2

gq4

–8h

Ren

al0.2

81.5

Usu

al1–2

gq6

–12h

1–2

gq1

2h

Cef

tazi

dim

e1–2

gq8

–12h

Ren

al0.2

11.8

1–2

gq4

8h

1g

q12–2

4h

0.5

gq2

4–4

8h

Cef

tria

xone

0.5

–2g

q12–2

4h

Ren

al, G

ut0.1

58

Usu

alU

sual

Usu

al

Imip

enem

/cila

stat

in0.5

–1g

q6h

Ren

al0.3

4–0

.45

1.0

0.5

gq6

–8h

0.5

gq8

–12h

0.2

5–0

.5g

q12h

Mer

open

em1

gq8

hR

enal

0.2

71

.0U

sual

0.5

gq1

2h

0.5

gq2

4h

Naf

cilli

n0

.5–2

gq4

–6h

Hep

atic

0.1

80

.5U

sual

Usu

alU

sual

Oxa

cilli

n0

.5–1

gq6

h3

01

–3g

q6h

Ren

al0

.5U

sual

Usu

alU

sual

Peni

cilli

n G

, cr

ysta

lline

12–1

8M

Uqd

Ren

al0.5

–0.6

70.5

Usu

alU

sual

��us

ual d

ose

peni

cilli

n V

0.4

–0.8

MU

q6h

60–7

00.6

7

aAl

l dat

a ha

ve b

een

extr

acte

d fr

om r

epre

sent

ativ

e pa

ckag

e in

sert

s an

d re

fere

nce

67.

Con

tdb

TDM

, th

erap

eutic

dru

g m

onito

ring.

Tabl

e 2.3

Sel

ecte

d ph

arm

acok

inet

ic fea

ture

s of

ant

i-inf

ecti

ve a

gent

sa–

cont

d

Pha

rmac

okin

etic

pro

pert

ies

Mai

nten

ance

reg

imen

in r

enal

fai

lure

b

Glo

mer

ular

filt

rati

on r

ate

Dru

gO

ral d

ose

Bio

avai

labi

lity

Intr

aven

ous

Maj

orV

dH

alf-l

ife

(%)

dose

excr

etio

n(l

/kg

)(h

)50–8

0m

l/10–5

0m

l/<10

ml/

path

way

min

min

min

Pipe

raci

llin/

tazo

bact

am3/0

.375

gq4

–6h

Ren

al0.1

41.0

Usu

al2/0

.25

gq6

h2/0

.25

gq8

h

Tica

rcill

in/c

lavu

lana

te3

gq4

–6h

Ren

al0

.16

1.2

Usu

al2

–3g

q6–8

h2

gq1

2h

Fluo

roqu

inol

ones

Cip

roflo

xaci

n0.2

5–0

.75

g70

0.4

gq8

–12h

Ren

al1.2

44

Usu

al0.2

5–0

.5g

q12h

0.2

5–0

.5g

q24h

q8–1

2h

Gat

iflox

acin

0.4

gq2

4h

96

0.4

gq2

4h

Ren

al1

.36

7–8

Usu

al0

.4g

q24

–48h

0.4

gq4

8h

Levo

floxa

cin

0.5

gq2

4h

98

0.5

gq2

4h

Ren

al1

.20

7U

sual

25

0m

gq2

4h

250

mg

q48h

Mox

iflox

acin

0.4

gq2

4h

89

Hep

atic

1.2

61

0–1

4U

sual

Usu

alU

sual

Oflo

xaci

n0.2

–0.4

gq1

2h

98

0.2

–0.4

gq1

2h

Ren

al1.1

47

Usu

al0.2

–0.4

gq2

4h

0.1

–0.2

gq2

4h

Linc

osam

ides

Clin

dam

ycin

0.1

5–0

.3g

q12h

0.9

00.3

–0.9

gq6

–8h

Hep

atic

0.8

52.4

Usu

alU

sual

Usu

al

Mac

rolid

es

Azith

rom

ycin

0.2

5g

q24h

37

0.5

gq2

4h

Hep

atic

6.6

12/6

8h

Usu

alN

o da

taN

o da

ta

Cla

rithr

omyc

in0.2

5–0

.5g

q12h

50

Hep

atic

/ren

al1.7

85–7

Usu

alU

sual

0.2

5–0

.5g

q24h

Eryt

hrom

ycin

(s)

0.2

5–0

.5g

q6h

18–4

50.5

–1g

q6h

Hep

atic

0.7

–1.5

2–4

Usu

alU

sual

Usu

al

Oxa

zolid

onon

es

Line

zolid

0.3

–0.6

gbi

d100

0.6

gq1

2h

Hep

atic

0.5

70.6

26

Usu

alU

sual

Usu

al

Str

epto

gram

ins

Qui

nupr

istin

/dal

fopr

istin

7.5

mg/

kgq8

–12h

Hep

atic

1.5

1.5

Usu

alU

sual

Usu

al

Vanc

omyc

in0.1

25–0

.25

gq8

h<1

0.5

–1g

q12h

Ren

al0.4

56–8

1g

q24h

1g

q3–1

0d

TDM

1g

q3–1

0d

TDM

Pha

rmac

okin

etic

pro

pert

ies

Mai

nten

ance

reg

imen

in r

enal

fai

lure

b

Glo

mer

ular

filt

rati

on r

ate

Dru

gO

ral d

ose

Bio

avai

labi

lity

Intr

aven

ous

Maj

orV

dH

alf-l

ife

(%)

dose

excr

etio

n(l

/kg

)(h

)50–8

0m

l/10–5

0m

l/<10

ml/

path

way

min

min

min

Sul

fona

mid

es

Sul

fisox

azol

e2–4

gq6

h70–8

0R

enal

2.1

43–7

Usu

al1

gq8

–12h

1g

q12–2

4h

Trim

etho

prim

/2–4

tab

s q2

4h

or90–1

00

3–5

mg/

kgq6

–12h

Ren

alS

: 0.2

2S

: 7–1

2U

sual

0.1

gq2

4h

Avoi

d

sulfa

met

hoxa

zole

1–2

DS

tab

s q2

4h

T: 2

.28

T: 2

4

Trim

etho

prim

0.1

gq1

2h

90

Ren

al2

.28

24

Usu

al0

.1q2

4h

Avoi

d

Tetr

acyc

lines

Dox

ycyc

line

0.1

gq1

2h

93

0.1

gq1

2h

Ren

al/g

ut0.7

14–2

5U

sual

Usu

alU

sual

Min

ocyc

line

0.1

–0.2

gq1

2h

95

0.1

gq1

2h

Hep

atic

1.4

216

Usu

alU

sual

Usu

al

Tetr

acyc

line

0.2

5–0

.5g

q6h

80

Ren

al1.7

88

Usu

alU

se d

oxyc

yclin

e

Oth

ers

Atov

aquo

ne0.7

5g

q12h

25–5

0G

ut0.6

70

Usu

alU

sual

No

data

Dap

sone

0.0

5–0

.1g

q24h

>85

Hep

atic

1.5

30

Usu

alU

sual

No

data

Met

roni

dazo

le0.2

5–0

.75

gq8

h90

0.5

gq6

–8h

Hep

atic

0.3

56–1

2U

sual

Usu

alU

sual

Pent

amid

ine

4m

g/kg

q24h

4m

g/kg

q24h

Non

-rena

l0.5

6–8

Usu

al4

mg/

kgq2

4–3

6h

4m

g/kg

q48h

Poly

myx

in B

7500–1

2500

U/

Ren

al6

7500–1

2500

U/

5625–1

2500

U/

3750–6

250

U/

kg/d

ayq1

2h

kg/d

ayq1

2h

kg/d

ayq1

2h

kg/d

ayq1

2h

Pyrim

etha

min

e25–7

5m

gq2

4h

>70

Hep

atic

3100

Usu

alU

sual

Usu

al

Rifa

butin

0.3

gq2

4h

53

Hep

atic

/fec

es45

Usu

alU

sual

Usu

al

Rifa

mpi

n0.6

g/da

y100

0.6

g/da

yH

epat

ic1.2

52–5

Usu

alU

sual

Usu

al

Trim

etre

xate

45

mg/

m2

q24h

Hep

atic

/gut

0.5

–211

Ant

ifung

als

Amph

oter

icin

B100

mg

q6h

<1

0.5

–1.5

mg/

kgq2

4h

Non

-rena

l4

Ser

um: 24;

Usu

alU

sual

Usu

al

deox

ycho

late

Amph

oter

icin

B3—

5m

g/kg

q24h

Non

-rena

l0.2

2S

erum

: 8.6

;U

sual

Usu

alU

sual

lipos

omal

(Am

Bis

ome)

Con

td

Tabl

e 2.3

Sel

ecte

d ph

arm

acok

inet

ic fea

ture

s of

ant

i-inf

ecti

ve a

gent

sa–

cont

d

Pha

rmac

okin

etic

pro

pert

ies

Mai

nten

ance

reg

imen

in r

enal

fai

lure

b

Glo

mer

ular

filt

rati

on r

ate

Dru

gO

ral d

ose

Bio

avai

labi

lity

Intr

aven

ous

Maj

orV

dH

alf-l

ife

(%)

dose

excr

etio

n(l

/kg

)(h

)50–8

0m

l/10–5

0m

l/<10

ml/

path

way

min

min

min

Amph

oter

icin

B5

mg/

kg24h

Non

-rena

l131

Ser

um: 10;

Usu

alU

sual

Usu

al

lipid

(AB

LC)

Amph

oter

icin

B5

mg/

kgq2

4h

Non

-rena

l4.1

Ser

um: 28.2

;U

sual

Usu

alU

sual

lipid

(AB

CD

)

Fluc

onaz

ole

0.1

5–0

.8g

q24h

0.9

00.4

–0.8

gq2

4h

Ren

al0.2

524

Usu

al��

dose

��do

se

Fluc

ytos

ine

37

mg/

kgq2

4h

>0.9

0R

enal

0.6

82.5

–6U

sual

37

mg/

kgq1

2–2

4h

TDM

Itrac

onaz

ole

0.2

gq1

2–2

4h

Hep

atic

11

30–4

0U

sual

Avoi

d i.v

.Av

oid

i.v.

Cap

sule

s0.1

–0.4

gq2

4h

10–4

5

Sol

utio

n0.2

gq1

2–2

4h

60–8

5

Ket

ocon

azol

e0.2

–0.4

gq1

2–2

4h

40–7

0

0.2

–0.4

gq1

2–2

4h

Hep

atic

28

Usu

alU

sual

Usu

al

(pH

-dep

ende

nt)

Nys

tatin

00

.4–1

MU

q6h

Usu

alU

sual

Usu

al

Terb

inafi

ne0.2

5g

q12–2

4h

>70

2.8

536

Usu

alN

o da

ta,

No

data

cons

ider

��do

se

Ant

iviral

s

Acyc

lovi

r200–8

00

mg

0.1

5–0

.35–1

0m

g/kg

q8h

Ren

al0.5

–0.8

2–2

.5U

sual

Usu

al200

mg

q12h

5�

day

Aman

tadi

ne0.1

g bi

d0.8

0R

enal

3–8

15–2

00.1

–0.1

5g

q24h

0.1

–0.2

g0.1

–0.2

gqw

eek

2–3

�w

eek

Cid

ofov

ir5

mg/

kgq2

wee

kR

enal

0.4

–0.5

17–6

5U

sual

Avoi

dAv

oid

Fam

cicl

ovir

0.1

25

g bi

d0

.77

Ren

al1

–22

.5U

sual

0.1

25

gq2

4h

0.1

25

gq4

8h

50

0m

g tid

0.5

gq1

2–2

4h

0.2

5g

q48h

Fosc

arne

tR

enal

0.4

–0.5

34

0–5

0m

g/kg

q8h

20

–30

mg/

kgq8

hAv

oid

Indu

ctio

n60

mg/

kgq8

h

Mai

nten

ance

90–1

20

mg/

kgqd

Pha

rmac

okin

etic

pro

pert

ies

Mai

nten

ance

reg

imen

in r

enal

fai

lure

b

Glo

mer

ular

filt

rati

on r

ate

Dru

gO

ral d

ose

Bio

avai

labi

lity

Intr

aven

ous

Maj

orV

dH

alf-l

ife

(%)

dose

excr

etio

n(l

/kg

)(h

)50–8

0m

l/10–5

0m

l/<10

ml/

path

way

min

min

min

Gan

cicl

ovir

0.7

4

Indu

ctio

n5

mg/

kgbi

di.v

.: R

enal

i.v.: 3

2.5

mg/

kgq1

2h

2.5

mg/

kgq2

4h

1.2

5m

g/kg

q24h

Mai

nten

ance

1000

mg

tid0.0

5O

ral:

GI

Ora

l: 2–7

500

mg

q8h

500

mg

q24h

500

µg3

�w

eek

Ose

ltam

ivir

75

mg

bid

0.7

57

5m

g bi

dR

enal

0.3

56

–10

Usu

al7

5m

gq2

4h

Avoi

d

Rib

aviri

n600–1

800

mg/

day

0.6

4H

epat

ic300

Usu

al?

Dos

e re

duct

ion

Avoi

d

(aer

osol

)

Rim

anta

dine

0.1

gbi

d0.9

6H

epat

ic24–3

0U

sual

Usu

al0.1

gq2

4h

Vala

cycl

ovir

1g

tidR

enal

3U

sual

1g

q12–2

4h

0.5

gq2

4h

Zana

miv

ir10

mg

bid

0.0

4–0

.17

Ren

al0.3

02.5

–5U

sual

No

data

No

data

(inha

latio

n)

(see Table 2.3). Therefore, decreases in kidneyfunction secondary to drug therapy (aminogly-cosides, amphotericin B, cyclosporin, antineo-plastic agents) or underlying disease states(sepsis, hypotension) can have a profoundinfluence on the overall clearance of antimicro-bial agents. In particular, the neutropenic popu-lation is affected since they are often receivinghigher dosages of anti-infectives and are atgreater risk for the development of nephrotoxi-city. For most antibacterial agents, dosingadjustments are not necessary until the creati-nine clearance is below 50 ml/min/70 kg (seeTable 2.3). However, for antibiotics that are pre-dominantly eliminated renally, major dose regi-men adjustments (e.g. one-half reductions) maybe necessary when the renal function is equal toor greater than half-normal. Specifically, inpatients with a creatinine clearance of less than30 ml/min/70 kg, patients should be dosedaccording to specific guidelines for the drug or,in the case of aminoglycosides, vancomycin,and flucytosine, on the basis of serum drug con-centration monitoring (see Table 2.3).

Dosage adjustments in patients with clini-cally significant hepatic dysfunction are lessclear, because liver disease is associated withchanges in multiple factors that can affect drugclearance (e.g. protein binding and Vd).38

Additionally, no clear clinical marker has beencorrelated with changes in drug clearance inhepatic dysfunction. The absence of a usefulmarker precludes specific dosage adjustmentcalculations. Generally, dosage adjustments areapproached on an individualized basis.Dosages should be decreased in patients withsignificant hepatic impairment (increases inclotting factors) or with patients receivingdrugs with a narrow therapeutic index (e.g.chloramphenicol).

Mucositis and graft-versus-host disease

Both mucositis and graft-versus-host disease(GVHD) are complications of bone marrowtransplantation and intensive chemotherapy

that may have an effect on the pharmacokinet-ics of antimicrobial agents. It is difficult todescribe, however, any consistent effect thatthese conditions will have on drug absorption,distribution, and elimination, partly owing toinstitution-specific differences in chemother-apy, infectious complications, and supportivecare. Although neutropenia itself does notappear to markedly alter the pharmacokineticsof �-lactams, aminoglycosides, quinolones, orazole antifungal agents,39–49 antibiotic bodyclearance may be increased in hyperdynamicstates such as sepsis or during periods of severestress.49 Moreover, fluid shifts and decreases inserum albumin may affect the Vd, drug clear-ance, and penetration of antibiotics into sometissues and fluids.22

Mucositis/stomatitis may either increase ordecrease the rate and extent of antibioticabsorption. If significant gut edema orachlorhydria is present, absorption may bedelayed or decreased for some antibiotics.49,50 Inpatients with grade II or grade III mucositis,absorption is unpredictable. Several studiesexamining the ability of oral non-absorbableantimicrobial agents to decontaminate the gas-trointestinal tract have documented significantdrug concentrations in the bloodstream inpatients with grade II–III mucositis.51–53

Oral antibiotic therapy is increasingly recog-nized as a simpler and more cost-effectiveoption in the treatment of some low-riskpatients with neutropenic fever.54,55 Mucositisand GVHD, however, should be considered rel-ative contraindications towards the use of oralantimicrobial therapy.26 The majority of studiesexamining the use of oral therapy in the man-agement of febrile neutropenia have excludedpatients with any evidence of mucositis orGVHD. Moreover, these patients are at higherrisk for breakthrough infections on oral therapycaused by streptococcal species, anaerobes, andCandida species. Further studies are required todocument the effects of mucositis and GVHDon antimicrobial pharmacokinetics.



Table 2.4 CNS penetration and protein binding of antimicrobial agents in febrile neutropeniaa

Drug Biliary CNS concentration % protein bindingconcentration (% serum)(% serum)

Antibacterials

Aminoglycosides

Amikacin 30 10–30 0–10

Gentamicin 10–60 10–30 0–10

Tobramycin Inhalation (TOBI) 10–60 10–30 0–10

�-lactamsAmoxicillin/clavulanate 100–3000 12 20–30

Ampicillin/sulbactam 100–3000 13–15 18–22

Aztreonam 114–405 2–5 56

Cefepime 5 10 20

Cefotaxime 15–75 10 30–51

Ceftazidime 20–40 <10

Ceftriaxone 200–500 8–16 85–95

Imipenem/cilastatin <2 8.5 15–25

Meropenem 3–300 21

Nafcillin 6 9–20 90

Oxacillin 5–10 94

Penicillin G, crystalline 500 5–10 (HD)b 65

penicillin VPiperacillin/tazobactam 100—6000 30 16—48

Ticarcillin/clavulanate 40 45/30

FluoroquinolonesCiprofloxacin 2–4 26 20–40

Gatifloxacin 4 36 20

Levofloxacin 5–6 30–50 24–40

Moxifloxacin 4–5 20–40 50

Ofloxacin 4–7 30–50 24–40

LincosamidesClindamycin 250–300 <1 85–94

a All data have been extracted from representative package inserts and reference 67.b HD, high dose.

Contd


Table 2.4 CNS penetration and protein binding of antimicrobial agents in febrile neutropeniaa – contd


MacrolidesAzithromycin Very high 2–13 12–50

Clarithromycin 7000 2–13 65–70

Erythromycin(s) High 2–13 70–74

OxazolidinonesLinezolid <6 31

StreptograminsQuinupristin/Dalfopristin 8–60 40–60

Vancomycin 50 7–14 (HD)b 10–55

SulfonamidesSulfisoxazole 40–70 80 40–60

Trimethoprim/sulfamethoxazole 100–200 40–50 40–70

Trimethoprim 100 40 40–70

TetracyclinesDoxycycline 200–3200 93

Minocycline 200–3200 76

Tetracycline 200–3200 20–67

OthersAtovaquone 99

Metronidazole 100 30–100 20

Pyrimethamine 85

Rifabutin 7–56

Rifampin 10 000 7–56 80

Trimetrexate 95

AntifungalsAmphotericin B 3 90deoxycholate

Amphotericin B 5–15liposomal (AmBisome)

Amphotericin B 0–3lipid (ABLC)


Table 2.4 CNS penetration and protein binding of antimicrobial agents in febrile neutropeniaa – contd


Amphotericin B 0–3

lipid (ABCD)Fluconazole 50–94 11–12

Flucytosine 60–100 2–4

Itraconazole 3–18 5 99CapsulesSolution

AntiviralsAcyclovir 1–2 0.5 9–33

Amantadine 0.5 67

Cidofovir <0.05 <6

Famciclovir >1 <25

Foscarnet 0.69 17InductionMaintenance

Ganciclovir 0.25–0.7 1–2InductionMaintenance

Oseltamivir

Ribavirin 100–1000 0.7 0

Rimantadine >1 0.4–0.6

Valacyclovir 9–33

SUMMARY

Antimicrobial therapy remains the most impor-tant medical intervention affecting survival inthe febrile neutropenic patient. With thegrowing armamentarium of new agents, it is important that clinicians consider thepharmacokinetic and pharmacodynamic prop-erties of drug therapy in addition to the indi-vidual antibiotic/pathogen MIC profile.

PK : PD ratios not only aid in the comparison ofanti-infective regimens with different potencies,but also provide a useful method for develop-ing institution-specific dosing strategies forempiric therapy. There are significant differ-ences among currently available antimicrobialagents with respect to distribution and penetra-tion into various tissues and bodily fluids. Foragents with low-to-moderate concentrations inthe lung tissue and CNS, higher dosages or

combination therapy should be considered forempiric regimens. Finally, neutropenia itselfhas not been shown to alter the pharmacoki-netic behavior of antibiotics in patients, but isoften associated with other conditions that maypreclude the use of oral therapy in low-riskpatients.

REFERENCES

1. Craig WA, Pharmacokinetic/pharmacodynamicparameters: rationale for antibacterial dosing ofmice and men. Clin Infect Dis 1998; 26: 1–10; quiz11–2.

2. Craig WA, Interrelationship between pharmaco-kinetics and pharmacodynamics in determiningdosage regimens for broad-spectrumcephalosporins. Diagn Microbiol Infect Dis 1995;22: 89–96.

3. Leggett J, Fantin B, Ebert S et al, Comparativeantibiotic dose–effect relations at several dosingintervals in murine pneumonitis and thigh-infection models. J Infect Dis 1989; 159: 281–92.

4. Vogelman B, Gudmundsson S, Leggett J et al,Correlation of antimicrobial pharmacokineticparameters with therapeutic efficacy in an ani-mal model. J Infect Dis 1988; 158: 831–47.

5. Craig WA, Gudmundson S, In Antibiotics inLaboratory Medicine (Lorain V, ed). Baltimore:Williams & Wilkins, 1996.

6. Turnidge JD, Prediction of antibiotic dosingintervals from in vitro susceptibility, pharmaco-kinetics and post-antibiotic effect: theoreticalconsiderations. Scand J Infect Dis Suppl 1990; 74:137–41.

7. Ramadan MA, Tawfik AF, Shibl AM, GemmellCG, Post-antibiotic effect of azithromycin anderythromycin on streptococcal susceptibility tophagocytosis. J Med Microbiol 1995; 42: 362–6.

8. Carbon C, Pharmacodynamics of macrolides,azalides, and streptogramins: effect on extracel-lular pathogens. Clin Infect Dis 1998; 27: 28–32.

9. Moore RD, Lietman PS, Smith CR, Clinicalresponse to aminoglycoside therapy: importanceof the ratio of peak concentration to minimalinhibitory concentration. J Infect Dis 1987; 155:93–9.

10. Blaser J, Stone BB, Groner MC, Zinner SH,Comparative study with enoxacin and netilmicin

in a pharmacodynamic model to determineimportance of ratio of antibiotic peak concentra-tion to MIC for bactericidal activity and emer-gence of resistance. Antimicrob Agents Chemother1987; 31: 1054–60.

11. Craig WA, Once-daily versus multiple-dailydosing of aminoglycosides. J Chemother 1995;7(Suppl 2): 47–52.

12. Li RC, Zhu M, Schentag JJ, Achieving an optimaloutcome in the treatment of infections. The roleof clinical pharmacokinetics and pharmacody-namics of antimicrobials. Clin Pharmacokinet1999; 37: 1–16.

13. Kashuba AD, Nafziger AN, Drusano GL, BertinoJS Jr, Optimizing aminoglycoside therapy fornosocomial pneumonia caused by Gram-negative bacteria. Antimicrob Agents Chemother1999; 43: 623–9.

14. Preston SL, Drusano GL, Berman AL et al,Pharmacodynamics of levofloxacin: a new para-digm for early clinical trials. JAMA 1998; 279:125–9.

15. Forrest A et al, Pharmacodynamics of intra-venous ciprofloxacin in seriously ill patients.Antimicrob Agents Chemother 1993; 37: 1073–81.

16. Fantin B, Leggett J, Ebert S, Craig WA,Correlation between in vitro and in vivo activityof antimicrobial agents against Gram-negativebacilli in a murine infection model. AntimicrobAgents Chemother 1991; 35: 1413–22.

17. Turnidge JD, The pharmacodynamics of beta-lactams. Clin Infect Dis 1998; 27: 10–22.

18. Roosendaal R, Bakker-Woudenberg IA, van denBerg JC, Michel MF, Therapeutic efficacy of con-tinuous versus intermittent administration of cef-tazidime in an experimental Klebsiella pneumoniaepneumonia in rats. J Infect Dis 1985; 152: 373–8.

19. Roosendaal R, Bakker-Woudenberg IA, van denBerghe-van Raffe M, Michel MF, Continuousversus intermittent administration of cef-tazidime in experimental Klebsiella pneumoniaepneumonia in normal and leukopenic rats.Antimicrob Agents Chemother 1986; 30: 403–8.

20. Craig WA, Ebert SC, Continuous infusion ofbeta-lactam antibiotics. Antimicrob AgentsChemother 1992; 36: 2577–83.

21. Craig WA, Andes D, Pharmacokinetics andpharmacodynamics of antibiotics in otitis media.Pediatr Infect Dis J 1996; 15: 255–9.

22. Drusano GL, Role of pharmacokinetics in theoutcome of infections. Antimicrob Agents


Chemother 1988; 32: 289–97.23. Bergeron MG, Tissue penetration of antibiotics.

Clin Biochem 1986; 19: 90–100.24. Nix DE, Goodwin SD, Peloquin CA et al,

Antibiotic tissue penetration and its relevance:impact of tissue penetration on infectionresponse. Antimicrob Agents Chemother 1991; 35:1953–9.

25. Klastersky J, Science and pragmatism in thetreatment and prevention of neutropenic infec-tion. J Antimicrob Chemother 1998; 41(Suppl D):13–24.

26. Hughes WT, Armstrong D, Bodey GP et al, 1997guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Infectious Diseases Society of America. ClinInfect Dis 1997; 25: 551–73.

27. Groll AH, Piscitelli SC, Walsh TJ, Clinical phar-macology of systemic antifungal agents: a com-prehensive review of agents in clinical use,current investigational compounds, and putativetargets for antifungal drug development. AdvPharmacol 1998; 44: 343–500.

28. Nagata MP, Gentry CA, Hampton EM, Is there atherapeutic or pharmacokinetic rationale foramphotericin B dosing in systemic Candida infec-tions? Ann Pharmacother 1996; 30: 811–18.

29. Saag MS, Graybill RJ, Larsen RA et al, Practiceguidelines for the management of cryptococcaldisease. Infectious Diseases Society of America.Clin Infect Dis 2000; 30: 710–18.

30. Wong-Beringer A, Jacobs RA, Guglielmo BJ,Lipid formulations of amphotericin B: clinicalefficacy and toxicities. Clin Infect Dis 1998; 27:603–18.

31. Janknegt R, de Marie S, Bakker-Woudenberg IA,Crommelin DJ, Liposomal and lipid formula-tions of amphotericin B. Clinical pharmacokinet-ics. Clin Pharmacokinet 1992; 23: 279–91.

32. Groll AH, Giri N, Petraitis V et al, Comparativeefficacy and distribution of lipid formulations ofamphotericin B in experimental Candida albicansinfection of the central nervous system. J InfectDis 2000; 182: 274–82.

33. Grant SM, Clissold SP, Fluconazole. A review ofits pharmacodynamic and pharmacokineticproperties, and therapeutic potential in superfi-cial and systemic mycoses. Drugs 1990; 39:877–916.

34. Dupont B, Drouhet E, Early experience with itra-conazole in vitro and in patients: pharmacoki-

netic studies and clinical results. Rev Infect Dis1987; 9(Suppl 1): S71–6.

35. Fletcher C, Sawchuk R, Chinnock B et al, Humanpharmacokinetics of the antiviral drug DHPG.Clin Pharmacol Ther 1986; 40: 281–6.

36. Brigden D, Bye A, Fowle AS, Rogers H, Humanpharmacokinetics of acyclovir (an antiviralagent) following rapid intravenous injection. JAntimicrob Chemother 1981; 7: 399–404.

37. Whitley RJ, Blum MR, Barton N, de Miranda P,Pharmacokinetics of acyclovir in humans follow-ing intravenous administration. A model for thedevelopment of parenteral antivirals. Am J Med1982; 73: 165–71.

38. Ritschel WA, Thompson GA, The one-pointmethod in predicting dosage regimen in case ofhepatic and/or renal failure in presence orabsence of change in volume of distribution. JClin Pharmacol 1979; 19: 350–6.

39. Bianco TM, Dwyer PN, Bertino JS Jr, Gentamicinpharmacokinetics, nephrotoxicity, and predic-tion of mortality in febrile neutropenic patients.Antimicrob Agents Chemother 1989; 33: 1890–5.

40. Daenen S, Erjavec Z, Uges DR et al, Continuousinfusion of ceftazidime in febrile neutropenicpatients with acute myeloid leukemia. Eur J ClinMicrobiol Infect Dis 1995; 14: 188–92.

41. Davis RL, Lehmann D, Stidley CA, Neidhart J,Amikacin pharmacokinetics in patients receivinghigh-dose cancer chemotherapy. AntimicrobAgents Chemother 1991; 35: 944–7.

42. Davis R, Markham A, Balfour JA, Ciprofloxacin.An updated review of its pharmacology, thera-peutic efficacy and tolerability. Drugs 1996; 51:1019–74.

43. Drusano GL, de Jongh C, Newman K et al,Moxalactam and piperacillin: a study of in vitrocharacteristics and pharmacokinetics in cancerpatients. Infection 1985; 13: 20–6.

44. Drusano GL, Plaisance KI, Forrest A et al,Steady-state pharmacokinetics of imipenem infebrile neutropenic cancer patients. AntimicrobAgents Chemother 1987; 31: 1420–2.

45. Janmohamed RM, Leyland MJ, Kelly J, Farrell I,Pharmacokinetics of imipenem/cilastatin in neu-tropenic patients with haematological malig-nancies. J Antimicrob Chemother 1990; 25: 407–12.

46. Lortholary O, Tod M, Rizzo N et al, Populationpharmacokinetic study of teicoplanin in severelyneutropenic patients. Antimicrob AgentsChemother 1996; 40: 1242–7.


47. MacGowan AP, Bedford KA, Blundell E et al,The pharmacokinetics of once daily gentamicinin neutropenic adults with haematological malig-nancy. J Antimicrob Chemother 1994; 34: 809–12.

48. Nyhlen A, Ljungberg B, Nilsson-Ehle I,Pharmacokinetics of meropenem in febrile neu-tropenic patients. Swedish Study Group. Eur JClin Microbiol Infect Dis 1997; 16: 797–802.

49. Fry DE, The importance of antibiotic pharmaco-kinetics in critical illness. Am J Surg 1996; 172:20S–25S.

50. Schafer-Korting M, Pharmacokinetic optimisa-tion of oral antifungal therapy. ClinPharmacokinet 1993; 25: 329–41.

51. Okuno SH, Foote RL, Loprinzi Cl et al, A ran-domized trial of a nonabsorbable antibioticlozenge given to alleviate radiation-inducedmucositis. Cancer 1997; 79: 2193–9.

52. Rohrbaugh TM, Anolik R, August CS et al,Absorption of oral aminoglycosides followingbone marrow transplantation. Cancer 1984; 53:1502–6.

53. Samaranayake LP, Robertson AG, McFarlane TWet al, The effect of chlorhexidine and benzy-damine mouthwashes on mucositis induced bytherapeutic irradiation. Clin Radiol 1988; 39: 291–4.

54. Rolston KV, New trends in patient management:risk-based therapy for febrile patients with neu-tropenia. Clin Infect Dis 1999; 29: 515–21.

55. Rubenstein EB et al, Outpatient treatment offebrile episodes in low-risk neutropenic patientswith cancer. Cancer 1993; 71: 3640–6.

56. Lewis RE, Klepser ME, Ernst EJ et al,Comparison of oral immediate-release (IR) andextended-release (ER) metronidazole bactericidalactivity against Bacteroides spp. using an in vitromodel of infection. Diagn Microbiol Infect Dis2000; 37: 51–5.

57. Nix DE, Tyrrell R, Muller M, Pharmacodynamicsof metronidazole determined by a time–killassay for Trichomonas vaginalis. Antimicrob AgentsChemother 1995; 39: 1848–52.

58. Klepser ME, Wolfe EJ, Pfaller MA, Antifungalpharmacodynamic characteristics of fluconazoleand amphotericin B against Cryptococcus neofor-mans. J Antimicrob Chemother 1998; 41: 397–401.

59. Lewis RE, Lund BC, Klepser ME et al,Assessment of antifungal activities of flucona-zole and amphotericin B administered alone andin combination against Candida albicans by usinga dynamic in vitro mycotic infection model.

Antimicrob Agents Chemother 1998; 42: 1382–6.60. Andes D, In vivo pharmacodynamics of ampho-

tericin B against C. albicans. In: Abstracts of the39th Interscience Conference on AntimicrobialAgents and Chemotherapy. Washington, DC:American Society for Microbiology, 1999: 79.

61. Andes D, van Ogtrop M, Characterization andquantitation of the pharmacodynamics of flu-conazole in a neutropenic murine disseminatedcandidiasis infection model. Antimicrob AgentsChemother 1999; 43: 2116–20.

62. Schulin T, Thauvin-Eliopoulos C, Moellering RCJr, Eliopoulos GM, Activities of the oxazolidi-nones linezolid and eperezolid in experimentalintra-abdominal abscess due to Enterococcus fae-calis or vancomycin-resistant Enterococcusfaecium. Antimicrob Agents Chemother 1999; 43:2873–6.

63. Dresser LD, Rybak MJ, The pharmacologic andbacteriologic properties of oxazoladinones, anew class of synthetic antimicrobials.Pharmacotherapy 1998; 18: 456–62.

64. Diekema DI, Jones RN, Oxazolidinones: areview. Drugs 2000; 59: 7–16.

65. Aeschlimann JR, Zervos MJ, Rybak MJ, Treatmentof vancomycin-resistant Enterococcus faecium withRP 59500 (quinupristin–dalfopristin) adminis-tered by intermittent or continuous infusion,alone or in combination with doxycycline, in an invitro pharmacodynamic infection model withsimulated endocardial vegetations. AntimicrobAgents Chemother 1998; 42: 2710–17.

66. Hershberger E, Aeschlimann JR, Moldovan T,Rybak MJ, Evaluation of bactericidal activities ofLY333328, vancomycin, teicoplanin, ampicillin–sulbactam, trovafloxacin, and RP59500 alone orin combination with rifampin or gentamicinagainst different strains of vancomycin-intermediate Staphylococcus aureus by time–killcurve methods. Antimicrob Agents Chemother1999; 43: 717–21.

67. ASHP, American Hospital Formulary Service.Bethesda, MD: American System of HealthSystem Pharmacists, 2000.

68. Diekema DJ, Pfaller MA, Jones RN et al, Surveyor bloodstream infections due to Gram-negativebacilli: frequency of occurrence and antimicro-bial susceptibility of isolates collected in theUnited States, Canada, and Latin America for theSENTRY Antimicrobial Surveillance Program,1997. Clin Infect Dis 1999; 29: 595–607.


3Controversies and quandaries: The design ofclinical trials in fever and neutropeniaLinda S Elting, Marianne Paesmans

INTRODUCTION

The evolution of clinical trial design in febrileneutropenia has paralleled the evolution of theantibiotic armamentarium. The initial studyestablishing the relationship between neu-trophil count and infection focused on mortal-ity, a common event in the 1960s before modernantibiotics were available.1 An early trial of car-benicillin, which documented the need forempiric therapy at the onset of infection, was a‘before-and-after’ design.2 Results from theintroduction of empiric therapy with an anti-pseudomonal penicillin were compared withhistorical data. In this case, the mortality ratewas so high prior to empiric and specific anti-pseudomonal therapy, and the clinicalexperience so consistent from one institution tothe next, that a randomized trial was notrequired to demonstrate efficacy. As the optionsfor antibiotic therapy have increased, increas-ingly stronger evidence from rigorouslydesigned clinical trials has been required tochange practice. Currently, clinical trialists arechallenged to design within-antibiotic-classcomparisons examining subtle differencesamong very similar agents.3 Despite over fourdecades of experience in clinical trials in febrileneutropenia, many issues remain unresolved,and new controversies have arisen with

advances in medicine and healthcare policy. Inthis chapter, we describe the current state of theart in design of clinical trials of febrile neu-tropenia, and discuss the controversies thatchallenge us in the field today.

THE STATE OF THE ART

Although controversies and challenges remain,there is consensus on many methodological cri-teria for clinical trials in febrile neutropenia.4

We discuss these in the following section.Clinical trials designed to establish the useful-ness of a new agent are divided into fourphases.5 Phase I trials examine the maximumtolerated dose of the agent. Phase II trialsroughly estimate its efficacy and toxicity todetermine if it is worthwhile to proceed withfurther research. Even when randomized, phaseII trials, are not conducted for comparison pur-poses. Phase III trials compare promisingagents with the best known standard of care orwith placebo if no standard exists. Phase IVtrials examine the effectiveness in large unse-lected populations. In this chapter, we focus onmethodological standards for phase II and IIItrials of the efficacy of empiric antimicrobialtherapy for febrile neutropenia.

Trial design

Clinical trials are conducted for the purpose ofgeneralizing their results to the treatment offuture patients.6 They should therefore bedesigned to maximize the generalizability ofthe results and to minimize random error,which has no preferred direction, and system-atic error, called bias. In the absence of bias, thedistribution of the values of the estimated treat-ment effect (in the case of repeated trials) is cen-tered around the true value. In the presence ofbias, estimates of treatment effect are not cen-tered around the true value.

The trial design should be described in amaster document (protocol) that includes thefollowing features.

• Description of the trial objectives The pri-mary objective should be clearly described.This objective will be used to determine thesample size. Secondary objectives, whichwill be analyzed descriptively for futurehypothesis generation, should also beincluded.

• Design In the description of the design,the phase (II or III) should be stated. Thestandard for phase III trials is the random-ized design. This design, which was firstused in clinical research about 50 yearsago,7 is the only design that reliably elimi-nates selection bias. In large samples, it alsoresults in comparability of the studygroups.6 Random assignment is analogousto random sampling, which is an assump-tion basic to statistical inference. It is theonly accepted design when comparing theefficacy of two (or more) treatments.

• Description of the targeted patientpopulation This description should includea list of the eligibility criteria both for inclu-sion and for exclusion from the trial, alongwith operational definitions of these cri-teria. At a minimum, definitions of feverand neutropenia should be included.(Guidelines for these definitions have beenprovided by the Immunocompromised

Host Society.8) Other factors that influenceoutcome may be used for inclusion orexclusion of patients, depending on theobjectives of the study. These include theunderlying neoplasm, age, risk class asdefined by validated risk models,9,10 andseverity of illness. These also should beclearly defined in the protocol. Some inves-tigators have suggested the use of expectedduration of neutropenia as an eligibility cri-terion. However, predictions of the dura-tion of neutropenia at trial entry aregenerally inaccurate. To avoid bias, all eli-gible patients should be included in thetrial. If not, the number of patients rejectedshould be recorded, along with the reasonfor rejection.

• Description of the treatment plan A cleardescription of dosing and administrationschedules should be included, as well asinstructions for duration of therapy, andconditions for modification or discontinua-tion of therapy. Descriptions of the numberand timing of clinical, laboratory, andmicrobiological examinations should alsobe specified.

• Description of the outcomes The most com-monly reported outcome in trials of febrileneutropenia is the response to the empiricregimen. Although opinions vary on thespecific definition of this outcome, what-ever the definition used, it should be speci-fied a priori, because different measuresmay yield different results.11 Other out-comes, such as mortality, incidence ofadverse events or toxicities, and time todefervescence, may also be useful. Timingof the assessments of outcome should beclearly delineated.

• Statistical design The objective of a clinicaltrial is formally addressed by a statisticaltest of one or more hypotheses. The nullhypothesis of no difference is typicallytested against an alternative hypothesis.The observed results of the trial lead torejection of the null hypothesis or to accep-tance of the null hypothesis. The decision is


based on reasoning with probability distrib-utions. Two types of random error mayoccur. The null hypothesis may be rejectedwhen it is, in fact, true (type I error) or thenull hypothesis may be accepted when it isactually false (type II error). The probabilityof making a type I error is controlled ineach statistical test, and is reported as the‘p value’. Classically, but often too dogmat-ically, a p value of less than 0.05 results inrejection of the null hypothesis. It should beinterpreted, case by case, considering theother characteristics of the trial, particularlythe number of hypothesis tests performed.The probability of a type II error is con-trolled only by an adequate sample size,which is achieved by using realistic esti-mates of the expected response rates andthe difference that would be consideredclinically significant. This should be con-sidered when interpreting a p value ofgreater than 0.05, which does not necessar-ily mean that the null hypothesis is true,but only that there is insufficient evidencethat it is false.12

In phase II trials, the primary outcome ofinterest is usually dichotomous – that is, it hasonly two possible values, such as success andfailure. The objective is to estimate the successrate and to determine whether further studywill be worthwhile by comparing the observedsuccess rate with a threshold rate specified apriori. In order to limit exposure of patients toan ineffective treatment, most designs use two(or more) stages. After accrual of an initialseries of patients, the trial is terminated if thestudy drug is clearly inferior. Otherwise, asecond series of patients is recruited to permitprecise estimation of the success rate and toinform the decision to proceed to phase IIIstudies. The most commonly accepted two-stage designs are those proposed by Gehan,13

Fleming,14 and Simon.15 There are two versionsof the Simon design, one minimizing the sam-ple size for treatments of low activity and oneminimizing the maximum sample size.15

Recently developed alternatives include three-stage designs,16 designs that evaluate efficacyand toxicity jointly,17 and Bayesian designs,which incorporate a priori estimates of theprobability of outcomes.18

We recommend a randomized design for allphase III trials. In order to eliminate investiga-tor bias, randomization should be organized sothat the investigator cannot know which treat-ment will be received before an individualpatient is registered on the trial. In multicenterstudies, randomization should be centralized inone coordinating center. The minimizationtechnique for treatment allocation results in aminimal value of the global imbalance function,and is a recommended choice.19

Most commonly, two treatments are com-pared, with the intention of detecting apreviously specified, clinically significant dif-ference,20 using either a fixed sample size or aplanned interim analysis with early trial termi-nation possible in the case of a treatment dif-ference greater than expected.21 In some cases,the goal is to show equivalence between thestudied treatments.22 For the primary compari-son, the probability of type I error is set to 5%and the probability of type II error to 10–20%.The decision between an equivalence trial and asuperiority trial must be made early in thedesign phase, because it has a major impact onboth sample size and on the hypothesis tested.During the design phase, the sample size is cal-culated and a plan of the statistical analysis isprepared. This plan includes the statistical teststhat will be used and the subsets that will beanalyzed separately (intention to treat and perprotocol).

Trial monitoring

Monitoring the progress of the trial is as impor-tant as monitoring the progress of individualpatients. Patient accrual, investigator com-pliance with the protocol, and safety monitor-ing should be done systematically. This isparticularly important for multicenter trials,

DESIGN OF CLINICAL TRIALS 47

and clear procedures for timely communicationof key monitoring information between investi-gators and the central data monitoring centershould be specified as part of the protocol. Nopreliminary analyses of the primary outcomeare conducted unless planned a priori, withappropriate statistical adjustment for the mul-tiple statistical tests. If interim analyses areplanned, the results are not communicated tothe investigators or to the scientific communityprior to completion of recruitment. Ideally,interim analyses are performed and interpretedby an independent data monitoring board,which advises the principal investigator onearly termination or continuation of the trial.23

Outcome assessment

The assessment of outcomes should be as con-sistent as possible. In multicenter trials, the datareview committee determines the eligibility andoutcome of each case. The reviewers areblinded to the treatment received. The datareview committee will also conduct site visits toverify protocol compliance and data against thesource documents.

Trial analysis and reporting

Standards for analyzing and reporting theresults of clinical trials have been described pre-viously.8,24–26 Briefly, the analysis should beginwith a description of the number of patientsregistered in the trial, the number of ineligiblepatients, with a description of the reasons forineligibility, and the number of inevaluablepatients (and reason for inevaluability). In ran-domized trials, these results should be reportedseparately for each treatment group. This analy-sis is followed by a description of the patients’characteristics, including, at least, demographiccharacteristics, sites of infection, and documen-tation of infection, pathogens and susceptibili-ties. The type and dosage of chemotherapyshould be specified for descriptive purposes,

although, at the present time, there are no sim-ple means to predict the occurrence and courseof febrile neutropenia on the basis of the type ofchemotherapy that has been given.

In the case of randomized trials, the compa-rability of the treatment groups for importantprognostic factors should be examined. Ifimbalances are discovered for important prog-nostic factors, statistical comparisons of thetreatment groups should be adjusted retrospec-tively. If the outcome is dichotomous, the logis-tic regression model can be used for thispurpose.27 Time-to-event distributions can beestimated using the Kaplan–Meier techniqueand compared using the logrank test.Examination of time-to-event outcomes withadjustment for covariates can be accomplishedusing Cox proportional hazards regression.28

Reports of the results of hypothesis testingshould be accompanied by reports of parameterestimates with confidence intervals (a confi-dence level of 95% is generally adequate).Ideally, interpretation of the data will includeexamination of both p values and confidenceintervals to permit assessment of clinical andstatistical significance.29

CONTROVERSIES AND QUANDARIES

Despite the sophistication of current trialdesigns, there are still a number of controver-sial and challenging issues to be resolved. Theperennial problem of ensuring comparability ofthe study groups and generality of resultsplagues studies of febrile neutropenia as it doesother studies. Blinding of trials, although theo-retically optimal, is clinically and operationallychallenging. In addition to these longstandingissues, methodological advances and changesin the healthcare marketplace have introducednew challenges in clinical trial design. Theseinclude the use of alternatives to classical fixed-sample-size designs and conducting cost–effec-tiveness analyses alongside clinical trials.Among the more controversial issues are thechoice of outcomes to measure and the hand-


ling of multiple entries and withdrawals. Wediscuss each of these below.

How can the generality of results beensured?

The sole reason for conducting clinical trials isto apply the observed outcomes of subjects intrials more broadly, to a larger population.Therefore, it is essential that the study sampleand the larger population be similar withrespect to factors that affect clinically importantoutcomes of the condition. Factors affecting theoutcomes of febrile neutropenia have been welldescribed.1,8,30 They include host factors,infection-related factors, and factors related toantineoplastic therapy (Table 3.1). Depending


Table 3.1 Prognostic factors in febrileneutropenia

Host factors

Severity of illnessComorbid conditionsStage of diseaseAge

Infection factors

Site of infection, especially complex tissueinfectionsPathogen and susceptibilityShock

Treatment factors

Depth of neutropeniaDuration of neutropeniaHigh-dose chemotherapyBone marrow transplantationPresence of cathetersProphylactic antibiotics

on the purpose of the study, any or all of thesemay be clinically important.

The prevalence of these factors varies signifi-cantly, depending on practice patterns, referralpatterns, and the local microbiological flora. Tothe extent that the prevalence in the study sam-ple does not match that in the overall or refer-ence population, the results of the study willnot be useful. In order to ensure that the resultsof trials are useful outside the study sample,stratified analyses accounting for those factorspresent at baseline are appropriate. (Subsetanalyses of factors that occur during the courseof therapy should be avoided, since their occur-rence may be affected by the study drugs. Thiswill almost certainly bias the results of thestudy.31) Univariate subset analyses by a fewvariables of relevance to the specific population(specified a priori) and multiple-variable mod-eling of overall response rates are recom-mended. However, numerous subset analysesshould be avoided, because of the risk ofobserving chance occurrences of statisticallysignificant differences. To compensate for thisproblem, a lower threshold for statistical signif-icance should be used.

By what method is comparability of groupsensured?

In comparative trials, it is also essential that thestudy groups be comparable at baseline withrespect to the prognostic factors mentionedabove. However, in studies of febrile neutrope-nia, there are substantial within-study hetero-geneities in prognostic and complicatingfactors. Randomization does not guaranteecomparability with respect to important prog-nostic factors – it leaves comparability tochance. As illustrated in Table 3.2, in some ran-domized trials, imbalance of important prog-nostic factors occurs. Stratification is used priorto randomization to ensure comparability for afew prognostic factors known at baseline.32–39

Depending on the hypotheses and populationbeing studied, these may include age, risk

group, severity of illness or comorbidity score,and study site (in multicenter studies).

Unfortunately, the most significant predic-tors of outcome of febrile neutropenia are notknown prior to randomization. In some cases,surrogates can be used. For example, growthfactor use, high-dose chemotherapy use, andbone marrow transplantation are reasonablygood surrogates for the depth and duration ofneutropenia. The presence of shock is a fair sur-rogate for the presence of systemic infection.However, none of these are perfect surrogates.Therefore, it is generally necessary to constructmultiple-variable models of outcome to accountfor prognostic factors present at baseline, butnot yet known.

Which outcomes should be measured andreported?

The outcomes of interest in clinical trials offebrile neutropenia have evolved as antibioticagents have become more effective and effect-ive agents have become more numerous. Mostagree that the primary outcome of interest isresponse to antibiotic therapy, but opinions dif-fer widely on the timing of its measurement.4

Some measure response at the end of initialtherapy, and others measure it after therapy hasbeen modified.40,41 In the absence of a standarddefinition for the primary outcome, meta-analyses and informal comparisons across trialsare virtually impossible. In the absence of con-

sensus, we favor reporting both outcomes.Infection-related mortality is an important out-come, but, in trials of modern antibiotic regi-mens, it is usually too rare an event to bepractical in clinical trials. All-cause mortality istoo non-specific to be useful in a cancer popu-lation. The incidence of toxicity or superinfec-tion may be useful in discriminating betweenantibiotic regimens of similar efficacy. A recentstudy suggests that time to response also showspromise in this regard, but its usefulness inpractice remains to be demonstrated.42

Outcomes in clinical trials also reflect recenttherapeutic developments and trends in thehealthcare industry. In trials of outpatient ther-apy of febrile neutropenia, response in the out-patient setting is as important as response toantibiotic therapy. Trends in quality of life andtools for measuring quality of life during febrileneutropenia are currently being studied.Finally, in today’s world, economic outcomes,such as duration of antibiotics, duration of hos-pitalization, and cost of care, are also important.

Is there a more efficient alternative to thetraditional fixed-sample-size randomizeddesign?

The fixed-sample-size randomized controlledclinical trial (RCCT) is the gold standard forcomparisons of therapies for febrile neutrope-nia and for all other conditions. However, thesetrials can be prohibitively difficult for uncom-


Table 3.2 Comparability of treatment groups in two randomized studies

Ref Regimen 1 Regimen 2 Regimen 3 p value% pneumonia (95% CI) % pneumonia (95% CI) % pneumonia (95% CI)

60 4 (2, 8) 4 (2, 8) — 0.6661 21 (15, 29) 10 (6, 16) 11 (6, 17) 0.009

CI, confidence interval.

mon problems, and they are costly for the mostcommon of problems. Designs requiringsmaller sample sizes would be preferable.Furthermore, fixed-sample-size RCCTs possessthe undesirable characteristic of exposing asmany subjects to an inferior therapy as thesuperior therapy.43 Ethicists and statisticianshave argued that in trials addressing uncom-mon or life-threatening conditions, the weightof the individual’s interests exceeds the collect-ive (research) interest of society.43 Therefore,exposure of individual subjects to an inferiortherapy should be minimized. For ethical andpractical reasons, alternative, data-dependentrandomized designs have been studied fordecades.43–49 These are used only rarely, buttrials of febrile neutropenia present a promisingvenue for such designs.

Data-dependent randomized designs are ofthree basic types. ‘Adaptive’ designs incorporateaccumulating outcome data to amend probabili-ties of treatment allocation in order to givepatients a better chance of receiving the superiortreatment. This is a randomized, play-the-winnerdesign. ‘Bayesian’ designs test prior estimates ofprobabilities against implied posterior probabili-ties, which are updated as evidence accumulates.It is generally unnecessary to continue such trialsto the prespecified accrual. ‘Sequential’ designsinvolve interim monitoring for predeterminedthreshold values of outcomes, which, whencrossed, lead to termination of the trial. Eachinterim test affects type I error (‘spends alpha’),and therefore adjustment of the overall signifi-cance level is required for the final analysis.These designs are only practical when individualsubjects’ outcomes are known rapidly (beforerecruitment of the remainder of the subjects).Bayesian designs are only practical when otherphase III or pilot study data are available onwhich to base prior estimates of probabilities.43

Should clinical trials be ‘blinded’?

If the RCCT is the gold standard in clinicalresearch, a blinded RCCT is the platinum stan-

dard, and a double-blinded RCCT is the jewelin the crown. Blinding is a mechanism by whichinvestigators and/or subjects are kept ignorantof which of the alternative therapies (investiga-tional or control) is received.32 It is achieved by‘packaging’ the investigational and controltherapies so that they appear identical to thesubjects and caregivers. The study report formsare also ‘packaged’ so that study drug assign-ment is unknown to those assessing response.

Blinding minimizes the risk that actions orjudgements on the part of investigators or sub-jects will bias the results of the trial. Blinding ofstudy subjects is essential when the outcomesthat will be measured are subjective, such assymptom severity, wellbeing, or quality of life.Investigator blinding is essential to allow unbi-ased assessment of outcomes. Even assessmentof mortality, which appears to be a straight-forward clinical event, can be biased inunblinded studies if knowledge of the treat-ment assignment influences attribution of causeof death. It is also important in studies that per-mit amendment of treatment plans based onclinical judgement. Although blinding adds tothe complexity and cost of studies, the primaryobjection is usually a clinical one. Sometimes,management of the treatment or its complica-tions requires knowledge of the specific treat-ment to which the subject has been assigned.On the basis of this clinical objection, blindingis commonly rejected as an option. However, insuch cases, it is often possible for a physicianexternal to the study to decide on the propercourse of action, without compromising thecare of subjects. Although it is occasionally inthe best interest of an individual subject tobreak the blinding, this rare event should notpreclude the use of blinding in clinical trials.

How should multiple episodes in the samepatient be handled?

Entry of multiple episodes of febrile neutrope-nia in a single patient violates the assumptionsof independence required by most statistical


tests. Practically speaking, multiple entries perpatient may introduce bias due to within-patient correlation for important prognostic fac-tors and for outcomes. The risk of this problemin febrile neutropenia is not trivial. When mul-tiple episodes are closely related in time, it islikely that severity of illness, use of growth fac-tors, high-dose chemotherapy and BMT, anddepth and duration of neutropenia will be vir-tually identical. Nevertheless, most infectiousdisease experts agree that inclusion of multipleentries is desirable, provided that sufficienttime (usually 14 days) has elapsed betweenepisodes to ensure that a new episode hasoccurred. Thus, multiple entry of individualpatients is commonplace in published trials offebrile neutropenia, and is endorsed by expertpanels.8 Two techniques may be used toaccount for the violation of assumptions intro-duced by this practice:

1. When multiple entries occur, a separateanalysis, using only one episode per patient(either the first episode or a randomly cho-sen episode) should be reported. If differ-ences between study drugs change in aclinically or statistically meaningful way,confounding due to multiple entries islikely and the results should be interpretedin that light.

2. A second option is to include all episodesfor all patients in a mixed, multiple-variablemodel, with patients nested within febrileepisodes. This is effectively a repeated-measures analysis, which accounts forwithin-patient correlation for prognosticfactors. Such analyses can now be com-puted with standard statistical packagessuch as SAS.

How should withdrawals be handled?

Studies of febrile neutropenia vary in the fre-quency with which subjects, once randomized,fail to receive the entire course of therapy speci-fied by the protocol. Two clinical phenomena –

‘early deaths’ and adverse reactions thatrequire discontinuation of the study drug –account for the majority of such cases. Studiesalso vary in the frequency with which subjects,initially considered eligible for the study, areproven to be ineligible, after more completediagnostic evaluation. Most studies of empirictherapy of febrile neutropenia include at least afew patients whose specific infections renderthem ineligible after the results of baseline cul-tures become available. Most studies alsoinclude a few patients whose fever is proven tobe due to problems other than infection. Thesepatients are typically considered inevaluable.Although removal of inevaluable and ineligiblecases from the analysis of eligible subjects whocomplete an entire course of therapy is intu-itively appealing, their removal can introducebias of unknown magnitude and direction.50

Several methods of handling withdrawalshave been suggested, from retaining all suchcases in the analysis to removing them. Basedon what is known about febrile neutropeniaand patients’ responses to antibiotic therapy,we favor retention of patients who receive inad-equate trials because of early deaths or adverseevents and classification of such cases as fail-ures to antibiotic therapy. Elimination of suchcases removes failures from the analysis; a trulyinferior antibiotic regimen presumably wouldhave more such cases than an optimal regimen.Thus, the results of a comparison would bebiased in favor of an inferior regimen.

The case of ineligible subjects is less straight-forward. Inclusion of patients with proven fun-gal infections in the analysis of a trial ofantibacterial antibiotics is counterintuitive.However, removal of subjects whose fungalinfections are documented late in the course ofantibiotic therapy biases the results in favour ofregimens that predispose to fungal superinfec-tion. Given the necessity of treating febrileneutropenia empirically, we recommendenrollment of patients presumed to be eligibleand subsequent withdrawal of ineligiblepatients based only on the results of diagnostic testsobtained at baseline. We further recommend


removal of patients considered inevaluable forother reasons, such as fever due to other causes.In such cases, two analyses should be presented– one an intention-to-treat analysis, withinevaluable subjects included, and a secondwith such subjects removed (per protocolanalysis).50

Should cost–effectiveness analyses beconducted alongside clinical trials?

Cost–effectiveness analyses alongside clinicaltrials are controversial.51–59 Such analyses pro-vide high internal validity because of the exten-sive clinical information obtained for the trial.Piggybacking cost–effectiveness studies on toclinical trials also may be a more efficient andless costly way of collecting such data.However, these analyses involve significanttradeoffs. High internal validity is obtained atthe cost of very questionable external validity.To what extent do patients enrolled on a clinicaltrial resemble all patients with febrile neutrope-nia? How generalizeable are costs and resourceutilization derived from protocol-driven care inacademic centers? In our view, the argumentthat tips the balance toward conducting suchtrials is that decisions about reimbursement fornew agents are made at the time products arelicensed. If cost–effectiveness analyses aredelayed until phase IV or prospectivecommunity-based data are available, thenmany patients may be denied access to newagents.

For these reasons, we recommend thatcost–effectiveness analyses be conducted along-side phase III clinical trials in febrile neutrope-nia. From a methodological standpoint, thereare two important issues in such trials. First,there is the problem of sample size. Because theratio of the effect size to the variance is typicallysmaller than that for clinical outcomes, samplesizes for phase III studies may need to beincreased to provide sufficient power.55,58 Werecommend that cost-of-therapy studies be con-ducted alongside phase II studies in order to

obtain preliminary estimates of the magnitudeand variance of cost in order to informdecisions about sample size for phase III stud-ies.

The second major methodological problem isoutcome measurement. Cost–effectiveness isgenerally expressed as the incremental cost perquality-adjusted life-year. Episodes of febrileneutropenia are short-term health states; there-fore, the results of such analyses are measuredin life-months or days. An incremental advant-age in quality-adjusted life-months couldpotentially be realized from (i) a lower mortal-ity rate, or (ii) higher quality because of morerapid response to therapy. As previously noted,mortality is extremely uncommon, and isunlikely to be useful in discriminating betweenantibiotic regimens. Thus, the challenge incost–effectiveness is to measure quality of lifeduring the short duration of febrile neutrope-nia. Although this issue is being studied, thereare currently no widely accepted tools for mea-suring quality of life during febrile neutropenia.

SUMMARY

Clinical trial methodology in febrile neutrope-nia has developed significantly since the initialtrials were conducted in the 1960s. However,important controversies still remain, and newdevelopments in the healthcare marketplacedemand innovative methodological solutions.Data-driven designs, methods for cost–effec-tiveness, and measurement tools for quality oflife during febrile neutropenia are particularlyfruitful topics for methodological research.Such research is critical to the success of futuretrials in febrile neutropenia.

REFERENCES

1. Bodey GP, Buckley M, Sathe YS, Freireich EJ,Quantitative relationships between circulatingleukocytes and infection in patients with acuteleukemia. Ann Intern Med 1966; 64: 328–40.


2. Schimpff S, Satterlee W, Young V et al, Empirictherapy with carbenicillin and gentamicin forfebrile patients with cancer and granulocytope-nia. N Engl J Med 1971; 284: 1061–5.

3. Klastersky J, Zinner SH, Calandra T et al,Empiric antimicrobial therapy for febrile granu-locytopenic cancer patients: lessons from fourEORTC trials. Eur J Cancer 1988; 24(Suppl 1):S35–45.

4. Feld R, Methodology for clinical trials in cancerpatients with febrile neutropenia. Support CareCancer 1998; 6: 423–4.

5. Pocock SJ, Clinical Trials: A Practical Approach.Chichester: Wiley, 1990.

6. Piantadosi S, Clinical Trials: A MethodologicPerspective. New York: Wiley, 1997.

7. Medical Research Council, Streptomycin treat-ment of pulmonary tuberculosis. BMJ 1948; ii:769–82.

8. Immunocompromised Host Society, The design,analysis and reporting of clinical trials on theempiric management of the neutropenic patient.J Infect Dis 1990; 161: 397–401.

9. Talcott JA, Siegel RD, Finberg R et al, Riskassessment in cancer patients with fever andneutropenia: a prospective, two-center valida-tion of a prediction rule. J Clin Oncol 1992; 10:316–22.

10. Klastersky J, Paesmans M, Rubenstein EB et alfor the Study Section on Infections of theMultinational Association for Supportive Care inCancer, The Multinational Association forSupportive Care in Cancer Risk Index: a multi-national scoring system for identifying low-riskfebrile neutropenic cancer patients. J Clin Oncol2000; 18: 3038–51.

11. Elliott C, Pater JL, The effect of different mea-sures of outcome on the results of studies ofempiric antibiotic therapy in febrile neutropenicpatients. Clin Invest Med 1988; 11: 327–30.

12. Altman DG, Bland MJ, Statistics notes. Absenceof evidence is not evidence of absence. BMJ 1995;311: 485.

13. Gehan EA, The determination of the number ofpatients required for a follow-up trial of a newchemotherapy agent. J Chron Dis 1961; 13:346–53.

14. Fleming TR, One-sample multiple testing pro-cedure for phase II clinical trials. Biometrics 1982;38: 143–51.

15. Simon R, Optimal two-stage designs for phase II

clinical trials. Control Clin Trials 1989; 10: 1–10.16. Chen TT, Optimal three-stage designs for phase

II cancer clinical trials. Stat Med 1997; 16:2701–11.

17. Bryant J, Day R, Incorporating toxicity consider-ations into the design of two-stage phase II clini-cal trials. Biometrics 1995; 51: 1372–83.

18. Stallard N, Sample size determination for phaseII clinical trials based on Bayesian decisiontheory. Biometrics 1998; 54: 279–94.

19. Pocock SJ, Simon R, Sequential assignment withbalancing for prognostic factors in the controlledclinical trial. Biometrics 1975; 31: 103–15.

20. Armitage P, Statistical Methods in MedicalResearch. Oxford: Blackwell Scientific, 1971.

21. Jennison C, Turnbull BW, Group SequentialMethods with Applications to Clinical Trials. BocaRaton, FL: Chapman and Hall/CRC, 2000.

22. Blackwelder WC, Proving the null hypothesis inclinical trials. Control Clin Trials 1982; 3: 345–53.

23. Wittes J, Behind closed doors: the DataMonitoring Board in randomized clinical trials.Stat Med 1993; 12: 419–24.

24. Chalmers TC, Smith H, Blackburn B et al, Amethod for assessing the quality of a random-ized control trial. Control Clin Trials 1981; 1:31–49.

25. Pater JL, Weir L, Reporting the results of ran-domized trials of empiric antibiotics in febrileneutropenic patients – a critical survey. J ClinOncol 1986; 4: 346–52.

26. Paesmans M, Statistical considerations in clinicaltrials testing empiric antibiotic regimens inpatients with febrile neutropenia. Support CareCancer 1998; 6: 438–43.

27. Hosmer DW, Lemeshow S, Applied LogisticRegression. New York: Wiley, 1989.

28. Marubini E, Valsecchi MG, Analysing SurvivalData from Clinical Trials and Observational Studies.Chichester: Wiley, 1995.

29. Braitman LE, Confidence intervals assess bothclinical significance and statistical significance.Ann Intern Med 1991; 114: 515–17.

30. Elting L, Rubenstein E, Rolston K et al,Outcomes of bacteremia in neutropenic cancerpatients: observations from two decades of epi-demiologic and clinical trials. Clin Infect Dis1997; 25: 247–59.

31. Byar DB, Simon RM, Friedewald WT et al,Randomized clinical trials. N Engl J Med 1976;295: 74–80.


32. Friedman LM, Furberg CD, DeMets DL,Fundamentals of Clinical Trials, 2nd edn. Littelton,MA: PSC, 1985.

33. Grizzle JE, A note on stratifying versus completerandom assignment in clinical trials. Control ClinTrials 1982; 3: 365–8.

34. Hill AB, The clinical trial. Br Med Bull 1951; 7:278–82.

35. Meier P, Stratification in the design of a clinicaltrial. Control Clin Trials 1981; 1: 355–61.

36. Peto R, Pike MC, Armitage P et al, Design andanalysis of randomized clinical trials requiringprolonged observation of each patient. Br JCancer 1976; 34: 585–612.

37. Pocock SJ, Simon R, Sequential treatment assign-ment with balancing for prognostic factors in thecontrolled clinical trial. Biometrics 1975; 31:103–15.

38. Pocock SJ, Allocation of patients to treatments inclinical trials. Biometrics 1979; 35: 183–97.

39. Elting LS, Stratification in clinical trials of febrileneutropenia. Support Care Cancer 1998; 6: 457–61.

40. Pizzo PA, Hathorn JW, Hiemen ZJ et al, A ran-domized trial comparing ceftazidime alone withcombination antibiotic therapy in cancer patientswith fever and neutropenia. N Engl J Med 1986;315: 552–8.

41. Kern WV, Cometta A, De Bock R et al, Oralintravenous empiric therapy for fever in patientswith granulocytopenia who are receiving cancerchemotherapy. N Engl J Med 1999; 341: 305–11.

42. Elting LS, Rubenstein EB, Rolston KVI et al,Time to clinical response: an outcome of anti-biotic therapy of febrile neutropenia with impli-cations for quality and cost of care. J Clin Oncol2000; 18: 3699–706.

43. Palmer CR, Rosenberger WF, Ethics and prac-tice: alternative designs for phase III randomizedclinical trials. Control Clin Trials 1999; 20: 172–86.

44. Simon R, Adaptive treatment assignmentmethods and clinical trials. Biometrics 1977; 33:743–9.

45. Rosenberger WF, Lachim JM, The use ofresponse-adaptive designs in clinical trials.Control Clin Trials 1993; 14: 471–84.

46. Brophy JM, Lawrence J, Placing trials in contextusing Bayesian analysis: GUSTO revisited byReverend Bayes. JAMA 1995; 273: 871–5.

47. Lilford RJ, Thornton JG, Braunholtz D, Clinicaltrials and rare diseases: a way out of a conun-drum. BMJ 1995; 311: 1621–5.

48. Lilford RJ, Braunholtz D, For debate: The statisti-cal basis of public policy: a paradigm shift isoverdue. BMJ 1996; 313: 603–7.

49. Brophy JM, Joseph L, Bayesian interim statisticalanalysis of randomized trials. Lancet 1997; 349:1166–8.

50. Tsiatis A, Analysis and interpretation of trialresults: intent-to-treat analysis. J Acquir ImmuneDefic Syndr 1990; 3: S120–3.

51. Drummond MF, O’Brien B, Stoddart GL,Torrance GW, Methods for the Economic Evaluationof Health Care Programmes, 2nd edn. Oxford:Oxford University Press, 1997.

52. Drummond M, Economic analysis alongsideclinical trials: problems and potential. JRheumatol 1995; 22: 1403–7.

53. Bennett CL, Westerman IL, Economic analysisduring phase III clinical trials: Who, what, when,where, and why. Oncology 1995; 9(Suppl):169–75.

54. Donaldson C, Hundley V, McIntosh E, Usingeconomics alongside clinical trials: why we can-not choose the evaluation technique in advance.Health Econ 1996; 5: 267–9.

55. Schulman KA, Llana T, Yabroff KR, Economicassessment within the clinical development pro-gram. Med Care 1996; 34: DS89–95.

56. Bennett CL, Waters TM, Economic analyses inclinical trials for cooperative groups: operationalconsiderations. Cancer Invest 1997; 15: 448–53.

57. Fayers PM, Hand DJ, Generalization from phaseIII clinical trials: survival, quality of life, andhealth economics. Lancet 1997; 350: 1025–7.

58. Schulman KA, Ohishi A, Park J et al, Clinicaleconomics in clinical trials: the measurement ofcost and outcomes in the assessment of clinicalservices through clinical trials. Keio J Med 1999;48: 1–11.

59. Van Enckevort PJ, TenVergert EM, Kingma J etal, Factors to consider when designing phase IIIclinical trials involving economic evaluations.Perceptual and Motor Skills 1999; 89: 1059–72.

60. Freifeld AG, Walsh T, Marshall D et al,Monotherapy for fever and neutropenia incancer patients: a randomized comparison of cef-tazidime versus imipenem. J Clin Oncol 1995; 13:165–76.

61. Jones PG, Rolston KVI, Fainstein V et al,Aztreonam therapy in neutropenic patients withcancer. Am J Med 1986; 81: 243–8.


4Current epidemiology of infections inneutropenic cancer patientsWinfried V Kern

INTRODUCTION

Fever is the most frequent – sometimes the only– sign of infection subsequent to neutropenia.Most epidemiologic studies have focused onfever episodes and have evaluated their causesand outcome. A rough approximation of feverincidence rates according to the duration ofneutropenia is shown in Figure 4.1. Severeinfections may sometimes develop withoutfever. Conversely, fever may not always resultfrom infection, and may remain unexplained.

In addition to the depth and duration of neu-tropenia, other variables may impact on the fre-quency and causes of infectious complicationsin neutropenic patients. These include age,comorbidities, activity and site of the under-lying disease (‘obstruction leads to infection’),preceding type of therapy (type and doses ofchemotherapeutic agents; radiation), variablesrelated to exposure to pathogens (includinghospital hygiene, intravenous catheter use andcare, diet, air filtration, etc.), use of antimicro-bial chemoprophylaxis (Figure 4.1), and others.Although fever in a severely neutropenic hostrequires prompt diagnostic work-up and initialempiric therapy, there is considerable hetero-geneity in causative pathogens, sites of infec-tion, clinical evolution, and risk of severecomplications.

80

60

40

20

0

Inci

denc

e of

fev

er (%

)

Days of neutropenia1 9 17 25

100

5 13 21

Antibacterialprophylaxis

? Young age

High-dose chemotherapy

(mucositis)

Figure 4.1 Schematic diagram estimating theincidence of fever according to the duration ofneutropenia (neutrophils <500/µl). Mucositis andantibacterial prophylaxis among other variables mayimpact on the fever incidence, as indicated by thearrows and the dashed lines.

The outcome of febrile neutropenia hasimproved after the adoption of the concept ofempiric antimicrobial therapy and with theavailability of broad-spectrum �-lactams for ini-tial therapy.1–3 Deaths from primary infection,however, continue to be observed. Complicatedsecondary infections due to drug-resistantmicroorganisms have now become more

common in patients with a long duration ofneutropenia, and reduce the likelihood of sur-vival.4,5 In an analysis of 3080 febrile neu-tropenic patients selected to participate inclinical trials of the European Organization forResearch and Treatment of Cancer (EORTC)International Antimicrobial Therapy Co-operative Group, the acute mortality rate was3.2%, while the mortality rate at the end of thefebrile neutropenic episode was 8.7%.6

EPIDEMIOLOGY

Acute leukemia

Bone marrow involvement and intensivechemotherapy render the patient with acuteleukemia highly vulnerable to infection.7 Often,neutropenia lasts 3 weeks or longer, andpatients with early onset of fever and pro-

longed periods of broad-spectrum antibioticsare at increased risks of developing fungalsuperinfections. In a retrospective study at theUniversity Hospital of Zürich, the incidence offebrile neutropenia among acute leukemiapatients (139 patients; 230 neutropenicepisodes) was 86%.8 In a retrospective analysisat Ulm University Hospital and Medical Center,the overall incidence of febrile neutropenia inhospitalized adult patients with acute leukemia(period 1990–1993, 221 patients, 539 neutropenicepisodes) was 71%. This rate had been similar inan earlier analysis, and did not change substan-tially in more recent years (Figure 4.2). Feverwas best predicted by the duration of severeneutropenia (<100 cells/µl) and the type ofleukemia (lymphoblastic, ALL, versus myeloid,AML) (Figure 4.2), while age, sex, and type ofchemotherapy were not predictive, and the status(relapsed/refractory versus de novo) was mar-ginally significant (unpublished observations).


100

0

Inci

denc

e of

fev

er (%

)

AML

ALL80

60

40

20

1984–1987

1990–1993

1997–June 2000

Median daysof neutropenia

20

13

17

14

17

13

Figure 4.2 Incidence of febrile neutropenia in adult patients with acute leukemia during neutropenic episodes.Data are from retrospective studies at Ulm University Hospital and Medical Center during three differentperiods.

Intensive consolidation with high-dose cyto-sine arabinoside (HDAC) has been associatedwith a high incidence of fever.9 Streptococcalbacteremia, sometimes associated with respira-tory distress syndrome, has been reported as acomplication. Complications of HDAC withdaunorubicin and with idarubicin respectivelymay lead to differing complications in thecourse of fever. HDAC–idarubicin combina-tions may be associated with greater gastroin-testinal damage, resulting in more cases withdiarrhea or typhlitis/enterocolitis.10 Theseobservations illustrate the impact of chemother-apeutic regimens on fever and infection.Patients with ALL (compared with AML)develop fever less often during remissioninduction, partly because the antileukemic regi-mens frequently contain steroids. This mayincrease the risk of fungal infections. Thesepatients also develop infections after neutrophilrecovery.11,12 The type of steroids (dexametha-sone or prednisone) may have an influence oninfectious complications.13 The incidence offebrile neutropenia in childhood leukemias islower than in adult acute leukemia patients,and the outcome better.14

Other hematologic malignancies

Not all chemotherapy regimens for Hodgkin´sdisease or non-Hodgkin´s lymphoma (NHL)produce profound neutropenia. Also, depend-ing on marrow reserves, neutropenia after com-monly used lymphoma regimens seldom lastslonger than 5–6 days. Accordingly, the inci-dence of febrile neutropenia may be quite low.15

Lymphoma patients may have other defects inhost defense in addition to neutropenia.Examples are multiple myeloma and chroniclymphocytic leukemia patients with functionalhypogammaglobulinemia. Hodgkin´s disease isassociated with T-cell defects, increasing thesusceptibility to opportunistic infections suchas listeriosis, cryptococcosis, and toxoplasmo-sis. Enhanced susceptibility to opportunisticinfections in Hodgkin´s disease and some NHL

patients may also be related to the wider use ofimmunosuppressive drugs such as steroids andpurine analogs. Agents such as fludarabine and2-chlorodeoxyadenosine (cladribine) enhancethe risk of opportunistic infections by inducinga long-lasting T-cell defect that can be meas-ured by enumerating CD4� T cells in theperipheral blood.16,17

Chemotherapy for solid tumors

The incidence of febrile neutropenia in solidtumor patients receiving chemotherapydepends on the dose intensity of the cytotoxicregimens. Many regimens do not produce pro-found neutropenia. Accordingly, febrile neu-tropenia may be quite uncommon. Feverincidences in small cell lung cancer are usuallyless than 50%.15 Fever was reported in 15 of 45ovarian cancer patients (33%) receiving 177cycles of paclitaxel with or without platinum.18

Fever after chemotherapy for testicular cancerwas also relatively uncommon (<20%).19,20 Theinitial cycle is more likely to put patients at riskof developing fever than are subsequent cycles.The reasons for this are manifold, and includehigh tumor burden leading to obstructionand/or reduced general status or evencachexia. Also, patients are often in hospitaland have undergone recent surgery and/orother invasive procedures when they arereceiving their initial chemotherapy, while sub-sequent cycles are given on an outpatient basis.Less exposure to nosocomial flora and a bettergeneral status then contribute to a better toler-ance of cytotoxic drugs.

Blood stem cell and bone marrowtransplantation

Numerous reports have assessed the frequencyand outcome of infectious complications afterhigh-dose chemotherapy with autologousperipheral blood stem cell reinfusion(autoPBSCT).21–26 The duration of neutropenia is

CURRENT EPIDEMIOLOGY OF INFECTIONS IN NEUTROPENIC CANCER PATIENTS 59

usually about 7–10 days, depending on marrowreserves, pretreatment with cytotoxic drugs,prior radiation, and the number of reinfusedCD34� cells. Many centers that offer thiscomparatively new approach use prophylacticantimicrobial regimens similar to those used forallogeneic peripheral blood or bone marrowtransplant (alloBMT) recipients. Nevertheless,reported fever incidence rates in many centersappear to be high: greater than 60%, and oftenapproaching 100%. Death rates, however,appear to be lower than those reported in acuteleukemia patients, most likely because of themuch shorter duration of neutropenia.Experience has been obtained with autoPBSCTpatients in ambulatory care settings.27–29 Thereis no indication that fever is more or less fre-quent in ambulatory patients compared withhospitalized patients. Based on the availableliterature, however, one should expect thatroughly half of the patients, if not more, requirereadmission – most because of febrile neutrope-nia that is not manageable on an outpatientbasis. A significant problem in these patientsmay be severe mucosal damage, leading tostomatitis, abdominal pain, and diarrhea.Bacterial translocation or endotoxinemia lead-ing to fever may thus be facilitated.

Although the use of allogeneic peripheralblood stem cells instead of bone marrow hasreduced the time to engraftment, fever inci-dence rates and infection rates during the earlyneutropenic phase have remained high in thissetting. Depending on the use of chemoprophy-laxis and on environmental exposure, docu-mented infections, including secondaryinfections, may be more common in alloBMTpatients than in autoPBSCT patients.30–32 (SeeChapter 7.)

Neutropenia due to other causes

Limited epidemiologic data are available on theincidence of fever in patients with neutropeniadue to myelodysplastic syndromes (MDS) andaplastic anemia, and in patients with non-

malignant chronic neutropenia.33–36 Case–con-trol studies in the setting of HIV-related neu-tropenia have estimated an adjusted risk ofgreater than 20-fold for Gram-negative bac-teremia in patients with a neutrophil count ofbelow 250 cells/µl compared with patients withlevels above 1000 cells/µl.37,38 Some of the infec-tious complications among such patients aredue to complex host defense deficits, such ashypocomplementemia, T-cell deficiencies, orqualitative defects in phagocyte function inaddition to neutropenia. It is important to rec-ognize that chronic neutropenia precedingintensive chemotherapy (as is often the case inMDS) or immunosuppressive therapy (as inaplastic anemia) predisposes the patient to sub-stantially increased risks of infection due toincreased colonization with fungi and/or anti-biotic-resistant bacteria.

IMPACT OF ANTIMICROBIAL PROPHYLAXISON FEVER IN PATIENTS RECEIVING CANCERCHEMOTHERAPY

Some studies of antibacterial prophylaxis haveshown a significant reduction in the incidenceof fever during neutropenia.39,40 One explana-tion may be that it is inherently difficult toshow effects on fever incidence rates on theextreme sides of neutropenia, i.e. in short-dura-tion and very long-duration neutropenia. It ismore likely to be able to show an effect, if itexists, in patients with intermediate duration ofneutropenia (Figure 4.1). Few studies have eval-uated the time to onset of fever. Delayed feveronset and reduced fever duration might well beacceptable endpoints of effective prophy-laxis.41–44 Prophylactic regimens targetingGram-positive bacteria (such as addition ofrifampin, macrolides, or penicillins) have notbeen more effective than regimens that primar-ily target Gram-negative bacteria (single-drugfluoroquinolone prophylaxis).39,45,46 In studies ofantifungal prophylaxis, fever has not com-monly been used as endpoint.


UNEXPLAINED FEVER VERSUS INFECTION,PRIMARY INFECTION VERSUSSUPERINFECTION

Unexplained fever (or fever of unknown origin,FUO) – i.e. negative cultures and no localizingsigns and symptoms – is common in neu-tropenic patients. The pathophysiology of unex-plained fever is poorly understood. Someinvestigators believe that such episodes areinfections with a low microbial load, which inthe absence of empiric therapy would eventu-ally develop into overt infection. Careful exami-nation may reveal mild localizing symptoms, orviral infection. Using eubacterial rRNA poly-merase chain reaction (PCR), one can find evid-ence of DNAemia (bacteremia with non-viableorganisms or below the threshold of conven-tional cultures) in approximately 25% of febrileneutropenic patients without culture-provenbacteremia.47 In unexplained fever episodes,there is usually a moderate cytokine responsemeasurable in the plasma that is not very differ-ent from that with non-bacteremic documentedinfections.48,49 Fungal DNAemia has also beenfound in patients with FUO who laterdeveloped documented invasive fungal infec-tion.50–52 The inability in some studies of pro-phylaxis to reduce the incidence of fever

despite a reduction in the incidence of docu-mented infection may be due to the drug’seffect of rendering cultures negative, i.e. sup-pressing viable organisms to counts below thelimit of detection. Blood culture studies usingresin media to inactivate residual drug activityin blood samples support this view.53,54 ThusFUO or fever with non-specific mild localizingsigns and symptoms often represents an earlyphase of infection that cannot be documentedby routine clinical examination and laboratorytests. It is plausible that the shorter the neu-tropenia, the more frequent is FUO, while docu-mented infections tend to become morecommon as neutropenia persists (Figure 4.3).Two recent studies among low-risk neutropenicpatients (median duration of neutropenia lessthan 5 days) reported a relative frequency ofunexplained fever of about 60–70%.55,56 Amongpatients with acute leukemia and alloBMTrecipients (median duration of neutropenialonger than 10 days) this proportion is usuallyless than 50%. The prognosis of FUO is excel-lent, both in terms of time to defervescencewith initial empiric therapy and in terms of sur-vival (Figure 4.4). However, while manypatients at initial presentation have unex-plained fever, FUO is a diagnosis that can beestablished only at the end of neutropenia.


100

0

Inci

denc

e (%

)

Days of neutropenia1 9 17 255 13 21

80

60

40

20

Unexplained fever

Other documentedinfectionBacteremia

Invasive fungalinfections

Figure 4.3 Schematic diagramestimating the incidence ofunexplained fever and documentedinfections (bacteremia, non-bacteremic focal infections,invasive fungal infections)according to the duration ofneutropenia (neutrophils <500/µl).

DOCUMENTED INFECTIONS

A commonly used classification differentiatesmicrobiologically documented infections (withor without bacteremia) from clinically docu-mented infections. Widely used in studies ofinitial empiric therapy,57 this classification haslimited prognostic implications – primarilybecause it does not recognize the focus of infec-tion. Review of a number of studies suggeststhat among non-bacteremic infections, pul-monary infection differs most in prognosticcharacteristics from infections at other sites.Also, the prognosis of bacteremic infection dif-fers substantially between cases with or with-


1.0

0.6

0.4

0.2

Days of neutropenia14 28 35

0.8

7 21

0.0

0

1.0

14 28 357 210.0

0

0.2

0.4

0.6

0.8

C

B

AD

E

F

Figure 4.4 Time to defervescence (in days) accordingto type of infection. Data are from a retrospectivestudy at Ulm University Hospital and Medical Centerevaluating the outcome of febrile neutropenicepisodes among adult patients with acute leukemia(period 1990–1993). The insert shows the time todefervescence for the subgroups of patients withbacteremia with (E) and without (F) pulmonaryinfiltrates.

out a clinical focus; and among the cases with aclinical focus, bacteremia with pneumonia maybe the most critical infection (Figure 4.2).Conversely, urinary tract infections (withoutbacteremia) or cases of mild tonsillopharyngitis(clinically documented infection) do not signifi-cantly differ in their prognosis from FUO.

Bacteremia

Bacteremia is one of the most frequent compli-cations of neutropenia. Classically, entericGram-negative rods have been the most fre-quent pathogens of bloodstream infections.Over the past three decades, considerablechanges have occurred in the types of bacteriacausing infection.58–65 As a consequence of long-dwelling intravascular devices, fluoroquinoloneprophylaxis, and high-dose chemotherapy-induced mucositis, there has been a shifttowards bacteremia due to Gram-positive cocci(Table 4.1).

Organisms enter the bloodstream viamucosal sites, skin, or intravascular catheters.Entry via the gastrointestinal tract is probably acommon event, and a number of unexplainedfevers may represent portal bacteremias. Themore aggressive the chemotherapy and result-ing mucosal damage, the more likely is bac-teremia with saprophytic organisms from theoropharyngeal microflora (which do not needto pass the liver). The quantity of the pathogenthat entered the bloodstream may make a dif-ference.66 Numerous, unusual blood culture iso-lates from neutropenic patients represent oralmicroflora constituents. Examples includeMicrococcus, Gemella, Stomatococcus, variousstreptococci, Leptotrichia, Actinomyces, andFusobacterium.

IncidenceThe proportion of bacteremic infections amongfebrile neutropenic episodes ranges between10% and 40%. Occasionally, higher rates havebeen reported.30 The overall incidence of bac-teremia per neutropenic episode ranges

Key: A, bacteremia; B, FUO; C, non-bacteremicnon-pulmonary focal infection; D, pneumonia.

between less than 3% and 30% (Figure 4.3).Without effective chemoprophylaxis, more thanhalf of the isolates are Gram-negative rods. Inpatients with acute leukemia and in BMT recip-ients, the expected incidence of Gram-negativebacteremia per neutropenic episode is about15–20%; in lymphoma patients receiving CHOP(cyclophosphamide, doxorubicin, vincristine,and prednisone) chemotherapy, it is about 3%per neutropenic episode. Strategies for chemo-prophylaxis and initial empiric therapy need toconsider these figures when – as is often thecase – Gram-negative aerobes are the primarytarget organisms.

Risk factors for the development of bac-teremia have been poorly defined.67,68 In onestudy, systemic antifungal prophylaxis wasidentified and later confirmed as a risk fac-tor.69,70 Rackoff and others have identified anabsolute monocyte count (<100/µl) plus highfever as predicting bacteremia.71,72 Procalcitoninserum levels may discriminate between bac-teremia and other documented infections and

FUO.73–75 Measurement at fever onset of inter-leukin (IL)-6 or IL-8 plasma or serum levelsmay be useful to predict the absence of Gram-negative bacteremia.48,76,77 None of these factorscan predict bacteremia accurately.

PrognosisThe prognosis of bacteremia is worse than thatof unexplained fevers. Polymicrobial bac-teremia is associated with a worse prognosisthan single-organism bacteremia. A study fromSpain examined prognostic factors influencingmortality in cancer patients with neutropeniaand bacteremia.78 The overall mortality ratewithin 30 days of the onset of bacteremia was24%. Shock at onset, pneumonia, uncontrolledcancer, and absence of quinolone prophylaxiswere independently associated with increasedmortality. Early evolution into septic shock isclearly related to Gram-negative bacteremia,whereas it is less frequent in Gram-positive bac-teremia, and particularly unusual in bacteremiadue to coagulase-negative staphylococci.11,69,70,78,79


Table 4.1 Bacteremia in clinical trials of the EORTC International Antimicrobial Therapy CooperativeGroup

Single-organism bacteremia

Trial Period No. of patients % Gram-negative % Gram-positive

I 1973–1976 145 71 29II 1977–1980 111 67 23III 1980–1983 141 59 41IV 1983–1985 219 59 41V 1986–1988 213 37 63VIII 1989–1991 151 31 69IX 1991–1993 161 33 67XI 1994–1996 199 31 69XII (low-risk) 1995–1997 39 59 41XIV (high-risk)a 1997–2000 186 47 53

a Prelimary data.

A rough estimate for the overall case-fatalityrate in Gram-negative bacteremia is 10–15%;the estimated rate in viridans streptococcal bac-teremia is about 10%, depending on host fac-tors.80 Elting and colleagues81 have shown thatmajor organ and tissue infection (such as pneu-monia) in neutropenic patients with bacteremiaseverely compromises response to initial ther-apy as well as the ultimate outcome –experience that is consistent with that of otherinvestigators.78,82,83 In the same study, shock,Pseudomonas aeruginosa as the organism, and invitro resistance to the therapeutic agent(s) wereother prognostic factors. Among neutropenicchildren with bacteremia, the overall mortalityrate was reported to be 15%.84

Bacteremia: selected organisms andassociated syndromes

Viridans streptococciThese organisms have been reported to causebacteremic infection among leukemia patientsand BMT recipients.85 The frequency of strepto-coccal infection appears to be dependant on thecumulative dose of the cytotoxic agent cytosinearabinoside or the use of anthracyclines in BMTpatients, in association with mucosal dam-age.9,80,86–93 Fluoroquinolone prophylaxis may beanother risk factor. Incidence rates vary widely,being as high as 48% in adults and 36% in chil-dren with high-dose cytosine arabinoside ther-apy, and 46% in BMT patients. 87 Seriouscomplications, including encephalopathy, respi-ratory distress syndrome (ARDS), and septicshock have been reported in association withthese organisms. The pathogenesis of thesecomplications is unknown. Polymicrobial infec-tion with an as yet unidentified anaerobe hasbeen one speculation about pathogenic eventsin this so-called ‘�-streptococcal shock syn-drome’. Other possible mechanisms are directtoxicity from cytotoxic agents and/or fromstreptococcal products such as toxins, super-antigens, or cell wall components, probably viaeffects on host immune cells.94 Penicillin resis-

tance has become a problem in some cen-ters.93,95,96 Death rates may be as high as 38% inadults and 24% in children, and occur mostoften in association with ARDS.

In contrast to the increased incidence of viri-dans streptococci, pneumococcal bacteremia issurprisingly rare in neutropenia. In trials of theEORTC International Antimicrobial TherapyCooperative Group, less then 25 cases weredocumented in more than 3000 patients(unpublished observations). Similarly, aSpanish study found only 17 episodes of pneu-mococcal bacteremia among 340 neutropeniccancer patients with bacteremia.97

Coagulase-negative staphylococciThe relative proportion of coagulase-negativestaphylococci (CNS) among blood culture iso-lates has been increasing since the late 1980s. Ina study from Spain, bacteremia caused by CNSamong neutropenic cancer patients increasedbetween 1988 and 1993 from 3 episodes per1000 admissions to 19 episodes per 1000 admis-sions.61 CNS are the classical vascular accessdevice-related pathogens.98,99 Although mostcases of bacteremia due to CNS are catheter-related, the reverse is not necessarily true.There are several studies reporting a large pro-portion (40–50%) of catheter-related infectionsdue to organisms other than CNS.100–103 Device-related sepsis may frequently be caused byAcinetobacter spp. and other non-fermenters,Enterobacter and Citrobacter spp., Bacillus spp.,Corynebacterium spp., Micrococcus, and others.Mucosal sites can be the origin of CNS bac-teremia in the neutropenic host.104,105 In the indi-vidual patient, clonal diversity of CNScolonizing the skin and mucosal surfaces isbeing reduced after hospital admission and inresponse to administration of antimicrobialdrugs.106,107 Catheters are rapidly colonized viaskin or hub, and organisms survive in biofilmsspread over the internal and external cathetersurfaces. Even in a given hematology–oncologyservice, there may be only a limited number ofclones of Staphylococcus epidermidis causingcatheter-related infections over extended peri-


ods of time. Consequently, infection due to rela-tively virulent and usually multiresistant iso-lates of CNS among hospitalized patients hasbecome very difficult to control. Most hematol-ogy–oncology departments report endemicoxacillin resistance in CNS isolates,108 resultingin the frequent use of glycopeptide antibi-otics.108,109

Most often, the course of bacteremia due toCNS is uncomplicated even when the catheterhas not been removed. Also, despite resistanceto oxacillin, patients with CNS bacteremiareceiving delayed glycopeptide therapy defer-vesce as rapidly as those receiving upfront gly-copeptides. This may be related to low-gradebacteremia, or to contamination rather thantrue bacteremia. Complications of CNS bac-teremia include catheter tunnel infections, cel-lulitis, septic thrombophlebitis, endocarditis,osteitis/osteomyelitis, and foreign body infec-tions at distant sites (e.g. joint prosthesis infec-tion). CNS are a group of more than 10 speciespotentially pathogenic for man. The most fre-quent isolates from febrile neutropenic patientsare S. epidermidis, S. haemolyticus, S. warneri,and S. hominis. Many laboratories no longer dif-ferentiate CNS to species level. Since suscepti-bility patterns differ among species, suchdifferentiation might be important in cases ofrelapsing bacteremia.

Staphylococcus aureus bacteremia is much lesscommon than CNS bacteremia.110 It is infre-quently reported in neutropenic patients, withnotable exceptions.59,60,111,112 The reason isunclear. It is likely that patients with solidtumors are at higher risk of S. aureus infectionthan patients with hematologic malignancies(who experience neutropenia more often)owing to increased age, less intensive antimi-crobial pretreatments, or more postoperativeinfections.113

Pseudomonas aeruginosa and other non-fermentersBacteremic infections due to Ps. aeruginosa havebeen a common, difficult-to-treat complicationin neutropenic patients.114 More than three

decades ago, reported case-fatality rates in neu-tropenic patients were over 50%. In the 1970s,the outcome improved with the use of combi-nation therapy consisting of carbenicillin (or other �-lactam antibiotics with anti-pseudomonal activity) plus gentamicin. Morerecently, cure rates of about 80% have beenreported.115,116 Currently, the relative frequencyamong blood culture isolates from neutropenicpatients is about 5%, and the overall incidenceof Ps. aeruginosa bacteremia among febrile neu-tropenic episodes, accordingly, is only about1%. At the University of Texas MD AndersenCancer Center, an incidence in acute leukemiapatients of 55 per 1000 registrations has beenreported (including non-neutropenic episodes).At that center, no major changes in incidencewere noted over the last 30 years, except that inmore recent years there were more community-acquired infections.115 In a large multicenterstudy of bloodstream infections among cancerpatients, infection due to Ps. aeruginosa wasidentified as an independent risk factor ofdeath.117

Initial clinical sites of infection are frequentlyobserved, with pneumonia being the most com-mon site (about 40%), and skin and soft tissueincluding the perianal/perirectal area thesecond most common site.115 Skin and soft tis-sue infections can be extensive and severe, andcan extend from the perirectal area to includethe perineum and scrotum as a rapidly spread-ing necrotizing infection.118 Ecthyma gan-grenosum occurs in less than 5% of patients.The frequency of development of shock may beas high as 20–30%, depending on the clinicalsite, evolution of neutrophil counts, and ade-quate therapy.

Glucose-non-fermenting Gram-negative rodsother than Ps. aeruginosa are a heterogenousgroup of organisms comprising Pseudomonasspp. ( fluorescens, putida, stutzeri, and others),related genera (for example Burkholderia,Stenotrophomonas, Flavimonas, Chryseomonas,Comamonas, Shewanella, and Methylobacterium),and the genera Acinetobacter, Achromobacter,Alcaligenes, Ochrobactrum, Agrobacterium,


Flavobacterium, Sphingobacterium, and Moraxella.With the exception of the latter genus, theseorganisms have been increasingly prevalentamong blood culture isolates from immuno-compromised patients over the last twodecades.119–123 The most common isolate in thisgroup is probably Stenotrophomonas maltophilia.Many of the organisms are widely distributedin the environment such as soil, water, organicmaterial. Some (e.g. Ps. fluorescens, Burkholderiapickettii and Ps. paucimobilis) are well known fortheir property to contaminate wet hospitalequipment, antiseptics, intravenous fluids, andcosmetics. Infections often follow instrumenta-tion, are associated with intravascular catheters,or follow administration of contaminated intra-venous fluids or blood products (includingbone marrow or stem cells).124–135 A complica-tion may be right-sided endocarditis. Rarely,other sites of infection can be observed, notablythe lung in the case of S. maltophilia andChryseobacterium meningosepticum.124,136,137 Theinfections have sometimes occurred in smallclusters of true infection or of pseudobac-teremia, presumably because of commonsources in the hospital and often unusualantimicrobial resistance patterns with associ-ated selection and transmission advan-tages.137–145 Examples of the unusual resistancepatterns of the organisms are the intrinsic resis-tance of S. maltophilia to carbapenems, theunpredictable resistance to aminoglycosides(often gentamicin-susceptible but tobramycin-and/or amikacin-resistant) in Comamonas,Alcaligenes, Agrobacterium, and others, and theunusually frequent susceptibility of many non-fermenters to trimethoprim–sulfamethoxazoleand tetracyclines. Since susceptibilities varygreatly between and within species, detailedtesting is mandatory. Unfortunately, disk diffu-sion tests have often been unreliable.

Unusual organismsMany rare microorganisms have been reportedto cause infection in neutropenic patients.146

Examples are nutritionally variant streptococcior Gemella among streptococci, rare species of

CNS, and rare species of Enterobacteriaceae suchas Kluyvera, Hafnia, and Rahnella.147 Many ofthese cause catheter-related infections: Bacillusspp.,148–151 diphtheroids, some of the aerobicactinomycetes, and Mycobacterium fortuitumand other non-tuberculous mycobacteria.152–157

Most others (Fusobacterium and other gram-negative anaerobes,158–160 Lactobacillus spp.,161,162

clostridia,163 anaerobic actinomycetes, Capno-cytophaga spp. group DF-1,164,165 and Stoma-tococcus spp.166–168) have been implicated incausing bacteremia in association with oral orintestinal mucositis.

Antibiotic-resistant EnterobacteriaceaeRecent reports have emphasized the risk of theemergence of fluoroquinolone- and broad-spec-trum �-lactam-resistant Enterobacteriaceae.169

Fluoroquinolone resistance appears to be theresult of the extensive use of fluoroquinolonesfor prophylaxis in cancer patients.170–172 Theprevalence of resistance in the cancer patientpopulation, however, is also highly dependanton the resistance rates among isolates from thecommunity.173 A study from Spain, for example,described a 20% rate of colonization with fluo-roquinolone-resistant Escherichia coli amongcancer patients on admission. This rateincreased after admission to more than 40%.174

In the Netherlands and Germany, colonizationrates are much lower, although fluoro-quinolones are continuously used for preven-tion of Gram-negative sepsis in high-riskneutropenic cancer patients.170,175 The incidenceof bacteremia due to fluoroquinolone-resistantE. coli among patients given prophylaxis there-fore ranges widely between less than 3% andmore than 10%.173–176 The latter rate indicatesthat the prophylactic trend towards reductionof Gram-negative sepsis has been lost. Based ondata from two recent EORTC InternationalAntimicrobial Therapy Cooperative Grouptrials, the proportion of patients with Gram-negative bacteremia during fluoroquinoloneprophylaxis increased between the periods1993–1994 and 1998–2000; an increase in theincidence of Gram-negative bacteremia,


however, was also observed among patientswithout fluoroquinolone prophylaxis (unpub-lished observations) (Table 4.2). Some hospitalswhere fluoroquinolone prophylaxis has beendiscontinued have experienced an increasingincidence of Gram-negative bacteremia, alongwith a reduction of fluoroquinolone resistanceamong the isolates.177,178

According to reports of large numbers ofGram-negative isolates from patients with neu-tropenia, �-lactam resistance rarely exceeds10%.179,180 Resistance rates are higher for aztre-onam and ceftazidime than for carbapenems,and appear to increase slowly in those centersthat use these drugs for empiric therapy.Extended-spectrum �-lactamase can be foundin common Gram-negative bacilli, notably in E.coli and Klebsiella pneumoniae.181 Small epidemicshave been reported.182 There is a well-docu-mented risk of development of �-lactam resis-tance in Enterobacter.180,183 The recent increase inGram-negative bacteremia in neutropenicpatients treated within EORTC InternationalAntimicrobial Therapy Cooperative Grouptrials is in part a result of more frequent infec-tions due to Klebsiella and Enterobacter.

Vancomycin-resistant enterococciVancomycin-resistant Enterococcus faecium(VRE) have been surprisingly common in manyhematology–oncology units. The organism isknown for its intrinsically decreased suscepti-bility to ampicillin. Endemic situations as wellas small outbreaks of bacteremic infection inoncology units have been reported.184–187

Outbreaks in cancer hospitals in associationwith antibiotic formulary changes have beendescribed.187 These outbreaks may initially bepolyclonal. New patients may become colo-nized with VRE within two weeks after admis-sion, and shed the organism over extendedperiods of time.185,188 In a given unit, the geneticdiversity of VRE becomes limited, and cross-contamination and cross-infection are fre-quent.189 An endemic situation follows. The useof third-generation cephalosporins togetherwith glycopeptides appears to be associatedwith the emergence of vancomycin-resistantenterococci. A study from the UK showed thatabout 50% of the patients became colonizedwith VRE during a period in which ceftazidimewas used as initial empiric regimen.186 Afterchanges to piperacillin–tazobactam as empiric


Table 4.2 Gram-negative bacteremia and fluoroquinolone prophylaxis in clinical trials of the EORTCInternational Antimicrobial Therapy Cooperative Group

Trial XIa Trial XIVa,b

(1994–1996) (1997–2000)

Number of patients 386 332Fluoroquinolone prophylaxis 212 (55%) 125 (38%)

Gram-negative bacteremiaPatients with fluoroquinolone prophylaxis 13/212 (6.1%) 12/125 (9.6%)Patients without fluoroquinolone prophylaxis 16/174 (9.2%) 30/207 (14.5%)

a The analysis includes only those 14 centers that participated in both trials.b Preliminary data.

�-lactam, colonization rates first fell to less than20%, increasing to 36% upon reintroduction ofthe cephalosporin. The incidence of clinicalinfection can be low (about 10%), despite com-mon intestinal colonization.184,185 A recent studydescribed a 33% risk of bacteremia in colonizedcancer patients.188 The case-fatality rate rangesfrom 10% to more than 70%, but the attributablemortality is probably much lower. VRE maysometimes be a surrogate marker for poor-prognosis cancer. Vancomycin-dependant ente-rococci are mutants of VRE. A small outbreakinvolving five BMT patients has recently beendescribed.190 Infection control measures andrevision of antibiotic policies (less use ofempiric vancomycin) were able to control thesituation.

Pneumonia

Pneumonia is one of the most critical infectionsin patients with neutropenia. Response to initialempiric therapy is often poor, and reported case-fatality rates range between 20% and more than60%.83,191–193 Patient age, etiology, presence of bac-teremia, severity of lung infection, and persistentneutropenia have a major impact on survival.Patients who develop respiratory insufficiencyhave a very poor prognosis; only 20% or less sur-vive. The relative frequency of pneumoniaamong febrile neutropenic episodes rangesbetween 10% and 30%.8,23,83 In children, the infec-tion is less frequent. Pneumonia often developsafter several days of empiric therapy83,191 andsometimes it is difficult to say whether it is theprimary infection or a superinfection.

Based on culture and histology, the etiologyof lung infections in the neutropenic host can bedefined in only 10–45% of cases,191–195 partlybecause patients have often been given empirictherapy before invasive diagnostic proceduressuch as bronchoscopy. The most common bac-terial pathogens have been Ps. aeruginosa,Klebsiella, pneumococci, and S. aureus.Pneumococcal, staphylococcal, and polymicro-bial infections may be more common in solid

tumor patients, but there is limited epidemio-logic data on this issue. Isolates from initialblood cultures are not necessarily the cause ofevolving pneumonia, even among patients withdocumented pulmonary infiltrates at the timeof bacteremia. Thus, using results of blood cul-tures overestimates the bacterial etiology ofpneumonia in the febrile neutropenic patient. Inparticular, cases of bacteremia due to entero-cocci, viridans streptococci, CNS, diphtheroids,and non-fermenters other than Ps. aeruginosawith subsequent pulmonary infiltrates remainsuspicious of a non-bacterial superinfection ofthe lung, toxic lung injury, or pulmonary hem-orrhage. Among the most important organismscausing severe lung infection in neutropenia inpresent times are filamentous fungi, notablyAspergillus spp., Legionella, mycobacteria,Pneumocystis carinii, protozoa, and viruses areinfrequent.191–193,196,197 Recently, cases ofChlamydia pneumoniae respiratory infection inacute leukemia patients have been described.198

Like many other pneumonias, they occurredlater during febrile neutropenia. A viral etiol-ogy outside the BMT setting is rare. Chest com-puted tomography (CT) scans havesubstantially improved the diagnostic sensitiv-ity and specificity, while there is increasingdebate about the role of routine chest radiogra-phy within 24 hours after the onset of fever inthe absence of localizing signs and symp-toms.199,200

Selected focal infections

Many sites of infection have been described infebrile neutropenia. The most frequent outsidethe lungs are oropharynx, skin/soft tissue andintravascular catheters, paranasal sinuses, gas-trointestinal tract/perianal area, and urinarytract. Unusual sites are the central nervous sys-tem/meninges, bones and joints, eyes and ears,heart, liver, biliary tract/pancreas, endocrineorgans, and lymph nodes. These unusual siteswhen involved are very often secondary sitesafter hematogenous dissemination.


Strict consensus definitions are not availablefor all of these focal infections, and a detailedcomparative epidemiologic analysis is thereforenot possible. An example is oropharyngealmucositis. Although there has been muchprogress in clinical grading of oral lesions,201 thequestion remains unanswered as to whethermucositis at any stage or only at more severestages should be considered an infection.Clinically, oral lesions in the neutropenicpatient with fever are diverse. They includeoropharyngeal candidiasis, buccal ulcers andnecrotizing ulcerative gingivitis (both sugges-tive of being caused by herpes simplex virus),soft palate ulcers, and necrotizing tonsillitis.Such lesions are common. In AML patients, forexample, their incidence approaches 100%, anddifferent entities often coexist. In the individualpatient, the grading of the lesions correlatesinversely with the neutrophil count. The inci-dence of urinary tract infections, in contrast,rarely exceeds 10%. Often, the diagnosis isbased solely on documentation of significantbacteriuria. Pyuria is not a reliable marker inthe neutropenic patient, and dysuria is rare.Individually, it is difficult to ascertain that thefever is, in fact, related to the bacteriuria.

Catheter-related infectionIntravascular catheter-related infection mayremain localized or may generalize afterhematogeous dissemination. Some epidemio-logic studies of catheter-related infection inneutropenic patients have analyzed only thefrequency and outcomes of bacteremic infec-tions related to central venous catheters. Theepidemiology of non-bacteremic catheter-related infections during neutropenia has beenless well studied. Definitions vary, even forcatheter-related bacteremia, and often a definitediagnosis cannot be established without micro-biologic examination of the explanted catheter.In a recent study, approximately half of thecatheter-related infections were non-bacteremic,including insertion site and tunnel infections.202

Up to 20% of all central venous catheters will bethe source of infection, depending on many fac-

tors such as intensity of catheter usage, infec-tion control measures during insertion andmanipulations, catheter type, and duration anddegree of neutropenia. Totally implantedcatheters are associated with fewer infectionsthan tunnelled catheters or non-tunnelled cen-tral venous catheters.203,204 For tunnelledcatheters, incidence rates range between 1 orless205,206 and about 5–7 per 1000 catheterdays.102,202,207,208 A large study of non-tunelledsilastic catheters found an infection rate of 1.3per 1000 catheter days.209 In children with ALL,the adjusted risk for infection (any type) wastwo- to fourfold higher when a central venouscatheter was in place.210

Neutropenic enterocolitis and antibiotic-associated colitis Acute abdomen is the clinical presentation ofacute neutropenic enterocolitis.211 This compli-cation typically begins 7–10 days afterchemotherapy, with fever, right lower quadrantor diffuse abdominal pain, and rebound tender-ness, sometimes with diarrhea or bloody diar-rhea. Patients without preceding chemotherapyare rarely involved. An incidence of about 1%was reported in autoPBSCT recipients.23 Amongacute leukemia patients, an incidence of 5–6%was reported;212,213 following taxane-basedchemotherapy, the incidence was 0.1%.214

Complications are pneumatosis intestinalis,bowel necrosis, perforation, peritonitis, andseptic shock. In a report from the 1980s, thecase-fatality rate appears to have been veryhigh (64%).215 More recently, rates of 45%213 orless212,214,216–218 were reported.

Neutropenic enterocolitis is a non-infectioustransmural inflammatory lesion following cyto-toxic chemotherapy, with secondary tissueinvasion, spread, and translocation of bacteria.In some cases, necrotizing infection due toclostridia has been implicated in the pathogene-sis;219 in others, necrotizing leukemic infiltratesin the bowel wall have been documented, andpseudomembranes or only mild mucosalinflammation have been found. Most fre-quently, the terminal ileum, the caecum


(‘typhilitis’), and ascending colon are involved.The cause of death is usually septic shock, withor without bacteremia. Occasionally, there isbowel perforation.

Clostridium difficile has become recognized asthe most frequent cause of antibiotic-associatedcolitis. This disease is characterized by acuteinflammation of the colonic mucosa, with forma-tion of macroscopic or microscopic pseudomem-branes. Atypical presentations are abdominaldistention, ascites, and hyperbilirubinemia with-out diarrhea. Asymptomatic carriage of theorganism has been described. Critical illnessleading to intensive care unit admission or deathcaused by the disease occur with a relative fre-quency of less than 5%. Although toxigenic C.difficile is the most frequent cause, occasionallyother pathogens have been implicated, notablyother clostridia, salmonellae, S. aureus, andKlebsiella oxytoca. The relative risks of specificantibacterial drugs inducing antibiotic-associ-ated colitis are not precisely known.Pretreatment with several cytotoxic drugs, suchas methotrexate or 5-fluorouracil, can also pre-dispose to development of the disease. Unlikeneutropenic enterocolitis, antibiotic-associatedcolitis most often affects the distal bowel. Morethan 80% of cases are nosocomially acquired.The clostridial spores can persist on fomites andsurfaces for several months. Nosocomial trans-mission via the hands of personnel or contami-nated environments is common. Among cancerpatients with diarrhea, 10–45% will have a posit-ive toxin test result and/or a positive culture ofthe organism.220–224 Fluoroquinolone prophylaxisduring neutropenia may lower the risk of devel-oping C. difficile colitis.225 The overall incidenceamong autoPBSCT patients in one recent studywas 7%.220 The outcome was good in all patients.Interestingly, there is an epidemiologic linkbetween C. difficile colitis and the emergence ofbacteremia due to VRE.226 A policy of initialreisolation after readmission of known carriers isable to reduce endemicity.227 Hospitalization,parenteral vancomycin in the last 2 months, andhigh-dose chemotherapy predicts an increasedrisk of C. difficile colitis.228

Perianal/perirectal cellulitisPerianal or perirectal cellulitis is rare.According to a recent observational study, theincidence among patients with acute leukemiawas 7%, and the case-fatality rate was 20%.229

The reported incidence among autoPBSCTpatients and alloBMT patients is much smaller(<1%).23,230 In the acute leukemia patients cohortof Ulm University Hospital, perianal cellulitiswas observed in 2% of febrile neutropenicepisodes (unpublished observations). Initially,the inflammation is characterized by painfulinduration and redness perianally. Progressioninto deeper tissue with involvement of theperirectal area and development of necroticlesions and fistula indicates a critical stage.Some patients may become bacteremic. E. coliand Ps. aeruginosa are the most common organ-isms involved, sometimes associated withanaerobes, enterococci, and CNS.

Invasive fungal infections

Based on several comparative trials evaluatingthe impact of prophylactic antifungals, the inci-dence of documented deep fungal infectionsamong patients with hematologic malignanciesgiven no systemic antifungal prophylaxis isabout 4–10% (Table 4.3).231–237 The rate is higheramong BMT patients. For example, 18% provensystemic fungal infections were reportedamong placebo recipients in a study conductedat the Fred Hutchinson Cancer Research Centerin Seattle.235 However, a significant proportionof invasive fungal infections in this patientpopulation occurred after engraftment (Figure4.5). The incidence of deep fungal infection dur-ing neutropenia usually does not exceed 10%even in this patient population. This contrastswith the frequent use of therapeutic ampho-tericin B and other antifungals, which is on theorder of about 20–25%, and 50% or more inBMT recipients. Reasons for the frequent use ofempiric or preemptive therapy are the poorprognosis of invasive fungal infections withcurrent therapeutic protocols and the inherent


difficulties in firmly establishing the diagnosis. One-half to two-thirds of the documented

infections among patients given no systemicprophylaxis are Candida infections; the othersare mould infections (Table 4.3). The latter arevery often termed ‘invasive aspergillosis’.However, many of these cases are not culture-confirmed, and Aspergillus cannot be differenti-ated histopathologically from most otherhyalohyphomycetes. Therefore, the term ‘inva-sive hyalohyphomycosis’ is preferable (Table4.4). The introduction of fluconazole into clini-cal practice has reduced the incidence of sys-temic Candida infections, and has shifted thespectrum of yeasts from C. albicans and C. tropi-calis to C. glabrata and C. krusei. As with bacterial


Table 4.3 Incidence of and mortality related to proven invasive fungal infections in patients withneutropenia. Data are from placebo arms of large, controlled clinical trials evaluating antifungalchemoprophylaxis

Proven invasive fungal infection

Authors Underlying disease Incidence No. of yeast/ Related(%) other fungal mortality

infections rate (%)

Goodman et al (1992)236 48% alloBMT 16 27/3 652% autoBMT

Slavin et al (1995)235 12% autoBMT 18 29/3 21a

88% alloBMT

Rotstein et al (1999)237 60% acute leukemia 17 22/1 544% autoBMT

Winston et al (1993)234 100% acute leukemia 8 7/4 0

Nucci et al (2000)233 80% acute leukemia 9 6/3 1

Menichetti et al (1999)232,b 76% acute leukemia 9 8/1 3

Harousseau et al (2000)231,b 70% acute leukemia 5 3/10 2

aMortality until day 110 after BMT.bPatients in the placebo arms received oral non-absorbable polyene antifungals.

20

0

Num

ber

of p

atie

nts

Days after BMT10

15

10

5

40 70 100 130 160

Autologous

Allogeneic

Figure 4.5 Time course of invasive aspergillosis inautologous and allogeneic BMT patients. Data arefrom Wald et al.276

infections, the spectrum of fungi causingsignificant infections in the neutropenic patienthas broadened to include rare and unusualorganisms that are often resistant to currentlyused antimicrobial drugs. Yeasts other than themore common Candida spp. listed above andmoulds other than Aspergillus, Fusarium (themost prevalent plant fungus worldwide),238,239

and zygomycetes (Figure 4.6) (Mucor spp. andrelated genera),240,241 however, are still rare inneutropenic patients. Rare and unusual fungimost recently reported include Trichosporon/Blastoschizomyces spp.,242–244 Candida dublinien-sis,245 Candida inconspicua,246 Paecilomyces lilacinus(coming from contaminated skin lotions),247

Metarrhizium anisopliae (used commercially forthe biocontrol of insects),248 Acremonium spp.,249

Scedosporium prolificans,250 Cunninghamellabertholletiae,251 Trichoderma longibrachiatum,252

Hormographiella aspergillata,253 Penicillium cit-rinum,254 Malassezia (associated with hyper-alimentation with intravenous lipids),dematiaceous (darkly pigmented) fungi causingphaeohyphomycoses (which are exceptional inthe setting of neutropenia), and some others.The list of unusual fungal pathogens is enlarg-ing much more rapidly in patients with chronicimmunodeficient states, including transplant

patients, than in patients with chemotherapy-induced (transient) neutropenia. The reason forthis is obvious – it is simply a result ofincreased exposure of the usually ambulatory,chronic immunosuppressed patient to particu-lar environments that may be specific habitatsof rare fungi.

FungemiaFungemia, most often candidemia, may origi-nate from gastrointestinal lesions, but also oftendevelops as an initial catheter-related infec-tion.255 C. parapsilosis and C. lusitaniae appearto be overrepresented in catheter-relatedfungemias. A risk factor for the development ofcandidemia is prior colonization at multiplemucosal sites.256 The risk of developing fungemiais much higher in patients colonized by C. tropi-calis than in those colonized by C. albicans.Another possible risk factor for fungemia is priorglycopeptide therapy and/or bacteremia.257–259

The case-fatality rate in many series of neu-tropenic patients exceeds 20%.258,260–263 Tissueinvasion and dissemination markedly decreasessurvival. Decreased survival has also beenobserved in infection due to C. glabrata and C.krusei when compared with other non-albicansspecies or C. albicans. The infection usually


Table 4.4 Selected agents of aspergillosis and of other hyalohyphomycoses causing invasive fungalinfection with similar histologic appearance in patients with neutropenia

Aspergillus spp. Fusarium spp. Penicillium spp.A. flavus F. chlamydosporum P. citrinumA. fumigatus F. moniliforme Scedosporium spp.A. nidulans F. oxysporum S. apiospermumA. niger F. solani S. prolificansA. ochraceus Neosartorya fischeri Scopulariopsis spp.A. terreus Paecilomyces spp. S. brevicaulisA. ustus P. lilacinus Scytalidium spp.

Acremonium spp. P. variotii S. dimidiatumEmmonsia spp. Hormographiella spp.

E. parva H. aspergillata

develops late after onset of neutropenia. It issometimes accompanied by non-specific respira-tory symptoms, myalgias, and characteristic skinand chorioretinal lesions.264 Some use the term‘acute disseminated candidiasis’ in such cases oftissue invasion. Severe sepsis and shock occurswith a similar frequency as in Gram-negativebacteremia.

Rare cases of fungemia are caused byTrichosporon, Rhodotorula, Fusarium, Malassezia,and others. Fusarium infections are exceptionalin that the organism is a filamentous fungus rel-atively often growing in blood culture whencausing disseminated infection. Hematogenousdissemination of Fusarium involves multiplesites and organs, including sinuses, lungs, skin,brain, bone, and joints. In contrast to aspergillo-sis, blood cultures are positive in 50% or moreof all cases. Fungemia due to other organisms isoften catheter-related, and may have a ratherbenign evolution after catheter removal.

Disseminated candidiasisSubacute or chronic disseminated candidiasis,formerly called ‘hepatosplenic candidiasis’, wasincreasingly diagnosed in the late 1980s andearly 1990s.265–268 In a Finnish center, incidence

rates increased from about 2% (1980–84) toabout 10% (1989–93) among patients with acuteleukemia.266 This can be explained partly byimproved imaging techniques; improved sur-vival of candidemia may be another reason.Strictly speaking, subacute or chronic disease isnot an infection of the neutropenic patient, butrather an infection after neutrophil recovery.The infection begins with fever during neu-tropenia, and the fever persists for weeksdespite empiric therapy (including antifungals)and neutrophil recovery. The host reaction isgranuloma formation, with a paucity ofmicroorganisms, and the organs predominantlyinvolved are liver and spleen. These chronicinfections are being seen much less frequentlysince the introduction of fluconazole into clini-cal practice.

AspergillosisMoulds of the genus Aspergillus are widespreadin the environment, being found in the air, soil,and water, and on plants, certain food, anddecomposing organic matter. Fungal sporecounts in outdoor air may be as high as100–1000 CFU per cubic meter of air.Pathogenic Aspergillus spp. represent about


Figure 4.6 Zygomycosis in aneutropenic patient. Necrotizingskin and soft tissue infectionextending into the tabulaexterna of the skull in a patientwith non-Hodgkin’s lymphoma,who was initially treated as anoutpatient. Absidia corymbiferawas isolated in pure culturefrom necrotic tissue.

1–10% of these spores. Inhalation of spores isthe major source of invasive infection. The levelof environmental contamination and host fac-tors determine the incidence of invasive infec-tion.269–271 Critical host factors are neutropeniaof long duration, graft-versus-host disease,and intensive immunosuppressive therapy.Macrophages and cellular immune responseplay an important role in controlling the infec-tion.272 AlloBMT patients are therefore at highrisk.273 Prior or concomitant bacteremia orGram-negative bacterial infection andcytomegalovirus disease increase the risk ofaspergillosis in these patients and/or worsenthe outcome.274,275 Among BMT patients, theonset of infection is bimodal. Peaks at 16 and 96days after transplant have been observed(Figure 4.5).276 For patients with early infection,underlying disease, donor type, season, andtransplant outside of air-filtered rooms wereassociated with significant risk for invasiveaspergillosis. The cumulative incidence amongalloBMT patients exceeds 10%. A. fumigatusremains the most frequent species, andaccounts for more than 50% of cases, A. flavus isthe second most frequent species. A. niger andA. terreus are isolated with increasedfrequency.277 The incubation time is unknown,and it is an open question how many casesbecoming manifest during admission are in factcommunity-acquired. Molecular typing of path-ogenic isolates and isolates from hospital envi-ronments suggest that half or more of theinfections may not be nosocomial.278–280 Theimpression of many physicians is that duringthe last decade, the incidence of invasiveaspergillosis has been increasing as a result ofless use of protected environments/reverse iso-lation for leukemia treatment, and of muchmore frequent discharges to home care.Nosocomial outbreaks have been described inassociation with environmental disturbances:hospital construction or construction in adja-cent areas; contaminated fireproofing materials,or air filters in the hospital ventilation system;contaminated carpeting.269,270,281–284 High-effi-ciency particulate air (HEPA) filtration in sealed

rooms with positive air pressure is protective insuch situations, leading to undetectable fungalspore counts (i.e. <0.1 CFU per cubic meter ofair).283 The potential for tap water and waterfrom shower heads to aerosolize molds needsto be studied in more detail.285,286

Clinical disease associated with invasiveaspergillosis in neutropenic patients includessinusitis, pulmonary infection, and dissemi-nated infection. Brain involvement is frequent(about 40–50% among alloBMT patients), andcarries a poor prognosis.287–290 Rare manifesta-tions include necrotizing cellulitis, necrotizingmucosal lesions, hematogenous osteitis (oftenspondylodiscitis), nephritis, and thyroiditis.The outcome is poor. Invasive infection limitedto the lungs responds to antifungal therapy inabout 30–50% of cases. The infection oftenbecomes chronic, and the patient remains athigh risk of relapse during subsequent neu-tropenic episodes or periods of increasedimmunosuppression.291,292 Among allogeneictransplant patients with an earlier invasiveaspergillosis, relapse rates of 40% have beenreported. Extensive granulomatous tissue reac-tions can be seen in chronic disease. A majorstep towards improved prognosis of patientswith invasive aspergillosis has been earlier pre-sumptive diagnosis by chest CT scan, whilescreening by serologic assays and nucleic acidamplification-based laboratory tests still await athorough evaluation of their cost–effectivenessagainst clinical and radiological assess-ment.293,294 A major diagnostic problem remainsthe difficulty in establishing a culture-con-firmed diagnosis. Bronchoalveolar lavage cul-tures in cases of pulmonary involvement arepositive in only 10–40% of cases. The paucity ofculture-confirmed cases with a majority of pre-sumptive cases or histologically diagnosed non-Aspergillus-specific hyalohyphomycoses may bea good reason for a more intense use of PCRand serological tests.


Viral infections during febrile neutropenia

Reactivation of latent herpesviruses, morespecifically of herpes simplex virus (HSV), is byfar the most common viral infection in adultpatients with neutropenia.295 Outside the bloodand marrow transplant setting, infections dueto varicella zoster virus (VZV), cytomegalovirus(CMV), Epstein–Barr virus (EBV), and otherherpesviruses (human herpesvirus (HHV)-6,HHV-7, and HHV-8) are uncommon, and therisk of reactivation of one of these is muchsmaller in autoPBSCT patients than in alloBMTpatients. Patients treated with fludarabine orother purine analogs, however, may form a spe-cific subgroup with a different risk profile. In arecent case report, for example, an AML patientdeveloped EBV-associated lymphoproliferativedisease after fludarabine/cytosine arabi-noside/granulocyte colony-stimulating factorchemotherapy, which is a very unusual compli-cation outside the alloBMT setting. Thepatient´s leukemia was in remission, and thepatient was no longer neutropenic at thattime.296 Also, in the autoPBSCT setting, patientswith CD34� cell-selected transplants (which areassociated with more pronounced CD4 lym-phopenia compared with unselected products)may have a significantly higher risk of viralinfection, specifically CMV infection.297

Primary infections with other viruses, suchas adenovirus, influenza and parainfluenzaviruses, respiratory syncytial virus (RSV), andothers, do occur. They usually reflect the gen-eral epidemiologic situation and exposurecharacteristics, rather than indicating a particu-lar risk specific to neutropenia (season, crowd-ing, etc.). It appears plausible that the severityof these infections among neutropenic patientsoutside the BMT setting is substantially differ-ent from that seen in the general population.The data supporting this view, however, areconflicting. Active untreated RSV infections, forexample, need not consistently lead to severecomplications during neutropenia, even amongautoPBSCT recipients.298 On the other hand,there are reports of outbreaks of severe infec-

tions among neutropenic patients, and neu-tropenic patients with RSV infection may be atincreased risk of developing pneumonia (asopposed to tracheobronchitis and upper respi-ratory tract infection) and of death.299–302 In theMD Anderson Cancer Center study, not alldeaths, however, were definitely related to theRSV infection.301

HSV reactivation during neutropenia is com-mon. Studies in acute leukemia patients haveshown that among seropositive subjects, about25–50% will have shedding of the virus in themouth, associated with herpes labialis and/orulcerative gingivostomatitis.42,303,304 The inci-dence is higher among alloBMT patients (about70% or more), reflecting the more severemucosal damage in this setting and additionalimmunosuppression. HSV lesions in neutro-penic patients may be severe and long-lasting.There has been some discussion of the role ofHSV lesions in facilitating entry of bacterial andfungal mircoorganisms into the bloodstream.Esophagitis is a significant complication, buthas become rare because of prompt therapy oforopharyngeal lesions. Other organs are veryrarely involved, although there have beenreports of HSV pneumonitis.

Complications due to other herpesviruses areunusual during neutropenia, and often occurafter neutrophil recovery in auto PBSCT oralloBMT patients. The most frequent have beenVZV infections, and the most problematic hasbeen CMV disease. CMV viremia or antigene-mia has been documented in about 3% ofautoPBSCT recipients versus about 60% ofalloBMT recipients. In one report, the incidenceof fatal CMV interstitial pneumonia inautoPBSCT patients was 0.8%.24 CMV disease inalloBMT patients before engraftment has beendocumented in a few cases.305 Most often, thelung has been the primary site of infection, but,in some cases, the histopathologic appearance ofthe lesions was atypical and/or significantcopathogens could be identified. Only occasion-ally, patients with hematologic malignancy orsolid tumors develop CMV disease during neu-tropenia – most often pneumonia and colitis.306,307


VZV infections have been surprisingly common(10–25%) in several series of autoPBSCTpatients.24,308 Clinical disease from reactivationduring neutropenia was probably suppressedby prophylactically administered acyclovir, andovert infection was not observed until afterneutrophil recovery (and discontinuation ofacyclovir). In children with ALL, the risk ofhaving chickenpox or herpes zoster is clearlyincreased.309 Severe cases with minimal skinand extensive extracutaneous involvement(lungs, liver, spleen, or central nervous system)have been reported.310 HHV-6 and HHV-7 arerecently discovered �-herpesviruses. DNA canbe detected in blood or bone marrow fromhealthy subjects.311,312 Infection with HHV-6 isvery common, approaching 100% in seropreva-lence. The virus appears to persist in low levelsin cells and tissues, and can reactivate duringcytotoxic chemotherapy and/or immunosup-pressive therapy. It has been associated withlymphopenia, exanthema, and hepatopathy inchildren with cancer.313 In BMT patients, bonemarrow suppression, interstitial pneumonitis,and encephalitis have been reported as compli-cations of HHV-6 reactivation.314–316 Primaryinfections with self-limited clinical symptomsand asymptomatic reactivation have also beenwell documented in alloBMT recipients.317

SUMMARY

The epidemiology of infections in neutropeniccancer patients is complex and undergoes peri-odic changes. A number of factors influencethe spectrum and severity of infection, includ-ing the underlying malignancy and associatedimmunologic deficits, geographic and local fac-tors, the use of strategies such as chemoprophy-laxis, the increasing use of catheters and otherforeign objects, and the widening applicationsof hematopoietic stem cell transplantation.Newer, opportunistic pathogens will continueto emerge, and widespread resistance amongbacterial, fungal, and viral pathogens will con-tinue to be a significant problem. Constant vig-

ilance, in order to detect epidemiologic shiftsearly, is essential. Rapid and more specificdiagnostic methods need to be developed,along with less immunosuppressive and myelo-suppressive antineoplastic treatment mod-alities. Until this happens, infections willcontinue to be a challenge in this unique groupof patients.

REFERENCES

1. Zembower T, Epidemiology of infectious com-plications in cancer patients. Cancer Treat Res1998; 96: 33–75.

2. Pizzo PA, Management of fever in patients withcancer and treatment-induced neutropenia.N Engl J Med 1993; 328: 1323–31.

3. Klastersky J, Science and pragmatism in thetreatment and prevention of neutropenic infec-tion. J Antimicrob Chemother 1998; 41(Suppl D):13–24.

4. Nucci M, Spector N, Bueno AP et al, Risk fac-tors and attributable mortality associated withsuperinfections in neutropenic patients withcancer. Clin Infect Dis 1997; 24: 575–9.

5. Engels EA, Ellis CA, Supran SE et al, Earlyinfection in bone marrow transplantation:quantitative study of clinical factors that affectrisk. Clin Infect Dis 1999; 28: 256–66.

6. Paesmans M, Factors predicting mortalityamong febrile neutropenic cancer patientstreated within a clinical trial: ten yearsexperience by the EORTC InternationalAntimicrobial Therapy Cooperative Group. In:Proceedings of 3rd International Symposium onFebrile Neutropenia, Brussels, Belgium, 1997: Abst97.

7. Chanock SJ, Pizzo PA, Infectious complicationsof patients undergoing therapy for acuteleukemia: current status and future prospects.Semin Oncol 1997; 24: 132–40.

8. Jud B, Gmür J, Follath F, Art und Häufigkeitschwerer Infektkomplikationen bei derBehandlung akuter Leukämien. Schweiz MedWochenschr 1994; 124: 2060–3.

9. Gamis AS, Howells WB, de Swarte-Wallace J etal, Alpha-hemolytic streptococcal infection dur-ing intensive treatment for acute myeloidleukemia: a report from the children´s cancer


group study CCG-2891. J Clin Oncol 2000; 18:1845–55.

10. Seipelt G, Hofmann WK, Martin H et al,Comparison of toxicity and outcome in patientswith acute myeloid leukemia treated with high-dose cytosine-arabinoside consolidation afterinduction with a regimen containing idarubicinor daunorubicin. Ann Hematol 1998; 76: 145–51.

11. Auletta JJ, O’Riordan MA, Nieder ML,Infections in children with cancer: a continuedneed for the comprehensive physical examina-tion. J Pediatr Hematol Oncol 1999; 21: 501–8.

12. Rahiala J, Perkkio M, Riikonen P, Infectionsoccurring during the courses of anticancerchemotherapy in children with ALL: a retro-spective analysis of 59 patients. Pediatr HematolOncol 1998; 15: 165–74.

13. Hurwitz CA, Silverman LB, Schorin MA et al,Substituting dexamethasone for prednisonecomplicates remission induction in childrenwith acute lymphoblastic leukemia. Cancer2000; 88: 1964–9.

14. Hann I, Viscoli C, Paesmans M et al, A compari-son of outcome from febrile neutropenicepisodes in children compared with adults.Results from four EORTC studies. Br J Haematol1997; 99: 580–8.

15. Bow EJ, Infection risk and cancer chemother-apy: the impact of the chemotherapeutic regi-men in patients with lymphoma and solidtissue malignancies. J Antimicrob Chemother1998; 41(Suppl D): 1–5.

16. Potter M, New anti-cancer therapies, newopportunities for infection. Curr Opin Infect Dis1999; 12: 359–63.

17. Morra E, Nosari A, Montillo M, Infectious com-plications in chronic lymphocytic leukaemia.Hematol Cell Ther 1999; 41: 145–51.

18. Carlson JW, Fowler JM, Mitchell SK et al,Chemoprophylaxis with ciprofloxacin in ovar-ian cancer patients receiving paclitaxel: a ran-domized trial. Gynecol Oncol 1997; 65: 325–9.

19. Hartlapp JH, Antimicrobial prophylaxis inimmunocompromised patients. Drugs 1987;34(Suppl 1): 131–3.

20. Counsell R, Pratt J, Williams MV,Chemotherapy for germ cell tumors: prophylac-tic ciprofloxacin reduces the incidence of neu-tropenic fever. Clin Oncol R Coll Radiol 1994; 6:232–6.

21. Holland HK, Dix SP, Geller RB et al, Minimal

toxicity and mortality in high-risk breast cancerpatients receiving high-dose cyclophos-phamide, thiotepa, and carboplatin plus autolo-gous marrow/stem-cell transplantation andcomprehensive supportive care. J Clin Oncol1996; 14: 1156–64.

22. Mossad SB, Longworth DL, Goormastic M et al,Early infectious complications in autologousbone marrow transplantation: a review of 219patients. Bone Marrow Transplant 1996; 18:265–71.

23. Salazar R, Sola C, Maroto P et al, Infectiouscomplications in 126 patients treated with high-dose chemotherapy and autologous peripheralblood stem cell transplantation. Bone MarrowTransplant 1999; 23: 27–33.

24. Offidani M, Corvatta L, Olivieri A et al,Infectious complications after autologousperipheral blood progenitor cell transplantationfollowed by G-CSF. Bone Marrow Transplant1999; 24: 1079–87.

25. Seropian S, Nadkarni R, Jillela AP et al,Neutropenic infections in 100 patients withnon-Hodgkin lymphoma or Hodgkin´s diseasetreated with high-dose BEAM chemotherapyand peripheral blood progenitor cell transplant:outpatient treatment is a viable option. BoneMarrow Transplant 1999; 23: 599–605.

26. Eucker J, Sezer O, Metzner B et al, Factorsinfluencing the rate of infectious complicationsin high-dose chemotherapy with autologousperipheral stem cell transplantation. SupportCare Cancer 2000; 8: 160 (Abst 61).

27. Herrmann RP, Trent M, Cooney J, Cannell PK,Infections in patients managed at home duringautologous stem cell transplantation for lym-phoma and multiple myeloma. Bone MarrowTransplant 1999; 24: 1213–17.

28. Meisenberg B, Gollard R, Brehm T et al,Prophylactic antibiotics eliminate bacteremiaand allow safe outpatient management follow-ing high-dose chemotherapy and autologousstem cell rescue. Support Care Cancer 1996; 4:364–9.

29. Westermann AM, Holtkamp MM, Linthorst GAet al, At home management of aplastic phasefollowing high-dose chemotherapy with stem-cell rescue for hematological and non-hemato-logical malignancies. Ann Oncol 1999; 10: 511–17.

30. Mullen CA, Nair J, Sandesh S, Chan KW, Feverand neutropenia in pediatric hematopoietic


stem cell transplant patients. Bone MarrowTransplant 2000; 25: 59–65.

31. Busca A, Saroglia EM, Giacchino M et al,Analysis of early infectious complications inpediatric patients undergoing bone marrowtransplantation. Support Care Cancer 1999; 7:253–9.

32. van Burik JA, Weisdorf DJ, Infections in recipi-ents of blood and marrow transplantation.Hematol Oncol Clin North Am 1999; 13: 1065–89.

33. Oguma S, Yoshida Y, Uchino H et al, Infectionin myelodysplastic syndromes before evolutioninto acute non-lymphoblastic leukemia. Int JHematol 1994; 60: 129–36.

34. Cunningham I, MacCallum SJ, Nicholls MD etal, The myelodysplastic syndromes: an analysisof prognostic factors in 226 cases from a singleinstitution. Br J Haematol 1995; 90: 602–6.

35. Higuchi T, Mori H, Niikura H et al, Prognosticimplications in myelodysplastic syndromes: areview of 62 cases. Leuk Lymphoma 1996; 21:479–84.

36. Sievers EL, Dale DC, Non-malignant neutrope-nia. Blood Rev 1996; 10: 95–100.

37. Hermans P, Sommereijns B, van Cutsem N,Clumeck N, Neutropenia in patients with HIVinfection: a case control study in a cohort of1403 patients between 1982 and 1993. JHematother Stem Cell Res 1999; 8(Suppl 1):S23–32.

38. Caperna J, Barber RE, Toerner JG, MathewsWC, Estimation of the effect of neutropenia onrates of clinical bacteremia in HIV-infectedpatients. Epidemiol Infect 1998; 120: 71–80.

39. Cruciani M, Rampazzo R, Malena M et al,Prophylaxis with fluoroquinolones for bacterialinfections in neutropenic patients: a meta-analysis. Clin Infect Dis 1996; 23: 795–805.

40. Engels E, Lau J, Barza M, Efficacy of quinoloneprophylaxis in neutropenic cancer patients: ameta-analysis. J Clin Oncol 1998; 16: 1179–87.

41. International Antimicrobial Therapy Co-operative Group of the EORTC, Reduction offever and streptococcal bacteremia in granulo-cytopenic patients with cancer. A trial of oralpenicillin V or placebo combined withpefloxacin. JAMA 1994; 272: 1183–9.

42. Bergmann OJ, Mogensen SC, Ellermann-Eriksen S, Ellegaard J, Acyclovir prophylaxisand fever during remission induction therapyof patients with acute myeloid leukemia: a ran-

domized, double-blind, placebo-controlled trial.J Clin Oncol 1997; 15: 2269–74.

43. Lopez A, Soler JA, Julia A et al, Prophylaxiswith ciprofloxacin in postchemotherapy neu-tropenia in acute myeloid leukemia. Med ClinBarc 1994; 102: 81–5.

44. Kern W, Kurrle E, Ofloxacin versus trimetho-prim-sulfamethoxazole for infection preventionin acute leukemia. Infection 1991; 19: 73–80.

45. Hidalgo M, Hornedo J, Lumbreras C et al, Lackof ability of ciprofloxacin–rifampin prophylaxisto decrease infection related morbidity in neu-tropenic patients given cytotoxic therapy andperipheral blood stem cell transplants.Antimicrob Agents Chemother 1997; 41: 1175–7.

46. Gomez-Martin C, Sola C, Hornedo J et al,Rifampin does not improve the efficacy ofquinolone antibacterial prophylaxis in neu-tropenic cancer patients: results of a random-ized trial. J Clin Oncol 2000; 18: 2126–34.

47. Ley BE, Linton CJ, Bennett DM et al, Detectionof bacteremia in patients with fever and neu-tropenia using 16S rRNA gene amplification bypolymerase reaction. Eur J Clin Microbiol InfectDis 1998; 17: 247–53.

48. De Bont ES, Vellenga E, Swaanenburg JC et al,Plasma IL-8 and IL-6 levels can be used todefine a group with low risk of septicemiaamong cancer patients with fever and neutrope-nia. Br J Haematol 1999; 107: 375–80.

49. Engel A, Mack E, Kern P, Kern WV, An analysisof interleukin-6, interleukin-8 and C-reactiveprotein serum concentrations to predict fever,gram-negative bacteremia and complicatedinfection in neutropenic cancer patients.Infection 1998; 26: 213–19.

50. Yamakami Y, Hashimoto A, Yamagata E et al,Evaluation of PCR for detection of DNA spe-cific for Aspergillus species in sera of patientswith various forms of pulmonary aspergillosis.J Clin Microbiol 1998; 36: 3619–23.

51. Hebart H, Löffler J, Meisner C et al, Early detec-tion of Aspergillus infection after allogeneic stemcell transplantation by polymerase chain reac-tion screening. J Infect Dis 2000; 181: 1713–19.

52. Skladny H, Buchheidt D, Baust C et al, Specificdetection of Aspergillus species in blood andbronchoalveolar lavage samples of immuno-compromised patients by two step PCR. J ClinMicrobiol 1999; 37: 1846–51.

53. Kern W, Kirchner S, Vanek E, Resin versus


standard blood culture medium used with thenew BACTEC automated infrared system. Anevaluation in febrile granulocytopenic patients.Zentralbl Bakteriol 1990; 273: 156–63.

54. Rozdzinski E, Feldkamp M, Schmeiser T, KernW, Evaluation of the BACTEC high volumeresin blood culture media in febrile neutropenicpatients. Med Microbiol Lett 1992; 1: 213–19.


56. Kern WV, Cometta A, de Bock R et al, Oral ver-sus intravenous empiric therapy for fever inpatients with granulocytopenia who are receiv-ing cancer chemotherapy. N Engl J Med 1999;341: 312–18.

57. Immunocompromised Host Society, Thedesign, analysis and reporting of clinical trialson the empirical antibiotic management of theneutropenic patient. J Infect Dis 1990; 161:397–401.

58. Kiehn TE, Armstrong D, Changes in the spec-trum of organisms causing bacteremia andfungemia in immunocompromised patients dueto venous access devices. Eur J Clin MicrobiolInfect Dis 1990; 9: 869-872.

59. Gunther G, Bjorkholm M, Bjorklind A et al,Septicemia in patients with hematological dis-orders and neutropenia. A retrospective studyof causative agents and their resistance profile.Scand J Infect Dis 1991; 23: 589–98.

60. Zinner SH, Changing epidemiology of infec-tions in patients with neutropenia and cancer:emphasis on Gram-positive and resistant bacte-ria. Clin Infect Dis 1999; 29: 490–4.

61. Gonzalez-Barca E, Fernandez-Sevilla A,Carratala J et al, Prospective study of 288episodes of bacteremia in neutropenic cancerpatients in a single institution. Eur J ClinMicrobiol Infect Dis 1996; 15: 291–6.

62. Funada H, Matsuda T, Changes in the incidenceand etiological patterns of bacteremia associ-ated with acute leukemia over a 25-year period.Intern Med 1998; 37: 1014–18.

63. Marie JP, Vekhoff A, Pico JL et al, Neutropenicinfections: a review of the French FebrileAplasia Study Group trials in 608 febrile neu-tropenuic patients. J Antimicrob Chemother 1998;41(Suppl D): 57–64.

64. Paganini H, Bologna R, Debbag R et al, Feverand neutropenia in children with cancer in onepediatric hospital in Argentina. Pediatr HematolOncol 1998; 15: 405–13.

65. Wehl G, Allerberger F, Heitger A et al, Trendsin infection morbidity in a pediatric oncologyward, 1986–1995. Med Pediatr Oncol 1999; 32:336–43.

66. Rolston KVI, Balakrishnan M, Elting L, TarrandJJ, Is quantitative variation in cancer patientswith bacteremic infections linked to severity ofinfection? In: Proceedings of 40th InterscienceConference on Antimicrobial Agents andChemotherapy. Washington, DC: AmericanSociety for Microbiology, 2000: 473 (Abst 729).

67. Yuen KY, Woo PC, Hui CH, Unique risk factorsfor bacteraemia in allogeneic bone marrowtransplant recipients before and after engraft-ment. Bone Marrow Transplant 1998; 21: 1137–43.

68. Pagano L, Tacconelli E, Tumbarello M et al,Bacteremia in patients with hematologicalmalignancies. Analysis of risk factors, etiologi-cal agents and prognostic indicators.Haematologica 1997; 82: 415–19.

69. Viscoli C, Bruzzi P, Castagnola E, Factors asso-ciated with bacteremia in febrile granulocy-topenic cancer patients. Eur J Cancer 1994; 30A:430–7.

70. Viscoli C, Paesmans M, Langenaeken J et al,Association between antifungal prophylaxisand occurrence of bacteremia in febrilenetropenic cancer patients. In: Proceedings of37th Interscience Conference on AntimcrobialAgents and Chemotherapy. Washington, DC:American Society for Microbiology, 1997; 380(Abst LM-83).

71. Rackoff WR, Gonin R, Robinson C et al,Predicting the risk of bacteremia in childrenwith fever and neutropenia. J Clin Oncol 1996;14: 919–24.

72. Klaassen RJ, Goodman TR, Pham B, Doyle JJ,Low-risk prediction rule for pediatric oncologypatients presenting with fever and neutropenia.J Clin Oncol 2000; 18: 1012–19.

73. Bernard L, Ferriere F, Casassus P et al,Procalcitonin as an early marker of bacterialinfection in severely neutropenic febrile adults.Clin Infect Dis 1998; 27: 914–15.

74. Engel A, Steinbach G, Kern P, Kern WV,Diagnostic value of procalcitonin serum levelsin neutropenic patients with fever: comparison


with interleukin-8. Scand J Infect Dis 1999; 31:185–9.

75. Ruokonen E, Nousianen T, Pulkki K, Takala J,Procalcitonin concentrations in patients withneutropenic fever. Eur J Clin Microbiol Infect Dis1999; 18: 283–5.

76. Lehrnbecher T, Venzon D, de Haas M et al,Assessment of measuring circulating levels ofinterleukin-6, interleukin-8, C-reactive protein,soluble Fc-gamma receptor type III, and man-nose-binding protein in febrile children withcancer and neutropenia. Clin Infect Dis 1999; 29:414–19.

77. Kern WV, Heiss M, Steinbach G, Prediction ofgram-negative bacteremia in patients withcancer and febrile neutropenia by means ofinterleukin-8 levels in serum: targeting empiri-cal monotherapy versus combination therapy.Clin Infect Dis 2001; 32: 832–5.

78. Gonzalez-Barca E, Fernandez-Sevilla A,Carratala J et al, Prognostic factors influencingmortality in cancer patients with neutropeniaand bacteremia. Eur J Clin Microbiol Infect Dis1999; 18: 539–44.

79. Aledo A, Heller G, Ren L et al, Septicemia andseptic shock in pediatric patients: 140 consecu-tive cases on a pediatric hematology-oncologyservice. J Pediatr Hematol Oncol 1998; 20: 215–21.

80. Kern W, Kurrle E, Schmeiser T, Streptococcalbacteremia in adult patients with leukemiaundergoing aggressive chemotherapy. Areview of 55 cases. Infection 1990; 18: 138–45.

81. Elting LS, Rubenstein EB, Rolston KV, BodeyGP, Outcomes of bacteremia in patients withcancer and neutropenia: observations from twodecades of epidemiological and clinical trials.Clin Infect Dis 1997; 25: 247–59.

82. Carratala J, Roson B, Fernandez-Sevilla A et al,Bacteremic pneumonia in neutropenic patientswith cancer: causes, empirical antibiotic ther-apy, and outcome. Arch Intern Med 1998; 158:868–72.

83. Link H, Maschmeyer G, Meyer P et al,Interventional antimicrobial therapy in febrileneutropenic patients. Study Group of the PaulEhrlich Society for Chemotherapy. Ann Hematol1994; 69: 231–43.

84. Viscoli C, Castagnola E, Giacchino M et al,Bloodstream infections in children with cancer:a multicentre surveillance study of the ItalianAssociation of Paediatric Haematology and

Oncology. Supportive Therapy Group –Infectious Diseases Section. Eur J Cancer 1999;35: 770–4.

85. Bochud PY, Calandra T, Francioli P, Bacteremiadue to viridans streptococci in neutropenicpatients: a review. Am J Med 1994; 97: 256–64.

86. Van der Lelie H, van Ketel RJ, von dem BorneAEGK et al, Incidence and clinical epidemiol-ogy of streptococcal septicemia during treat-ment of acute myeloid leukemia. Scand J InfectDis 1991; 23: 163–8.

87. Donnelly JP, Dompeling EC, Meis JF, de PauwBE, Bacteremia due to oral viridans streptococciin neutropenic patients with cancer: cytostaticsare a more important risk factor than antibacte-rial prophylaxis. Clin Infect Dis 1995; 20: 469–70.

88. Richard P, Amador Del Valle G, Moreau P et al,Viridans streptococcal bacteraemia in patientswith neutropenia. Lancet 1995: 345: 1607–9.

89. Villablanca JG, Steiner M, Kersey J et al, Theclinical spectrum of infections with viridansstreptococci in bone marrow transplantpatients. Bone Marrow Transplant 1990; 6:387–93.

90. Martino R, Subira M, Manteiga R et al, Viridansstreptococcal bacteremia and viridans strepto-coccal shock syndrome in neutropenic patients:comparison between children and adultsreceiving chemotherapy or undergoing bonetransplantation. Clin Infect Dis 1995; 20: 476–7.

91. Martino R, Manteiga R, Sanchez I et al, Viridansstreptococcal shock syndrome during bonemarrow transplantation. Acta Haematol 1995; 94:69–73.

92. Ruescher T, Sodeifi A, Scrivani SJ et al, Theimpact of mucositis on alpha-hemolytic strepto-coccal infection in patients undergoing autolo-gous bone marrow transplantation forhematologic malignancies. Cancer 1998; 82:2275–81

93. Bilgrami S, Feingold JM, Dorsky D et al,Streptococcus viridans bacteremia followingautologous peripheral blood stem cell trans-plantation. Bone Marrow Transplant 1998; 21:591–5.

94. Engel A, Kern P, Kern WV, Cytokines andcytokine inhibitors in the neutropenic patientwith �-streptococcal shock syndrome. ClinInfect Dis 1996; 23: 785–9.

95. Carratala J, Alcaide F, Fernandez-Sevilla A et al,Bacteremia due to viridans streptococci that are


highly resistant to penicillin: increase amongneutropenic patients with cancer. Clin Infect Dis1995; 20: 1169–73.

96. Bochud PY, Eggiman P, Calandra T et al,Bacteremia due to viridans streptococcus inneutropenic patients with cancer: clinical spec-trum and risk factors. Clin Infect Dis 1994; 18:25–31.

97. Carratala J, Marron A, Fernandez-Sevilla A etal, Treatment of penicillin-resistant pneumococ-cal bacteremia in neutropenic patients withcancer. Clin Infect Dis 1997; 24: 148–52.

98. Raad II, Bodey GP, Infectious complications ofindwelling vascular catheters. Clin Infect Dis1992; 15: 197–210.

99. Raad I, Alrahwan A, Rolston K. Staphylococcusepidermidis. Clin Infect Dis 1998; 26: 1182–7.

100. Nouwen JL, Wielenga JJ, van Overhagen Het al, Hickman catheter-related infections inneutropenic patients: insertion in the operatingtheater versus insertion in the radiology suite. J Clin Oncol 1999; 17: 1304.

101. Castagnola E, Garaventa A, Viscoli C et al,Changing pattern of pathogens causing broviaccatheter-related bacteraemias in children withcancer. J Hosp Infect 1995; 29: 129–33.

102. Elishoov H, Or R, Strauss N, Engelhard D,Nosocomial colonization, septicemia andHickman/Broviac catheter-realted infections inbone marrow transplant recipients. Medicine1998; 77: 83–101.

103. Groeger JS, Lucas AB, Thaler HAT et al,Infectious morbidity associated with long-termuse of venous access devices in patients withcancer. Ann Intern Med 1993; 119: 1168–74.

104. Lina B, Forey F, Troncy J et al, Oral source ofStaphylococcus epidermidis septicemia in a neu-tropenic patient. Eur J Clin Microbiol Infect Dis1994; 13: 773–5.

105. Kennedy HF, Morrison D, Kaufmann ME et al,Origins of Staphylococcus epidermidis andStreptococcus oralis causing bacteraemia in abone marrow transplant patient. J Med Microbiol2000; 49: 367–70.

106. Nouwen JL, van Belkum A, de Marie S et al,Clonal expansion of Staphylococcus epidermidisstrains causing Hickman catheter-related infec-tions in a hemato-oncologic department. J ClinMicrobiol 1998; 36: 2696–702.

107. Lyytikainen O, Valtonen V, Sivonen A et al,Molecular epidemiology of Staphylococcus epi-

dermidis isolates in a hematological unit duringa 4-month survey. Scand J Infect Dis 1995; 27:575–80.

108. Jones RN, Contemporary antimicrobial suscep-tibility patterns of bacterial pathogens com-monly associated with febrile patients withneutropenia. Clin Infect Dis 1999; 29: 495–502.

109. Feld R. Vancomycin as part of initial empiricalantibiotic therapy for febrile neutropenia inpatients with cancer: pros and cons. Clin InfectDis 1999; 29: 503–7.

110. Morrison VA, Peterson BA, Bloomfield CD,Nosocomial septicemia in the cancer patient:the influence of central venous access devices,neutropenia, and type of malignancy. MedPediatr Oncol 1990; 18: 209–16.

111. Escande MC, Herbrecht R, Prospective study ofbacteraemia in cancer patients. Results of aFrench multicentre study. Support Care Cancer1998; 6: 273–80.

112. Rubio M, Palau L, Vivas JR et al, Predominanceof Gram-positive microorganisms as a cause ofsepticemia in patients with hematologicalmalignancies. Infect Control Hosp Epidemiol 1994;15: 101–4.

113. Gopal AK, Fowler VGJr, Shah M et al,Prospective analysis of Staphylococcus aureusbacteremia in nonneutropenic adults withmalignancy. J Clin Oncol 2000; 18: 1110–15.

114. Rolston KV, Tarrand JJ. Pseudomonas aeruginosa– still a frequent pathogen in patients withcancer: 11-year experience at a comprehensivecancer center. Clin Infect Dis 1999; 26: 463–4.

115. Chatzinikolaou I, Abi-Said D, Bodey GP et al,Recent experience with Pseudomonas aeruginosabacteremia in patients with cancer:Retrospective analysis of 245 episodes. ArchIntern Med 2000; 160: 501–9.

116. Todeschini G, Franchini M, Tecchio C et al,Improved prognosis of Pseudomonas aeruginosabacteremia in 127 consecutive neutropenicpatients with hematologic malignancies. Int JInfect Dis 1998–99; 3: 99–104.

117. Coullioud D, van der Auwera P, Viot M, LassetC, Prospective multicentric study of the etiol-ogy of 1051 bacteremic episodes in 782 cancerpatients. CEMIC (French–Belgian Study Club ofInfectious Diseases in Cancer). Support CareCancer 1993; 1: 34–46.

118. Martinelli G, Alessandrino EP, Bernasconi P etal, Fournier´s gangrene: a clinical presentation


of necrotizing fasciitis after bone marrow trans-plantation. Bone Marrow Transplant 1998; 22:1023–6.

119. Kern W, Rozdzinski E, Schmeiser T et al,Emerging bacterial pathogens in patients withacute leukemia: viridans streptococci and non-fermentative gram-negative bacilli. In: AcuteLeukemias IV – Prognostic Factors and TreatmentStrategies (Büchner T, Hiddemann W, WörmannB et al, eds). Berlin: Springer-Verlag, 1994:795–8.

120. Martino R, Martinez C, Pericas R et al,Bacteremia due to glucose non-fermentinggram-negative bacilli in patients with hemato-logical neoplasias and solid tumors. Eur J ClinMicrobiol Infect Dis 1996; 15: 610–15.

121. Martino R, Santamaria A, Munoz L et al,Bacteremia by Gram-negative bacilli in patientswith hematologic malignancies. Comparison ofthe clinical presentation and outcome on infec-tions by enterobacteria and non-glucose-fer-menting Gram-negative bacilli. Acta Haematol1999; 102: 7–11.

122. Castagnola E, Conte M, Venzano P et al,Broviac catheter-related bacteraemias due tounusual pathogens in children with cancer: casereports with literature review. J Infect 1997; 34:215–18.

123. Williamson EC, Millar MR, Steward CG et al,Infections in adults undergoing unrelateddonor bone marrow transplantation. Br JHaematol 1999; 104: 560–8.

124. Micozzi A, Venditti M, Monaco M et al,Bacteremia due to Stenotrophomonas maltophiliain patients with hematologic malignancies. ClinInfect Dis 2000; 31: 705–11.

125. Edmond MB, Riddler SA, Baxter CM et al,Agrobacterium radiobacter: a recently recognizedopportunistic pathogen. Clin Infect Dis 1993; 16:388–91.

126. Hernandez JA, Martino R, Pericas R et al,Achromobacter xylosoxidans bacteremia inpatients with hematologic malignancies.Haematologica 1998; 83: 284–5.

127. Sader HS, Jones RN, Pfaller MA. Relapse ofcatheter-related Flavobacterium meningosepticumbacteremia demonstrated by DNA macrore-striction analysis. Clin Infect Dis 1995; 21:997–1000.

128. Hornei B, Luneberg E, Schmidt-Rotte H et al,Systemic infection of an immunocompromised

patient with Methylobacterium zatmanii. J ClinMicrobiol 1999; 37: 248–50.

129. Fernandez M, Dreyer Z, Hockenberry-Eaton M,Baker CJ, Methylobacterium mesophilica as acause of persistent bacteremia in a child withlymphoma. Pediatr Infect Dis J 1997; 16: 1007–8.

130. Brown MA, Greene JN, Sandin RL et al,Methylobacterium bacteremia after infusion ofcontaminated autologous bone marrow. ClinInfect Dis 1996; 23: 1191–2.

131. Duggan JM, Goldstein SJ, Chenoweth CE et al,Achromobacter xylosoxidans bacteremia: report offour cases and review of the literature. ClinInfect Dis 1996; 23: 569–76.

132. Knippschild M, Schmid EN, Uppenkamp M etal, Infection by Alcaligenes xylosoxidans subsp.xylosoxidans in neutropenic patients. Oncology1996; 53: 258–62.

133. Kern WV, Oethinger M, Kaufhold A et al,Ochrobactrum anthropi bacteremia: report of fourcases and short review. Infection 1993; 21:306–10.

134. Yu WL, Wang DY, Lin CW, Tsou MF, EndemicBurkholderia cepacia bacteraemia: clinical fea-tures and antimicrobial susceptibilities of iso-lates. Scand J Infect Dis 1999; 31: 293–8.

135. Salazar R, Martino R, Sureda A et al, Catheter-related bacteremia due to Pseudomonas paucimo-bilis in neutropenic cancer patients: report oftwo cases. Clin Infect Dis 1995; 20: 1573–4.

136. Bloch KC, Nadarajah R, Jacobs R,Chryseobacterium meningosepticum: an emergingpathogen among immunocompromised adults.Report of 6 cases and literature review. Medicine1997; 76: 30–41.

137. Labarca JA, Leber AL, Kern VL et al, Outbreakof Stenotrophomonas maltophilia bacteremia inallogenic bone marrow transplant patients: roleof severe neutropenia and mucositis. Clin InfectDis 2000; 30: 195–7.

138. Klausner JD, Zukerman C, Limaye AP, Corey L,Outbreak of Stenotrophomonas maltophilia bac-teremia among patients undergoing bone mar-row transplantation: association with faultyreplacement of handwashing soap. InfectControl Hosp Epidemiol 1999; 20: 756–8.

139. Rogues AM, Sarlangue J, de Barbeyrac B et al,Agrobacterium radiobacter as a cause ofpseudobacteremia. Infect Control Hosp Epidemiol1999; 20: 345–7.

140. Ezzedine H, Mourad M, van Ossel C et al, An


outbreak of Ochrobactrum anthropi bacteraemiain five organ transplant patients. J Hosp Infect1994; 27: 35–42.

141. Kappstein I, Grundmann H, Hauer T, NiemeyerC, Aerators as a reservoir of Acinetobacter junii:an outbreak of bacteraemia in paediatric oncol-ogy patients. J Hosp Infect 2000; 44: 27–30.

142. Namnyak S, Hussain S, Davalle J et al,Contaminated lithium heparin bottles as asource of pseudobacteraemia due to Pseudo-monas fluorescens. J Hosp Infect 1999; 41: 23–8.

143. Hsueh PR, Teng LJ, Pan HJ et al, Outbreak ofPseudomonas fluorescens bacteremia amongoncology patients. J Clin Microbiol 1998; 36:2914–17.

144. Chetoui H, Melin P, Struelens MJ et al,Common-source outbreak of Burkholderia picket-tii bacteremia. J Clin Microbiol 1997; 35:1398–403.

145. Luk WK, An outbreak of pseudobacteraemiacaused by Burkholderia pickettii: the critical roleof an epidemiological link. J Hosp Infect 1996; 34:59–69.

146. Beebe JL, Koneman EK, Recovery of uncommonbacteria from blood: association with neoplasticdisease. Clin Microbiol Rev 1995; 8: 336–56.

147. Hoppe JE, Herter M, Aleksic S et al, Catheter-related Rahnella aquatilis bacteremia in a pedi-atric bone marrow transplant recipient. J ClinMicrobiol 1993; 31: 1911–12.

148. Thuler LC, Velasco E, de Souza-Martins CA etal, An outbreak of Bacillus species in a cancerhospital. Infect Control Hosp Epidemiol 1998; 19:856–8.

149. Christenson JC, Byington C, Korgenski EK et al,Bacillus cereus infections among oncologypatients at a children’s hospital. Am J InfectControl 1999; 27: 543–6.

150. Musa MO, Al-Douri M, Khan S et al, Fulminantsepticaemic syndrome of Bacillus cereus: threecase reports. J Infect 1999; 39: 154–6.

151. Arnaout MK, Tamburro RF, Bodner SM et al,Bacillus cereus causing fulminant sepsis andhemolysis in two patients with acute leukemia.J Pediatr Hematol Oncol 1999; 21: 431–5.

152. Skiest DJ, Levi ME, Catheter-related bacteremiadue to Mycobacterium smegmatis. South Med J1998; 91: 36–7.

153. Esteban J, Fernandez-Roblas R, Roman A etal, Catheter-related bacteremia due to Myco-bacterium aurum in an immunocompromised

host. Clin Infect Dis 1998; 26: 496–7.154. Moreno A, Llanos M, Gonzalez A, Batista N,

Mycobacterium fortuitum bacteremia in animmunocompromised patient with a long-termvenous catheter. Eur J Clin Microbiol Infect Dis1996; 15: 423–4.

155. Rodgers GL, Mortensen JE, Blecker-Shelly D etal, Two case reports and review of vascularcatheter-associated bacteremia caused by non-tuberculous Mycobacterium species. Pediatr InfectDis J 1996; 15: 260–4.

156. Holland DJ, Chen SC, Chew WW, Gilbert GL,Mycobacterium neoaurum infection of a Hickmancatheter in an immunosuppressed patient. ClinInfect Dis 1994; 18: 1002–3.

157. Levendoglu-Tugal O, Munoz J, Brudnicki A etal, Infections die to nontuberculous mycobacte-ria in children with leukemia. Clin Infect Dis1998; 27: 1227–30.

158. Vidal AM, Sarria JC, Kimbrough RC 3rd, KeungYK, Anaerobic bacteremia in a neutropenicpatient with oral mucositis. Am J Med Sci 2000;319: 189–90.

159. Patel JB, Clarridge J, Schuster MS et al,Bacteremia caused by a novel isolate resem-bling Leptotrichia species in a neutropenicpatient. J Clin Microbiol 1999; 37: 2064–7.

160. Landsaat PM, van der Lelie H, Bongaerts G,Kuijper EJ, Fusobacterium nucleatum, a new inva-sive pathogen in neutropenic patients? Scand JInfect Dis 1995; 27: 83–4.

161. Fruchart C, Salah A, Gray C et al, Lactobacillusspecies as emerging pathogens in neutropenicpatients. Eur J Clin Microbiol Infect Dis 1997; 16:681–4.

162. Cooper CD, Vincent A, Greene JN et al,Lactobacillus bacteremia in febrile neutropenicpatients. Clin Infect Dis 1998; 26: 1247–8.

163. Valentine EG, Nontraumatic gas gangrene. AnnEmerg Med 1997; 30: 109–111.

164. Maury S, Leblanc T, Rousselot P et al,Bacteremia due to Capnocytophaga species inpatients with neutropenia: high frequency ofbeta-lactamase-producing strains. Clin Infect Dis1999; 28: 1172–4.

165. Kristensen B, Schonheyder HC, Peterslund NAet al, Capnocytophaga (Capnocytophaga ochraceagroup) bacteremia in hematological patientswith profound granulocytopenia. Scand J InfectDis 1995; 27: 153–5.

166. Gruson D, Hilbert G, Pigneux A et al, Severe


infection caused by Stomatococcus mucilaginosusin a neutropenic patient: case report and reviewof the literature. Hematol Cell Ther 1998; 40:167–9.

167. Henwick S, Koehler M, Patrick CC,Complications of bacteremia due toStomatococcus mucilaginosus in neutropenic chil-dren. Clin Infect Dis 1993; 17: 667–71.

168. Goldman M, Chaudhary UB, Greist A, FauselCA, Central nervous system infections due toStomatococcus mucilaginosus in immunocompro-mised hosts. Clin Infect Dis 1998; 27: 1241–6

169. Zinner SH, Changing epidemiology of infec-tions in patients with neutropenia and cancer:emphasis on Gram-positive and resistant bacte-ria. Clin Infect Dis 1999; 29: 490–4.

170. Kern WV, Andriof E, Oethinger M et al,Emergence of fluoroquinolone-resistantEscherichia coli at a cancer center. AntimicrobAgents Chemother 1994; 38: 681–7.

171. Carratala J, Fernandez-Sevilla A, Tubau F et al,Emergence of quinolone-resistant Escherichiacoli bacteremia in neutropenic patients withcancer who have received prophylactic nor-floxacin. Clin Infect Dis 1995; 20: 557–60.

172. Yeh SP, Hsueh EJ, Yu MS et al, Oralciprofloxacin as antibacterial prophylaxis afterallogeneic bone marrow transplantation: a reap-praisal. Bone Marrow Transplant 1999; 24:1207–11.

173. Kern WV, Epidemiology of fluoroquinolone-resistant Escherichia coli among neutropeniccancer patients. Clin Infect Dis 1998; 27: 235–7.

174. Carratala J, Fernandez-Sevilla A, Tubau F et al, Emergence of fluoroquinolone-resistantEscherichia coli in fecal flora of cancer patientsreceiving norfloxacin prophylaxis. AntimicrobAgents Chemother 1996; 40: 503–5.

175. Van Kraaij MG, Dekker AW, Peters E etal, Emergence and infectious complications ofciprofloxacin-resistant Escherichia coli in haema-tological cancer patients. Eur J Clin MicrobiolInfect Dis 1998; 17: 591–2.

176. Perea S, Hidalgo M, Arcediano A et al,Incidence and clinical impact of fluoro-quinolone-resistant Escherichia coli in the faecalflora of cancer patients treated with high dosechemotherapy and ciprofloxacin prophylaxis.J Antimicrob Chemother 1999; 44: 117–20.

177. Martino R, Subira M, Altes A et al, Effect of dis-continuing prophylaxis with norfloxacin in

patients with hematologic malignancies andsevere neutropenia. A matched case–controlstudy of the effect on infectious morbidity. ActaHaematol 1998; 99: 206–11.

178. Guerrini G, Calmaggi A, Sallaber S et al, Impactof different strategies to control antibiotic resis-tance in a oncohematological unit. Clin InfectDis 2000; 31: 283 (Abst 413).

179. Rolston KV, Elting L, Waguespack S et al,Survey of antibiotic susceptibility among Gram-negative bacilli at a cancer center. Chemotherapy1996; 42: 348–53.

180. Jacobson K, Rolston K, Elting L et al,Susceptibility surveillance among Gram-negat-ive bacilli at a cancer center. Chemotherapy 1999;45: 325–34.

181. Ariffin H, Navaratnam P, Mohamed M et al,Ceftazidime-resistant Klebsiella pneumoniaebloodstream infection in children with febrileneutropenia. Int J Infect Dis 2000; 4: 21–25.

182. Knowles S, Herra C, Devitt E et al, An outbreakof multiply resistant Serratia marcescens: theimportance of persistent carriage. Bone MarrowTransplant 2000; 25: 873–7.

183. Johnson MP, Ramphal R. �-lactam resistantEnterobacter bacteremia in febrile neutropenicpatients receiving monotherapy. J Infect Dis1990; 162: 981–3.

184. Kapur D, Dorsky D, Feingold JM et al,Incidence and outcome of vancomycin-resistantenterococcal bacteremia following autologosperipheral blood stem cell transplantation. BoneMarrow Transplant 2000; 25: 147–52.

185. Henning KJ, Delencastre H, Eagan J et al,Vancomycin-resistant Enterococcus faecium on apediatric oncology ward: duration of stoll shed-ding and incidence of clinical infection. PediatrInfect Dis J 1996; 15: 848–54.

186. Bradley SJ, Wilson ALT, Allen MC et al, Thecontrol of hyperendemic glycopeptide-resistantEnterococcus spp. on a haematology unit bychanging antibiotic usage. J AntimicrobChemother 1999; 43: 261–6.

187. Lisgaris MV, Hoyen C, Salata RA et al, An out-break of vancomycin-resistant enterococcus col-onization and bacteremia after a formularychange on an adults oncology unit. Clin InfectDis 2000; 31: 215 (Abst 20).

188. Zaas AK, Song X, Perl TM, Tucker PC,Nosocomial vancomycin-resistant enterococcalblood stream infection: development in oncol-


ogy patients colonized with vancomycin-resis-tant enterococci. Clin Infect Dis 2000; 31: 278(Abst 382).

189. Mannan G, Fuchs A, Molavi A et al,Vancomycin-resistant Enterococcus faecium bac-teremia on an oncology unit from 1995–1999.Clin Infect Dis 2000; 31: 278 (Abst 383).

190. Kirkpatrick BD, Harrington SM, Smith D et al,An outbreak of vancomycin-dependentEnterococcus faecium in a bone marrow trans-plant unit. Clin Infect Dis 1999; 29: 1268–73.

191. Maschmeyer G, Link H, Hiddemann W et al,Pulmonary infiltrations in febrile patients withneutropenia. Risk factors and outcome underempirical antimicrobial therapy in a randomizedmulticenter study. Cancer 1994; 73: 2296–304.

192. Ewig S, Glasmacher A, Ulrich B et al,Pulmonary infiltrates in neutropenic patientswith acute leukemia during chemotherapy: out-come and prognostic factors. Chest 1998; 114:444–51.

193. Fernandez-Aviles F, Batlle M, Ribera JM et al,Pneumonias in patients with malignanthemopathies. Their etiology, response to treat-ment and prognostic factors in 69 patients (88episodes). Med Clin Barc 1999; 112: 321–5.

194. Gruson D, Hilbert G, Valentino R et al, Utilityof fiberoptic bronchoscopy in neutropenicpatients admitted to the intensive care unit withpulmonary infiltrates. Crit Care Med 2000; 28:2224–30.

195. Gruson D, Hilpert G, Portel L et al, Severe res-piratory failure requiring ICU admission inbone marrow transplant recipients. Eur Respir J1999; 13: 883–7.

196. Sepkowitz KA, Brown AE, Telzak EE et al,Pseumocystis carinii pneumonia among patientswithout AIDS at a cancer hospital. JAMA 1992;267: 832–7.

197. Kulke MH, Vance EA, Pneumocystis cariniipneumonia in patients receiving chemotherapyfor breast cancer. Clin Infect Dis 1997; 25:215–18.

198. Heinemann M, Kern WV, Bunjes D et al, SevereChlamydia pneumoniae infection in patients withneutropenia: case reports and literature review.Clin Infect Dis 2000; 31: 181–4.

199. Heussel CP, Kauczor HU, Heussel G et al, Earlydetection of pneumonia in febrile neutropenicpatients: use of thin-section CT. AJR 1997; 169:1347–53.

200. Korones DN, Hussong MR, Gullace MA,Routine chest radiography of children withcancer hospitalized for fever and neutropenia:Is it really necessary? Cancer 1997; 80: 1160–4.

201. Peterson DE, Research advances in oral mucosi-tis. Curr Opin Oncol 1999; 11: 261–6.

202. Nouwen JL, Wielenga JJ, van Overhagen H etal, Hickman catheter-related infections in neu-tropenic patients: insertion in the operating the-ater versus insertion in the radiology suite.J Clin Oncol 1999; 17: 1304.

203. Stone S, Abdelmalak S, Laquaglia M et al,Central venous catheter infections among chil-dren with cancer. Clin Infect Dis 1999; 29: 1045(Abst 473).

204. Biffi R, de Braud F, Orsi F et al, Totallyimplantable central venous access ports forlong-term chemotherapy. A prospective studyanalyzing complications and costs of 333devices with a minimum follow-up of 180 days.Ann Oncol 1998; 9: 767–73.

205. Astagneau P, Maugat S, Tran-Minh T et al,Long-term central venous catheter infection inHIV-infected and cancer patients: a multicentercohort study. Infect Control Hosp Epidemiol 1999;20: 494–8.

206. Howell PB, Walters PE, Donowitz GR, Farr BM.Risk factors for infection of adult patients withcancer who have tunnelled central venouscatheters. Cancer 1995; 75: 1367–75.

207. Newman, KA, Reed WP, Schimpff SC et al,Hickman catheters in association with intensivecancer chemotherapy. Support Care Cancer 1993;1: 92–7.

208. Rotstein C, Brock L, Roberts RS, The incidenceof first Hickman catheter-related infection andpredictors of catheter removal in cancerpatients. Infect Control Hosp Epidemiol 1995; 16:451–8.

209. Raad I, Davis S, Becker M et al, Low infectionrate and long durability of nontunneled silasticcatheters. A safe and cost-effective alternativefor long-term venous access. Arch Intern Med1993; 153: 1791–6.

210. Rackoff WR, Ge J, Sather HN et al, Centralvenous catheter use and the risk of infection inchildren with acute lymphoblastic leukemia:a report from the Children´s Cancer Group.J Pediatr Hematol Oncol 1999; 21: 260–7.

211. Sayfan J, Shoavi O, Koltun L, Benyamin N,Acute abdomen caused by neutropenic entero-


colitis: surgeon’s dilemma. Eur J Surg 1999; 165:502–4.

212. Buyukasik Y, Ozcebe OI, Haznedaroglu IC et al,Neutropenic enterocolitis in adult leukemias.Int J Hematol 1997; 66: 47–55.

213. Jain Y, Arya LS, Kataria R, Neutropenic entero-colitis in children with acute lymphoblasticleukemia. Pediatr Hematol Oncol 2000; 17:99–103.

214. Kouroussis C, Samonis G, Androulakis N et al,Successful conservative treatment of neu-tropenic enterocolitis complicating taxane-based chemotherapy: a report of five cases. AmJ Clin Oncol 2000; 23: 309–13.

215. Wade DS, Nava HR, Douglass HO Jr,Neutropenic enterocolitis. Clinical diagnosisand treatment. Cancer 1992; 69: 17–23.

216. Song HK, Kreisel D, Canter R et al, Changingpresentation and management of neutropenicenterocolitis. Arch Surg 1998; 133: 979–82.

217. Baerg J, Murphy JJ, Anderson R, Magee JF,Neutropenic enteropathy: a 10-year review.J Pediatr Surg 1999; 34: 1068–71.

218. Gomez L, Martino R, Rolston KV, Neutropenicenterocolitis. Clin Infect Dis 1998; 27: 695–9.

219. Coleman N, Speirs G, Khan J et al, Neutropenicenterocolitis associated with Clostridium tertium.J Clin Pathol 1993; 46: 180–3.

220. Bilgrami S, Feingold JM, Dorsky D et al,Incidence and outcome of Clostridium difficileinfection following autologous peripheral bloodstem cell transplantation. Bone MarrowTransplant 1999; 23: 1039–42.

221. Schuller I, Saha V, Lin L et al, Investigation andmanagement of Clostridium difficile colonisationin a paediatric oncology unit. Arch Dis Child1995; 72: 219–22.

222. Husain A, Aptaker L, Spriggs DR, Barakat RR,Gastrointestinal toxicity and Clostridium difficilediarrhea in patients treated with paclitaxel-con-taining chemotherapy regimens. Gynecol Oncol1998; 71: 104–7.

223. Gorschlüter M, Glasmacher A, Hahn C et al,Clostridium difficile infections in neutropenicpatients. Onkologie 2000; 23(Suppl 7): 165 (Abst629).

224. Alangaden GJ, Jamma K, Chandrasekar PH etal, Clostridium difficile associated diarrhea dur-ing hemopoietic stem cell transplantation. ClinInfect Dis 2000; 31: 280.

225. Lew MA, Kehoe K, Ritz J et al, Ciprofloxacin

versus trimethoprim/sulfamethoxazole for pro-phylaxis of bacterial infections in bone marrowtransplant recipients: a randomized, controlledtrial. J Clin Oncol 1995; 13: 239–50.

226. Roghmann MC, McCarter RJ Jr, Brewrink J et al,Clostridium difficile infection is a risk factor forbacteremia due to vancomycin-resistant entero-cocci (VRE) in VRE-colonized patients withacute leukemia. Clin Infect Dis 1997; 25: 1056–9.

227. Boone N, Eagan JA, Gillern P et al, Evaluationof an interdisciplinary re-isolation policy forpatients with previous Clostridium difficile diar-rhea. Am J Infect Control 1998; 26: 584–7.

228. Hornbuckle K, Chak A, Lazarus HM et al,Determination and validation of a predictivemodel for Clostridium difficile diarrhea in hospi-talized oncology patients. Ann Oncol 1998; 9:307–11.

229. Buyukasik Y, Ozcebe OI, Sayinalp N et al,Perianal infections in patients with leukemia:importance of the course of neutrophil count.Dis Colon Rectum 1998; 41: 81–5.

230. Cohen JS, Paz IB, O’Donnell MR, Ellenhorn JD,Treatment of perianal infection following bonemarrow transplantation. Dis Colon Rectum 1996;39: 981–5.

231. Harousseau JL, Dekker AW, Stamatoullas-Bastard A et al, Itraconazole oral solution forprimary prophylaxis of fungal infections inpatients with hematological malignancy andprofound neutropenia: a randomized, double-blind, double-placebo, multicenter trial compar-ing itraconazole and amphotericin B. AntimicrobAgents Chemother 2000; 44: 1887–93.

232. Menichetti F, Del Favero A, Martino P et al,Itraconazole oral solution as prophylaxis forfungal infections in neutropenic patients withhematologic malignancies: a double-blind, mul-ticenter trial. Clin Infect Dis 1999; 28: 250–5.

233. Nucci M, Biasoli I, Akiti T et al, A double-blind,randomized, placebo-controlled trial of itra-conazole capsules as antifungal prophylaxis forneutropenic patients. Clin Infect Dis 2000; 30:300–5.

234. Winston DJ, Chandrasekar PH, Lazarus HM etal, Fluconazole prophylaxis of fungal infectionsin patients with acute leukemia. Results of arandomized placebo-controlled, double-blind,multicenter trial. Ann Intern Med 1993; 118:495–503.

235. Slavin M, Osborne B, Adams R et al, Efficacy


and safety of fluconazole prophylaxis for fungalinfections after marrow transplantation – aprospective, randomized, double-blind study.J Infect Dis 1995; 171: 1545–52.

236. Goodman JL, Winston DJ, Greenfield RA et al,A controlled trial of fluconazole to prevent fun-gal infections in patients undergoing bone mar-row transplantation. N Engl J Med 1992; 326:845–51.

237. Rotstein C, Bow EJ, Laverdiere M et al,Randomized placebo-controlled trial of flucona-zole prophylaxis for neutropenic cancerpatients: benefit based on purpose and intensityof cytotoxic therapy. The Canadian FluconazoleProphylaxis Study Group. Clin Infect Dis 1999;28: 331–40.

238. Musa MO, Aleisa, A, Halim M et al, the spec-trum of Fusarium infection in immunocomp-romised patients with haematologicalmalignancies and in non-immunocompromisedpatients: a single institution experience over 10years. Br J Haematol 2000; 108: 544–8.

239. Krcmery VJr, Jesenska Z, Spanik S et al,Fungemia due to Fusarium spp. in cancerpatients. J Hosp Infect 1997; 36: 223–8.

240. Kontoyiannis DP, Wessel VC, Bodey GP,Rolston KVI, Zygomycosis in the 1990s in a ter-tiary-care cancer center. Clin Infect Dis 2000; 30:851–6.

241. Ribes JA, Vanover-Sams CL, Baker DJ,Zygomycetes in human disease. Clin MicrobiolRev 2000; 13: 236–301.

242. Kwon-Chung KL, Bennett JE, Medical Mycology:Infections Due to Trichosporon and OtherMiscellaneous Yeast-Like Fungi. Philadelphia: Lee& Febiger, 1992.

243. Tashiro T, Nagai H, Nagaoka H et al,Trichosporon beigelii pneumonia in patients withhematologic malignancies. Chest 1995; 108:190–5.

244. Krcmery VJr, Mateicka F, Kunova A et al,Hematogenous trichsporonosis in cancerpatients: report of 12 cases including 5 duringprophlyaxis with itraconazole. Support CareCancer 1999; 7: 39–43.

245. Meis JF, Ruhnke M, de Pauw BE et al, Candidadubliniensis candidemia in patients withchemotherapy-induced neutropenia and bonemarrow transplantation. Emerg Infect Dis 1999;5: 150–3.

246. D´Antonio D, Violante B, Mazzoni A et al, A

nosocomial cluster of Candida inconspicua infec-tions in patients with hematological malig-nancies. J Clin Microbiol 1998; 36: 792–5.

247. Orth B, Frei R, Itin PH et al, Outbreak of inav-sive mycoses caused by Paecilomyces lilacinusfrom a contaminated skin lotion. Ann InternMed 1996; 125: 799–806.

248. Burgner D, Eagles G, Burgess M et al,Disseminated infection due to Metarrhiziumanisopliae in an immunocompromised child.J Clin Microbiol 1998; 36: 1146–50.

249. Brown NM, Blundell EL, Chown SR et al,Acremonium infection in a neutropenic patient.J Infect 1992; 25: 73–6.

250. De Batlle J, Motje M, Balanza R et al,Disseminated infection caused by Scedosporiumprolificans in a patient with acute multilinealleukemia. J Clin Microbiol 2000; 38: 1694–5.

251. Kontoyianis DP, Vartivarian S, Anaissie EJ et al,Infections due to Cunninghamella bertholletiae inpatients with cancer: report of three cases andreview. Clin Infect Dis 1994; 18: 925–8.

252. Richter S, Cormican MG, Pfaller MA et al, Fataldisseminated Trichoderma longibrachiatum infec-tion in an adult bone marrow transplantpatient: species identification and review of theliterature. J Clin Microbiol 1999; 37: 1154–60.

253. Verweij PE, van Kasteren M, van de Nes J et al,Fatal pulmonary infection caused by the basid-iomycete Hormographiella aspergillata. J ClinMicrobiol 1997; 35: 2675–8.

254. Mok T, Koehler AP, Yu MY et al, FatalPenicillium citrinum pneumonia with pericardi-tis in a patient with acute leukemia. J ClinMicrobiol 1997; 35: 2654–6.

255. Abi-Said D, Anaissie E, Uzon O et al, The epi-demiology of hematogenous candidiasis causedby different Candida species. Clin Infect Dis1997; 24: 1122–8.

256. Guiot HF, Fibbe WE, van ‘t Wout JW. Risk fac-tors for fungal infection in patients with malig-nant hematologic disorders: implications forempirical therapy and prophylaxis. Clin InfectDis 1994; 18: 525–32.

257. Richet HM, Andremont A, Tancrede C et al,Risk factors for candidemia in patients withacute lymphocytic leukemia. Rev Infect Dis 1991;13: 211–15.

258. Pagano L, Antinori A, Ammassari A et al,Retrospective study of candidemia in patientswith hematological malignancies. Clinical fea-


tures, risk factors and outcome of 76 episodes.Eur J Haematol 1999; 63: 77–85.

259. Marr KA, Seidel K, White TC et al, Candidemiain allogeneic blood and marrow transplantrecipients: evolution of risk factors after theadoption of prophylactic fluconazole. J InfectDis 2000; 181: 309–16.

260. Viscoli C, Girmenia C, Marinus A et al,Candidemia in cancer patients: a prospective,multicenter surveillance study by the InvasiveFungal Infection Group of the EuropeanOrganization for Research and Treatment ofCancer (EORTC). Clin Infect Dis 1999; 28:1071–9.

261. Anaissie EJ, Rex JH, Uzun O, Vartivarian S,Predictors of adverse outcome in cancerpatients with candidemia. Am J Med 1998; 104:238–45.

262. Meunier F, Aoun M, Bitar N, Candidemia inimmunocompromised patients. Clin Infect Dis1991; 14(Suppl 1): S120–5.

263. Goodrich JM, Reed EC, Mori M et al, Clinicalfeatures and analysis of risk factors for invasivecandidal infection after marrow transplanta-tion. J Infect Dis 1991; 164: 731–40.

264. Ribeiro S, Sousa AB, Nunes O et al, Candidemiain acute leukemia patients. Support Care Cancer1997; 5: 249–51.

265. Bodey GP, Anaissie EJ, Chronic systemic can-didiasis. Eur J Clin Microbiol Infect Dis 1989; 8:855–7.

266. Anttila VJ, Elonen E, Nordling S et al,Hepatosplenic candidiasis in patients withacute leukemia: incidence and prognostic impli-cations. Clin Infect Dis 1997; 24: 375–80.

267. Martino R, Sureda A, Brunet S, Disseminatedcandidiasis in patients with acute leukemia.Clin Infect Dis 1998; 26: 245–6.

268. Bodey GP, Luna MA, Disseminated candidiasisin patients with acute leukemia: two diseases.Clin Infect Dis 1998; 27: 238.

269. Fridkin SK & Jarvis WR, Epidemiology of noso-comial fungal infections. Clin Microbiol Rev1996; 9: 499–511.

270. Vandenbergh MFG, Verweij PE, Voss A,Epidemiology of nosocomial fungal infections:invasive aspergillosis and the environment.Diagn Microbiol Infect Dis 1999 ; 34: 221–7.

271. Warnock DW, Fungal infection in neutropenia:current problem and chemotherapeutic control.J Antimicrob Chemother 1998; 41: 95–105

272. Latgé JP, Aspergillus fumigatus and aspergillosis.Clin Microbiol Rev 1999; 12: 310–50.

273. O’Donnell MR, Schmidt GM, Tegtmeier BR etal, Prediction of systemic fungal infection inallogeneic marrow recipients. Impact of ampho-tericin prophylaxis in high risk patients. J ClinOncol 1994; 12: 827–34.

274. Sparrelid E, Hagglund H, Remberger M et al,Bacteraemia during the aplastic phase after allo-geneic bone marrow transplantation is associ-ated with early death from invasive fungalinfection. Bone Marrow Transplant 1998; 22:795–800.

275. Wahiduzzaman M, Chandrasekar PH,Alangaden GJ, High frequency of Gram-negat-ive bacterial pneumonia in bone marrow/stemcell transplant recipients with graft-versus-hostdisease and pulmonary aspergillosis. Clin InfectDis 1999; 29: 1014 (Abst 296).

276. Wald A, Leisenring W, van Burik JA, BowdenRA, Epidemiology of Aspergillus infections in alarge cohort of patients undergoing bone mar-row transplantation. J Infect Dis 1997; 175:1459–66.

277. Denning DA, Invasive aspergillosis. Clin InfectDis 1998; 26: 781–805.

278. Girardin H, Sarfati J, Traore F et al, Molecularepidemiology of nosocomial invasiveaspergillosis. J Clin Microbiol 1994; 32: 684–90.

279. Chazalet V, Debeaupuis JP, Sarfati J et al,Molecular typing of environmental and patientisolates of Aspergillus fumigatus from varioushospital settings. J Clin Microbiol 1998; 36:1494–500.

280. Leenders ACAP, van Belkum A, Behrendt M etal, Density and molecular epidemiology ofAspergillus in air and relationship to outbreaksof aspergillus infection. J Clin Microbiol 1999; 37:1752–7

281. Arnow PM, Sadigh M, Costas C et al, Endemicand epidemic aspergillosis associated with in-hospital replication of Aspergillus organisms.J Infect Dis 1991; 164: 998–1002.

282. Thio CL, Smith D, Merz WG et al, Refinementsof environmental assessments during an out-break investigation of invasive aspergillosis in aleukemia and bone marrow transplant unit.Infect Control Hosp Epidemiol 2000; 21: 18–23.

283. Withington S, Chambers ST, Beard ME et al,Invasive aspergillosis in severely neutropenicpatients over 18 years: impact of intranasal


amphotericin B and HEPA filtration. J HospInfect 1998; 38: 11–18.

284. Manuel RJ, Kibbler CC, The epidemiology andprevention of invasive aspergillosis. J HospInfect 1998; 39: 95–109.

285. Anaissie EJ, Stratton SL, Summerbell RC et al,Aspergillus species aerosols in hospitals: show-ering as a potential mode of exposure. In:Proceedings of 39th Interscience Conference onAntimicrobial Agents and Chemotherapy, SanFrancisco. Washington, DC: American Societyfor Microbiology, 1999: 579 (abst).

286. Warris A, Gaustad P, Meis J et al, Water as asource of filamentous fungi in a childhood bonemarrow transplantation unit. In: Proceedings of39th Interscience Conference on AntimicrobialAgents and Chemotherapy, San Francisco.Washington DC: American Society forMicrobiology, 1999: 579 (abst).

287. Abbasi S, Shenep JL, Hughes WT et al,Aspergillosis in children with cancer: a 34-yearexperience. Clin Infect Dis 1999; 29: 1210–19.

288. Denning DW, Marinus A, Cohen J et al, AnEORTC multicentre prospective survey of inva-sive aspergillosis in haematological patients:diagnosis and therapeutic outcome. J Infect1998; 37: 173–80.

289. Pagano L, Ricci P, Montillo M et al, Localizationof aspergillosis to the central nervous systemamong patients with acute leukemia: report of14 cases. Gruppo Italiano MalattieEmatologiche dell’Adulto Infection Program.Clin Infect Dis 1996; 23: 628–30.

290. Ribaud P, Chastang C, Latge JP et al, Survivaland prognostic factors of invasive aspergillosisafter allogeneic bone marrow transplantation.Clin Infect Dis 1999; 28: 322–30.

291. Martino R, Nomededéu J, Altés A et al,Successful bone marrow transplantation inpatients with previous invasive fungal infec-tions: report of four cases. Bone MarrowTransplant 1994; 13: 265–9.

292. Offner F, Cordonnier C, Ljungman P et al,Impact of previous aspergillosis on the outcomeof bone marrow transplantation. Clin Infect Dis1998; 26: 1098–103.

293. Caillot D, Casasnovas O, Bernard A et al,Improved management of invasive pulmonaryaspergillosis in neutropenic patients using earlythoracic computed tomographic scan andsurgery. J Clin Oncol 1997; 15: 139–47.

294. Blum U, Windfuhr M, Buitrago-Tellez C et al,Invasive pulmonary aspergillosis. MRI, CT, andplain radiographic findings and their contribu-tion for early diagnosis. Chest 1994; 106:1156–61.

295. Wood MJ, Viral infections in neutropenia – cur-rent problems and chemotherapeutic control. JAntimicrob Chemother 1998; 41(Suppl D): 81–93.

296. Tinchon C, Auner H, Linkesch W, EBV-associ-ated lymphoproliferative disease after therapywith fludarabine in AML. Onkologie 2000;23(Suppl 7): 163 (Abst 623).

297. Holmberg LA, Boeckh M, Hooper H et al,Increased incidence of cytomegalovirus diseaseafter autologous CD34-selected peripheral bloodstem cell transplantation. Blood 1999; 94: 4029–35.

298. Aslan T, Fassas AB, Desikan R et al, Patientswith multiple myeloma may safely undergoautologous transplantation despite ongoingRSV infection and no ribavirin therapy. BoneMarrow Transplant 1999; 24: 505–9.

299. Fouillard L, Mouthon L, Laporte JP et al, Severerespiratory syncytial virus pneumonia afterautologous bone marrow transplantation: areport of three cases and review. Bone MarrowTransplant 192; 9: 97–100.

300. Harrington RD, Hooton TM, Hackman RC et al,An outbreak of respiratory syncytial virus in abone marrow transplant center. J Infect Dis 1992;165: 987–93.

301. Whimbey E, Couch RB, Englund JA et al,Respiratory syncytial virus pneumonia in hos-pitalized adult patients with leukemia. ClinInfect Dis 1995; 21: 376–9.

302. Vettenranta K, Ukkonen P, Saarinen UM, RSVinfection complicating the therapy of pediatricmalignancies: report of six cases. Med PediatrOncol 1996; 26: 261–3.

303. Bergmann OJ, Ellermann-Eriksen S, MogensenSC, Ellegaard J, Acyclovir given as prophylaxisagainst oral ulcers in acute myeloid leukaemia:randomised, double blind, placebo controlledtrial. BMJ 1995; 310: 1169–72.

304. Lönnqvist B, Palmblad J, Ljungman P et al, Oralacyclovir as prophylaxis for bacterial infectionsduring induction therapy for acute leukaemiain adults. Support Care Cancer 1993; 1: 139–44.

305. Limaye AP, Bowden RA, Myerson D, BoeckhM, Cytomegalovirus disease occurring beforeengraftment in marrow transplant recipients.Clin Infect Dis 1997; 24: 830–5.


306. Van den Brande J, Schrijvers D, Colpaert C,Vermorken JB, Cytomegalovirus colitis afteradministration of docetaxel–5-fluorouracil–cis-platin chemotherapy for locally advancedhypopharyngeal cancer. Ann Oncol 1999; 10:1369–72.

307. Nagafuji K, Eto T, Hayashi S et al, Fatalcytomegalovirus interstitial pneumonia follow-ing autologous peripheral blood stem cell trans-plantation. Bone Marrow Transplant 1998; 21:301–3.

308. Peters WP, Ross M, Vredenburgh JJ et al, High-dose chemotherapy and autologous bone mar-row support as consolidation afterstandard-dose adjuvant therapy for high-riskprimary breast cancer. J Clin Oncol 1993; 11:1132–43.

309. Poulsen A, Schmiegelow K, Yssing M.Veraicella zoster infections in children withacute lymphoblastic leukemia. Pediatr HematolOncol 1996; 13: 231–8.

310. Rowland P, Wald ER, Mirro JR Jr et al,Progressive varicella presenting with pain andminimal skin involvement in children withacute lymphonblastic leukemia. J Clin Oncol1995; 13: 1697–703.

311. Ohyashiki JH, Abe K, Ojima T et al,Quantification of human herpesvirus-6 in

healthy volunteers and patients with lympho-proliferative disorders by PCR–ELISA. Leuk Res1999; 23: 625–30.

312. Gautheret-Dejean A, Dejean O, Vastel L et al,Human herpesvirus-6 and human herpesvirus-7 in the bone marrow from healthy subjects.Transplantation 2000; 69: 1722–3.

313. Michalek J, Horvath R, Benedik J, Hrstkova H,Human herspesvirus-6 infection in childrenwith cancer. Pediatr Hematol Oncol 1999; 16:423–30.

314. Johnston RE, Geretti AM, Prentice HG et al,HHV-6-related secondary graft failure follow-ing allogeneic bone marrow transplantation.Br J Haematol 1999; 105: 1041–5.

315. Wang FZ, Linde A, Hagglund H et al, Humanherpesvirus-6 DANN in cerebrospinal fluidspecimens from allogeneic bona marrow trans-plant patients: Does it have clinical signifi-cance? Clin Infect Dis 1999; 28: 562–8.

316. Singh N, Paterson DL, Encephalitis caused byhuman herpesvirus-6 in transplant recipients:relevance of a novel neurotropic virus.Transplantation 2000; 69: 2474–9.

317. Cone RW, Huang ML, Corey L et al, Humanherpesvirus-6 infections after bone marrowtransplantation: clinical and virologic manifes-tations. J Infect Dis 1999; 179: 311–18.


5Infections in patients with solid tumorsKenneth VI Rolston

INTRODUCTION

Patients with cancer develop infection far moreoften than individuals without cancer.1 Thefunction of the immune system is a major factorin determining the frequency and nature ofinfection, and the overall response to therapy,once an infection has developed. Patients withhematologic malignancies or aplastic anemiaand those receiving immunosuppressive ther-apy following bone marrow transplantationoften have prolonged periods of neutropenia,defects in phagocytosis, and impaired cellularand/or humoral immunity – each associatedwith an increased frequency and a distinctspectrum of infection. In contrast, most patientswith solid tumors are not significantlyimmunosuppressed, but are predisposedtowards infection as a result of damage to nor-mal anatomic barriers such as the skin andmucosal surfaces, obstructive phenomena (e.g.lung carcinoma and biliary and pancreatictumors), procedures such as surgery and radia-tion, central nervous system dysfunction,and the use of medical devices such asshunts, catheters, and prostheses. Althoughchemotherapy-induced neutropenia does occurin patients with solid tumors, it is often short-lived, does not have the same impact as it doesin patients with hematologic malignancies, and

is associated with a lower frequency or risk ofdeveloping infection.

Infections in patients with hematologicmalignancies and in bone marrow transplantrecipients have been studied in great detail,and many of the principles for the managementof infections in cancer patients havebeen developed in this patient population.2

However, solid tumors account for the vastmajority of cancers in adults. Data published bythe American Cancer Society indicate thatapproximately 1.14 million new cases of solidtumors are diagnosed each year in the USA.3

The spectrum, clinical features, diagnosis, andmanagement of infection in these patients issubstantially different, and treatment strategiesspecific for these patients need to be developed.This chapter will review various aspects ofinfections that occur commonly in, or areunique to, patients with solid tumors.

RISK FACTORS FOR INFECTION

Several factors contribute to the risk of infectionin patients with solid tumors. The presence ofmultiple risk factors in the same patient is notuncommon, and contributes towards increasedrisk. These factors are summarized in Table 5.1and include neutropenia, disruption of normal

anatomic barriers, obstruction, procedures anddevices employed in therapy of the tumor, aswell as a number of other factors.

Neutropenia

Neutropenia is induced most often by antineo-plastic chemotherapy. Varying degrees of neu-tropenia are also seen after radiation therapy,

after the administration of agents such as ganci-clovir, and occasionally after extensive infiltra-tion of the marrow by tumor. Unlike patientswith hematologic malignancies, patients withsolid tumors usually have normally functioningneutrophils, and conventional chemotherapyrarely produces severe neutropenia that lastsfor more than 7–10 days.4 Thus the ‘at-risk’period is generally short, and many solid tumorpatients who develop a febrile episode while


Table 5.1 Risk factors predisposing towards infection in patients with solid tumorsa

Risk factor(s) Additional comments

Neutropenia Antineoplastic chemotherapy, radiation therapy, infiltrationof bone marrow with tumor, other agents (ganciclovir)

Disruption of normal anatomic barriers Chemotherapy (mucositis), radiation therapy, diagnostic ortherapeutic surgical procedures, catheters and otherdevices

Obstruction Primary or metastatic tumorAirways Post-obstructive pneumonia/lung abscess/empyemaBiliary tract Ascending cholangitisUrinary tract Urinary tract infection/prostatitisBowel Bowel obstruction/perforation, peritonitis/hemorrhage

Procedures and devicesVascular access catheters Catheter-related infectionShunts Shunt infectionProsthetic devices Infected prosthesisDiagnostic/therapeutic surgery Local/disseminated infection

Miscellaneous factorsLoss of gag reflex/cord compression Central nervous system, tumors → aspiration, impaired

micturition → urinary tract infectionsAge, malnutrition, antibiotic usage Increased risk and severity of infection, selection of

resistant pathogens

a Multiple factors in the same patient increase the risk of infection.

they are neutropenic are considered ‘low-risk’(see Chapter 8).

Disruption of normal anatomic barriers

Normal anatomic barriers, which include intactskin, oropharyngeal, respiratory, gastrointesti-nal, and genitourinary mucosal surfaces,provide an important defense mechanismagainst invasion by microorganisms. Cancerchemotherapy often damages mucosal surfaces,increasing the risk of infections caused byorganisms that colonize these surfaces. Agentsthat are particularly prone to causing mucositisinclude chlorambucil, cisplatin, cytarabine(cytosine arabinoside, Ara-C), doxorubicin(Adriamycin), 5-fluorouracil (5-FU), andmethotrexate. Damage to mucosal barriers canalso be caused by radiation therapy, surgicalprocedures, and the use of medical devices.When such patients are hospitalized, the risk ofacquiring serious nosocomial infections (oftencaused by multidrug-resistant microorganisms)increases. (See Chapter 1.)

Obstruction

Obstruction caused by rapidly expanding pri-mary or metastatic lesions is fairly common inpatients with solid tumors. Bronchogenic carci-nomas (or metastatic pulmonary lesions) oftencause partial airway obstruction, leading to thedevelopment of post-obstructive pneumonia.Empyema may occasionally complicate post-obstructive pneumonia. Biliary tract obstructionin patients with hepatobiliary pancreatic tumorsresults in ascending cholangitis. Ureteral obstruc-tion resulting in urinary tract infection is seen inpatients with carcinoma of the cervix, whereasureteral obstruction causing urinary tract infec-tion and/or prostatitis is seen in patients withcarcinoma of the prostate. In all these situations,mixed or polymicrobial infections are common,and the etiologic agents are generally those thatcolonize the site of obstruction.

Procedures and devices

Surgery, medical procedures, radiation therapy,and the widespread use of catheters and otherdevices (shunts, stents, prostheses) are oftenassociated with the development of infection.The use of multiple-lumen vascular accesscatheters (e.g. Hickman or Broviac catheters)has become commonplace, and greatly facili-tates the drawing and/or administration ofblood or blood products, and the administra-tion of chemotherapy or antimicrobial agentsand other supportive medications. Infection isthe major complication associated with thesecatheters. The organisms causing catheter-related infections are listed below in Table 5.3(see Chapter 10). Approximately 80% are Gram-positive, with Staphylococcus spp. being pre-dominant. Urinary catheters are usedfrequently when obstruction or urinary inconti-nence is present. Local involvement of the blad-der or ureters with the malignancy oftenrequires the creation of surgical diversions intoileal or colonic segments. Bacteremia progress-ing to acute or chronic pyelonephritis withintestinal microorganisms is not uncommon.Many patients with central nervous system(CNS) tumors require the placement of cere-brospinal fluid shunts. When infected, the CNSend of the shunt produces symptoms such asheadache, mental status changes, andmeningismus, whereas the distal ends of suchshunts, which are generally located in thepleural or peritoneal cavities, give rise to symp-toms of pleuritis or peritonitis. Surgicallyimplanted prosthetic devices are used fre-quently in patients with osteosarcoma andother bone tumors. Infection is the most com-mon complication associated with thesedevices, and is caused most often by organismscolonizing the skin.

Miscellaneous factors

Patients with primary CNS tumors or metasta-tic brain lesions often develop partial loss of the

INFECTIONS IN PATIENTS WITH SOLID TUMORS 93

gag reflex, predisposing towards aspiration.Neurologic abnormalities resulting in impairedmicturition also occur. Damage to ciliary func-tion in the respiratory tract, most often theresult of radiation, increases the likelihood ofdeveloping pneumonia. Many solid tumorsoccur in the elderly, in whom immunologicdeficits caused by ageing, malnutrition, andcancer cachexia may all influence the frequencyand severity of infection, and the ultimateresponse to therapy. Previous and concurrentantibiotic usage can influence the spectrum ofinfection by selecting resistant organisms. Forexample, excessive vancomycin (oral and par-enteral) usage has been associated withincreased isolation rates of glycopeptide-resistant organisms such as vancomycin-resistant enterococci (VRE), Leuconostoc andPediococcus spp.5 The primary drawback ofquinolone prophylaxis is the development ofresistant Gram-negative bacilli (Escherichia coli,Pseudomonas aeruginosa, etc.).6 Prophylactic andempiric antimicrobial regimens are used lessoften and for shorter durations in patients withsolid tumors than in other, more immunosup-pressed, patients. However, practice patternsvary, and must be taken into considerationalong with local susceptibility/resistance pat-terns when evaluating patients for infection.

PREDOMINANT SITES OF INFECTION

Predominant sites of infection depend upon thelocation and size of the primary tumor ormetastatic lesions, and the site and nature ofmedical devices and surgical procedures. Theseare summarized in Table 5.2.

As indicated previously, patients with CNStumors often have partial or complete loss ofthe gag reflex, predisposing them to aspirationpneumonia. Impaired micturition and urinaryretention as a result of neurological impairmentleads to urinary tract infection. Followingsurgery for tumor resection and/or the place-ment of shunts, surgical wound infections, epi-dermal and subdural infections, cerebral

abscesses, meningitis, and shunt infection candevelop. Infections of the upper respiratorytract – including sinusitis, pneumonia (includ-ing aspiration and ventilator-associated pneu-monia), and local cellulitis or necrotizinginfections following surgical excision andreconstruction – are the most common sites inpatients with head and neck tumors. Infectedmasses extending along the soft tissues in theneck can give rise to airway obstruction.

Patients with carcinoma of the lung developpulmonary infections such as post-obstructiveand/or necrotizing pneumonia, lung abscess,empyema, and surgical wound infections.Localized infections may lead to the develop-ment of bacteremia or disseminated infections.Cellulitis following axillary lymph node dissec-tion is the most common site in patients withbreast cancer. Mastitis and breast abscesses areless common.7 Cholangitis with or without bac-teremia, solitary or multiple hepatic abscesses,and peritonitis are not infrequent in patientswho have hepatobiliary–pancreatic tumors.8

Abscesses in the pancreatic bed and subdi-aphragmatic abscesses can occur followingextensive surgical resection. Patients receivingintra-arterial chemotherapy for hepatic tumorsare also at risk for such infections. Patients withcolonic or gynecologic tumors develop abdomi-nal or pelvic abscesses, occasionally after fistulaformation or perforation of a viscus. Ureteralobstruction resulting in urinary tract infection isrelatively common in patients with carcinomaof the cervix, and is caused most often by localextension of tumor, and occasionally by radia-tion damage. Osteomyelitis, osteoradionecrosis,and infected prosthetic devices – with adjacentbone, joint, or soft tissue infections – predomi-nate in patients with osteosarcoma and otherbone neoplasms.

SPECTRUM OF INFECTION

Most infections in patients with solid tumorsare caused by the patients’ own residentmicroflora. The acquisition of nosocomial


pathogens occurs after hospitalization, particu-larly following prolonged or multiple antibioticexposure(s). The distribution of causativeorganisms, therefore, generally mirrors the nor-mal flora at a particular site of infection, or thenosocomial flora of a particular unit or institu-tion. For example, surgical wound infections

and catheter-related infections are caused mostoften by organisms colonizing the skin (coag-ulase-negative staphylococci, Staphylococcusaureus, Streptococcus spp., Bacillus spp.,Corynebacteria), although certain opportunisticpathogens such as Acinetobacter spp., theEnterobacteriaceae, Ps. aeruginosa, and Candida


Table 5.2 Predominant sites of infection in cancer patients with solid tumors

Tumor Common sites of infection

Brain (CNS) Wound infection; epidural and/or subdural infection; brain abscess;meningitis/ventriculitis; shunt-related infection; urinary tract infection;pneumonia (aspiration)

Head and neck Cellulitis/wound infection; deep facial space infection; mastoiditis;sinusitis; aspiration/nosocomial pneumonia; cavernous (or other)sinus thrombosis; meningitis; brain abscess; retropharyngeal andparavertebral abscesses; osteomyelitis

Upper gastrointestinal Mediastinitis; tracheo-esophageal fistula with pneumonitis; gastricperforation and abscess; feeding-tube-related infections

Breast Surgical wound infection; cellulitis/lymphangitis following axillary nodedissection; mastitis; breast abscess

Hepatobiliary–pancreatic Surgical wound infection; peritonitis; ascendingcholangitis � bacteremia; hepatic, pancreatic, or subdiaphragmaticabscess

Lower gastrointestinal and pelvic Wound infection, peritonitis; intra-abdominal or pelvic abscess; acuteor chronic urinary tract infection; necrotizing fasciitis; typhlitis;enterocolitis (radiation-induced); perianal/perirectal infection;sacral/coccygeal osteomyelitis

Genitourinary and prostate Acute and chronic pyelonephritis � bacteremia; prostatitis; catheter-related complicated urinary tract infection; wound infection

Bone, joints, cartilage Surgical wound infection; skin and skin structure infection; bursitis;synovitis; septic arthritis; osteomyelitis; infected prosthesis

spp. are significant pathogens in the nosocomialsetting. Similarly, most respiratory infectionsare caused by the resident oropharyngeal flora(Streptococcus pneumoniae, Haemophilus influen-zae, mouth anaerobes, etc.), with Staphylococcusspp. and Gram-negative bacilli gaining pre-dominance in the hospital. Enteric Gram-negative bacilli, intestinal anaerobes, and theenterococci dominate abdominal and pelvicsites of infection. Polymicrobial infections occurcommonly when there is tissue involvement.9

Examples include pneumonia, complicated orextensive wound infections, neutropenic ente-rocolitis (typhlitis), perirectal infections, andother skin and skin structure infections.Catheter-associated infections may also bepolymicrobial in nature. Gram-positive cocci,Gram-negative bacilli, anaerobes, and yeast(Candida spp.) are commonly isolated, depend-ing upon the site of infection. Occasionally, bac-terial, fungal, and/or viral infections maycoexist. Candida spp. frequently colonize debili-tated hospitalized patients, particularly thosewho have received multiple or prolongedcourses of broad-spectrum antibacterial ther-apy. Candiduria and candidemia are notuncommon in this setting, although dissemi-nated candidiasis is distinctly uncommon insolid tumor patients. Colonization with Candidaspp., therefore, is not sufficient reason for anti-fungal therapy in such patients. However, colo-nization at multiple sites in the same patientdoes increase the likelihood of disseminatedinfection and pre-emptive therapy might beindicated in some patients who are heavily col-onized. The emergence of resistant Candida spp.such as C. krusei, C. glabrata, and C. tropicalis isof great concern.

Localized fungal infections such as primarycutaneous aspergillosis (associated with vascu-lar catheters) or nailbed infection by Fusariumand other fungi are rare, and seldom progressto more invasive/disseminated infections.Invasive mold infections are rare. Localdebridement and a short course of antifungaltherapy usually produces satisfactory responserates. Viral infections (cytomegalovirus (CMV),

varicella zoster virus (VZV), Epstein–Barr virus(EBV), respiratory syncytial virus (RSV), respi-ratory viruses) and parasitic infections (toxo-plasmosis, Strongyloidiasis) are also quite rare insolid tumor patients. There are increasingreports of the occurrence of Pneumocystis cariniipneumonia (PCP) in patients with breastcancer, and other solid tumor patients receivingcorticosteroid or other immunosuppressivetherapies.10–12 A breakdown of predominantpathogens according to site of infection is pro-vided in Table 5.3.

CLINICAL FEATURES AND DIAGNOSIS

The predominant clinical features encounteredwith specific infections depend largely on thesite and nature of the infection. As a generalrule, patients who are severely neutropenic orare receiving corticosteroid or other type ofimmunotherapy have a blunted inflammatoryresponse, leading to a paucity of clinical signsand symptoms.13 In contrast, most patients withsolid tumors who develop an infection have anormal, vigorous inflammatory response, mak-ing clinical evaluation and diagnosis a little eas-ier to accomplish.

There is no substitute for a careful anddetailed history and a thorough physical exami-nation as part of the initial evaluation. Since theinstitution of broad-spectrum empiric therapyis generally not as critical as in neutropenicpatients with hematologic malignancies, timespent on obtaining pertinent historical informa-tion and conducting a physical examination canoften lead to the identification of a specificfocus. For example, a history of travel to or resi-dence in areas endemic for specific infections(tuberculosis, endemic mycoses, parasitic dis-eases) is important, and might help draw atten-tion to them as the patient is being evaluated.Knowledge of prior surgical procedures andimplanted prosthetic devices is also an impor-tant historical factor in such patients, since itmay help identify a specific focus of infection.The use of prior antibiotics and over-the-


counter medications can alter the nature andspectrum of subsequent infection, and shouldbe determined when interviewing the patient.An immunization history (e.g. pneumococcalvaccine) might also be useful. Seizure activity

might suggest an intracranial process and/oraspiration pneumonia. A persistent cough, pro-ductive of large amounts of foul-smelling spu-tum, is consistent with the presence of a lungabscess or post-obstructive pneumonitis. The


Table 5.3 Predominant organisms by site of infection

Site Predominant organisms

Bloodstream Coagulase-negative staphylococci, Staphylococcusaureus, Enterococcus spp., Streptococcus spp.,Enterobacteriaceae, Ps. aeruginosa, Candida spp.

Central nervous system (including shunt- Coagulase-negative staphylococci, S. aureus, enteric related and post-surgical infections) Gram-negative bacilli, Streptococcus pneumoniae,

Haemophilus influenzae, mouth anaerobes, Listeriamonocytogenes, Cryptococcus neoformans

Respiratory tract (upper respiratory infections, S. pneumoniae, H. influenzae, Moraxella catarrhalis,lower respiratory tract infections, lung abscess, Enterobacteriaceae, Ps. aeruginosa, coagulase-negativeempyema) staphylococci, S. aureus, mouth anaerobes

(Peptococcus, Peptostreptococcus, Fusobacterium,etc.)

Biliary tract Enteric Gram-negative bacilli, Enterococcus spp.,enteric anaerobes, Candida spp.

Intra-abdominal/pelvic Enteric Gram-negative bacilli, Enterococcus spp.,enteric anaerobes (Bacteroides spp., Clostridium spp.),Candida spp.

Catheter-related Coagulase-negative staphylococci, S. aureus, Bacillusspp., Corynebacterium jeikeium, Ps. aeruginosa,Acinetobacter spp., Stenotrophomonas maltophilia,enteric Gram-negative bacilli, mycobacteria (rapidgrowers), Candida spp.

Skin/skin structure S. aureus, Streptococcus spp., Ps. aeruginosa, entericGram-negative bacilli, anaerobes

expression of urine or fecal material throughthe vagina indicates the presence of a vesico-vaginal or recto-vaginal fistula. During physicalexamination, close attention needs to befocused on the oropharyngeal cavities, thegroin and perirectal region, sites where obstruc-tion might occur, surgical wounds, prostheticdevices, irradiated areas, the skin (including thenail beds), catheter insertion sites, and theparanasal sinuses.

There is nothing unique about the initial lab-oratory evaluation of patients with solidtumors. A basic evaluation that includes chem-ical analysis of the blood and urine, a completeand differential blood count, tests for hepaticand renal function, and all appropriate microbi-ological cultures should be conducted on allpatients. In patients with diarrhea, testing forClostridium difficile toxin is often recommendedas the first step. If this is negative and the diar-rhea is suspected to be of infectious etiology,the stools should be tested for specific bacteria(e.g. Aeromonas, Campylobacter, Plesiomonas,Salmonella, Shigella, and Yersinia), protozoa(amoeba, Cryptospordium, Giardia, etc.), viruses(rotavirus, CMV), and also mycobacteria(Mycobacterium avium complex: MAC). Urinecultures are indicated if the patient is sympto-matic, the urinalysis is abnormal, a urinarycatheter is in place, or surgery involving theurinary tract has been performed. Examinationof the cerebrospinal, pleural, and ascitic fluidsshould be performed when clinically indicated,and the specimens should be sent for bacterial,fungal, and other appropriate cultures.Radiographic evaluation of the chest is notuseful on a routine basis, but is indicated whenprimary or metastatic lung disease or pul-monary symptoms (cough, sputum, dyspnea,chest pain, or hemoptysis) are present.Radiographic imaging studies (computedtomography, magnetic resonance imaging) maybe particularly useful in evaluating the CNS,paranasal sinuses, pulmonary, abdominal, andpelvic foci. Indium-labeled leukocyte scans,bone scans, and gallium scans are often done,but provide useful diagnostic information infre-

quently. Doppler or venous flow studies areuseful for the evaluation of deep venous throm-bophlebitis that is often not clinically apparent.Serologic studies are generally not very usefulunless a specific pathogen that elicits a serologicresponse is suspected.

Cutaneous lesions should be biopsied forGram staining, staining for other pathogens(fungi, mycobacteria, and protozoa), culture,and cytologic examination. Other invasive pro-cedures, such as biopsies of the lung, liver,bone, brain, lymph nodes, and bone marrow,should be performed expeditiously when clini-cally indicated, and handled in a similar fash-ion. In general, such procedures are easier toperform in patients with solid tumors than inpatients with hematologic malignancies, sincemost patients are not thrombocytopenic, andhemostasis is not a significant problem.

The role of serial microbiological surveillancecultures in patients with solid tumors has notbeen established. They can occasionally provideuseful information; however, the predictiveyield of such cultures is low. Knowledge oflocal microflora, however, is an important fac-tor when considering the use of surveillancecultures. In institutions where resistant organ-isms such as VRE, methicillin-resistant S. aureus(MRSA), Ps. aeruginosa, Stenotrophomonas mal-tophilia, C. krusei, etc. are relatively common,prior knowledge of colonization with suchorganisms can help in the choice of appropriatetherapy when infection develops, and in theprevention of nosocomial transmission of theseorganisms from patient to patient.

THERAPY

Experience from the National Cancer Institute(Bethesda, MD) and from institutions that par-ticipate in the IATCG/EORTC trials indicatesthat the frequency of bacteremias is lower inneutropenic patients with solid tumors than inthose with hematologic malignancies (12% ver-sus 25%), whereas episodes of unexplainedfever are more common (65% versus 40–50%).


In contrast, at the University of Texas MDAnderson Cancer Center, clinically and micro-biologically documented infections are encoun-tered more often in patients with solid tumors(50–60% versus 40%) and episodes of unex-plained fever are less common in these patients(40% versus 60%).4,14 Therapeutic strategies,therefore, are largely institution-dependant,and will be discussed under two broad cat-egories: treatment of unexplained fever, andtreatment of specific infections.

TREATMENT OF UNEXPLAINED FEVER

The principles of the management of unex-plained fever in solid tumor patients with neu-tropenia are similar to those for patients withhematologic malignancies.4,15,16 Empiric therapygenerally consists of the administration of par-enteral broad-spectrum antibiotics, while thepatient is monitored in the hospital. Severalchoices for initial therapy are available, but spe-cific regimens need to be tailored to localmicroflora and susceptibility patterns. Thesechoices include the following:

combination regimens• aminoglycoside plus �-lactam;• glycopeptide plus �-lactam;• glycopeptide plus quinolone;• aminoglycoside plus quinolone.

monotherapy• extended-spectrum anti-pseudomonal

cephalosporin;• carbapenem.

Initial usage of glycopeptides should be con-sidered only when the likelihood of infectionwith resistant Gram-positive organisms is high(MRSA, viridans streptococci, coagulase-negative staphylococci, and Corynebacteriumjeikeium). The routine use of these agents shouldbe avoided, since this has been associated withthe emergence of VRE and other glycopeptide-resistant microorganisms.17,18 Of concern is the

increasing level of glycopeptide resistanceamong organisms such as Bacillus spp. andRhodococcus spp.19 If used empirically, gly-copeptide therapy should be discontinuedpromptly when relevant microbiological cul-tures are negative for resistant Gram-positiveorganisms. Various combination regimens andbroad-spectrum agents used as monotherapyhave been associated with overall responserates of 65–95%.15

Changes or alternations of the initial regimenare indicated for failure to respond and/or pro-gressive infection, new clinical developments,or microbiological data and susceptibilityinformation that indicate the need for a change.Frequent alternations or modifications includethe following:

• addition of a glycopeptide if not used ini-tially, when Gram-positive coverage needsto be strengthened;

• addition of a second drug with potentGram-negative activity (if only one wasincluded in the initial regimen), especiallyif a documented Gram-negative infection isnot responding adequately;

• additional anaerobic coverage (clin-damycin, metronidazole) for infectionssuch as necrotizing gingivitis, neutropenicenterocolitis, perirectal abscesses, or otherintra-abdominal pelvic sites;

• addition of empiric antifungal agents (flu-conazole, itraconazole, amphotericin B orits lipid formulations);

• surgical intervention (e.g. drainage of anabscess or debridement for suspected/doc-umented fungal infections of devitalizedtissue) or removal of foreign bodies such asan infected catheter is occasionally neces-sary.

Standard therapy for febrile neutropenia inpatients with solid tumors does not differ muchfrom that in patients with hematologic malig-nancies, except that the duration of treatment isgenerally shorter owing to the shorter ‘at-risk’period. Recently, clinical and statisticallyderived risk-prediction models have enabled


clinicians to simplify the treatment of ‘low-risk’neutropenic patients (most of whom arepatients with solid tumors receiving conven-tional chemotherapy), with the use of par-enteral, sequential (intravenous → oral), or oralregimens that enable early discharge from thehospital, or outpatient therapy for the entireduration of the febrile episode.20–23 This ‘risk-based’ approach to therapy might be more cost-effective, and might result in improved qualityof life for patients and their care-givers, thanthe standard approach of hospital-based ther-apy. Early discharge from hospital or outpa-tient therapy also decreases the frequency of‘healthcare-associated’ infections (many ofwhich are caused by multidrug-resistant organ-isms), and reduces other hazards of hospital-based care.24–26 (See Chapter 9.)

THERAPY OF SPECIFIC INFECTIONS

Gram-positive infections

Response to standard antibacterial therapy insolid tumor patients who develop Gram-positive infections exceeds 95%.27 Agents otherthan glycopeptides (vancomycin, teicoplanin)to which the specific pathogen isolated is sus-ceptible are generally appropriate. Theseinclude anti-staphylococcal penicillins (naf-cillin, oxacillin), other �-lactams, trimetho-prim/sulfamethoxazole (TMP/SMX), the tetra-cyclines, the macrolides, clindamycin, andsome newer quinolones (gatifloxacin, moxi-floxacin).28,29 Glycopeptides are indicated whenan organism resistant to other antimicrobialagents is isolated, or in patients who are allergicto �-lactams or other antimicrobial agents, butshould not be used solely for the sake of conve-nience of administration. Occasionally, combi-nation regimens that interact synergistically arepreferable. Examples include an aminoglyco-side plus a �-lactam, and vancomycin plus anaminoglycoside or rifampin.

New agents have recently become availablefor the treatment of glycopeptide-resistant

organisms, particularly VRE. These include theoxazolidinone linezolid, and the quinipristin/dalfopristin combination Synercid.30,31 Althoughthese agents are active against a wide spectrumof organisms, they should not be used indis-criminately, since resistance to them is alreadybeing encountered.

Many Gram-positive bacteremic infectionsare related to the presence of an indwellingcatheter. Many of these infections, especiallythose caused by coagulase-negative staphylo-cocci, can be treated with antimicrobial agentsalone, without removal of the catheter.However, persistent bacteremia or fever,significant infection at the catheter entry site, orthe isolation of certain microorganisms (Bacillusspp., S. aureus, C. jeikeium) might necessitatecatheter removal. The duration of therapy is10–14 days unless evidence of endocarditis,septic thrombophlebitis, or other signs of dis-semination is present.

Gram-negative infections

A large number of antimicrobial agents withpotent activity against commonly isolated Gram-negative bacilli are currently available. The mostcommonly used are the extended-spectrumcephalosporins (ceftazidime, cefepime), anti-pseudomonal penicillin/�-lactamase combina-tions (ticarcillin/clavulanate, piperacillin/tazobactam), the carbapenems (imipenem,meropenem), the monobactam (aztreonam), theaminoglycosides (gentamicin, tobramycin,amikacin), and the quinolones (ciprofloxacin,gatifloxacin). Agents such as TMP/SMX andrifampin also have useful activity against manyGram-negative pathogens, although they are sel-dom used as ‘first-line’ agents. The aminoglyco-sides are not appropriate for the monotherapyof Gram-negative infections in neutropenicpatients, even if in vitro susceptibility of thecausative pathogen is demonstrated. They arebest utilized in combination with other classes ofantimicrobial agents, particularly if such combi-nations are synergistic.


The appropriate management of infectionscaused by Ps. aeruginosa continues to bedebated. Some authorities recommend the useof combination regimens (preferably synergis-tic) without exception, particularly in patientswith severe neutropenia. However, in a largereview of Ps. aeruginosa bacteremia in cancerpatients from the MD Anderson Cancer Center,the most critical factors for a favorable responsewere the timing of therapy (i.e. any delayadversely affected outcome), and susceptibilityof the organism to the non-aminoglycosidecomponent of the therapeutic regimen.32 Thisexperience has been confirmed in a follow-upstudy from the same institution.33 Althoughcombination therapy may not always be neces-sary, Ps. aeruginosa is known to be an aggressivepathogen, and is associated with considerablemorbidity and mortality. In light of recent dataindicating that major organ and/or tissueinvolvement is associated with poorer out-comes, particularly in patients with Ps. aerugi-nosa bacteremia, it might be prudent toadminister combination regimens to patientswith such complicated infections.34

Unlike Gram-positive bacteremias, mostGram-negative bacteremias are not catheter-related. However, Acinetobacter spp., Pseudo-monas spp., and S. maltophilia are more likely tocause catheter-related infections than otherGram-negatives, and catheter removal, in addi-tion to appropriate antimicrobial therapy,might occasionally be necessary.

Anaerobic infections

Many empiric regimens contain adequateanaerobic coverage, since they include agentssuch as the carbapenems and piperacillin/tazobactam. However, monotherapy withextended-spectrum cephalosporins or combina-tion therapy with cephalosporin/aminoglyco-side does not provide adequate anaerobiccoverage. In patients receiving such regimens,the addition of agents such as clindamycin ormetronidazole or a change to a broad-spectrum

agent with anaerobic activity is indicated whenanaerobes are isolated or strongly suspected tobe present. Surgical intervention may be life-saving in some infections involving anaerobes,and surgical consultants should be involved inthe management of necrotizing anaerobic infec-tions from the onset.

Fungal and viral infections

Fluconazole is effective for the treatment of can-didemia caused by susceptible species (i.e. mostCandida spp. except C. krusei.35 Species otherthan C. albicans might require high-dosefluconazole therapy (800–1200 mg/day) foradequate response to occur, owing to ‘dose-dependent susceptibility’ of these iso-lates.36 Removal or exchange of infectedcatheters shortens the duration of candidemia,and hastens response.37 Amphotericin B is alsoeffective, but its use is limited by substantialtoxicity. The lipid formulations of amphotericinB are much less toxic, but far more expensive,and are indicated when toxicity or refractoryinfections make the use of amphotericin Bdeoxycholate unrealistic or impractical.38

Fluconazole is also effective for more localizedinfections (thrush, esophagitis, vaginitis, can-diduria). Cryptococcal meningitis is seen occa-sionally – particularly in patients receivingprolonged corticosteroid therapy. Other, dis-seminated fungal infections are rare in solidtumor patients, and their treatment is standard.

Patients with solid tumors are not at particu-lar risk of developing disseminated viral infec-tions, and there is nothing unique about theirmanagement, when documented in this patientpopulation. Acyclovir remains the agent ofchoice for the treatment of infections caused byherpes simplex virus (HSV) and VZV. For local-ized lesions or where parenteral therapy is nolonger necessary, the newer oral agents (valaci-clovir, famciclovir) provide greater bioavailabil-ity than oral acyclovir. CMV, EBV, and humanherpesvirus-6 (HHV-6) are rarely encounteredin solid tumor patients.


Community respiratory viruses (influenza,parainfluenza, RSV) have recently been shownto be important causes of morbidity and mor-tality in recipients of bone marrow transplantsand patients with hematologic malignancies.They have not been studied extensively inpatients with solid tumors, and, in all likeli-hood, do not have the same impact in this set-ting. Management according to currentlyestablished standards is adequate.

Many solid tumor patients develop elevationof aminotransferase (transaminase) levels, indi-cating hepatic injury, but it is often difficult todetermine whether the disease is viral or drug-induced. The presence of hepatitis may result inconsiderable delays in the administration ofantineoplastic therapy, since it is hazardous toadminister hepatotoxic drugs to patients whoseliver functions are already impaired. Severalreports have focused on the phenomenon ofreactivation of quiescent liver disease due tohepatitis B virus following cytotoxic orimmunosuppressive therapy.39,40 The clinicalpicture is that of fulminant hepatitis, which hasled to the requirement for liver transplantationin some patients.

Parasitic infections (toxoplasmosis, cryp-tosporidosis, etc.) are uncommon in solid tumorpatients. Standard therapeutic measures areindicated.

Mycobacterial infections

Mycobacterial infections occur more frequentlyin patients with cancers than in the generalpopulation. In a review of 201 cases of tubercu-losis from the Memorial Sloan-Kettering CancerCenter, patients with lung cancers had the high-est prevalence of tuberculosis (920 per 100 000)among solid tumor patients.41 Lung cancer andhead and neck cancer patients presented moreoften with tuberculosis at the time of cancerdiagnosis, whereas patients with other malig-nancies developed tuberculosis more oftenwhile receiving cancer chemotherapy. Morerecent data from the MD Anderson Cancer

Center indicate that tuberculosis remainsparticularly common in patients with head andneck cancers42 (Figure 5.1). Fifty percent ofpatients were receiving antineoplastic therapyand 14.2% were, or had recently been, on corti-costeroids. The various manifestations of tuber-culosis described in this report includepulmonary tuberculosis, adenitis, chest walland psoas abscess, pleuritis, meningitis, andwidely disseminated infection. No multidrug-resistant isolates were encountered. Despitethis, the mortality rate was 25%.

Two mechanisms of tuberculosis reactivationhave been postulated in this setting. Tumornecrosis can cause the breakdown of pre-existing granulomas, liberating sequesteredmycobacteria. Alternatively, chemotherapy-induced immunosuppression, corticosteroids,or malignancy-associated cachexia can impaircell-mediated immunity to such a degree thatreactivation of tuberculosis occurs. Since ther-apy for many solid tumors has increasinglybecome more intensive, the possibility of anincrease in the incidence of tuberculosis reacti-vation exists. Consequently, a high index ofsuspicion for tuberculosis needs to be main-tained, particularly if a patient’s history orepidemiologic background suggests priorexposure to or active treatment of tuberculosis.In such patients, the development of pul-monary symptoms and/or radiographic find-ings should prompt an evaluation for thepresence of tuberculosis (tuberculin skin test;sputum, bronchoalveolar lavage, or lung biopsysamples for acid-fast bacillus (AFB) smears andcultures). Prophylaxis with daily isoniazid isindicated for a period of 6–12 months inpatients with a positive tuberculin test and noevidence of active tuberculosis. Newer prophy-lactic regimens (rifampin plus pyrazinamide)that can be administered for a shorter periodhave recently been evaluated – primarily inpatients with AIDS – but have not been wellstudied in patients with neoplastic diseases.43

Whenever AFBs are identified in smears orcultures of respiratory specimens or on histo-pathology or cytology, therapy for presumed



Figure 5.1 Right upper lobe cavitary tuberculosis ina Latin American male with primary nasopharyngealcarcinoma. The patient responded to standard anti-tubercular chemotherapy.

active pulmonary tuberculosis should be insti-tuted. Upon final identification of the organ-isms, therapy should be continued (iftuberculosis is confirmed), modified (if otherpathogenic mycobacteria such as M. kansasii orMAC are identified), or discontinued (if conta-minants such as M. gordonae are identified).

Mycobacterium kansasii has traditionally beenconsidered to be the most virulent non-tuberculous mycobacterium, and infectionscaused by M. kansasii have also been describedwith increased frequency in cancer patients. Ina recently published report from the MDAnderson Cancer Center, the incidence of M.kansasii infection was 25 cases per 100 000cancer patient registrations.44 The infection wasactually more common in patients withleukemia than in solid tumor patients (115 ver-

sus 14 cases per 100 000), and pleuropulmonarydisease was predominant. Two patients (8%)had disseminated infection. Most patients weretreated with rifampin-based regimens.Although 60% died within 20 months of M.kansasii isolation, death was attributed to theprimary neoplasm and not to M. kansasii inmost instances.

Rapidly growing mycobacteria have alsobeen reported to cause pulmonary, catheter-related, and disseminated infections in cancerpatients.45–48 These organisms are isolated lessfrequently than M. tuberculosis and M. kansasii,and are encountered predominantly in patientswith solid tumors.45 They are resistant to stan-dard antitubercular drugs, but are susceptibleto agents such as TMP/SMX, the quinolones,and the macrolides. As in the case of othermycobacterial infections, combination regimensare recommended.

Similarly, MAC bacteria cause pulmonary, ordisseminated infections in cancer patients, pre-dominantly in solid tumor patients.49 Theseorganisms are also resistant to standard antitu-bercular agents, and require combination ther-apy with agents to which they are susceptible.50

SPECIAL CONSIDERATIONS

Post-obstructive pulmonary infections

Lung cancer patients often develop focal pul-monary infections. In most instances, this isdue to partial obstruction of the bronchial treeleading to atelectasis and post-obstructivepneumonitis. Occasionally, the infection mayprogress to abscess formation and evenempyema (Figure 5.2). In a small number ofcases, infection occurs within an area of tumornecrosis, rather than as a result of airwayobstruction. These infections are predominantlypolymicrobial (staphylococci, Gram-negativebacilli, anaerobes), and, in addition to pro-longed broad-spectrum antibiotic therapy,methods to overcome the obstruction (antineo-plastic therapy, radiation, endobronchial

brachytherapy stent placement) are usuallynecessary to ensure adequate drainage of theinfected lung.

Infections associated with breast cancersurgery

Breast cancer is the most commonly diagnosedmalignancy in women in the USA, accountingfor 182 800 of 600 400 new cancer cases inwomen in 2000.51 Most of these patientsundergo some surgical procedure of theinvolved breast, along with resection of ipsilat-eral lymph nodes, and radiation, which predis-poses them to poor wound healing and various


Figure 5.2 This 48-year-old man with primaryadenocarcinoma of the lung presented with fever,chills, and a cough productive of foul-smellingsputum. Chest radiograph revealed extensiveinfiltration in the left lung, post-obstructivepneumonia, progressing to lung abscess formation.Note the air–fluid level in the abscess.

infectious complications (both early and late),primarily involving the skin and soft tissue.52–54

These range from post-operative wound-infection, post-operative hematoma or seromaformation followed by secondary infection,breast cellulitis, lymphedema and lymphangi-tis, mastitis, and breast abscess formation(Figure 5.3). In a matched case–control studydesigned to identify risk factors for the devel-opment of breast cellulitis after breast conserva-tion therapy, the following six factors weresignificant:55

• drainage of a hematoma;• postoperative ecchymosis• presence of lymphedema;• volume of resected breast tissue;• previous number of biopsies;• number of breast seroma aspiration proce-

dures.

Patients with these risk factors may be at life-long risk for developing infections that are slowto respond and are often recurrent. Even minortrauma in patients with lymph node dissectioncan lead to cellulitis or more invasive infections.Patients should be instructed to minimize suchevents by avoiding phlebotomy, vascular accesscatheters, blood pressure monitoring or otherroutine procedures on the involved extremity.Early and aggressive therapy at the earliest signof infection can prevent local skin breakdownand invasion. A selected group of patients withrecurrent infections might benefit from chronicsuppressive antimicrobial therapy.

Ommaya-reservoir-related infections

The treatment of diffuse leptomeningeal diseaseor neoplastic meningitis includes chemotherapydelivered through an Ommaya reservoir, sinceplacing an Ommaya reservoir allows directaccess to the ventricular system for both fluidanalysis and drug delivery. These reservoirs,however, can act as a focus of infection,particularly in patients in whom frequent


Figure 5.3 This 60-year-oldwoman underwent surgery for leftbreast carcinoma, including theremoval of 16 lymph nodes. Shedeveloped massive edema, pain,and cellulitis in her left upperextremity followingchemotherapy. Note thedifference between the normaland infected extremity.

ventricular access is required. Coagulase-negative staphylococci, and other commoninhabitants of the skin, are the most frequentpathogens in this setting. Systemic and intra-ventricular installation of antimicrobial agentsto which the offending pathogens are suscepti-ble usually produces a satisfactory response. Insome cases, removal of the infected Ommayareservoir is necessary, in addition to antimicro-bial therapy.

Hepatobiliary infection

Obstruction of the biliary tract as a result ofhepatobiliary and/or pancreatic tumors resultsin the development of ascending cholangitis.8

On rare occasions, single or multiple hepaticabscesses might also develop. Ascendingcholangitis might also be the initial manifesta-tion of local malignancy, and may lead to thisdiagnosis during evaluation. Finally, hepaticabscesses have been reported after invasiveprocedures for hepatocellular carcinoma,

including the administration of intra-arterialchemotherapy. Most of these infections arepolymicrobial in nature, with enteric Gram-negative bacilli, anaerobes and Enterococcusspp. predominating. Broad-spectrum antimicro-bial therapy and percutaneous drainage areoften necessary in order to achieve an adequateresponse and prevent recurrent infection.

Gynecologic-cancer-associated infections

Local obstruction caused by tumor, tumornecrosis, and therapeutic modalities (includingchemotherapy, surgery, and radiation) all con-tribute to infections in patients with gyneco-logic malignancy. Tumor-related infectionsdepend on the site and size of the tumor. Forexample, infections complicating stage I cervi-cal cancer generally involve the surfaces of thetumor and are usually limited to the vagina.56

As tumors enlarge, obstruction to variousorgans results in the development of urinarytract infections, tubo-ovarian abscesses, and

pyometra. Rupture of tubo-ovarian abscesses orpyometra can lead to the development of acuteperitonitis.57,58 These complications are rare,since most gynecologic cancers are detectedand treated at an earlier stage.

Infections related to the treatment modalitiesused in patients with gynecologic cancersdepend on the mode of treatment. The generalprinciples relating to the management of febrileneutropenic patients also apply to patientswith gynecologic cancers when they receivechemotherapy and become neutropenic. How-ever, documented intrapelvic/abdominal sitesare more common than in patients with othercancers, and empiric antimicrobial coverageneeds to include potent Gram-negative andanaerobic activity. In addition to routine post-surgical infections (e.g. wound infections), theremoval of pelvic organs and tissue results inthe creation of spaces that are filled by bloodand serum, with a high potential for infection.Complications of radiation include bowelobstruction or perforation, and fistula forma-tion. All these complications are life-threatening, and require prompt and aggressiveantimicrobial therapy in conjunction withappropriate surgical intervention.

Infections mimicking cancer

Certain infections can occasionally produceclinical manifestations and radiographic imagesthat are indistinguishable from those producedby neoplasms. The most common site of suchlesions is the lung. Some patients may be totallyasymptomatic, and pulmonary lesions may beidentified on routine yearly physician visits, orduring medical evaluations required by newemployers or insurance companies. Mostlesions identified in this manner do turn out tobe neoplastic. However, in a large study con-ducted at the MD Anderson Cancer Center, 37of 2908 patients (1.3%) with pulmonary lesionswho were referred to ‘rule out’ lung cancer hadan infection instead.59 In none of these patientswas an infection strongly suspected during the

primary evaluation. Fungal infections (histo-plasmosis, coccidioidomycosis, and cryptococ-cosis) accounted for 46% of these infections,and mycobacterial infections for 27%. Bacterialand parasitic infections were uncommon.

Lesions in other parts of the body (liver,bone, thyroid, lymph nodes, breast, etc.) canalso simulate cancer and create diagnostic andtherapeutic challenges.60–63 A recent report high-lights a series of patients with actinomycosis,who presented with mandibular and pelviclesions that were thought to be neoplasticon initial presentation and evaluation.64

Documentation of a specific diagnosis bymicrobiological and/or histologic techniques ismandatory for the proper management of thesepatients.

Many of these infections can also presentdiagnostic challenges in patients with cancerwho have been effectively treated. When newlesions are detected in such patients, the mostcommon suspicion is that of metastatic or recur-rent neoplastic disease. These lesions also needto be evaluated carefully, and a specific diagno-sis made, since the management of recurrentneoplastic disease is totally different from thatof infection.

SUMMARY

Patients who have hematologic malignanciesoften develop life-threatening infections, espe-cially when they are severely neutropenic. Sincethis is a relatively homogeneous group, it hasbeen the subject of intense study, and a largenumber of well-designed trials have beeninstrumental in developing general and specificprinciples for the management of such patients.In contrast, patients with solid tumors areextremely heterogeneous, and infections insuch patients have been less well studied. Theydo, however, represent the majority of newcases of cancer diagnosed each year, anddevelop a large number of infectious complica-tions. Some are related to the tumor itself, andsome to local phenomena such as obstruction or


disruption of normal anatomic barriers. Othersare related to various treatment modalities(chemotherapy, surgery, or radiation) and thenature of these infections depends on the treat-ment modality and the type and site of tumorbeing treated. Since most tumors are diagnosedearlier than they used to be owing to improve-ments in screening programs and diagnostictechniques, infections related to the tumor itselfare becoming less common. In contrast, patientswith solid tumors are receiving increasinglyintensive antineoplastic therapy (often employ-ing multiple modalities in the same patient) inorder to achieve better antitumor responses.Consequently, infections related to these treat-ment modalities have become more frequent. Inorder to better understand the diversity ofinfections seen and to develop managementstrategies specifically for the different solidtumor groups (CNS tumors, breast cancer, lungcancer, etc.), carefully designed studies focus-ing on the predisposing factors, epidemiology,manifestations, diagnosis, and treatment ofthese infections need to be conducted. Suchstudies will provide the information needed toappropriately manage patients with solidtumors who develop infections, rather thanapplying management strategies that have beendeveloped for, and are more pertinent in,patients with hematologic malignancies.

REFERENCES

1. Bodey GP, Infection in cancer patients: a contin-uing association. Am J Med 1986; 81(Suppl A-1):11–26.

2. Pizzo PA, Management of fever in patients withcancer and treatment induced neutropenia. NEngl J Med 1993; 328: 1323–32.

3. Greenlee RT, Murray T, Bolden S, Wingo PA,Cancer statistics, 2000. CA Cancer J Clinicians2000; 50: 7–33.

4. Rolston KVI, Infections in patients with solidtumors. In: Management of Infections in Immuno-compromised Patients. (Pizzo P, Glauser MP, eds).London: WB Saunders Company Ltd, 2000;117–40.

5. Rolston KVI, Raad I, Whimbey E, Bodey GP, Thechanging spectrum of bacterial infections infebrile neutropenic patients. In: FebrileNeutropenia (Klastersky JA, ed). Berlin: Springer-Verlag, 1998: 53–6.

6. Rolston KVI, Commentary: Chemoprophylaxisand bacterial resistance in neutropenic patients.Infect Dis Clin Pract 1998; 7: 202–4.

7. Keidan Rd, Hoffman JP, Weese JL et al, Delayedbreast abscesses after lumpectomy and radiationtherapy. Am Surg 1990; 56: 440–4.

8. Rolston KVI, Dholakia N, Rodriguez S,Rubenstein EB, Nature and outcome of febrileepisodes in patients with pancreatic and hepato-biliary cancer. Support Care Cancer 1995; 3:414–17.

9. Elting LS, Bodey GP, Fainstein V, Polymicrobialsepticemia in the cancer patient. Medicine 1986;65: 218–22.

10. Kulke MH, Vance EA, Pneumocystis carinii pneu-monia in patients receiving chemotherapy forbreast cancer. Clin Infect Dis 1997; 25: 215–18.

11. Sepkowitz KA, Brown AE, Talzak EE et al,Pneumocystis carinii pneumonia among patientswithout AIDS at a cancer hospital. JAMA 1992;267: 832–7.

12. Koibuchi Y, Iino Y, Yokoe T et al, Pneumocystiscarinii pneumonia during treatment for recurrentbreast cancer. A case report. Jpn J Clin Oncol1995; 25: 218–21.

13. Sickles EA, Greene WH, Wiernik PH, Clinicalpresentation of infection in granulocytopenicpatients. Arch Intern Med 1975; 135: 715.

14. Rodriguez SJ, Rolston KVI, Infections in cancerpatients with solid tumors. In: Proceedings of 37thAnnual Meeting of the Infectious Diseases Society ofAmerica, Philadelphia, PA, November 18–21, 1999:Abst 456.

15. Hughes WT, Armstrong D, Bodey GP et al, 1997guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

16. NCCN Practice Guidelines for Fever andNeutropenia, Oncology 1999; 13: 197–257.

17. Murray BE, Vancomycin-resistant enterococcalinfections. N Engl J Med 2000; 342: 710–21.

18. Van der Auweera P, Pensart N, Korten V et al,Influence of oral glycopeptides on the fecal floraof human volunteers: selection of highlyglycopeptide-resistant enterococci. J Infect Dis1996; 173: 1129–36.


19. Diekema DJ, Coffman SL, Marshall SA et al,Comparison of activities of broad spectrum �-lac-tam compounds against 1128 Gram-positive coccirecently isolated in cancer-treatment centers.Antimicrob Agents Chemother 1999; 43: 940–3.

20. Talcott JA, Finberg R, Mayer RJ et al, The med-ical course of cancer patients with fever and neu-tropenia. Arch Intern Med 1988; 148: 2501–68.

21. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low risk cancer patients withfever and neutropenia: a pilot study of 30patients based on validated prediction rule. JClin Oncol 1994; 12: 107–14.

22. Klastersky J, Paesman M, Rubenstein EB et al,The Multinational Association for SupportiveCare in Cancer Risk Index: a multinational scor-ing system for identifying low-risk febrile neu-tropenic cancer patients. J Clin Oncol 2000; 18:3038–51.

23. Rolston K, New trends in patient management:risk-based therapy for febrile patients with neu-tropenia. Clin Infect Dis 1999; 29: 515–21.

24. Rubenstein EB, Rolston K, Benjamin RS et al,Outpatient treatment of febrile episodes in lowrisk neutropenic cancer patients. Cancer 1993; 71:3640–6.

25. Gerberding JL, Preventing antimicrobial-resistant healthcare infections: beyond 2000.Clinical Updates in Infectious Diseases, Vol V, Issue2, National Foundation for Infectious Diseases,Bethesda, MD, 2000.

26. Kohn L, Corrigan J, Donaldson M (eds), To Err isHuman: Building a Safer Health System.Committee on Quality of Health Care inAmerica. Institute of Medicine Report. NationalAcademy Press: Washington, DC, 2000.

27. Rubin M, Hathorn JW, Marshall D et al, Gram-positive infections and the use of vancomycin in550 episodes of fever and neutropenia. AnnIntern Med 1988; 108: 30–5.

28. Rolston KVI, LeBlanc B, Ho DH, In vitro activityof gatifloxacin against Gram-positive isolatesfrom cancer patients. In: Proceedings of 39thAnnual Interscience Conference on AntimicrobialAgents and Chemotherapy, San Francisco, September26–29, 1999: Abst 360.

29. Rolston KVI, Streeter HD, LeBlanc BM, In-vitroactivity of a new carbapenem (MK-0826) againstGram-positive organisms (GPO) from hospital-ized cancer patients. In: Proceedings of 1stInternational Symposium – Resistant Gram-Positive

Infections, San Antonio, TX, December 3–5, 2000(abst).

30. Mollering Rec, Linden PK, Reinhardt J et al, Theefficacy and safety of quinupristin/dalfopristinfor the treatment of infections caused byvancomycin-resistant Enterococcus faecium. JAntimicrob Chemother 1999; 44: 251–61.

31. Noskin GA, Siddiqui F, Stosor V, Successfultreatment of persistent vancomycin-resistantEnterococcus faecium bacteremia with linezolidand gentamicin. Clin Infect Dis 1999; 28: 689–90.

32. Bodey GP, Jadeja L, Elting L, Pseudomonas bac-teremia: retrospective analysis of 410 episodes.Arch Intern Med 1985; 145: 1621–9.

33. Chatzinikolaou I, Abi-Said D, Bodey GP et al,Recent experience with Pseudomonas aeruginosabacteremia in patients with cancer. Retrospectiveanalysis of 245 episodes. Arch Intern Med 2000;160: 501–9.

34. Elting LS, Rubenstein EB, Rolston KVI, BodeyGP, Outcomes of bacteremia in patients withcancer and neutropenia: observations from twodecades of epidemiological and clinical trials.Clin Infect Dis 1997; 25: 247–59.

35. Rex JH, Walsh TJ, Sobel JC et al, Practice guide-lines for the treatment of candidiasis. Clin InfectDis 2000; 30: 662–78.

36. Rex JH, Pfaller MA, Galgiani NJ et al,Development of interpretive breakpoints for anti-fungal susceptibility testing: conceptual frame-work and analysis of in vitro–in vivo correlationdata for fluconazole, itraconazole, and Candidainfections. Clin Infect Dis 1997; 24: 235–47.

37. Rex JH, Bennett JE, Sugar AM et al, Intravascularcatheter exchanges and the duration of can-didemia. Clin Infect Dis 1995; 21: 994–6.

38. Wong-Beringer A, Jacobs RA, Guglielmo BJ,Lipid formulations of amphotericin B: clinicalefficacy and toxicities. Clin Infect Dis 1998; 27:603–18.

39. Rolston KVI, Bodey GP, Infections in patientswith cancer. In: Cancer Medicine, 5th edn(Holland JF, Frei E, eds). Hamilton, Ontario: BCDecker, 2000: 2407–32.

40. Galbraith RM, Eddleston AL, Williams R,Zuckerman AJ, Fulminant hepatic failure inleukemia and choriocarcinoma related to with-drawal of cytotoxic drug therapy. Lancet 1975; ii:528.

41. Kaplan MH, Armstrong D, Rosen P,Tuberculosis complicating neoplastic disease: a


review of 201 cases. Cancer 1974; 33: 850–8.42. DeLaRosa G, Jacobson K, Whimbey E,

Kontoyiannis D, Tuberculosis in the ’90s in a ter-tiary cancer center. In: Proceedings of the 38thAnnual Meeting of the Infectious Diseases Society ofAmerica, New Orleans, LA, September 7–10, 2000:Abst 552.

43. Gordin F, Chaisson RE, Matts JP et al, Rifampinand pyrazinamide vs. isoniazid for prevention oftuberculosis in HIV-infected persons. JAMA2000; 283: 1445–50.

44. Jacobson K, Teira R, Libshitz H et al,Mycobacterium kansasii infections in patients withcancer. Clin Infect Dis 2000; 30: 965–9.

45. Jacobson K, Garcia R, Libshitz H, Clinical andradiological features of pulmonary diseasecaused by rapidly growing mycobacteria incancer patients. Eur J Clin Microbiol Infect Dis1998; 17: 615–21.

46. Rolston KVI, Jones PG, Fainstein V, Bodey GP,Pulmonary disease caused by rapidly growingmycobacteria in patients with cancer. Chest 1985;4: 503–6.

47. Raad II, Vartivarian S, Khan A, Bodey G,Catheter-related infections caused by theMycobacterium fortuitum complex: 15 cases andreview. Rev Infect Dis 1991; 13: 1120–5.

48. Hoy JF, Rolston KVI, Hopfer L, Bodey GP,Mycobacterium fortuitum bacteremia in patientswith cancer and long-term venous catheters. AmJ Med 1987; 83: 213–17.

49. Bhamani A, Elting L, Tarrand J, Rolston K,Mycobacterium avium-complex (MAC) infection incancer patients. In: Proceedings of the 35th AnnualMeeting of the Infectious Diseases Society of America,San Francisco, CA, September 13–16, 1997: Abst 708.

50. Dunne M, Fessel J, Kumar P et al, A randomized,double-blind trial comparing azithromycin andclarithromycin in the treatment of disseminatedMycobacterium avium infection in patients withhuman immunodeficiency virus. Clin Infect Dis2000; 31: 1245–52.

51. Simon MS, Cody RL, Cellulitis after axillarylymph node dissection for carcinoma of thebreast. Am J Med 1992; 93: 543–8.

52. Rescigno J, McCormick B, Brown AE,Myskowski PL, Breast cellulitis after conservat-ive surgery and radiotherapy. Int J Radiat OncolBiol Phys 1994; 29: 163–8.

53. Dixon JM, Breast infection. BMJ 1994; 309: 946–9.54. Brewer VH, Hahn KA, Rohrbach BW et al, Risk

factor analysis for breast cellulitis complicatingbreast conservation therapy. Clin Infect Dis 2000;31: 654–9.

55. Brooker DC, Savage JE, Twiggs LB et al,Infectious morbidity in gynecologic cancer. Am JObstet Gynecol 1987; 156: 515–20.

56. Barton DPJ, Fiorica JV, Hoffman MS et al,Cervical cancer and tubo-ovarian abscesses: areport of three cases. J Reprod Med 1993; 38:561–4.

57. Imachi M, Tanaka S, Ishikawa S, Matsuo K,Spontaneous perforation of pyometra presentingas generalized peritonitis in a patient with cervi-cal cancer. Gynecol Oncol 1993; 50: 384–8.

58. Rolston KVI, Rodriguez S, Dholakia N et al,Pulmonary infections mimicking cancer: a retro-spective, three year review. Support Care Cancer1997; 5: 90–3.

59. Chandrasoma PT, Intramedullary cord tubercu-losis resembling glioma. Neurol India 1976; 24:164–6.

60. Cleaton-Jones P, Oral tuberculosis – its similarityto oral carcinoma. J Can Dent Assoc 1971; 37:388–9.

61. Kumar S, Chandrasekar D, Rao MS, BanerjeeCK, Solitary paravesical tuberculosis mas-querading as bladder carcinoma. Tubercle 1981;62: 143–4.

62. Matthews JI, Matarese SL, Carpenter JL,Endobronchial tuberculosis simulating lungcancer. Chest 1984; 86: 642–4.

63. Shindy SR, Chandawarkar RY, Deshmukh SP,Tuberculosis of the breast masquerading as car-cinoma: a study of 100 patients. World J Surg1995; 19: 379–81.

64. LaRosa AM, Weigand M, Rolston KVI,Actinomycosis: an unusual infection presentingas or complicating malignancy. Infect Dis ClinPract 2000; 9: 331–3.


6Infections in patients with hematologicmalignanciesBen E De Pauw

INTRODUCTION

An intact defense system offers protectionagainst infections through a complex interrela-tionship of protecting surfaces, cells, and solu-ble factors. A good general condition, optimalnutritional status, and normal organ function,together with all components of the cellular andhumoral immune system, provide adequateprotection against pathogenic microorganisms.There are fundamental differences betweenhematologic malignancies and solid tumorswhich affect the incidence as well as the sever-ity of infectious complications concerned.Leukemias and lymphomas reside by definitionwithin the immune system itself, exerting adual deleterious effect. The malignant popu-lation interferes with and supplants theimmunocompetent elements at their originallocation. Hemorrhages, inevitable during thecourse of acute leukemia, may impede organfunction and facilitate the growth of microor-ganisms that may be present. The effects of thevarious noxious events that occur while treat-ing a hematologic malignancy differ in severityand in primary targets. Moreover, such haz-ardous events exercise their impact dynami-cally as the degree of disturbance varies withtime during or after a course of treatment.There is, in fact, a reciprocity: better supportive

care allows more aggressive therapy to achievebetter cure rates at the price of peculiar, hith-erto rare, infectious complications. Therefore,the survival of patients with a hematologicmalignancy depends heavily on the quality ofsupportive care. Neutropenia is the mostimportant risk factor, there being an inversecorrelation between the number of circulatingneutrophils and the frequency of infection. Allpatients with a neutrophil count of less than100/µl for more than three weeks will developfever, whereas only one-fifth of patients whoare not neutropenic become febrile.1

The signs and symptoms of infection aremuted owing to the absence of neutrophils, anddiagnostic procedures may be very problematicin immunocompromised hosts.2 No microbio-logical explanation for the fever will be foundin about 60% of patients who become febrilewhile neutropenic, but the fact that more thanthree-quarters of them will improve clinicallyafter treatment with broad-spectrum antibacte-rials suggests an occult bacterial source as thecause of fever.3,4 A small inoculum, enough tocause symptoms of infection in a patient with adefective defense mechanism, might stay belowthe detection limit of standard blood culturetechniques, particularly if marginal samples forcultures are taken. Ideally, close cooperationought to be established between all disciplines

involved in the care of the patient: nurses,hematologists or oncologists, microbiologists,radiologists, pulmonologists, pathologists, andspecialists in infectious diseases.

A vast majority of the organisms responsiblefor infection during chemotherapy-inducedneutropenia arise from the patient’s endoge-nous microbial flora, particularly from the gas-trointestinal tract and cutaneous surfaces.3

Interestingly, potentially pathogenic organismsthat belong to the patient’s flora on admissionare unlikely to cause serious infections.Conversely, Fainstein et al5 showed that theoropharyngeal and fecal flora are altered dur-ing hospital visits, and this change in coloniza-tion has clinical consequences. These acquiredorganisms show a greater propensity to invadethe body, and are frequently responsible forlife-threatening, disseminated infections. Butworldwide, there is presently no predominantcausative pathogen. Prior to 1960, Staphylo-coccus aureus was most commonly involved infatal infections, whereas during the 1960s and1970s, Gram-negative rods prevailed.6 Thespectrum of organisms responsible for infec-tions in the neutropenic patient is constantlychanging in conjunction with alterations in themanagement of the underlying disease.6–8

Prophylactic use of antibacterials, mainly aimedat Gram-negative organisms, has certainlyreduced the number of culture-proven infec-tions by aerobic Gram-negative rods.9 In addi-tion, the hematopoietic colony-stimulatingfactors reduce the length of neutropenia, whichallows the use of higher doses of cytotoxicdrugs, and this promotes other complicationssuch as severe mucositis. Extensive mucosaldamage is often accompanied by impaired pro-duction of saliva, and mucin, if produced at all,may be extremely viscous and difficult to eitherswallow or cough up. Under these circum-stances, viridans streptococci have become thepredominant pathogens in patients who aretreated for a hematologic malignancy.10,11

Intravenous catheters are often essential for thesuccessful management of immunocompro-mised patients. Subcutaneously implanted

venous access devices are seldom used, becausesurgery is a dangerous procedure in patientswith a bleeding tendency. Hence, a transcuta-neously inserted catheter with or without a sub-cutaneous tunnel tract is the standard. Suchcatheters provide the single most effectivemeans of breaching the skin barrier and creat-ing ready access for microorganisms such asstaphylococci, particularly Staphylococcus epider-midis, and, to a lesser extent, Candida andStenotrophomonas spp.12–14 In patients whoundergo modern, very intensive treatment,even sepsis with Clostridium perfringens and C.septicum has been described. Under these cir-cumstances, the ‘non-pathogens’ Staphylococcusepidermidis, JK bacteria, and Corynebacteriumparvum or ‘diphtheroids’ cannot be dismissedas a possible cause of septicemia or organ infec-tion.15 Anaerobic organisms rarely feature assingle pathogens; they play a role in mixedpolymicrobial infections and represent 6–13%of all bacteremic episodes. Such multiple-organism infections often reflect an unremittingunderlying disease, and carry a bad prognosisin spite of adequate antibiotic therapy.

Since the mid 1970s, antimicrobial agentshave been given to patients in an attempt toreduce infectious complications arising duringneutropenia.9 Decontamination of the digestivetract was employed in an attempt to eliminatepotentially pathogenic organisms from the ali-mentary tract, the major reservoir for Gram-negative bacilli, but later observationssuggested that the systemic action ofabsorbable antibiotics may be more importantthan any local effect on the gut flora. Althoughtotal decontamination in combination with ster-ile food has generally been abandoned as a bur-den on patients’ quality of life withoutimprovement of survival rate, selective decont-amination of the digestive tract still has sup-porters, mainly in Europe.9 Considering onlythose prophylactic studies that are placebo-controlled, it is by no means clear what benefitprophylactically administered antibiotics offer.Bacteremia is reduced, but the rate of fever andthe need to employ empiric therapy is not influ-


enced. It is therefore unfortunate that an ade-quate placebo-controlled study with sufficientpower is lacking. Furthermore, probablybecause of the practice of giving promptempiric therapy, the mortality ultimately attrib-utable to Gram-negative rods, the principal tar-get of prophylaxis, is identical to that inpatients not receiving antibiotics prophylacti-cally. Nevertheless, it is tempting to pursue thisapproach in neutropenic patients who areclearly colonized with virulent organisms suchas Enterobacter cloacae, Pseudomonas aeruginosa,or yeasts. Regular monitoring of importantbody sites as well as the patient’s environmentwould be mandatory to support this strategy.

It must be underscored that inadequatehygiene by visitors, nurses, doctors, and otherpersonnel is invariably a prominent source ofinfection. Organisms such as Legionella pneu-mophila, Ps. aeruginosa and Aeromonas hydrophilacan reach potentially dangerous concentrationsin watery environments such as air-conditioning systems, sinks, bathrooms, andwater for flowers and plants. Hence, many cen-ters do not allow flowers or home-cooked foodin patients’ rooms. Whilst it is mandatory toeliminate notorious sources of infection withinthe hospital and at the patient’s home, masks,gowns, gloves, and isolation of patients in asterile environment are not required unlessthere is a specific indication, such as the car-riage of virulent organisms. Laminar airflow orHEPA-filtered rooms can be recommended asan adjunct to care, if high concentrations of fun-gal spores in the air of a given ward dictate this.Besides, it is irrational that prescription of H2-receptor antagonists and other antacids hasgained such popularity, considering the pos-sible colonization of the gut by Gram-negativebacilli and organisms such as Listeria monocyto-genes.

EMPIRIC THERAPY: WHY, WHEN, WHAT?

As there is no reliable way to distinguish afever of infectious origin from fever due to non-infectious causes, the possible presence of a life-threatening infection must be presumedwhenever fever occurs in a neutropenicpatient.2 Of course, the possibility that a febrileepisode is associated with drugs, such as allop-urinol, antibiotics, bleomycin, or cytarabine, orwith the underlying disease should always becontemplated, but such a connection is usuallyquite apparent. If left untreated, 40% of patientswho are neutropenic and bacteremic will diewithin the first 48 hours after the onset offever.16 Hence, antimicrobial therapy should becommenced within an hour of the first signs orsymptoms of infection. This strategy of promptintravenous administration of broad-spectrumantibacterials in maximal therapeutic dosageshas reduced the mortality associated with bac-teremia caused by Gram-negative rods toapproximately 10%, and has become a widelyaccepted principle of infection managementduring neutropenia for almost 30 years.16,17 It isvirtually impossible to cover for every conceiv-able pathogen, but, in view of their virulence, S.aureus, Pseudomonas spp., and other Gram-negative rods are among the primary targets ofempiric antimicrobial therapy. However, partlybecause of the changing pattern of infection andthe continuous marketing of new anti-infectiveagents, the question of what constitutes theoptimal regimen for the febrile neutropenicpatient has never been answered.18,19 The dis-cussion focuses on the number and kind ofantibiotics to use for empiric purposes.20,21 Itsimply illustrates that no uniformly superiorcombination has been found – and nor will onebe in the future, because differences will alwaysexist between individual patients, centers, andclinical circumstances.3,20,22 It is essential toselect a regimen on the basis of the most com-monly encountered infectious pathogens, aswell as the resistance pattern of causativemicroorganisms, because empiric use of a �-lac-tam antibiotic, either alone or in combination,

INFECTIONS IN PATIENTS WITH HEMATOLOGIC MALIGNANCIES 113

to which a significant rate of resistance has beenfound must be avoided.7,17

Basically, three different strategies have beenuniversally accepted for empiric antibiotic ther-apy: (a) traditional combinations of either a�-lactam with an aminoglycoside or two �-lac-tams; (b) monotherapy with an extended-spectrum �-lactam; (c) both of these optionssupplemented by a glycopeptide from the onsetof fever (see Table 6.1).22–29 Depending on thecriteria applied, the response rates vary from30% to 70% and the overall survival rate ismore than 90%. An aminoglycoside plus eithera broad-spectrum penicillin or cephalosporinoffers possible synergism and, theoretically,minimal emergence of resistant strains;disadvantages are limited activity against someGram-positive bacteria and the risk of nephro-toxicity, hypokalemia, and ototoxicity, espe-cially when other drugs with a similar toxicity


Table 6.1 Established empiric antibioticregimens

Combination therapy�-lactam (ceftazidime, cefepime, cefpirome, orpiperacillin) with an aminoglycoside (amikacinor tobramycin) or a second �-lactam of adifferent class

Single-agent therapyCeftazidime, cefepime, cefpirome,imipenem–cilastatin, meropenem, and(probably) piperacillin–tazobactam

Addition of a glycopeptide to the combinationor single agentVancomycin or (outside North America)teicoplanin

profile are used concurrently. Such combina-tions have been tested at length in many trialsemploying numerous different antibacterials,and many physicians intuitively considered thisthe most suitable regimen for patients at highrisk for Gram-negative infections.19,22,23 Double�-lactam regimens appear to be an acceptablealternative if nephrotoxicity has to be avoided;however, because the targets are similar, it isconceivable that resistance may develop,although broad-spectrum penicillins have beencombined with clavulanate or tazobactam toextend their spectrum to include the �-lacta-mase-producing organisms.24–26 There are someindications that double �-lactam combinationscan prolong the duration of neutropenia,whereas their high sodium content may be aburden for elderly patients. With regard to theindividual drugs to be used in combination reg-imens, there is a choice of drugs rather thandrugs of choice. The presently most frequentlyused aminoglycosides are amikacin andtobramycin, whereas from the broad-spectrumpenicillins piperacillin–tazobactam and fromthe cephalosporins ceftazidime and cefepimeare favored, although these compounds can bereplaced by other antibiotics from the respec-tive classes on the basis of local patterns ofresistance. The availability of new broad-spectrum antibiotics has encouraged severalinvestigators to assess the feasibility of singleagents for empiric purposes. Another reason forchanging the rationale in antibiotic manage-ment is the significant improvement in antineo-plastic response rate, making the occurrence ofantibiotic-related toxicity even less tolerable. Itis also a fact that some of the traditional regi-mens involved up to 13 drug administrationsper day, compounding costs and the potentialfor medication errors and sometimes causingdelays in the administration of other parenteralmedication. Initially, single agents were onlyacknowledged to constitute adequate empirictherapy for unexplained fevers and not for doc-umented infections or episodes with prolongedneutropenia. However, randomized controlledtrials of empiric monotherapy, employing the

third-generation cephalosporin ceftazidime,have shown no difference in efficacy in compar-ison with traditional combinations, not even incases of Gram-negative bacteremia.28 It is note-worthy that the addition of an aminoglycosideor vancomycin to ceftazidime was necessary inless than 15% of the episodes of fever and neu-tropenia.28 Trials assessing the value ofcefepime and cefpirome showed successes simi-lar to those obtained with ceftazidime or theclassic combinations. Imipenem–cilastatin andmeropenem, with their extended spectrum,appear to be the most suitable candidates giventhe increasing challenge posed by Gram-positive infections.7,18 These compounds, with-out being definitely superior, fulfilled mostexpectations as far as efficacy was con-cerned.26,27,30,31 Stenotrophomonas maltophilia andother non-aeruginosa pseudomonads wereresponsible for treatment failures or break-through bacteremia, and therefore it seems pru-dent to avoid administration of these antibioticsin patients who are colonized with these organ-isms or in centers where these are prevalentpathogens. The occasionally occurring seizuresand nausea associated with the maximum doseof imipenem–cilastatin are a cause of concern,particularly when the patient has a brain lesionor if drugs such as cyclosporin A and cisplatinare used concomitantly. In such cases,meropenem may serve as a safe alternative.30,31

As empiric regimens for febrile neutropenicpatients must always contain reliable anti-pseudomonal activity, other currently availablebroad-spectrum antibiotics should not be used,although, on theoretical grounds, piperacillin–tazobactam would be considered by someinvestigators. Finally, it should be emphasizedthat extensive use of single-drug therapyrequires vigilance, since success depends upon continued susceptibility to the drug inquestion.

While the choice of a basic empiric regimenis relevant, the complex issue of whether andwhen to add a glycopeptide stays controversial.The glycopeptides seem to be the drugs ofchoice for these pathogens, but two opposing

opinions prevail as to their inclusion in the ini-tial regimen. One contends that drugs such asvancomycin can be added later when Gram-positive bacteria have been isolated or if noresponse is seen, with the powerful argumentthat glycopeptides do not contribute to the ulti-mate chances of cure in the vast majority ofpatients, because early mortality due to infec-tions with most Gram-positive organisms isvery low. The second opinion claims that addi-tion from the start will provide earlier effectivetreatment and thus reduce overall morbidity, inspite of the fact that by following this approach,as many as seven out of ten cases will be over-treated. The results of several prospective stud-ies do suggest that there is rarely need to utilizevancomycin as part of the front-line therapeuticregimen.32–34 If, however, there is reason to sus-pect the possible presence of a methicillin-resistant S. aureus (MRSA) on the basis of localpatterns of resistance, a glycopeptide must beincluded in the initial regimen to avoid unnec-essary mortality. The choice between the avail-able glycopeptides is trivial; teicoplanin, whichis not available in North America, is easy toadminister and has very few side-effects,whereas vancomycin has more reliable activityagainst certain subtypes of the coagulase-negative staphylococci.

Some of the problems associated with thetraditional combinations may be circumventedby substitution of a �-lactam by a quinolone.The fluorinated quinolones show no cross-resistance with �-lactams, and they are highlyactive against the rapidly fatal Entero-bacteriaceae. Their potential for use in an empiricsetting is constrained by their commissioningfor prophylactic purposes and their suboptimalactivity against Gram-positive pathogens,notably viridans streptococci. In combinationwith a specific anti-Gram-positive agent, theefficacy of the quinolones appears comparableto that achieved with established regimens,and, after initial clinical improvement, a switchto an oral formulation seems feasible.35

Experience with the monobactam aztreonam islimited, and the results achieved are too


conflicting to recommend employment of thisantibiotic for first-line therapy, but it can play arole in the treatment of febrile neutropenicpatients who are allergic or resistant to �-lac-tam antibiotics, provided that adequate Gram-positive cover is added.36

INDIVIDUAL ADAPTATIONS

Although the empirical regimen in a given hos-pital is usually identical for all febrile patients,it is evident that there is no regimen that will beappropriate for all febrile neutropenic patients.Approximately 20–30% of the febrile episodesin neutropenic patients are due to bacteremias,20% to clinically documented infections, and20% to non-bacteremic microbiologically docu-mented infections, while the remaining 30–40%are possible or doubtful infections.3 Amongstthe clinically documented infections, the lowerrespiratory tract is the site of infection in about55%, the upper respiratory tract and skin andsoft tissue contribute approximately 20% each,whereas the other sites are restricted to 5%.3,4

Patients with an obvious focus of infectionclearly represent a population that is more diffi-cult to treat than do those without any focus atall. Infectious death occurs in one-fifth ofepisodes with a focus of infection, in compari-son with less than 5% for episodes withoutone.22,27,37 Hence, it is reasonable to assume thata substantial number of febrile neutropenicpatients might benefit from an individually tai-lored empiric approach.20,22 Clinical and labora-tory investigations can help to select theoptimal empiric coverage (see Table 6.2).Studying the case notes can help to identify thecause of some fevers as being tissue damage bycytotoxic agents, the use of pyrogenic drugs,administration of blood products, or graft-versus-host disease. The history of a patientwith regard to previous infections might revealimportant information on possible drug allergyand on the actual infectious complication, sincefever may represent a recrudescence of aninfection acquired during a previous aplastic


Table 6.2 Important actions when feveroccurs in a neutropenic patient

• History, including details of infections duringprevious neutropenic episodes and ofconcomitant drugs

• Physical examination, with special attentionto catheter tunnel tract, lungs, perianalregion, skin, and mouth

• Assess the state of the underlying disease• Perform cultures from blood and clinically

suspicious lesions (in the case of a centralvenous line, blood from the line – all lumens– as well as peripheral blood)

• Check the results of possible surveillancecultures

• Consider concurrent infections in otherpatients in the same ward

• Chest X-ray (and, in case of any suspicion ofan abnormality, computed tomography (CT)scan)

• Assess kidney function, liver function,plasma minerals

episode. For bacterial infections, this informa-tion has, unfortunately, only very limited pre-dictive value. On the other hand, despite thecommon lack of physical signs and symptoms,a careful physical examination should be per-formed, paying special attention to vital signssuch as blood pressure, pulse, and respiratoryrate, and to oropharynx, lungs, venous accessdevices, perianal areas, and the course of thetemperature during the preceding days.Imaging techniques should be used whenappropriate. These diagnostic procedures areimportant, because different types of infectionare preferentially associated with distinctcausative organisms, and the results may helpto consider an individual adaptation of thestandard empiric scheme.22,38

Skin, soft tissue, abdominal, and catheter-related infections

The clinical spectrum of catheter-related infec-tions ranges from asymptomatic bacteremia asa manifestation of intraluminal colonization ora process confined to the site of insertion, tomarked inflammation of the tunnel tract andsepticemia with metastatic emboli in the skinand other organs.12,13 Malfunction of thecatheter, as revealed by the impossibility ofdrawing blood from the line, is often the firstsign of an infectious problem. Ecthyma gan-grenosum with extensive necrosis represents acharacteristic entity in patients with Ps. aerugi-nosa sepsis, but these cutaneous manifestationsare reported in only 2% of cases.6 Sincepathogens such as Candida spp. and Mucoralesorder can cause similar lesions, a needle aspira-tion or biopsy should be performed to ascertaina definite and accurate diagnosis as early aspossible in the course of the disease. Mucositis,gingivitis, and dental-related problems mayoccur in up to 85% of patients. Micro-biologically documented infections are fre-quent, featuring Candida albicans, viridansstreptococci, enterococci, or anaerobes. Herpessimplex virus (HSV) may also play a role.Mixed and polymicrobial infections are more orless the rule. Given the prevalence of Gram-positive organisms in these patients, selectionof a regimen with improved anti-streptococcalactivity or early addition of a glycopeptideappears legitimate. Life-threatening varicellazoster virus (VZV) infection or chickenpox,sometimes with visceral dissemination, is afeared entity among leukemic children.

Dysphagia or odynophagia in hematologicpatients may be due to chemotherapy or gastricreflux, but esophagitis is of infectious origin inthe majority of cases, HSV, either alone ortogether with Candida spp., being the mostlikely causative organism. Colitis or typhlitis inpatients with acute leukemia is accompanied bya combination of profuse diarrhea and severeabdominal pain with virtually no bowel move-ments. It may create a very alarming situation,

and since unnecessary surgical interventionsmay be detrimental, it is essential for physiciansto be aware of the existence and symptomatol-ogy of this entity. This syndrome is typicallychemotherapy-induced, but can be the result ofother different etiologic factors.39 Pseudo-membranous colitis from Clostridium difficile canbe severe and even fatal. Stools should be cul-tured and tested immediately for the presenceof this microorganism and its toxin if the diag-nosis is suspected. Relapses are frequent, andmay follow cancer chemotherapy or courseswith antibiotics such as clindamycin.Disproportional bacterial overgrowth in thegastrointestinal tract of patients with a dam-aged mucosa can serve as a source of bac-teremia by normally exclusively entericpathogens such as C. septicum. Considering theprobable involvement of anaerobes, a car-bapenem and the addition of metronidazole orvancomycin to a standard empiric regimen areobvious options when fever is accompanied byabdominal symptoms. Diagnostic problems areheld accountable for underrating enteric virusesas causative agents in gastrointestinal infec-tions. Perirectal cellulitis with painful lesionswithout abscess formation caused by Gram-negative rods, particularly Escherichia coli, withor without anaerobes and HSV, has become lesscommon.

Pulmonary infections

Management of pulmonary infiltrates is com-plex, given that as many as 40% may be ofnon-infectious origin.38,40,41 Bacterial infectionsaccount for most of the pulmonary infiltratesthat appear as segmental shadows respectingthe normal anatomic borders of the lung tis-sue.38 Surveys have shown that response ratesin pneumonia due to Gram-negative bacilli orS. aureus do not exceed 45%. An ominous find-ing in a patient with pneumonia is the concomi-tant presence of polymicrobial bacteremia. Themajority of focal infiltrates are caused by fungi,in contrast to diffuse abnormalities, which are


usually not of direct bacterial or fungal origin.When a new infiltrate appears and progressesin patients who remain neutropenic, particu-larly in conjunction with fever and chest pain,Aspergillus fumigatus should be the leadingdiagnostic consideration, with the therapeuticconsequence of early systemic antifungal ther-apy. Next to the adverse effects of cytotoxictherapy or irradiation, a number of causativemicroorganisms as well as pulmonary hemor-rhage must be considered in the case of a dif-fuse infiltrate. Pneumocystis carinii used to bethe leading pathogen, but now pneumonitisfollowing bacteremia with viridans streptococciconstitutes a more prominent problem.11,42

Partly as a result of diagnostic limitations,infections with viruses such as respiratory syn-cytial virus (RSV), influenza, and adenovirusesseem to be rare. They are often complicated byeither viral pneumonitis or secondary bacterialpneumonia. RSV presents with rhinorrhea,nasal congestion, sore throat, and cough. Inmost leukemic patients, a clinical course of viralinfections comparable to that in immunocom-petent adults is seen and this, in combinationwith the apparently low incidence, does notwarrant a frontline place in the therapeutic con-siderations unless the actual symptoms areparticularly prototypical.

Miscellaneous items

Urinary tract infections are rare in the absenceof urinary catheters, and they are virtually allcaused by Gram-negative rods.

Patients who receive potentially nephrotoxicdrugs are the most obvious candidates forempiric therapy with a single agent.Conversely, many specialists recommendcoverage with two appropriate antibacterialagents that do not exhibit cross-resistance forpatients who are known to be colonized byresistant Gram-negative bacilli. Occurrence of ashock syndrome often reflects the presence ofGram-negative rods, streptococci or S. aureus inthe bloodstream; under these circumstances,

aminoglycosides are perceived to be necessaryby some experts, in spite of the increased risk ofnephrotoxicity.

Insidious onset of fever, accompanied byheadache and confusion, is indicative of menin-gitis in patients with perturbed cellular immu-nity. The predominant pathogens include L.monocytogenes, Cryptococcus neoformans andToxoplasma gondii. If a patient complains ofseizures and headache, localization of leukemiaor lymphoma, intracerebral cryptococcoma orcerebral abscesses caused by organisms such asstaphylococci, T. gondii, Nocardia, or mycobacte-ria have to be considered.

There is increasing evidence showing thatlow-risk patients – for example those who arenot very ill, and have unexplained fever and anincreasing granulocyte count – can be treated as outpatients with home-administered intra-venous antimicrobial therapy. Even an oral regimen based on a fluoroquinolone or co-trimoxazole (trimethoprim–sulfamethoxazole)can prove feasible, provided that there are noother adverse prognostic factors and that com-munication between attending physician andpatient is optimal, with easy access to the hos-pital in case of emergency.43–45 This may alsoapply to these patients who have receivedchemotherapy for acute leukemia with signs ofincipient bone marrow recovery.

With such variation in prevalent organismsdepending on the clinical presentation, and theavailability of wide scale of broad-spectrumantibiotics with disparate properties, it is justi-fied to consider a more individually tailoredempiric approach (see Table 6.3). It is clear thatan individual strategy related to clinical symp-toms requires the clinician, who is daily attend-ing the patient, to play a pivotal role.

ANTIBIOTIC MODIFICATIONS EARLY AFTERTHE EMPIRIC PHASE

The basic aim with the use of empiric anti-biotics in febrile neutropenic patients is to pre-vent mortality from septicemia due to


Gram-negative rods and S. aureus during thefirst 2–3 days after the onset of fever when theresults of the microbiological investigations arenot yet available.16,17 By the end of the trulyempiric phase, the clinical condition will havedeteriorated in 10% of patients, improved in25%, and stabilized in 65%.46 An initial responserate of about 35% may be expected amongpatients with shock, compared with 70% amongpatients without shock. Since, irrespective of

the initial regimen, a substantial number ofpatients will not respond adequately, modifica-tions are inevitable. For this purpose, aplanned–progressive approach involving modi-fication of the antimicrobial regimen every 2–3days according to a predetermined scheduleuntil the patient becomes afebrile is ill advised,since it ignores the individual differencesbetween various febrile neutropenic patients.8 Itcan also instill a false sense of security precisely


Table 6.3 Individually tailored empiric antibiotic therapy

Clinical situation Empiric regimen

Concurrent diseases:Impaired kidney function Monotherapy with ceftazidime or fourth-generation cephalosporinAllergy to a given antibiotic Avoid this class of drugsHeart dysfunction Avoid drugs with high sodium content

Concomitant potentially nephro- Monotherapy with ceftazidime, fourth-generation cephalosporin, and ototoxic drugs or meropenem

Colonization by resistant organism Specifically targeted prophylaxis; selection on the basis ofsusceptibility

No focus of infection present Monotherapy with ceftazidime, fourth-generation cephalosporin, orcarbapenem

Shock present Consider addition of aminoglycosidePatient in remission, not ill Consider home antibiotic therapy

Focus of infection present:Upper respiratory tract Carbapenem/piperacillin–tazobactamLower respiratory tract Ceftazidime, fourth-generation cephalosporin or carbapenem;

consider aminoglycoside; early addition of an antifungal agentSkin and soft tissue Carbapenem/fourth-generation cephalosporin,(including central venous line) piperacillin–tazobactam, but consider

adding a glycopeptideUrinary tract Ceftazidime, fourth-generation cephalosporinAbdominal symptoms Carbapenem

because the regimens chosen offer an increasingspectrum of activity, encouraging the lamenta-ble belief that further attempts at diagnosis canbe omitted. In contrast, it is imperative to havea standardized approach to the microbiologicalevaluation of the neutropenic patient withfever. Blood, sputum, and urine for bacterialand fungal cultures should be collected at theonset of fever and at regular intervals in thepersistently febrile patient. If a central venousline is present, an extra blood specimen shouldbe taken from each lumen of the catheter. Whilelysis centrifugation may be too expensive foruse on all blood cultures, this technique pro-duces superior results in detecting fungi inpatients at risk. For patients with suspectedwound or soft tissue infections, it is alwayspreferable to obtain tissue samples. As this fre-quently proves unrealistic, swab samples ofaspirates may be collected and transportedimmediately to the laboratory. Technical per-sonnel should be instructed not to discard spu-tum samples on the basis of low numbers ofleukocytes. The value of serodiagnosis for viralinfections in the acutely ill patient is seriouslyrestricted by the lag time between the infectionand the immunologic response. Since informa-tion on this issue may become important laterduring the course of the disease, samplesshould be taken and stored.

A retrospective survey of 1951 cases showedthat in patients with a lower respiratory tractinfection, the modification rate was 69% whilstadjustments were deemed necessary in 51% ofcases with a skin and soft tissue infection, in44% of the febrile episodes accompanied byabdominal complaints, and in 37% of the upperrespiratory tract infections.47,48 The selection ofadditional antibiotics, if necessary, can beguided by clinical circumstances. Such anapproach is validated by the fact that variouscategories of infection, as mentioned previ-ously, are associated with different causativeorganisms.22,37,38 This strategy, which puts agreater emphasis on diagnosis than empiricinterventions, requires daily meticulous assess-ment of each case, but drugs that are potentially

toxic can be added with more confidence once apositive diagnosis has been made.

Pulmonary infections, either as the primaryfocus or as a complication of septicemia, pre-sent a dismal prospect, and have been heldresponsible for 70% of all fatal infections aftercytotoxic therapy.38,40,41 Typically, chest radio-graphs performed early in the evolution ofinfection fail to show infiltrates; it may takemore than 3 days for the infection to generateenough damage or for the few remaining gran-ulocytes to concentrate around the infectiousnidus to permit recognition on a radiograph.The critical decision faced by the clinician at thebedside of patients with pulmonary infiltratesis whether or not to perform invasive proce-dures such as bronchoscopy with or withoutbronchoalveolar lavage, transbronchial biopsy,transthoracic aspiration, or open lung biopsy.The value of these diagnostic approaches forthe optimal management of patients remainscontroversial, because the yield depends on thecollaboration and skill of various specialists.Besides, concurrent thrombocytopenia con-strains invasive diagnostic interventions inmost patients.

Coagulase-negative staphylococci and Coryne-bacterium jeikeium have to be isolated fromat least two sets of blood cultures to beconsidered clinically significant, but singleblood cultures that are positive for S. aureus,Streptococcus pneumoniae, or Enterococcus faecalisshould be regarded significant. Although viri-dans streptococci are common blood contami-nants in the general population, positive bloodcultures in patients with oromucositis shouldnot be disregarded, certainly not whenStreptococcus mitis is isolated.10,11 These strepto-cocci can cause life-threatening infections inabout 10% of cases, including septic shock andpneumonitis with an adult respiratory distresssyndrome, with a mortality of around 60%despite aggressive antibiotic therapy – particu-larly if chemotherapy involved the use of high-dose cytarabine. These so-called ‘alpha-strepsyndromes’ are almost certainly multifactorialin origin, and the streptococcal infection prob-


ably triggers off a sepsis syndrome when thereis pre-existing tissue damage and alteration inthe systemic or local immunity. Therefore, acombination of specific antibacterials with corti-costeroids, rather than merely additional antibi-otics, should be considered to manage patientsaffected by this complication.42 Most catheter-related uncomplicated Gram-positive bac-teremias can be easily eradicated by aglycopeptide-containing regimen but oneshould be prepared for relapses. In patientswith insertion-site infections, tunnel infections,and septic emboli, removal of the line isvirtually inevitable in the vast majority of cases.It is also advisable to remove the catheter inpatients with atypical mycobacterial infections,fungemias, and bacteremias due to pathogenscausing rapidly fatal infections. Infections withClostridium spp. certainly require the additionof drugs such as penicillin G and vancomycinto amplify the antibiotic cover. If double- ortriple-lumen catheters are being used, theantimicrobial therapy must be delivered inrotation to each of the lumen ports.

Persisting fever without any sign of clinicaldeterioration is a very questionable indicator ofinfection. It is generally contended that clini-cally or microbiologically defined infectionscannot be expected to respond to therapywithin 72 hours. Indeed, it has been demonstra-ted that more than half of the patients who arestill febrile after 3 days of antibiotic therapy willdefervesce without any alteration of the anti-biotic regimen.28,34 It is therefore remarkablethat, particularly in cases where cultures fail toyield a pathogen, the temptation to escalatetherapy by adding more drugs appears almostirresistible without there being any evidence ofclinical deterioration, persistence of a pathogen,or development of a new site of infection.47

Besides, if fever persists for 72 hours after ade-quate broad-spectrum antibacterial treatment inpatients without any clinical or laboratory evid-ence of bacterial infection, the increased tem-perature is unlikely to be of bacterial origin. Asfar as possible, the number of changes to ther-apy should be kept to a minimum, because,


Table 6.4 Reasons for modifying an empiricregimen

• Deterioration of vital signs, such as bloodpressure and ventilation

• Development of a new clinical focus withoutclinical improvement

• Progression of an existing clinical focusduring persisting neutropenia

• Persistence of a causative pathogen incultures taken during therapy

• In vitro resistant pathogen in the initialculture in the absence of clinicalimprovement

• Isolation of a new pathogen during therapy• Occurrence of a new fever spike• Unexplained fever for more than 5 days• Adverse event attributable to an antibiotic of

the empiric regimen• Typical symptoms in conjunction with a

known local epidemic with unusualmicroorganisms, such as Legionellapneumophilia

rather than improving outcome, the liberal useof antibiotics actually enhances the risk oforgan toxicity and the development of resis-tance, and generates excessive costs.47,49

Modifications ought to be based on firmgrounds, such as deterioration of vital signs,isolation of a relevant pathogen resistant to theantibiotics given, an antibiotic-related adverseevent, or the occurrence of a new focus ofinfection or progression of an existing focus(Table 6.4).

When results from blood cultures takenbefore initiation of empiric therapy becomeavailable, changes should be considered accord-ing to the susceptibility pattern of the offendingpathogen. Decisions should be guided by the

evolving clinical condition of the patient. It hasto be underscored that as long as the patientremains febrile and neutropenic, antimicrobialcover should never be restricted to antibioticsthat are active only against Gram-positivepathogens to avoid rapidly fatal breakthroughGram-negative bacteremia.

Second infections emerge proportionally tothe duration of granulocytopenia, and furtherfebrile episodes may occur in one-fifth ofpatients with neutropenia lasting more than 28days and in more than half of cases with neu-tropenia exceeding 4 weeks. Next to prosthetic-device-associated infections, these secondaryfebrile events mainly involve pulmonary infil-trates. Although bacteria account for more than90% of culture-documented infections, invasivefungi have become prominent pathogens,particularly in patients who have protractedperiods of neutropenia.50 Consequently, empiricantifungal therapy is considered a mandatorymodification whenever unexplained fever per-sists for more than 4 or 5 days or when a typicalpulmonary infiltrate occurs.

There are no objective arguments against thecurrent American and European guidelines forthe use of hematopoietic growth factors as anadjunct to antimicrobial therapy in febrile neu-tropenic.51,52 However, it is widely assumed thatstimulation of granulopoiesis is beneficial inconditions where a long delay in marrow recov-ery is potentially disastrous. This pertains topneumonias, severe cellulitis, and invasive fun-gal infections.

Discontinuation of antimicrobial therapy isrecommended if granulocyte recovery ensues.Alternatively, if the persistently neutropenicpatient has no complaints, and exhibits no clini-cal, radiological, or laboratory evidence of infec-tion, stopping antibiotic therapy or a change toorally administered antibacterials should beconsidered after four days without symptoms.Any new fever or episode of clinical deteriora-tion should prompt a recommencement ofantimicrobial therapy, since infection may haveonly been suppressed, not eradicated.

REFERENCES


2. Sickles AE, Greene WH, Wiernik PH, Clinicalpresentation of infection in granulocytopenicpatients. Arch Intern Med 1975; 135: 715–19.

3. Pizzo PA, Robichaud KJ, Wesley R, Commers JR,Fever in the pediatric and young adult patientwith cancer. A prospective study of 1001episodes. Medicine 1982; 61: 153–65.

4. Dompeling EC, Donnelly JP, Raemaekers JMMet al, Evolution of the clinical manifestations ofinfection during the course of febrile neutrope-nia in patients with malignancy. Infection 1998;26: 349–54.

5. Fainstein V, Rodriguez V, Turck M et al, Patternsof oropharyngeal and fecal flora in patients withacute leukemia. J Infect Dis 1981; 144: 82–6.

6. Bodey GP, The treatment of febrile neutropenia:from the dark ages to the present. Support CareCancer 1997; 5: 351–7.

7. Klastersky J, Zinner SH, Calandra T et al(European Organization for Research andTreatment of Cancer Antimicrobial TherapyCooperative Group), Empiric antimicrobial ther-apy for febrile granulocytopenic cancer patients:lessons from four EORTC trials. Eur J Cancer ClinOncol 1988; 24(Suppl 1): S35–45.

8. Klastersky J, Empirical antibiotic therapy in neu-tropenic cancer patients. Eur J Cancer 1993;29A(Suppl 1): S6–10.

9. Donnelly JP, Selective decontamination of thedigestive tract and its role in antimicrobial pro-phylaxis. J Antimicrob Chemother 1993; 31: 813–29.

10. Cohen J, Donnelly JP, Worsley AM et al,Septicaemia caused by viridans streptococci inneutropenic patients with leukaemia. Lancet1983; 321: 1452–4.

11. Kern W, Kurrle E, Vonek H, High risk of strepto-cocci septicemia after high dose of cytosine ara-binoside treatment for acute myelogenousleukemia. Klin Wochenschr 1987; 67: 773–80.

12. Press OW, Petersen SR, Larson EB et al,Hickman catheter infections in patients withmalignancies. Medicine 1984; 63: 189–200.

13. Raad II, Bodey GP, Infectious complications ofindwelling vascular catheters. Clin Infect Dis1992; 15: 197–210.


14. Elting LS, Bodey GP, Septicemia due toXanthomonas species and non-aeruginosaPseudomonas species: increasing incidence ofcatheter-related infections. Medicine 1990; 69:296–306.

15. Whimbey E, Kiehn TE, Brannon P et al,Bacteremia and fungemia in patients with neo-plastic disease. Am J Med 1987; 82: 723–30.

16. Schimpff SC, Satterlee W, Young VM, Serpick A,Empiric therapy with carbenicillin and gentam-icin for febrile patients with cancer and granulo-cytopenia. N Engl J Med 1971; 284: 1061–5.

17. Bodey GP, Evolution of antibiotic therapy forinfection in neutropenic patients: studies atMD Anderson Hospital. Rev Infect Dis 1989;11(Suppl 7): S1582–90.

18. The EORTC International Antimicrobial TherapyCooperative Group, Gram-positive bacteraemiain granulocytopenic cancer patients. Eur J Cancer1990; 26: 569–74.

19. Hughes WT, Armstrong D, Bodey GP et al, 1997Guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

20. De Pauw BE, Donnelly JP, Elves A et al, Towardsindividually tailored empiric antibiotic therapyin febrile granulocytopenic patients. Neth J Med1990; 37: 111–19.

21. Klastersky J, Treatment of neutropenic infection:trends towards monotherapy? Support CareCancer 1997; 5: 365–70.

22. Maschmeyer G, Hiddeman W, Link H et al,Management of infections during intensive treat-ment of hematologic malignancies. Ann Hematol1997; 75: 9–16.

23. The EORTC International Antimicrobial TherapyCooperative Group, Ceftazidime combined witha short or long course of amikacin for empiricaltherapy of Gram-negative bacteremia in cancerpatients with granulocytopenia. N Engl J Med1987; 317: 1692–8.

24. Anaissie EJ, Fainstein V, Bodey GP et al,Randomized trial of beta-lactam regimens infebrile neutropenic cancer patients. Am J Med1988; 84: 581–9.

25. Bodey GP, Fainstein V, Elting LS et al, �-lactamregimens for the febrile neutropenic patient.Cancer 1990; 65: 9–16.

26. Winston DJ, Ho WG, Bruckner DA, ChamplinRE, Beta-lactam antibiotic therapy in febrilegranulocytopenic patients – a randomized trial

comparing cefoperazone plus piperacillin, cef-tazidime plus piperacillin, and imipenem alone.Ann Intern Med 1991; 115: 849–59.

27. Rolston KVI, Berkey P, Bodey GP et al, A com-parison of imipenem to ceftazidime with andwithout amikacin as empiric therapy in febrileneutropenic patients. Arch Intern Med 1992; 152:283–91.

28. De Pauw BE, Deresinski SC, Feld R et al,Ceftazidime compared with piperacillin andtobramycin for the empiric treatment of fever inneutropenic patients with cancer – a multicenterrandomized trial. Ann Intern Med 1994; 120:834–44.

29. The EORTC International Antimicrobial TherapyCooperative Group and National CancerInstitute of Canada, Vancomycin added toempirical combination antibiotic therapy forfever in granulocytopenic cancer patients. J InfectDis 1991; 163: 951–8.

30. De Pauw BE, for the Meropenem Study Group ofLeuven–London–Nijmegen, Meropenem andceftazidime are equally effective as single agentsfor empirical therapy of the febrile neutropenicpatient. J Antimicrob Chemother 1995; 36: 185–200.

31. Cometta A, Calandra T, Gaya H et al,Monotherapy with meropenem versus combina-tion therapy with ceftazidime plus amikacin asempiric therapy for fever in granulocytopenicpatients with cancer. Antimicrob AgentsChemother 1996; 40: 1108–15.

32. The EORTC International Antimicrobial TherapyCooperative Group and National CancerInstitute of Canada, Vancomycin added toempirical combination antibiotic therapy forfever in granulocytopenic cancer patients. J InfectDis 1991; 163: 951–8.

33. Nováková IRO, Donnelly JP, De Pauw BE,Ceftazidime as monotherapy or combined withteicoplanin for initial empiric treatment of pre-sumed bacteremia in febrile granulocytopenicpatients. Antimicrob Agents Chemother 1991; 35:672–8.

34. Ramphal R, Bolger M, Oblon DJ et al,Vancomycin is not an essential component of theinitial empiric treatment regimen for febrile neu-tropenic patients receiving ceftazidime – a ran-domized prospective study. Antimicrob AgentsChemother 1992; 36: 1062–7.

35. Kelsey SM, Weinhardt B, Collins PW, NewlandAC, Teicoplanin plus ciprofloxacin versus


gentamicin plus piperacillin in the treatment offebrile neutropenic patients. Eur J Clin MicrobiolInfect Dis 1992; 11: 509–14.

36. Rolston KVI, Bodey GP, Elting L, Aztreonam inthe prevention and treatment of infection in neu-tropenic cancer patients. Am J Med 1990;88(Suppl 3C): S24–9.

37. Donnelly JP, Nováková IRO, Raemaekers JMM,De Pauw BE, Empiric treatment of localizedinfections in the febrile neutropenic patientswith monotherapy. Leuk Lymphoma 1993; 9:193–203.

38. Nováková IRO, Donnelly JP, De Pauw B,Potential sites of infection that develop in febrileneutropenic patients. Leuk Lymphoma 1993; 10:461–7.

39. Gomez L, Martino R, Rolston KV, Neutropenicenterocolitis: spectrum of the disease and com-parison of definite and possible cases. Clin InfectDis 1998; 27: 695–9.

40. Commers JR, Robichaud KJ, Pizzo PA, New pul-monary infiltrates in granulocytopenic patientsbeing treated with antibiotics. Pediatr Infect Dis1984; 3: 423–8.

41. Maschmeyer G, Link H, Hiddeman W et al,Pulmonary infiltrations in febrile patients withneutropenia. Cancer 1994; 73: 2296–304.

42. Dompeling EC, Donnelly JP, Raemaekers JMM,De Pauw BE, Pre-emptive administration of cor-ticosteroids prevents the development of ARDSassociated with Streptococcus mitis bacteremiafollowing chemotherapy with high-dose cytara-bine. Ann Hematol 1994; 69: 69–72.

43. Malik IA, Abbas Z, Karim M, Randomised com-parison of oral ofloxacin alone with combinationof parenteral antibiotics in neutropenic febrilepatients. Lancet 1992; 339: 1092–6.

44. Rubinstein EB, Rolston K, Benjamin RS et al,Outpatient treatment of febrile episodes in low-risk neutropenic patients with cancer. Cancer1993; 71: 3640–6.

45. Kern WV, Cometta A, de Bock R et al, Oral ver-sus intravenous empirical antimicrobial therapyfor fever in patients with granulocytopenia whoare receiving cancer chemotherapy. N Engl J Med1999; 341: 312–18.

46. Dompeling EC, Donnelly JP, Raemaekers JMMet al, Evolution of the clinical manifestations ofinfection during the course of febrile neutrope-nia in patients with malignancy. Infection 1998;26: 349–54.

47. De Pauw BE, Dompeling EC, Antibiotic strategyafter the empiric phase in patients treated for ahematological malignancy. Ann Hematol 1996; 72:273–9.

48. De Pauw BE, Raemaekers JMM, Schattenberg T,Donnelly JP, Empirical and subsequent use ofantibacterials agents in the febrile neutropenicpatient. J Intern Med 1997; 242: 69–77.

49. O’Hanley P, Easaw J, Rugo H, Easaw S,Infectious disease management of adultleukemic patients undergoing chemotherapy:1982 to 1986 experience at Stanford UniversityHospital. Am J Med 1989; 87: 605–13.

50. Bodey GP, Bueltman B, Duguid W et al, Fungalinfections in cancer patients: an internationalautopsy survey. Eur J Clin Microbiol Infect Dis1992; 11: 99–109.

51. Anaissie EJ, Vartivarian S, Bodey GP et al,Randomized comparison between antibioticsalone and antibiotics plus granulocyte–macro-phage colony-stimulating factor (Escherichia-coli-derived) in cancer patients with fever andneutropenia. Am J Med 1996; 100: 17–23.

52. Bennett CL, Weeks JA, Somerfield MR et al, forthe Health Services Research Committee of theAmerican Society for Clinical Oncology, Use ofhematopoietic colony-stimulating factors: com-parison of the 1994 and 1997 American Society ofClinical Oncology surveys regarding ASCO clin-ical practice guidelines. J Clin Oncol 1999; 17:3676–81.


7Evaluation and management of fever in theneutropenic hematopoietic stem celltransplant patientJames C Wade

INTRODUCTION

Current estimates of the numbers of stem celltransplants performed each year range from20 000 to 40 000, and continue to increase.Indications for transplantation now includemalignant, non-malignant, and congenital dis-eases. The development of non-myeloablativeapproaches for allogeneic transplantation haveexpanded the ability to test transplantation astherapy for common solid tumors (e.g. renal,melanoma, prostate, breast, cervical, etc.), andto offer transplantation to individuals whobecause of age or major organ dysfunction wereotherwise not candidates for a conventionaltransplant.1,2 The sources of stem cells for trans-plantation include bone marrow, mobilizedperipheral blood, cord blood, and fetal liver,which can be obtained from autologous orrelated and unrelated allogeneic sources.Marrow transplantation is now more appropri-ately referred to as hematopoietic stem celltransplantation (HSCT).

Increased susceptibility to infection remainsa major obstacle and cause of mortality forpatients undergoing HSCT.3,4 The past decadehas led to a better understanding of the patho-genesis of infection syndromes, the develop-ment of new diagnostic techniques, theintroduction of new, more effective antimicro-

bials, and the adoption of empiric or pre-emptive therapy strategies have enhancedpatient survival. However, the problem ofinfection for HSCT recipients remains adynamic one, with shifts in the patterns of iso-lated pathogens, changes in the antimicrobialsusceptibility of pathogens, introduction of newtransplant conditioning regimens, and theincreasing use of alternative donors. The inci-dence of infection-related death within the first30 days after transplant ranges from 5% to 10%,but the susceptibility to infection for somepatients may persist for months and years afterHSCT.4,5

There is a characteristic pattern of immuno-deficiency and immune reconstitution thataccompanies an HSCT.5 Four periods are nowrecognized:

1. Pretransplant, corresponding to the periodof infection risk just prior to transplanta-tion. This risk is secondary to the patient’sprevious therapy and underlying disease.

2. Pre-engraftment, corresponding to the weeksprior to marrow engraftment. This is theperiod of greatest risk for patients whoreceive an autologous graft, but is only thefirst risk period for allogeneic HSCT recipi-ents.5–7

3. Mid-recovery, corresponding to the period

from engraftment to approximately day 100post-engraftment.

4. Late recovery, corresponding to the timeinterval beyond 3 months after transplant.

Pretransplant

This is a critical period for all HSCT recipients.The increase in intensity of therapy given priorto transplantation makes the presence of sub-clinical vascular access device infection, occulthepatosplenic candidiasis, or subclinical inva-sive sinus or pulmonary fungal infections morecommon. Effective pretransplant treatment andpost-transplant infection suppression is depen-dent on the identification of such infectionsbefore the conditioning therapy is started. Thisis also the time in which the presence of impor-tant latent herpes group viruses, or activehepatitis virus should be detected through sero-logic evaluations.

Pre-engraftment

This risk period begins with the onset of theconditioning regimen, and continues untilapproximately day 30 after transplantation.Neutropenia is the primary predisposing factor,but the absence of neutrophils and other phago-cytes is usually accompanied by alimentarytract mucositis, the presence of central venouscatheters, and microbial flora shifts.6,7 HSCTrecipients are compromised by the ability oftheir endogenous flora to invade through dis-rupted mucosal and cutaneous barriers andcause systemic bacterial and fungal infections.8,9

These risks are further increased in patientswith delayed engraftment after transplantation.

Mid-recovery

This period begins with engraftment, and con-tinues until approximately day 100, when earlyB- and T-cell function begins to recover. This

period is characterized by the reappearance ofneutrophil function, and therefore infectionsassociated with neutropenia are uncommonunless the marrow function remains fragile,due to graft rejection, disease relapse, marrow-suppressing factors such as medications(e.g. ganciclovir), or viral infection (e.g.cytomegalovirus (CMV) human herpesvirus-6(HHV-6), or HHV-8).10–12 Cellular immune dys-function is the primary immune defect duringthis period. Late-onset aspergillosis occurs dur-ing this period, and affects 10–15% of allogeneictransplant recipients.13

Late-recovery

This post-transplant infection risk period beginswith the 4th month after transplant, and con-tinues until the patient has successfully beentapered off of all immunosuppression and isfree of chronic graft-versus-host disease(GVHD). Persistent cellular and humoralimmune dysfunction may lead to recurrentviral, bacterial, and fungal infections.14–17

The late-recovery period is also the time ofgreatest risk of relapse of the patient’s primarydisease. Relapse is often associated with therapid onset of marrow failure and resultantneutropenia. Management of these patients isparticularly complicated, because of the refrac-tory nature of their primary disease, and theaccompanying GVHD that may have been pre-sent prior to relapse or is induced as a mechan-ism to enhance primary disease control.18,19

EVALUATION, PREVENTION, ANDMANAGEMENT OF INFECTION

This chapter will focus on the infections inHSCT recipients that occur primarily duringperiods of neutropenia. Many of the evaluationand management principles for patients withfever and neutropenia are applicable to HSCTrecipients.6,7 Consequently, this chapter willreview evaluation and management approaches


in general, and discuss in greater depth thoseissues that are unique to HSCT recipients.

Neutropenia is primarily a consequence ofthe transplant-conditioning regimen or pre-transplant marrow failure state induced by thepatient’s primary disease. Fever and neutrope-nia are most common during the pretransplantand pre-engraftment period, although neutrope-nia may occur later as a consequence of infec-tion (e.g. CMV, HHV-6, HHV-8, or parvovirusB19), medication toxicity (e.g. ganciclovir), graftrejection, or post-transplant relapse of the pri-mary illness. The risk of neutropenia-associatedinfection for HSCT patients is enhanced by thepresence of mucosal or integumentary barrierdisruption induced by cytotoxic chemotherapy,irradiation, GVHD or the presence of long-termindwelling vascular access devices. Medication-induced central nervous system dysfunction,and the microbial floral shifts that accompanysevere illness or the administration of antibioticsalso enhance the infection risk.4,5,8,9

The primary sources of pathogens for neu-tropenic HSCT recipients are their endogenousbacterial and fungal flora, airborne molds andrespiratory viruses, microorganisms on thehands of healthcare providers, and latentviruses. Sites of infection for HSCT patients aresimilar to those for other neutropenic hosts.Infections originate primarily from the alimen-tary tract (i.e. mouth, pharynx, esophagus, largeand small bowel, and rectum), sinuses, lungs,and skin.6,7 Most febrile episodes that occurduring the pre-engraftment period are infec-tious in origin.6,20,21 Bacterial infections accountfor more than 90% of the first infection duringneutropenia. Herpes simplex virus (HSV) andthe respiratory viruses (e.g. influenza A and B,parainfluenza, and respiratory syncytial virus(RSV)) are also identified as first-infectionpathogens22 (Table 7.1). Antibiotic-resistant bac-teria, yeast, molds, and viruses are commoncauses of subsequent infections.23 The initialbacterial pathogens for HSCT recipients are

FEVER IN THE NEUTROPENIC HSCT PATIENT 127

Table 7.1 Infectious syndromes after HSCT pre-engraftment period

Syndrome Relative Relative life-frequencya threatening

potentiala

First fever:Staphylococci 3� 1�

Viridans streptococci 1� 2�

Gram-negative bacilli 1� 3�

Respiratory virus 1� 3�

Subsequent infection:Antibiotic-resistant bacteria 2�

Gram-positive cocci 1�

Gram-negative bacilli 2–3�

Fungi 2–3� 3–4�

Respiratory virus 1� 3�

a Frequency and life-threatening potential increase from 1 to 4�.

usually Gram-positive cocci.24 Of these, coagu-lase-negative staphylococci, streptococci, andenterococci predominate, but infections withStaphylococcus aureus and Corynebacteria jeikeiumremain important.25–29 Despite the perceiveddecreased virulence of many of these Gram-positive pathogens, there is attributable mortal-ity ascribed to such infections. Wenzel andco-workers30–32 reported that attributable mor-tality for coagulase-negative staphylococcalbloodstream infections was 13.6% (95% confi-dence interval (CI) 4.2–22.9%) and 37.1% (95%CI 10–64%) for bloodstream infections causedby vancomycin-resistant enterococci (VRE). Arecent report from the Mayo clinic noted thatcolonization with VRE increased the risk ofdeveloping a VRE bloodstream infection, andwas also associated with an increased risk ofdeath post-transplant for allogeneic transplantrecipients.33 Several large series have reportedrates of bloodstream infection with viridansstreptococci of 15–25% among HSCTrecipients.26–28 Approximately 10% of viridansstreptococcal infections are associated with a‘toxic-shock’-like syndrome that can be rapidlyfatal even with the institution of appropriateantibiotics. Thus, while empiric therapydirected at such pathogens may not always berequired, appropriate pathogen-directed ther-apy is mandatory when such organisms are iso-lated.

Gram-negative bacteria are the most virulentbacterial pathogens during the neutropenicperiod, and historically have been responsiblefor the highest rates of morbidity andmortality.6,7 The common Gram-negative bacilliremain Escherichia coli, Klebsiella spp. andPseudomonas aeruginosa. The increased use ofbroad-spectrum cephalosporins and the car-bapenem antibiotics has increased the fre-quency of isolation of �-lactam resistantEnterobacter spp. and Stenotrophomonas mal-tophilia.34–36

Increasing resistance to multiple antibioticsamong many of these pathogens poses asignificant problem in planning empiric treat-ment for HSCT patients. It is critical to be aware

of each institution’s specific antibiotic suscepti-bility patterns and to remember that pathogensand antibiotic susceptibility patterns are adynamic process and change over time.Antibiotic resistance remains most prominentamong the Gram-positive pathogens, specifi-cally the coagulase-negative staphylococci.Transplant recipients at the Fred HutchinsonCancer Research Center (FHCRC) in the calen-dar year 1999 experienced 374 bloodstreaminfections, for a rate of 0.718 bloodstream infec-tions/100 patient days (Table 7.2). Of theseinfections 359 were caused by a singlepathogen, 63% occurred during the pre-engraftment period and more than half werecaused by coagulase-negative staphylococci.The antibiotic resistance pattern for selectedcoagulase-negative staphylococci and viridansstreptococci show a high level of resistance tostandard �-lactam antibiotics. However, despitethe almost exclusive referral nature of theFHCRC patient population, the incidence ofbloodstream infections due to antibiotic-resistant S. aureus and VRE remains low.

The pathogens responsible for the subsequentor second infections include antibiotic-resistantbacteria, fungal pathogens, and viruses.Wingard et al23 have reported that Staphylococcusepidermidis is responsible for 50% of the secondor subsequent infections. Gram-negative bacilliare responsible for an additional 10%, with themajority of remaining pathogens being predom-inately fungal. These infections, with the excep-tion of coagulase-negative staphylococci, aredifficult to diagnose, more resistant to treat-ment, and associated with the highest rates ofmorbidity and mortality.13,37–39

EVALUATION AND MANAGEMENT OFNEUTROPENIA-ASSOCIATED INFECTIONS

There are several clinical guidelines for the pre-vention and empiric treatment of infection inthe neutropenic host: the National CancerComprehensive Network (NCCN) ClinicalGuidelines for Fever and Neutropenia; the


Infectious Diseases Society of America (IDSA);and www.cancernetwork.com; www.cdc.gov.20,21

There is also a diversity of infection manage-ment practices among the different transplantcenters. However, effective care for HSCTrecipients relies on a coordinated approach thatutilizes center-wide infection control, infectionprevention, and a combination of empiric andpre-emptive therapy. Transplant centers mustcustomize their practices based on their ownpatient population, cytotoxic regimens used,and local infection and susceptibility patterns.

Infection control

The single most powerful preventive measurefor the neutropenic HSCT patient is handwash-ing performed by the healthcare staff and otherindividuals who come into contact with thesepatients. Handwashing, while effective, con-tinues to be difficult to implement and a chal-lenge to maintain compliance. Hand soap that

contains chlorhexidine adds residual antimicro-bial effect to the mechanical cleansing thatoccurs with the physical washing. Gloves, ifused, should be put on only after entering apatient’s room and the handwashing has beencompleted. Gloves should be removed prior toleaving the patient’s room, and should never beutilized for more than one patient contact.Antimicrobial hand rubs can be used if soap,water, and sinks are not easily accessible. Thenew alcohol-containing waterless productsappear to be well tolerated and effectiveadjuncts to recurrent handwashing. Keepingstaff and patient visitors who have respiratorysymptoms (e.g. uncontrolled cough and respi-ratory secretions, conjunctivitis, or systemicsymptoms) from having contact with patientsmay decrease the risk of patients acquiring apotentially serious respiratory viral infection.Annual vaccination of healthcare workersagainst viral influenza is also important.

HSCT patients are no longer routinely caredfor during the pre-engraftment period in total


Table 7.2 Bacterial bloodstream isolates, FHCRC 1999

Single-organism bloodstream infectionsTotal 359

Coagulase-negative staphylococci 193Enterococci 41

E. faecalis 22E. faecium 11VREa 8

Streptococcus spp. 52Staphylococcus aureus 13

MRSAb 2Gram-negative bacilli 33Non-tuberculosis Mycobacteria 8Other 19

a Vancomycin-resistant enterococci.b Methicillin-resistant S. aureus.

sterile environments (laminar-airflow (LAF)units). This change in practice does not suggestthat measures to decrease the exposure ofpatients to potential pathogens are not benefi-cial, but rather that less intrusive infection con-trol measures can be implemented and stillaccomplish the outcome goals. HSCT patientsshould only be hospitalized in single-bedrooms, and care should be delivered by ahealthcare team that is familiar with transplantprocedures and understands the importance ofminimizing the nosocomial transmission ofpotential pathogens.40 Inpatient and preferablyoutpatient clinics or day-hospital facilitiesshould be ventilated with air that has beencleansed by high-efficiency air (HEPA) filters.Patient areas should also include a ventilationsystem that can deliver at least 15 air exchangesper hour. This level of clean-air ventilation isrecommended with the intent of decreasing theincidence of mold and other airborne infectionsfor HSCT patients. However, the protectionwith these measures is not complete. Increasedproportions of the pre-engraftment period mayoccur outside of these ‘clean’ environments,and, even when hospitalized, HSCT recipientsspend time away from the transplant unit whileundergoing imaging or endoscopic procedures.

Barrier isolation has been relegated to thosesituations where one is attempting to minimizethe spread by contact of potential pathogens.Isolation of patients known to be colonized orinfected with a pathogen that is multiplyantibiotic-resistant, or is known to have anincreased propensity for nosocomial transmis-sion (i.e. respiratory viruses), appears to beimportant.22,41

The presence of subclinical, or latent, infec-tion must be determined prior to the initiationof the transplant-conditioning regimen. Somepatients may be actively infected when they arereferred for HSCT, and these infections mayhave a direct effect on the patient’s outcome.The patient’s history of infections during priortreatment may reveal important informationregarding risks of infection that may occur dur-ing future periods of neutropenia. An infected

HSCT candidate should not be excluded fromtransplantation unless the infection cannot beadequately controlled pretransplant and mea-sures to maintain infection control during thepost-transplantation period are unavailable. Aprior history of either invasive mold (e.g. asper-gillosis) or disseminated candidiasis (e.g.hepatosplenic candidiasis) often raises concernsabout the patient’s suitability for transplanta-tion. A retrospective review conducted at theFHCRC reported 15 patients with documentedhepatosplenic candidiasis pretransplant.42 Allpatients had been treated with amphotericin Btherapy before transplantation, and all hadimprovement in their infection documented bycomputed tomography (CT) scans before thetransplant-conditioning therapy was initiated.These patients had no appreciable increasedrisk of recurrent yeast bloodstream infectionafter transplantation. Bjerke et al42 concludedthat a history of disseminated candidiasis wasnot an absolute contraindication to transplanta-tion.

Transplantation for patients with a previ-ously documented invasive mold infection hasbeen more difficult. Offner et al43 reported 48patients who underwent HSCT with a pretrans-plant history of documented or presumedaspergillosis. Of these 48 patients, 11 receivedan autologous HSCT. One-third of thesepatients experienced a documented recurrenceof aspergillosis post-transplant, with recurrentaspergillosis having an 88% mortality rate. Theauthors noted a trend toward a lower incidenceof recurrent infection and improved survivalfor patients with a longer interval between thedevelopment of invasive aspergillosis andtransplantation. Patients who received pre-emptive antifungal therapy or underwent anautologous transplant also had a lower fre-quency of recurrent infection. Surgical resectionof disease pretransplantation provided no sur-vival advantage for these patients.

The unpublished experience at the FHCRCfor patients with a prior history of aspergillosisis equally discouraging. We have transplanted34 such patients with conventional allogeneic


transplantation. All patients had been exten-sively treated with amphotericin B (>8 weeks),and prior to transplantation had shown clinicaland radiological improvement or resolution oftheir infection. All received amphotericin Bduring the period of maximum immunosup-pression. Of the 34 patients, 15 experienced arecurrent aspergillus infection post-transplant,and 12 others died from transplant-related mor-tality without a documented mold recurrence.Of the 6 patients who survived, 5 had an inter-val between diagnosis of aspergillosis andtransplantation of at least 9 months (range 9months–3.5 years). Recurrent aspergillosis wasuniformly fatal. While some patients with aprevious history of aspergillosis may be candi-dates for autologous transplantation, allogeneictransplantation is associated with a high inci-dence of recurrent infection and mortalitydespite pre-emptive antifungal therapy.

PREVENTION OF INFECTION

Bacterial

Antibacterial prophylaxis for patients with neu-tropenia remains controversial. Data support-ing the effectiveness of chemoprophylaxis withantibiotic(s) are balanced by a similar numberof reports that fail to show true efficacy. Ameta-analysis of 19 randomized trials of fluoro-quinolone prophylaxis in patients with neu-tropenia revealed a decrease in the number ofdocumented Gram-negative bacillary infec-tions, but no effect on the frequency of febrileepisodes, febrile morbidity, or infection-associated mortality.44 Many of these trials haveshown an increase in Gram-positive infectionsamong patients who received fluoroquinolonesor trimethoprim/sulfamethoxazole prophy-laxis. The use of prophylactic antibiotics hasalso been reported to increase the risk of subse-quent or secondary fungal infections.21 Few ofthese prophylactic trials have been performedexclusively in HSCT recipients, but this pro-phylactic approach has nonetheless been used

liberally for HSCT recipients. Patients trans-planted at the FHCRC are routinely given pro-phylactic, systemic antibacterial antibioticswhen their neutrophil count decreases below500/µl. This practice is based on a previouslypublished study that compared the pre-emptiveuse of broad-spectrum antibiotics during thepre-engraftment period versus LAF unit isola-tion.40 Infection morbidity and patient survivalwere similar for both groups. The pre-emptiveantibiotic regimen used at the FHCRC hasevolved over time, but presently consists of oralciprofloxacin or intravenous ceftazidime.Whether pre-emptive antibiotic therapy trulyincreases patient survival if compared withwithholding antibiotic treatment until the firstsign of infection (fever) has not been tested.

Fungal

The incidence of fungal infections has increasedamong HSCT recipients during the last decade.Centers now report an incidence that variesbetween 10% and 20%.4,5,13,37–39 Diagnosis andtreatment remain suboptimal, and preventingexposure or blocking colonization with thesepathogens is difficult. Candida colonization ispresent in as many as 80% of HSCT recipientspretransplantation, and persists throughout thefirst three months after transplantation unlesspatients receive azole suppression.38 When thefungal pathogen originates from the environ-ment, the protection provided by measuressuch as sterile environments and HEPA filtra-tion are limited to the time period whenpatients are being cared for in these clean facili-ties. The true incidence of colonization withAspergillus among HSCT recipients is unknown,but Wald et al13 in their study of 2496 consecu-tive HSCT patients reported that 2% of suchpatients became colonized at some time posttransplant. However, only 21% of patients whoultimately developed invasive aspergillosis haddocumented preinfection colonization. Yet,when Aspergillus colonization was detected, itwas associated with a 60% positive predictive


value for the development of invasiveaspergillosis. The positive predictive valueincreased to 94% if the HSCT patient was neu-tropenic.13

The use of masks by patients or healthcareworkers has not consistently affected the inci-dence of fungal infections. This lack of efficacymay be in part due to poor patient tolerance, orthe fact that the masks do not fit tightly enoughto be an effective barrier against fungal spores.The use of aerosolized or intravenous ampho-tericin B has provided inconsistent preventionresults.45–50 The study by Perfect et al,48 whichtested low-dose intravenous amphotericin Bprophylaxis, reported a similar incidence ofinvasive fungal infection for both the treatmentand placebo groups (8.8% versus 14.3%, respec-tively). The potential efficacy of prophylaxiswith lipid formulations of amphotericin B wasalso tested by Tollemar et al.49 They reportedthat this therapy decreased fungal colonizationbut provided no mortality benefit. The need forintravenous administration and the potentialtoxicity and cost of the lipid formulations ofamphotericin B makes them poorly suited forprophylaxis. Itraconazole is potentially usefulfor the prevention of a variety of fungal infec-tions, including aspergillosis, but its efficacy asprophylaxis in HSCT patients has not beenproven. Itraconazole has been difficult toadminister to HSCT recipients because oferratic absorption, numerous drug–drug inter-actions (e.g. with cyclosporine and tacrolimus),and medication-associated gastrointestinalintolerance. Concomitant administration ofcyclosporine and itraconazole causes alteredcyclosporine metabolism, and results in theneed to decrease the daily cyclosporine dosageby 20–50% (J Wingard, personal communica-tion, 2000). The availability of a new oralcyclodextran–itraconazole formulation, plus anintravenous itraconazole preparation, maymake this drug more efficacious.

Fluconazole has been shown to be highlyeffective for preventing Candida albicans infec-tion among allogeneic HSCT recipients.51,52 Adose of 400 mg daily, given from the time of

conditioning therapy to day 75 after transplant,significantly reduced superficial infection andinvasive disease and decreased the use ofamphotericin B. In the trial reported by Slavinet al,5 fluconazole not only decreased the inci-dence of C. albicans infections but alsoimproved patient survival. Long-term follow-up of this study cohort has recently beenreported by Marr et al.52 After 8 years of follow-up, survival remains significantly better for flu-conazole recipients (68/152 versus 41/148,respectively), and the incidence of invasive can-didiasis and death due to candidiasis remainlower for fluconazole recipients. Patientstreated with fluconazole also had a lower inci-dence of severe gut GVHD and death fromGHVD complications.

The benefit of fluconazole for autologousHSCT patients remains controversial.4 Thesepatients appear to have a risk of developingcandidiasis that is similar to that of patientswith acute leukemia who undergo inductionor reinduction therapy. Fluconazole has notbeen shown to provide a consistent benefit forsuch patients.53 In general, fluconazole prophy-laxis is probably not necessary for patientsreceiving an autologous transplant unless theconditioning regimen is expected to causesevere mucositis.

Viral

Viral infections that occur during the early post-transplant period include respiratory virus (e.g.RSV, parainfluenza, adenovirus, and influenza)and HSV. Early post-transplant CMV infectionand disease do occur, but infrequently.54

Infection control measures are critical for theprevention of respiratory virus infections. Thismay even require delaying patients’ transplan-tation, if their underlying disease is stable andthe incidence of such viral infections are epi-demic in the surrounding community or exces-sive among the transplant center’s healthcareteam. When HSCT patients, primary caregivers,or family members are exposed to influenza,


prompt prophylaxis with zanamivir aerosol,10 mg by inhalation daily, or oral oseltamivir,75 mg orally twice daily, should be considered.

The use of acyclovir (5 mg/kg intravenouslytwice daily or 400–800 mg orally twice daily),famciclovir 500 mg twice daily, or valacyclovir500 mg twice daily have all been shown to behighly effective in the prevention of HSV infec-tions.4,5,55 Suppression of HSV during the pre-engraftment period may decrease the mucosaldisruption cause by this viral recurrence, and,as reported by Baglin et al,56 the management ofHSV infections can minimize the duration andfrequency of febrile episodes.

ADJUNCTIVE MEASURES

Peripheral blood stem cells

The transition from bone marrow to growth-factor-mobilized peripheral blood stem cells(PBSC) as the hematopoietic stem cell producthas been an important improvement. The use ofPBSC has become standard practice for mostautologous transplants. A randomized trialconducted by Weaver et al57 reported that theinfusion of at least 5 � 106 CD34� stem cellsresulted in a median duration of post-transplant neutropenia (<500 polymorphonu-clear leukocytes (PMN)/µl) of 11 days. Thisshortened period of neutropenia may poten-tially decrease a patient’s risk of neutropenia-associated infection. However, the Seattlegroup has reported that when autologous PBSCproducts are preferentially selected for CD34�

cells, immune reconstitution is delayed.58,59 Thisdelay has resulted in an increased incidence ofCMV disease, varicella zoster virus (VZV) reac-tivation, and invasive bacterial, Candida, orsevere respiratory viral infections.

The observation that PBSC speeds engraft-ment among autologous HSCT recipients led tothe testing of such an approach for HSCT recip-ients who were to receive matched relateddonor transplants.60–62 Bensinger et al61 reportedthat the use of allogeneic PBSC decreases the

duration of neutropenia (<500 PMN/µl) from amedian of 21 days with bone marrow to 16days with PBSC. This decrease in the durationof neutropenia was not associated with a statis-tically significant decrease in the incidence offever or death from infection, but death fromthe idiopathic pneumonia syndrome wasdecreased. Most importantly, PBSC recipientshad improved disease-free and overall survival.These results are consistent with otherreports.60,62 The retrospective International BoneMarrow Transplant Registry (IBMTR) reviewsuggested that PBSC resulted in a decrease induration of neutropenia and shorter hospitalstays.60 The appropriate dose for an allogeneicPBSC is unclear, but engraftment is enhancedwith a dose of at least 5 � 106 CD34� cells/kg.The incidence of acute GHVD appears to besimilar for both bone marrow and PBSC, butthe incidence of chronic GHVD increases whenthe cell dose exceeds 8 � 106 CD34� cells/kg.61

Chronic GVHD occurring among PBSC recipi-ents may also be more severe and difficult totreat.63 The finding of an increase in treatment-refractory chronic GVHD does not have a directimpact on the pre-engraftment period, but clini-cians must be cautious when techniques such asPBSC minimize the immunosuppression (neu-tropenia) of one post-transplant period (pre-engraftment), but potentially increase theinfection risk for a later period (late recovery).

Growth factors

The American Society of Clinical Oncology(ASCO) has recently updated their guidelinesfor the use of growth factors.64 Granulocyteand granulocyte–macrophage colony-stimulatingfactors (G-CSF and GM-CSF) have consistentlybeen shown to hasten hematopoietic cell recov-ery, but this increase in neutrophil recovery hasnot translated into a consistent decrease inneutropenia-associated infections or improve-ment in survival.64,65 The ASCO guidelines nowrecommend limited ‘primary neutropenia pro-phylaxis’, and the standard practice at the


FHCRC is only to use them in cases of delayedengraftment (<500 PMN/µl) persisting beyondday 21 after transplantation.

Granulocyte transfusions

The primary immune defect during the pre-engraftment period is the absence of neu-trophils. Early trials showed benefit forGram-negative bacillary infections when donorgranulocytes were transfused.66,67 However, theroutine use of transfused granulocytes has beencompromised by the inability to collect ade-quate numbers, and these transfusions wereoften complicated by acute infusion reactions,an increased risk of alloimmunization, and thepotential transmission of CMV infection fromthe donor to the transplant recipient.68 Therehas recently been a resurgence of interest ingranulocyte transfusions with the observation69

that donor priming using growth factors andcorticosteroids can increase the granulocyteyield to (8–10) � 1010. This dose of infused gran-ulocytes can result in the transfusion recipienthaving a 1 hour post-transfusion white bloodcell count of (2.6 � 2.6) � 103 PMN/µl. Growth-factor-mobilized granulocyte transfusions todate have been tested primarily among HSCTpatients with severe antibiotic-resistant fungalor bacterial infections.70 While the efficacy ofthis approach has not yet been established, theSeattle group has shown that this type of trans-fusion can be effectively accomplished using arelated or an unrelated community donor.71

Adkins et al72 have published results from theonly prophylactic G-CSF-mobilized granulocytetrial in HSCT recipients. Patients received fourtransfusions on days 2, 4, 6 and 8 after autolo-gous transplantation. Leukocyte compatibilityof donor and recipient was determined byscreening for lymphocytoxicity against a panelof HLA-identified cells. All recipients receivedG-CSF-mobilized granulocytes from a singledonor. Leukocyte incompatibility adverselyaffected the recipients’ neutrophil incrementsafter transfusion, and resulted in a delay in

neutrophil engraftment, and increased thenumber of days of fever, platelet transfusions,and intravenous antibiotics. While the numberof granulocytes that can be collected has beenincreased with growth-factor mobilization, theclinical benefit of such a prophylactic or thera-peutic approach has not yet been demonstrated.

EVALUATION AND TREATMENT OFINFECTION

Most episodes of fever during periods of neu-tropenia are infectious in origin. These infec-tions have the potential to be rapidly fatal if notempirically treated.9 First infections are mostlikely caused by bacterial pathogens, while sub-sequent infections are usually caused by fungalpathogens, antibiotic-resistant bacteria, or viralpathogens.23 The initial evaluation and manage-ment of HSCT patients is not significantly dif-ferent than that of patients undergoingremission-induction or intensive consolidationtreatment for acute leukemia. Guidelines for themanagement of such patients have beendeveloped by several groups.20,21 The NCCNguidelines provide an excellent road map tomanage HSCT patients during their pre-engraftment period.20 In the latest version ofthese practice guidelines, a neutrophil count of<500 PMN/µl has been defined as neutropenia,and a single temperature of >38.0°C requiresclinical intervention. The patient’s evaluationshould be focused on determining the causativeorganism and potential site of infection. HSCTrecipients experience increased morbidity withrespiratory viruses, so determining exposure toill family members or caregivers is important.Laboratory assessment should focus on studiesthat help to define the functional status of theliver, kidneys, and lungs. Imaging procedures(e.g. chest radiographs, etc.) should be con-sidered when patients have any site-specificsymptoms. We routinely obtain a chest X-ray atthe onset of fever, and consider a CT scan of thelungs for patients who remain febrile, have pul-monary symptoms, or have room-air oxygen


saturation levels below 90%. Sinuses areimaged most efficiently with a CT scan.

Specimens for culture should be collectedduring or immediately after completion of thepatient’s examination. Blood should beobtained prior to initiation of antibiotics, butantibiotic administration should not be delayedwhile radiographs or other site-specific culturesare obtained. There is general consensus thatthe volume of blood cultured is the most impor-tant variable in optimizing microbial recoveryfor adult patients.73,74 It is recommended that atleast two blood cultures, or 20–40 ml of blood,be collected. The IDSA clinical care guidelinesrecommend that two blood cultures beobtained – one from a peripheral site and asecond from the central venous catheter.21 Thejustification for this recommendation is thebelief that disparity between the peripheralblood and the catheter blood culture may helpidentify catheter-related versus non-catheter-related infections. While drawing blood cul-tures from two different venipuncture sites mayhelp to distinguish between clinically importantand contaminant microorganisms, a meta-analysis of previously published studies hasshown little utility for obtaining blood for cul-ture from both the central indwelling venouscatheter and a peripheral vein.75 Thus, in manytransplant centers, blood for culture is onlydrawn from the patient’s vascular catheter.Quantitative blood cultures may be performed,but they are not routinely recommendedbecause of cost and the limited impact theyhave on clinical care.76,77

Site-specific cultures are important for HSCTrecipients. Cultures and biopsies from thesinus, lungs, and alimentary tract can be per-formed safely when they are coupled withappropriate platelet support and experiencedsubspecialty physicians (e.g. pulmonary crit-ical/care, gastroenterologists, surgeons, andotolaryngologists). One must have a lowthreshold to biopsy cutaneous lesions thatdevelop in the setting of fever and neutropenia.Histologic and microbiologic investigation ofthese cutaneous lesions may define the offend-

ing bacterial, fungal, or viral pathogen.Diarrhea is common after transplantation.

Diarrhea that occurs during the first 1–2 weeksafter completion of the transplant-conditioningtherapy usually results from treatment-inducedmucosal injury. This type of diarrhea will usu-ally resolve by day 10–14 after transplantation,but does signify additional mucosal injury thatcan be a portal for infection. Acute GVHDrarely occurs prior to day 14 post allogeneictransplant, but can be a major cause of diarrhea.Enteric infections that cause diarrhea are rare inthe pre-engraftment period. Cox et al78 prospec-tively studied 296 consecutive HSCT patients,and found that diarrhea occurred in 43%.However, diarrhea due to infection was uncom-mon, and accounted for only 13% of theseepisodes. In this study, organisms responsiblefor diarrhea included viruses (CMV, aden-ovirus, astrovirus, rotavirus) and bacteria(Clostridium difficile and Aeromonas). CMVenteritis is rare in the pre-engraftment period,and its overall incidence has decreased with theadvent of CMV-specific prophylaxis or preemp-tive therapy.4,5 Bacterial pathogens responsiblefor intestinal infection in the normal host (i.e.Salmonella, Shigella, Campylobacter, and Yersinia)almost never cause diarrhea in the hospitalizedtransplant recipient. Diarrhea secondary tointestinal parasites (e.g. Cryptosporidium, Giardialamblia, and Entamoeba histolytica) is alsounusual, but has been reported in center-specific outbreaks. The most common infectiouscause of diarrhea during the pre-engraftmentperiod is C. difficile colitis. The prevalence ofthis pathogen as part of a patient’s endogenousalimentary flora varies. It is believed to be pre-sent in less than 3% of normal hosts but its inci-dence is higher among HSCT recipients.79 Aprospective study of all new patients arrivingfor transplantation at the FHCRC found that21% of these patients had pre-existing coloniza-tion with C. difficile. These colonized individu-als provide a center-specific reservoir oforganisms that can be nosocomially transmittedto other susceptible patients, or for the colo-nized patient there can be future progression to


colitis once they are treated with antibiotictherapy. Optimal management of colonizedpatients is unclear, but infection control mea-sures are imperative, and evaluation of diar-rhea must always consider this pathogen.

INITIAL EMPIRIC THERAPY

HSCT recipients should be treated empiricallywith high-dose broad-spectrum antibiotics atthe first sign (fever) of infection. At present, alarge number of highly effective antibiotics areavailable. Despite many previous studies, it isnot possible to recommend a single antibiotic orantibiotic combination as initial treatment forHSCT patients with fever and neutrope-nia.20,21,80–83 The selection of antibiotics must takeinto consideration the following factors:

• The most common potential infectingpathogens.4,5,36

• The potential sites of infection.• The antimicrobial susceptibility patterns of

isolated pathogens. All infectious disease islocal, and knowledge of local antimicrobialsusceptibility patterns is critical in con-structing the appropriate empiric regimen.

• Empiric regimens must provide broad-spectrum antibiotic coverage. Antibioticregimens must provide a high-level ofactivity against Gram-negative bacilli,including Ps. aeruginosa, plus activityagainst Gram-positive organisms such as S.aureus and viridans streptococci.

• Pre-existing organ dysfunction. This is crit-ical for HSCT patients, who often have beenexposed to previous organ-damaging treat-ments (e.g. amphotericin B) or are receivingtreatment that has inherent renal and livertoxicity (e.g. cyclosporine and tacrolimus).In general, aminoglycosides, with theirinherent nephrotoxicity, should be usedwith caution in HSCT recipients.

There is considerable debate about theempiric use of vancomycin or antibiotics withincreased broad-spectrum Gram-positive activ-

ity (e.g. Linezolid and Synercid) in patientswith fever and neutropenia.20,21 The primaryjustification for empiric therapy is the know-ledge that a small number of Gram-positivepathogen infections can be rapidly fatal if nottreated promptly with the appropriate antibi-otics.26,28,29 However, the only single largeprospective randomized trial conducted inpatients with fever and neutropenia failed toshow a true clinical advantage for the use ofempiric vancomycin.84 This issue has not beendirectly studied in HSCT patients, but theEuropean Organization for the Research andTreatment of Cancer (EORTC) trial reported thatempiric use of vancomycin resulted in adecreased duration of fever, but did notimprove survival and was associated withexcess of renal and hepatic toxicity.

The primary barrier to the use of empiricvancomycin is the emergence of vancomycin-resistant pathogens. The development of colo-nization with vancomycin-resistant enterococci(VRE) has been associated with the increaseduse of vancomycin, although other antibioticsare also important.30,85 Because of this risk ofvancomycin-resistant pathogens, the initialempiric use of vancomycin or other new broad-spectrum Gram-positive antibiotics (Linezolidand Synercid) should be limited to HSCTpatients who develop fever and have one ormore of the following additional clinical factors.

1. Serious, clinically apparent catheter-relatedinfections. Many of these infections are dueto coagulase-negative staphylococci thathave a very high (80%) level of �-lactamantibiotic resistance.

2. Substantial mucosal damage coupled witha high-risk of infection with penicillin-resistant viridans streptococci. Significantmucosal disruption is of constant concernfor HSCT recipients, and, according to theSCOPE (Support of Commission Objectivesand Project Environment) surveillance data,18–29% of viridans streptococci isolatedfrom blood cultures will be resistant to�-lactam antibiotics.86,87


3. Blood cultures positive for Gram-positivebacteria prior to final pathogen identifica-tion and susceptibility testing.

4. Known colonization with �-lactam-resis-tant pneumococci, methicillin-resistant S.aureus (MRSA), or VRE.

5. Previous prophylaxis with quinoloneantibiotics or trimethoprim/sulfamethoxa-zole. Both of these agents have been associ-ated with an increased risk of Gram-positive infections.44,88,89 Recent molecular-epidemiology studies in HSCT recipientsand patients with acute leukemia have con-firmed the importance of alimentary tractcolonization with coagulase-negativestaphylococci as a risk factor for the devel-opment of blood stream infections withthese pathogens.90

6. The development of hypotension or thesepsis syndrome without an identifiedpathogen.

Empiric vancomycin could be considered inany of these six clinical situations, but if van-comycin therapy is initiated, it should be dis-continued after 3–4 days of treatment if anantibiotic-resistant organism is not identified.The empiric use of new agents such asLinezolid and Synercid is discouraged. Cost,lack of controlled trial experience, and potentialmarrow toxicity with Linezolid (personal com-munication Pharmacia/Upjohn) and significantmusculoskeletal toxicity with Synercid providereasons for caution when using these agentsempirically in HSCT recipients.

PATIENTS WITH DOCUMENTED INFECTIONSITES OR PATHOGENS

Identification of the causative pathogen allowsthe clinician the ability to optimize the antimi-crobial regimen and use therapy with a lowerincidence of adverse effects and costs. Theduration of treatment for documented infec-tions depends on the following factors:

• neutrophil recovery;

• rapidity of response to the antimicrobialtherapy;

• the site of infection and isolated pathogen;• status of the patient’s engraftment;• the patient’s need for additional immuno-

suppression (e.g. corticosteroids).

In general, most uncomplicated skin and ali-mentary tract mucosal infections are adequatelytreated with 5–7 days of treatment.20 For mostbacterial bloodstream infections, 1–2 weeks oftherapy are usually adequate, but fungal blood-stream infections require more prolonged ther-apy.20,21,91 Three to four weeks of therapy areneeded to effectively control bacterial sinus andlung infections, but a more prolonged antimi-crobial treatment is required if the causativepathogen is Ps. aeruginosa or a mold.20

Catheter-associated infections

Catheter-associated infections remain problem-atic. Long-term venous access devices were ini-tially developed for use in HSCT recipients, andare the standard of care for almost all HSCTpatients. The long-term indwelling venousaccess device allows the clinician to administerhigh-dose multi-agent therapy, provide consis-tent venous access for blood product support,administer parenteral nutrition and antimicro-bial therapy, and function as a portal for with-drawal of blood for physiologic monitoring,microbiologic evaluation, or PBSC collection.Catheter-associated infections are categorizedas entry-site infections, tunnel infections orbloodstream infections (Table 7.3). The delin-eation of an entry-site infection from a tunnelinfection can be clinically challenging, but theoccurrence of an apparent entry-site infectionplus a bloodstream infection usually indicatesthat the catheter tunnel is also infected. It isnow believed that the majority of entry-siteinfections can be managed effectively withantimicrobial therapy alone. Tunnel infectionsrequire catheter removal and culture, withmodification of the empiric antibiotics based on


culture and antibiotic susceptibility results.Determination of the true role of the venouscatheter in a bloodstream infection is difficultunless there is evidence of tunnel or entry-siteinflammation caused by the same organism thatis recovered from blood cultures. The majorityof bloodstream infections that occur in HSCTpatients who have indwelling catheters can bemanaged effectively with antimicrobial ther-apy, and do not require catheter removal.20,21

Immediate catheter removal is recommendedfor patients with bloodstream infections due tofungi, non-tuberculosis mycobacteria (Mycobac-

terium fortuitum complex and M. chelonae/abscessus group) and VRE.30,39,92 Persistentbloodstream infections are more frequent if thecatheter is not removed and the bloodstreampathogen is a Bacillus sp., C. jeikeium, S. aureus,Ps. aeruginosa, or S. maltophilia.29,93–95 All otherbloodstream infections can be initially treatedwith pathogen-specific antibiotics. Catheterremoval in these latter cases should be con-sidered if the bloodstream infection persistsbeyond 48 hours and no other site of infection isidentified. It is also important to consider thepossibility that the venous catheter is the pri-


Table 7.3 Vascular-catheter-associated infections in HSCT recipients

Infection Treatment/action

Entry-site infection • Pathogen-specific therapy(consider empiric vancomycin)

Tunnel infection • Catheter removal/culture• Pathogen-specific therapy

(consider empiric vancomycin)

Bloodstream infectionFungi (yeast or mold) • Immediate catheter removalNon-tuberculosis mycobacteria • Pathogen-specific therapyVancomycin-resistant enterococci

Corynebacterium jeikeium • Consider early catheter removalBacillus spp. • Pathogen-specific therapyStaphylococcus aureusPseudomonas aeruginosaStenotrophomonas maltophilia

All other positive blood cultures • Pathogen-specific therapy• Consider catheter removal for persistanta infection

a Positive blood culture >48 hours, no other site of infection.

mary site of bloodstream infections that recurafter a full course of antimicrobial therapy.20,96

There is not substantial data to support therecommendation that antibiotic administrationbe alternated through the different catheterlumens. One must believe that if the catheter istruly the source of infection, then antibioticsalone have a very low chance of sterilizing thevenous access device.

LACK OF CLINICAL RESPONSE TO INITIALEMPIRIC THERAPY

Management of HSCT patients with infectionwho do not clinically respond to antimicrobialtherapy is challenging.20 The lack of responsemay represent an infection with a pathogenresistant to the empiric antimicrobial regimenbeing administered, inadequate serum level ortissue levels of the antibiotics, infection at a vas-cular site (e.g. a catheter), closed-space infec-tion, the emergence of a second infection, or anunusually slow clinical response. It is importantto remember that the resolution of fever inpatients who are neutropenic is frequentlydelayed. A recent review of 488 episodes offever and neutropenia treated at the Universityof Texas MD Anderson Cancer Center revealedthat the median time to fever resolution rangedfrom 5 to 7 days.97 Less than 40% of patientsbecame afebrile before day 5 of therapy.Patients with Gram-negative bloodstream infec-tions were febrile for a mean of 6.6–8.2 days,while the time to fever defervescence was evenlonger for patients with Gram-positive blood-stream infections (range 6.6–12.4 days). Thesefindings are consistent with the results pub-lished by Freifeld et al,80 who reported that themean time for patients to become afebrile was 4days with ceftazidime therapy and 3 days forpatients treated with imipenem. These studiesprovide evidence that the time to fever defer-vescence can be prolonged, but that the feverresponse may also vary depending upon thespecific antibiotic regimen, site of infection, andinfecting pathogen.

Patients in whom fever persists beyond 4–5days of initial antimicrobial therapy shouldundergo reassessment. Broad-spectrum antibi-otics should be maximized, but, if possible,antibiotic combinations that minimize organtoxicity should be used. Clinical evaluationmust include a thorough daily evaluation,review of previous culture results, and furthersite-specific investigation as clinically indicated.Special attention should be given to the lungsand sinuses as occult sites of infection. Themanagement of such patients may require serialantimicrobial changes, and infectious diseasesconsultation can be helpful.

The need for change in therapy should bebased on the patient’s clinical status and thelikelihood of early marrow engraftment.Although fever resolution may be slow, persis-tent fever raises concern about an inadequatelytreated infection.97 For HSCT patients who arepersistently febrile and clinically unstable,additional Gram-negative bacillary coverage,empiric vancomycin, or empiric amphotericin Bshould be considered. The clinically stablebut febrile patient can be followed carefullywithout alteration of antimicrobial therapy.However, most physicians will consider the useof empiric amphotericin B or Ambisome (lipo-somal amphotericin B) if fever persists beyondday 6 of antibiotics.98 Walsh et al99 have recentlyreported on a multicenter trial that comparedvoriconazole (a new second-generation triazole)with Ambisome as empiric treatment forpatients with neutropenia and persistent fever.In this trial, voriconazole was comparable toAmbisome in therapeutic success, but superiorin reducing breakthrough fungal infections,infusion-related toxicity and nephrotoxicity.These encouraging results warrant furtherstudy in HSCT patients.

EVALUATION AND MANAGEMENT OFFUNGAL INFECTIONS

Invasive fungal infections have become anincreasingly important problem for HSCT


patients. Several studies have reported yearlyincreases in the incidence of candidiasis,aspergillosis, and some other mold infec-tions.3–5,13,37–39,100,101 With the progress that hasbeen made in the management of CMV infec-tion and pneumonia, aspergillosis is now thenumber one cause of infection death after allo-geneic transplantation.3,4,13,37 Fungal organismsare categorized into three general categories.The first group comprises the yeasts, which aredistinguished by their inability to form truehyphae, and their propensity to colonizemucosal surfaces. Candida spp. are the mostcommon, and were reported to occur in 10–20%of HSCT patients prior to the routine use ofazole prophylaxis.38,39,100 The second group com-prises the molds, which are characterized bytheir ability to form true hyphae, and are pri-marily acquired by the inhalation of aerosolizedspores. In the past, this was believed to be pri-marily contaminated air, but recent studieshave also raised the importance of the inhala-tion of aerosolized contaminated watersources.102 The most common molds areAspergillus spp., but Mucorales order, Fusarium,Bipolaris, and Pseudoallescheria are being morecommonly isolated. The incidence of Aspergillusinfections has been reported to range from 6%to 20% based on the transplant center and theyear of HSCT. The final group of fungi com-prises the dimorphic fungi, which have both ayeast and a hyphal form. These are oftenreferred to as the endemic fungi (e.g.Histoplasma and Coccidiodomycetes). Infectionswith these fungi are uncommon after HSCT,but should always be considered in patientswith a significant history of previous exposure.

The diagnosis of candidiasis and mold infec-tions continues to rely on recovery of thespecific pathogen from blood culture, oridentification by culture or histology from tis-sue samples. Rapid diagnosis of these infectionsusing serum antigen detection as a surrogatemarker for invasive disease remains a high pri-ority. Galactomannan detection for the diagno-sis of aspergillosis appears to be highly specificfor invasive aspergillosis, but test sensitivity

has varied from 43% to 83%, depending uponthe test cut-off index.103 The group fromBelgium prospectively collected sera fromHSCT recipients and then assessed the efficacyof galactomannan detection to predict invasiveaspergillosis.104 They reported that serial testingin HSCT recipients could identify invasiveaspergillus a median of 6 days (range 0–14days) before clinical parameters dictatedempiric antifungal therapy, and a median of 17days before the diagnosis of aspergillosis wasconfirmed. Clinical usefulness of this non-culture technique requires further prospectiveevaluation, but could potentially improve theability to diagnose invasive aspergillosis andmonitor the response to treatment.

Candidiasis

The frequency and occurrence of candidiasishas been well described.38,39,100 Goodrich et al39

reviewed the occurrence of candidiasis at theFHCRC before azole prophylaxis and the lipid-based amphotericin B products were beingused. In this report, 1.4% of transplant recipi-ents developed invasive candidiasis. The mostcommon organisms were C. albicans (62%) andC. tropicalis (21%). The median time to thedevelopment of a Candida bloodstream infec-tion was 15 days after transplantation, andthese infections had an attributable mortality of39%. Infection mortality was further increasedto 88% if Candida tissue invasion was also docu-mented. Autologous and allogeneic HSCTrecipients had a similar risk of candidiasis dur-ing the pre-engraftment period, but the risk ofdeveloping candidiasis persisted for allogeneictransplant recipients despite the return of gran-ulocyte function. Risk factors for this periodincluded neutropenia, older age, HLA-mis-matched donor, acute GVHD, and corticos-teroid use (Table 7.4). The occurrence ofbloodstream infections by Candida spp. wasrecently reassessed at the FHCRC after theintroduction of fluconazole prophylaxis.38

Forty-four percent of patients were colonized


with Candida spp. at some time either prior toor during the first 75 days after transplant. C.albicans was more likely to be recovered beforetransplantation, with non-albicans species morelikely to be recovered during the period ofazole use. C. albicans resistance to fluconazolewas modest (5.3% of isolates), with the mostcommon bloodstream isolates being C. glabrata,C. parapsilosis, and C. krusei. Candida blood-stream infections occurred in 4.7% of patients, amedian of 28 days after transplantation, andwere associated with a 20% mortality rate. Ofnote in this study, the use of fluconazole pro-phylaxis negated the importance of neutropeniaas a risk factor, but had no impact on infectionswith non-albicans species (Table 7.4).

Aspergillosis

Aspergillus infections primarily involve the res-piratory tract. Invasive infection of the lungappears more common, but the true incidenceof sinus involvement in HSCT recipients is

known. There is a suggestion that differentAspergillus spp. may have different tissue tro-pisms, with A. flavus more likely to cause sinusinfection than is A. fumigatus. It has becomeclear that Aspergillus infection after HSCToccurs in three distinct pathophysiologic con-ditions.13 One group of patients will developtheir infection during the pre-engraftmentperiod with the primary risk factor being neu-tropenia (Table 7.4). The second group devel-ops their infection later, between days 40 and120 post transplantation, and are predisposedto infection because of a persistent cellularimmune defect. The third group experienceAspergillus infections very late after transplanta-tion, and these infections appear to be highlycorrelated with delayed immune reconstitution,CMV disease and chronic GVHD (K Marr, per-sonal communication, 2000). Extrapulmonaryspread of Aspergillus is more common withneutropenia-associated infections, while non-neutropenic Aspergillus pneumonia amongHSCT recipients is more likely to present withprogressive diffuse pulmonary infiltrates.105 The


Table 7.4 Risk factors (multivariate analysis) for candidiasis and aspergillosis after HSCT

Candidiasis Aspergillosis13,52

Pre-azole39 Azole treatment52 Early Late

Increased age Yes Yes No YesUnrelated donor Yes No No YesGVHD Yes No No YesCorticosteroids Yes No No YesNeutropenia Yes No Yes NoSeason (summer) No No Yes NoConcomitant infection Yes Yes Yes YesLaminar airflow units No No Yes NoConstruction No No No YesCMV-positive Yes Yes Yes Yes

mortality rate for Aspergillus infection develop-ing early or late after transplant remains veryhigh, ranging from 60% to 88%.13,38 Median sur-vival after diagnosis is usually short (36 days)for patients with early and late aspergillosis,but can be longer when it occurs very late aftertransplantation. Diagnosis, prevention, andtreatment of Aspergillus infections remain inad-equate. Diagnosis is dependent on recovery ofthe infecting organism by histology or microbi-ology from tissue specimens or pulmonary orsinus lavage. Results of galactomann assays areencouraging.103,104 While the use of surveillancecultures continues to be controversial in thecare of HSCT recipients, the finding of A. fumi-gatus or A. flavus in respiratory specimens hasbeen shown to be highly predictive of futureinvasive aspergillosis.13 Mucosal eschars alongthe nasal septum are an important clinical clueto the diagnosis of Aspergillus sinusitis. Biopsyand culture of such lesions are always indi-cated. If nasal lesions are not observed, sinusaspirate and biopsy may establish the diagnosisand preclude the need for further diagnosticprocedures.

Bronchoalveolar lavage (BAL) is the standardapproach for evaluating an HSCT patient withpulmonary lesions. BAL is a less sensitive pro-cedure if the pulmonary lesions are focal, small,and/or peripherally located. The sensitivity ofBAL to diagnose Aspergillus is only 50–60%.Thus a negative procedure does not exclude thediagnosis of Aspergillus. Open lung biopsy isusually reserved for patients with a negativeBAL, or for patients in whom the disease is pro-gressive and a diagnosis must be immediatelyestablished. A thoracoscopic approach is nowroutinely utilized when possible because oflowered procedure morbidity. Biopsies of bothperipheral and central areas of abnormal lungare recommended because of the focal nature ofAspergillus infections, and the wide distributionof organisms within the pneumonic process.

Treatment of established Aspergillus infec-tions that occur in HSCT recipients remainsinadequate, and is in a state of evolution. Earlydiagnosis and treatment remains critical.

Prolonged (8–10 weeks) high-dose (1.0–1.5 mg/kg/day) amphotericin B or an equivalent doseof a lipid formulation of amphotericin B is thestandard initial treatment.20,21 Additional sup-pressive therapy with an oral agent such as itra-conazole is often given if the patient remainsimmunosuppressed.20 To date, the lipid formu-lations of amphotericin B have not been shownto increase survival, although the incidences ofboth acute infusion-related toxicity and nephro-toxicity are decreased.98,106 Usage of lipid for-mulations of amphotericin B must be based onthe risk of developing nephrotoxicity and thecost of treatment.107,108 The use of antifungalcombinations remains controversial, and somein vitro studies have suggested the possibilityof clinically relevant antagonism if azoles suchas itraconazole are combined with ampho-tericin B.

Voriconazole, a new triazole with enhancedactivity against Aspergillus, is available in bothan oral and an intravenous formulation, andappears encouraging as treatment of aspergillo-sis. Other new antifungal agents (e.g. posacona-zole, ravuconazole, caspofungin, and liposomalnystatin) are being tested in phase II-III trials astreatment of invasive Aspergillus infections inHSCT recipients. These agents show promise,with response rates of 25–40% among patientswho have failed to respond to amphotericinB. Efforts to enhance infected patients’ immunestatus have been recommended, but the use ofadjunct growth factors (G-CSF and GM-CSF),or G-CSF-mobilized granulocyte transfusionhave not been shown to be beneficial.64,109 Therole of surgery as an adjunct to antimicrobialtreatment for pulmonary Aspergillus infectionsremains controversial, but a preliminary analy-sis from the FHCRC (D Weiss, personal com-munication, 2000) shows little survival benefit ifsurgical resection is performed.

Respiratory viruses

RSV, parainfluenza, and infuenza type A or Bare now recognized as important pathogens for


HSCT recipients during the pre-engraftmentperiod.22,110–113 These infections are usually sea-sonal (winter months), but outbreaks may per-sist within a transplant center beyond the timewhen healthy individuals in the community areno longer being infected. The true incidence ofthese infections is unknown, but some centershave reported an attack rate among HSCTrecipients of 20%.112 The progression of upperrespiratory tract infection to lower tract dis-ease varies among the specific viruses.Approximately 50% of patients who becomeinfected with RSV will develop pneumonia.22,112

The rate of progression is lower for parain-fluenza (32%), and rare for influenza. The RSVpneumonia mortality rate has remained high(27–82%), but mortality may be even moresevere when the infection and pneumoniadevelop pre-engraftment.22 The mortality ratefrom parainfluenza pneumonia has variedbetween 32% and 57%, with the mortality rateamong HSCT recipients who develop influenzapneumonia being in a similar range. There is noproven effective therapy for any of these respi-ratory virus infections after HSCT, althoughaerosolized ribavirin alone or in combinationwith either polyclonal or monoclonal RSVimmunoglobulin has been used.112 Supportivecare and appreciation of the potential complica-tions such as secondary bacterial and fungalpneumonia are the mainstay of treatment.Intravenous RSV immunoglobulin and pre-emptive therapy with aerosolized ribavirin arecurrently under investigation for the preventionand treatment of RSV pneumonia. Infectioncontrol efforts to minimize acquisition andtransmission of these pathogens are critical.

SITE OF CARE FOR HSCT PATIENTS WITHFEVER AND NEUTROPENIA

There is increasing acceptance that certainpatients with fever and neutropenia can besafely managed as outpatients, or with short-ened hospital stays.20,114–117 Several investigatorshave developed prospective models to help

predict a population of patients who would beat low risk of developing infection-associatedcomplications, and would therefore be candi-dates for outpatient therapy.117,118 In general,most models have excluded HSCT recipientsfrom the low-risk patient group because of thebelief that their underlying immunosuppres-sion and intensity of treatment inherentlymakes them a high-risk population.117,118

The FHCRC has not attempted to managethe first episode of fever and neutropenia dur-ing the pre-engraftment period on an outpa-tient basis. This has not been feasible becausemost patients experience significant comorbidillness during the pre-engraftment period (e.g.severe mucositis, or renal or hepatic dysfunc-tion). However, approximately 5–10% of con-ventional allogeneic transplant recipients andoccasional autologous HSCT recipients willdevelop a period of neutropenia (<500PMN/µl) during periods following initialengraftment. Many of these neutropenic peri-ods are complicated by the development offever and infection. These neutropenic periodsare caused by graft failure, disease relapse, andmarrow suppression caused by infection (e.g.CMV, HHV-6, or HHV-8) or medications (e.g.ganciclovir). In general, we do not now admitall of these patients to the hospital for broad-spectrum antibiotics. Patients are evaluated atthe first sign of infection (fever), with a thor-ough physical examination, blood cultures,chest X-ray, and pertinent physiologic assess-ments (e.g. hematological, kidney, liver, andlung). Patients who are clinically stable, and donot have hypotension, hypoxia (SaO2

< 90% onroom air), a stool volume >500 cm3/24 h, orrenal insufficiency (serum creatinine > 2.0 mg/dl)are then treated with intravenous ceftazidimeand ciprofloxacin in our day hospital/ambula-tory facility and observed for 6–8 hours.Patients who remain stable are then allowed tocontinue their antibiotic therapy as outpatients.All patients return to the clinic the followingmorning for reassessment. Patients are assesseddaily until clinically improved, antibiotics areadjusted based on culture results, and patients


who show signs of clinical deterioration areimmediately hospitalized for more intensemonitoring. The Center’s infectious diseaseteam evaluates all patients who remain febrilebeyond day 3 of broad-spectrum antibiotic ther-apy. Experience with this pilot approach is lim-ited, but critical issues for the success of thistreatment strategy are willing, educatedpatients, full-time caregivers (part of the rou-tine FHCRC transplant procedure), follow-upcare available 24 hours/day, 7 days a week, anda low threshold for hospitalization if needed.

SUMMARY

The morbidity and mortality of infections forHSCT recipients during the pre-engraftmentperiod have dramatically decreased, but duringthis same time period there has been the emer-gence of antibiotic-resistant bacteria, and asignificant increase in the incidence of fungalinfections. More potent and less toxic antibi-otics have been developed. PBSC transplantshave decreased the duration of pre-engraftmentneutropenia, and progress has been made in thearea of prophylaxis of C. albicans infections.However, patients now are at an increased riskof non-albicans Candida infections, and the mor-tality from mold infections (e.g. Aspergillus)remains high. The promise of new diagnostictechniques, additional antimicrobial agents,and strategies for treatment and prophylaxishold the potential that HSCT may be safer inthe future. It will be critical for the transplantteam to remember that HSCT-associated infec-tions are a dynamic process, with change aguarantee. Transplant physicians must not for-get the value and importance of standardinfection-control measures. It will be a chal-lenge to implement such infection-controlapproaches as the care of HSCT patientsbecomes more focused in an ambulatory/dayhospital facility.

REFERENCES

1. Maris MB, Sandmaier BM, Niederwieser D etal, Comparison of donor chimerism, graft rejec-tion, and GVHD after hematopoeitic stem-celltransplants (HSCT) from HLA matched siblingsand unrelated donors using conditioning with2 GY TIB with and without fludarabine (FLU).Blood 2000; 96: 520a.

2. Childs R, Chernoff A, Contentin N et al,Regression of metastatic renal-cell carcinomaafter nonmyeloblative allogeneic peripheral-blood stem-cell transplantation. N Engl J Med2000; 343: 750–8.

3. Walter E, Bowden R, Infection in the bone mar-row transplant recipient. Infect Dis Clin NorthAm 1995; 9: 823–47.

4. VanBurik JH, Weisdorf DJ, Infections in recipi-ents of blood and marrow transplantation.Hematol Oncol Clin North Am 1999; 13: 1065–89.

5. Van Burk J-A, Weisdorf D, Infections in recipi-ents of blood and marrow transplantation. In:Principles and Practice of Infectious Diseases, 5thedn (Mandell GL, Bennett JE, Dolin R, eds).Philadelphia: Churchill Livingstone, 2000:3136–47.

6. Pizzo PA, Management of fever in patients withcancer and treatment-induced neutropenia. NEngl J Med 1993; 328: 1323–32.

7. Wade JC, Management of infection in patientswith acute leukemia. Hematol Oncol Clin NorthAm 1993; 7: 293–315.

8. Bodey GP, Buckley M, Sathe YS et al,Quantitative relationships between circulatingleukocytes and infection in patients with acuteleukemia. Ann Intern Med 1966; 64: 328–40.

9. Schimpff SC, Empiric antibiotic therapy forgranulocytopenic cancer patients. Am J Med1986; 80(Suppl 5c): 13–20.

10. Imbert-Marcille BM, Tang XW, Lepelletier D etal, Human herpesvirus-6 infection after autolo-gous or allogeneic stem cell transplantation: asingle-center prospective longitudinal study of92 patients. Clin Infect Dis 2000; 31: 881–6.

11. Wang F-Z, Linde A, Hagglund H et al, Humanherpesvirus 6 DNA in cerebrospinal fluid speci-mens form allogeneic bone marrow transplantpatients: Does it have clinical significance? ClinInfect Dis 1999; 28: 562–8.

12. Luppi M, Barozzi P, Schulz TF et al, Bone mar-row failure associated with human herpesvirus


8 infection after transplantation. N Engl J Med2000; 343: 1378–85.


14. Sullivan KM, Nims J, Leisenring W et al,Determinants of late infection following mar-row transplantation for aplastic anemia andmyelodysplastic syndrome. Blood 1995;86(Suppl 1): 213a.

15. Sullivan KM, Mori M, Sanders J et al, Late com-plications of allogeneic and autologous marrowtransplantation. Bone Marrow Transplant 1992;10: 127–34.

16. Hoyle C, Goldman JM, Life-threatening infec-tions occurring more than 3 months after BMT.Bone Marrow Transplant 1994; 14: 247–52.

17. Sullivan KM, Agura E, Anasetti C et al, Chronicgraft-versus-host disease and other late compli-cations of bone marrow transplantation. SeminHematol 1991; 28: 250–9.

18. Anderson KC, Weinstein HJ, Transfusion-associated graft-versus-host disease. N Engl JMed 1990; 323: 315–21.

19. Flowers ME, Leisenring W, Beach K et al,G-CSF given to donors before apheresis doesnot prevent aplasia in patients treated withdonor leukocyte infusions for recurrent CMLafter bone marrow transplantation. Biol BloodMarrow Transplant 2000; 6: 321–6.

20. Wade JC, Rubenstein EB, and the NCCNGuidelines Committee, Clinical practice guide-lines for fever and neutropenia. Oncology 1999;13: 197–257.

21. Hughes WT, Armstrong D, Bodey GP et al, 1997guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

22. Harrington RD, Houton TM, Hackman RC et al,An outbreak of respiratory syncytial virus in abone marrow transplant center. J Infect Dis1992; 165: 987–93.

23. Wingard JR, Santos GW, Saral R, Differencesbetween first and subsequent fevers duringprolonged neutropenia. Cancer 1987; 59: 844–9.

24. Glauser M, Boogaerts M, Cordonnier C et al,Empiric therapy of bacterial infections in severeneutropenia. Clin Microbiol Infect 1997;3(Suppl 1): S77–86.

25. Engelhard D, Elishoov H, Strauss N et al,Nosocomial coagulase-negative staphylococcalinfections in bone marrow transplantation recipi-ents with central vein catheter – a 5-year prospec-tive study. Transplantation 1996; 61: 430–4.

26. Villablanca JG, Steiner M, Kersey J et al, Theclinical spectrum of infections with viridansstreptococci in bone marrow transplant recipi-ents. Bone Marrow Transplant 1990; 6: 387–93.

27. Classen DC, Burke JP, Ford CD et al,Streptococcus mitis sepsis in bone marrow trans-plant patients receiving oral antimicrobial pro-phylaxis. Am J Med 1990; 89: 441–6.

28. Elting LS, Bodey GP, Keefe BH, Septicemia andshock syndrome due to viridans streptococci: acase–control study of predisposing factors. ClinInfect Dis 1992; 14: 1201–7.

29. Stamm WE, Tompkins LS, Wagner KF et al,Infection due to Corynebacterium species in mar-row transplant recipients. Ann Intern Med 1979;91: 167–73.

30. Edmond MB, Ober JF, Dawson JD et al,Vancomycin-resistant enterococcal bacteremia:natural history and attributable mortality. ClinInfect Dis 1996; 23: 1234–9.

31. Martin MA, Pfaller MA, Wenzel RP, Coagulase-negative staphylococcal bacteremia: mortalityand hospital stay. Ann Intern Med 1989; 110:9–16.

32. Wenzel RP, Perspective: Attributable mortality– the promise of better antimicrobial therapy. JInfect Dis 1998; 178: 917–19.

33. Johnston PB, Litzow MR, Elliott MA et al,Colonization with vancomycin-resistant entero-coccus correlates with poor outcome in patientsundergoing allogeneic blood and marrow trans-plants. Blood 2000; 96: 786a.

34. Brown AE, Kiehn TE, Armstrong DE, Bacterialresistance in patients with neoplastic disease.Infect Dis Clin Pract 1995; 4(Suppl 3): S136–44.

35. Koll BS, Brown AE, The changing epidemiologyof infections at a cancer hospital. Clin Infect Dis1993; 17(Suppl 2): S322–8.

36. Elting LS, Rubenstein EB, Rolston KV et al,Outcomes of bacteremia in patients with cancerand neutropenia. Clin Infect Dis 1997; 25:247–59.

37. Ribaud P, Chastang C, Latage JP et al, Survivaland prognostic factors of invasive aspergillosisafter allogeneic bone marrow transplantation.Clin Infect Dis 1999; 28: 322–30.


38. Marr KA, Seidel K, White TC et al, Candidemiain allogeneic blood and marrow transplantrecipients: evolution of risk factors after adop-tion of prophylactic fluconazole. J Infect Dis2000; 181: 309–16.

39. Goodrich JM, Reed EC, Mori M et al, Clinicalfeatures and analysis of risk factors for invasivecandidal infection after marrow transplanta-tion. J Infect Dis 1991; 164: 731–40.

40. Petersen FB, Buckner CD, Clift RA et al,Infectious complications in patients undergoingmarrow transplantation: a prospective random-ized study of the additional effect of decontami-nation and laminar air flow isolation amongpatients receiving prophylactic systemic antibi-otics. Scand J Infect Dis 1987; 19: 559–67.

41. Hospital Infection Control Practices AdvisoryCommittee (HIPAC), Recommendations forpreventing the spread of vancomycin-resistance. Infect Control Hosp Epidemiol 1995; 16:105–13.

42. Bjerke J, Meyers JD, Bowden RA, Hepatospleniccandidiasis – a contraindication to marrowtransplantation? Blood 1994; 84: 2811–14.

43. Offner F, Cordonnier C, Ljungman P et al,Impact of previous aspergillosis on the outcomeof bone marrow transplantation. Clin Infect Dis1998; 26: 1098–103.


45. Conneally E, Cafferkey MT, Daly PA et al,Nebulized amphotericin B as prophylaxisagainst invasive aspergillosis in granulocy-topenic patients. Bone Marrow Transplant 1990;5: 403–6.

46. O’Donnell MR, Schmidt GM, Tegtmeier BR etal, Prediction of systemic fungal infection inallogeneic marrow recipients: impact of ampho-tericin prophylaxis in high-risk patients. J ClinOncol 1994; 12: 827–34.

47. Rousey SR, Russler S, Gottlieb M, Ash RC,Low-dose amphotericin B prophylaxis againstinvasive aspergillus infections in allogeneicmarrow transplantation. Am J Med 1991; 91:484–92.

48. Perfect JR, Klotman ME, Gilbert CC et al,Prophylactic intravenous amphotericin B inneutropenic autologous bone marrow trans-plant recipients. J Infect Dis 1992; 165: 891–7.

49. Tollemar J, Ringden O, Andersson S et al,Prophylactic use of liposomal amphotericin B(Ambisome) against fungal infections: a ran-domized trial in bone marrow transplant recipi-ents. Transplant Proc 1993; 25: 1495–7.


51. Slavin MA, Osborne B, Adams R et al, Efficacyand safety of fluconazole for fungal infectionsafter marrow transplant – a prospective, ran-domized, double-blind study. J Infect Dis 1996;171: 1545–52.

52. Marr KA, Seidel K, Slavin MA et al, Prolongedfluconazole prophylaxis is associated withpersistent protection against candidiasis-related death in allogeneic marrow transplantrecipients: long-term follow-up of a random-ized, placebo-controlled trial. Blood 2000; 96:2055–61.

53. Rotstein C, Bow EJ, Laverdiere M et al,Randomized placebo-controlled trial of flucona-zole prophylaxis for neutropenic cancerpatients: benefit based on purpose and intensityof cytotoxic therapy. Clin Infect Dis 1999; 28:331–40.

54. Broers AEC, van der Holt R, van Esser JWJ et al,Increased transplant-related morbidity andmortality in CMV-seropositive patients despitehighly effective prevention of CMV diseaseafter allogeneic T-cell-depleted stem cell trans-plantation. Blood 2000; 95: 2240–5.

55. Wade JC, Newton B, Flournoy N et al, Oral acy-clovir for the prevention of herpes simplexvirus reactivation after marrow transplantation.Ann Intern Med 1984; 100: 823–8.

56. Baglin TP, Gray JJ, Marcus RE, Wreghitt TG,Antibiotic resistant fever associated with herpessimplex virus infection in neutropenic cancerpatients with haematological malignancy. J ClinPathol 1989; 42: 1255–8.

57. Weaver CH, Schulman KA, Wilson-Relyea B etal, Randomized trial of filgrastim, sar-gramostim, or sequential sargramostim and fil-grastim after myelosuppressive chemotherapyfor the harvesting of peripheral-blood stemcells. J Clin Oncol 2000; 18: 43–53.

58. Holmberg LA, Boeckh M, Hooper H et al,Increased incidence of cytomegalovirus disease


after autologous CD34-selected peripheralblood stem cell transplantation. Blood 2000; 94:4029–35.

59. Crippa F, Holmberg L, Hooper H et al,Infections after autologous CD34 selectedperipheral blood stem cell transplantation.Blood 2000; 96: 586a.

60. Champlin RE, Schmitz N, Horowitz MM et al,Blood stem cells compared with bone marrowas a source of hematopoietic cells for allogeneictransplantation. Blood 2000; 95: 3702–9.

61. Bensinger WI, Martin PJ, Storer B et al,Transplantation of bone marrow as comparedwith peripheral-blood cells from HLA-identicalrelatives in patients with hematologic cancers.N Engl J Med 2001; 344: 175–81.

62. Powles R, Mehta J, Kulkarni S et al, Allogeneicblood and bone-marrow stem-cell transplanta-tion in haematological malignant diseases: arandomised trial. Lancet 2000; 355: 1231–7.

63. Flowers MED, Matos AVB, Storer B et al,Clinical manifestations of chronic graft-versus-host disease (cGVHD) after transplantation ofperipheral blood stem cell (PBSC) compared tobone marrow. Blood 2000; 96: 203a.

64. Ozer H, Armitage JO, Bennett CL et al, 2000update of recommendations for the use ofhematopoietic colony-stimulating factors:evidence-based, clinical practice guidelines. JClin Oncol 2000; 18: 3558–85.

65. Hartmann LC, Tchetter LK, Habermann TM etal, Granulocyte colony-stimulating factor insevere chemotherapy-induced afebrile neu-tropenia. N Engl J Med 1997; 336: 1776–80.

66. Clift RA, Sanders JE, Thomas ED et al,Granulocyte transfusions for the prevention ofinfection in patients receiving bone marrowtransfusions. N Engl J Med 1978; 298: 1052–7.

67. Bhatia S, McCulough J, Perry EH et al,Granulocyte transfusions: efficacy in treatingfungal infections in neutropenic patientsfollowing bone marrow transplantation.Transfusion 1994; 34: 226–32.

68. Winston D, Ho WG, Young LS, Prophylacticgranulocyte transfusions during human bonemarrow transplantation. Am J Med 1980; 68:893–9.

69. Price TH, Bowden RA, Boeckh M et al, PhaseI/II trial of neutrophil transfusions from donorsstimulated with G-CSF and dexamethasone fortreatment of patients with infections in

hematopoietic stem cell transplantation. Blood2000; 95: 3302–9.

70. Pena E, Narvios A, Lichtiger B, Therapeuticgranulocyte transfusions: retrospective study of34 patients. Blood 2000; 96: 63a.

71. Hubel K, Carter R, Liles WC et al, Communitydonors vs. related donors for granulocyte trans-fusion therapy in hematopoietic transplanta-tions: a comparative analysis of feasibility andoutcome. Blood 2000; 96: 63a.

72. Adkins DR, Goodnough LT, Shenoy S et al,Effect of leukocyte compatibility on neutrophilincrement after transfusion of granulocytecolony-stimulating factor-mobilized prophylac-tic granulocyte transfusions and on clinical out-comes after stem cell transplantation. Blood2000; 95: 3605–12.

73. Reimer LG, Wilson ML, Weinstein MP, Updateon detection of bacteremia and fungemia. ClinMicrobiol Rev 1997; 10: 444–65.

74. Weinstein MP, Current blood culture methodsand systems: clinical concepts, technology, andinterpretation of results. Clin Infect Dis 1996; 23:4046.

75. Siegman-Igra Y, Anglim AM, Shapiro DE et al,Diagnosis of vascular catheter-related blood-stream infection: a meta-analysis. J ClinMicrobiol 1997; 35: 928–36

76. Whimbey E, Wong B, Kiehn TE et al, Clinicalcorrelations of serial quantitative blood culturesdetermined by lysis-centrifugation in patientswith persistent septicemia. J Clin Microbiol 1984;19: 766–71.

77. Benezra D, Kiehn TE, Gold JWM et al,Prospective study of infections in indwellingcentral venous catheters using quantitativeblood cultures. Am J Med 1988; 85: 495–8.

78. Cox GJ, Matsui SM, Lo RS et al, Etiology andoutcome of diarrhea after marrow transplanta-tion: a prospective study. Gastroenterology 1994;107: 1398–407.

79. McFarland LV, Mulligan ME, Kwok RYY et al,Nosocomial acquisition of Clostridium difficileinfection. N Engl J Med 1989; 17: 484–90.

80. Freifeld AG, Walsh T, Marshall D et al,Monotherapy for fever and neutropenia incancer patients: a randomized comparison ofceftazidime vs. imipenem. J Clin Oncol 1995; 13:165–76.

81. Cometta A, Calandra T, Gaya H et al,Monotherapy with meropenem vs. combination


therapy with ceftazidime plus amikacin asempiric therapy for fever in granulocytopenicpatients with cancer. Antimicrob AgentsChemother 1996; 40: 1108–15.

82. Oblon D, Ramphal R, A randomized trial ofcefepime vs. ceftazidime as initial therapy forpatients with prolonged fever and neutropeniaafter intensive chemotherapy. Proc Am AssocCancer Res 1993; 34: 1362a.

83. Cometta A, Zinner S, de Bock R et al,Piperacillin–tazobactam plus amikacin vs. cef-tazidime plus amikacin as empiric therapy forfever in granulocytopenic patients with cancer:the International Antimicrobial TherapyCooperative Group of the EuropeanOrganization for Research and Treatment ofCancer. Antimicrob Agents Chemother 1995; 39:445–52.

84. Winston DJ, Lazarus HM, Beveridge RA et al,Randomized, double-blind, multicenter trialcomparing clinafloxacin with imipenem asempirical monotherapy for febrile granulocy-topenic patients. Clin Infect Dis 2001; 32: 381–90.

85. Donskey CJ, Hanrahan JA, Hutton RA, Rice LB,Effect of parenteral antibiotic administration onpersistence of vancomycin-resistant Entero-coccus faecium in the mouse gastrointestinaltract. J Infect Dis 1999; 180: 384–90.

86. Pfaller M, Jones R, Marshall S et al, Nosocomialstreptococcal blood stream infections in theSCOPE Program: species occurrence andantimicrobial resistance: the SCOPE HospitalStudy Group. Diagn Microbial Infect Dis 1997; 29:259–63.

87. Edmond MB, Wallace SE, McClish DK et al,Nosocomial bloodstream infections in UnitedStates hospitals: a three-year analysis. Clin InfectDis 1999; 29: 239–44.

88. Bow EJ, Loewen R, Vaughan D, Reducedrequirement for antibiotic therapy targetingGram-negative organisms in febrile, neu-tropenic patients with cancer who are receivingantibacterial chemoprophylaxis with oralquinolones. Clin Infect Dis 1995; 20: 907–12.

89. Ward TT, Thomas RG, Fye CL et al,Trimethoprim–sulfamethoxazole prophylaxis ingranulocytopenic patients with acute leukemia:evaluation of serum antibiotic levels in a ran-domized, double-blind, placebo-controlledDepartment of Veterans Affairs cooperativestudy. Clin Infect Dis 1993; 17: 323–32.

90. Herwaldt LA, Hopis RJ, Boyken LD et al,Molecular epidemiology of coagulase-negativestaphylococci isolated from immunocompro-mised patients. Infect Control Hosp Epidemiol1992; 13: 86–92.

91. Wendt C, Messer SA, Hollis RJ et al, RecurrentGram-negative bacteremia: incidence and clini-cal patterns. Clin Infect Dis 1999; 28: 611–17.

92. Roy V, Weisdorf D, Mycobacterial infectionsfollowing bone marrow transplantation. A 20year retrospective review. Bone MarrowTransplant 1997. 19: 467–70.

93. Vassilomanolakis M, Plataniotis G, Koumakis Get al, Central venous catheter-related infectionsafter bone marrow transplantation in patientswith malignancies: a prospective study withshort-course vancomycin prophylaxis. BoneMarrow Transplant 1995; 15:r 77–80.

94. Banerjee C, Bustamante CI, Wharton R et al,Bacillus infections in patients with cancer. ArchIntern Med 1988; 148: 1769–74.

95. Press OW, Ramsey PG, Larsen EB et al,Hickman catheter infections in patients withmalignancy. Medicine 1984; 63: 189–200.

96. Engellard D, Elishoov H, Strauss N et al,Nosocomial coagulase-negative staphylococcalinfections in bone marrow transplantationrecipients with central vein catheter. A 5 yearprospective study. Transplantation 1996; 61:430–4.

97. Elting LS, Rubenstein EB, Rolston K et al, Timeto clinical response: an outcome of antibiotictherapy of febrile neutropenia with implica-tions for quality and cost of care. J Clin Oncol2000; 21: 3699–706.

98. Walsh TJ, Finberg R, Arndt C et al, Liposomalamphotericin B is effective for empirical anti-fungal therapy in cancer patients with feverand neutropenia. N Engl J Med 1999; 340:764–71.

99. Walsh T, Pappas P, Winston D et al,Voriconazole versus liposomal amphotericin Bfor empirical antifungal therapy of persistentlyfebrile neutropenic patients: a randomized,international, multicenter trial. In: Proceedingsof 40th Interscience Conference on AntimicrobialAgents and Chemotherapy, San Francisco, CA,2000: 20 (Abst).

100. Viscoli C, Girmenia A, Marinus A et al,Candidemia in cancer patients: a prospectivemulticenter surveillance study by the Invasive


Fungal Infection Group (IFIG) to the EuropeanOrganization for the Research and Treatment ofCancer (EORTC). Clin Infect Dis 1999; 28:1071–9.

101. Kontoyiannis DP, Sumoza D, Tarrand J et al,Significance of aspergillemia in patients withcancer: a ten-year study. Clin Infect Dis 2000; 31:188–9.

102. Anaissie E, Stratton S, Summerbell R et al,Opportunistic moulds in a hospital water sys-tem: a 3-year prospective study. Blood 2000; 96:787a.

103. Marr KA, Balajee A, Leisenring W et al, Utilityof galactomannan detection for the diagnosis ofinvasive aspergillosis in HSCT recipients. Blood2000; 96: 786a.

104. Maertens JA, Verhaegen J, Van Eldere J,Boogaerts M, Detection of circulating galac-tomannan (GM) is an early and sensitive indica-tor of invasive aspergillosis (IA) in allogeneicstem cell transplant recipients (ASCTR). Blood2000; 96: 585a.

105. Marr KA, Carter K, Myerson D et al,Aspergillosis in HSCT recipients: evidence fortwo distinct pathophysiologic conditions asso-ciated with engraftment status. Blood 2000; 96:787a.

106. Walsh TJ, Hiemenez JW, Seibel NL et al,Amphotericin B lipid complex for invasive fun-gal infections: analysis of safety and efficacy of556 cases. Clin Infect Dis 1998; 26: 1383–96.

107. Wingard JR, Kubilis P, Lee L et al, Clinicalsignificance of nephrotoxicity in patientstreated with amphotericin B for suspected orproven aspergillosis. Clin Infect Dis 1999; 29:1402–7.

108. Cagnoni PJ, Walsh TJ, Prendergast MM et al,Pharmacoeconomic analysis of liposomalamphotericin B versus conventional ampho-tericin B in the empirical treatment of persis-tently febrile neutropenic patients. J Clin Oncol2000; 18: 2476–83.

109. Gaviria JM, van Burik JH, Dale DC et al,

Comparison of interferon-�, granulocytecolony-stimulating factor, and granulocyte-macrophage colony-stimulating factor for prim-ing leukocyte-mediated hyphal damage ofopportunistic fungal pathogens. J Infect Dis1999; 179: 1038–41.

110. Wendt CH, Weisdorf DJ, Jordan MC et al,Parainfluenza virus respiratory infection afterbone marrow transplantation. N Engl J Med1992; 326: 921–6.

111. Aschan J, Ringden O, Ljungman P et al,Influenza B in transplant patients. Scand J InfectDis 1989; 21: 349–50.

112. Bowden RA, Respiratory virus infections aftermarrow transplant: the Fred HutchinsonCancer Research Center experience. Am J Med1997; 102: 27–30.

113. Englund JA, Sullivan CJ, Jordan MC et al,Respiratory syncytial virus infection inimmunocompromised adults. Ann Intern Med1988; 109: 203–8.

114. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low-risk cancer patients withfever and neutropenia: a pilot study of 30patients, based on a validated prediction rule. JClin Oncol 1994; 12: 107–14.


116. Rolston KV, New trends in patient management:risk-based therapy for febrile patients with neu-tropenia. Clin Infect Dis 1999; 29: 515–19.

117. Talcott J, Siegel R, Finberg R et al, Risk assess-ment in cancer patients with fever and neu-tropenia: a prospective, two-center validation ofa prediction rule. J Clin Oncol 1992; 10: 316–22.

118. Klastersky J, Paesmans M, Rubenstein EB et al,The Multinational Association for SupportiveCare in Cancer Risk Index: a multinational scor-ing system for identifying low-risk febrile neu-tropenic cancer patients. J Clin Oncol 2000; 18:3038–51.


8Initial clinical evaluation and risk assessmentof the febrile neutropenic cancer patientJames A Talcott, Edward B Rubenstein

INTRODUCTION

Bodey et al,1 by noting the relationship betweenabsolute neutrophil counts (ANC) less than1000/mm3 and an increased risk of infection,especially serious infection, identified fever andneutropenia as a high-risk clinical state.Although assessing patient risk is an integralaspect of a physician’s role, formal discussionsof risk assessment are rare in medical text-books. However, risk assessment has been cen-tral to the management of cancer patients withfever and neutropenia since its original descrip-tion by Bodey and colleagues. Since that report,many papers have identified factors that eitherclinically or statistically are associated with out-comes in patients with fever and neutropenia.In this chapter, we will provide an overview ofthe initial clinical approach to the neutropeniccancer patient who presents with fever, providea framework for risk assessment as it relates tothe care of patients with fever and neutropenia,and detail some of the characteristics thatdefine low- and high-risk subgroups that can beused to help the clinician make initial decisionsregarding site of care (inpatient versus outpa-tient) and route for antibiotic therapy (oral ver-sus intravenous).

Defining the clinical problem

The signs and symptoms of infection are oftensubtle in the neutropenic patient, but feverremains the cardinal sign of early infection. Asingle temperature of greater than 38.0°C(100.4°F) in the absence of an obvious environ-mental cause is generally considered a fever.Neutropenic patients who have just receivedblood products should not be considered tohave simple pyrogenic febrile reactions fromtheir transfusion if they remain febrile two ormore hours after transfusion. These patientsmust be presumed to be infected and shouldbe treated with broad-spectrum antibiotics.Similarly, initial febrile episodes in neutropenicpatients should never be attributed to non-infectious causes such as tumor fever or drugfever. Numerous studies have shown that morethan half of neutropenic patients who becomefebrile, if carefully evaluated, will end up hav-ing either a microbiologically documentedinfection (MDI) or a clinically documentedinfection (CDI).2–7 Approximately 10–20% ormore of febrile patients with neutrophil countsof less than 100/µl will have documented bac-teremias.8 The purpose of the initial evaluationis to try and elucidate the focus of the patient’sinfection so that the initial empiric antimicro-bial regimen is appropriately selected based

upon the organism most likely to be responsiblefor the initial infection. Primary sites of infec-tion include the alimentary tract from theoropharynx to the perirectum, including theperiodontal area, mouth, pharynx, esophagus,large and small bowel and rectum, the sinuses,lungs, and skin. Common sites of infectioninclude the skin and soft tissues surroundingcentral venous catheters or the catheters them-selves (see Chapter 10). Patients with neutrope-nia and signs or symptoms of infection whopresent without fever should be evaluated andtreated as if they were febrile, including thosepatients who are elderly or debilitated andthose who have received corticosteroids, whichmay also initially blunt the febrile response.Neutropenic patients with infections due toClostridium septicum may also present withoutfever.9

Essential elements of the history,examination, and evaluation

Since fever in the neutropenic cancer patient istraditionally considered an oncologic emer-gency, it is important for the clinician to rapidlyperform a focused history and physical exami-nation. Baseline history should include thetiming of the onset of the fever, associatedsymptoms such as true rigors (which suggestbloodstream infection), chest pain, dyspnea orcough (which suggest respiratory tract infec-tion), and other site-specific history and evalua-tions as outlined in Table 8.1. Althoughuncommon, the presence of headache and pho-tophobia with or without neck stiffness shouldstill suggest meningitis in a neutropenic febrilecancer patient.

Laboratory testing should include a completeblood count with differential and platelet count,serum electrolytes, blood urea nitrogen, serumcreatinine, evaluation of liver transaminases(aminotransferases), bilirubin, alkaline phos-phatase, and appropriate site-specific cultures.Initially, two blood cultures should be drawn,with a minimum of 20 cm3 of blood per culture.

If the patient has an indwelling vascular accessdevice (VAD), the site should be examined forerythema, induration, or purulence. RoutineVAD site cultures are not indicated. At leastone blood culture should be taken through thelumen of the VAD, and, if the patient will allowa peripheral venous sample, then another bloodculture from a vein should also be obtained.10

For patients with urinary symptoms such asfrequency or dysuria, Gram stain and culture isindicated. Routine urinalysis is rarely helpful,since neutropenic patients do not usually havepyuria. Routine chest radiographs in theabsence of chest symptoms have a low yield,and a normal chest radiograph does notexclude the possibility of pneumonia, since theneutropenic patient cannot mount a vigorousinflammatory response.11 The chest examina-tion is important, however. Crackles or rales, ifheard, are strongly suggestive of a pneumonicprocess, and should prompt further investiga-tion even in the presence of a normal chest film.Chest computed tomography (CT) has beenshown to be useful in the evaluation of poten-tial fungal pneumonia in neutropenic trans-plant and leukemia patients.12 The baselineclinical evaluation is important for planningnon-infectious disease supportive care therapysuch as blood product transfusions and hydra-tion therapy, and for documenting the patient’sclinical status at presentation. Patients who ini-tially present without an obvious focus of infec-tion may end up having subsequently docu-mented infections if carefully evaluated andreassessed daily.13

SITE-SPECIFIC ASSESSMENT

Sinus and nasal passages

The sinuses are a common site of infection inneutropenic cancer patients. Predisposing fac-tors may include prior chronic sinusitis andnasal polyps. Patients may complain ofheadache or unilateral facial pain involvingthe frontal, temporal, or occipital areas.


INITIAL CLINICAL EVALUATION AND RISK ASSESSMENT 153

Table 8.1 History, examination and evaluation in the neutropenic cancer patient

Site of infection Signs/symptoms Diagnostic evaluation Organisms to consider

Sinus/nasal passages Unilateral pain/tearing Limited CT of sinuses Streptococcus pneumoniaeDysesthesias ENT consult Staphylococcus aureusPeriorbital cellulitis Consider sinus drainage, biopsy Gram-negative bacteriaRhinorrhea Appropriate stains/cultures Anaerobic bacteria

Aspergillus spp. and othermolds/fungi

Skin/soft tissues/wounds Pain Biopsy/aspiration Coagulase-negative Erythema Stains/cultures staphylococciCellulitis S. aureusVesicular lesions Gram-negative bacteria:Ecthyma gangrenosum Pseudomonas aeruginosaNodules/abscesses Herpes simplex virus (HSV)

Varicella zoster virusCandida spp.Rapidly growingmycobacterium (non-tuberculous)

Mouth/oropharynx Pain, odynophagia Cultures/stains HSVErythema, mucositis Consider dental oncology Candida spp.Plagues evaluation StreptococciGingivitis AnaerobesNecrotizing/vesicular Gram-negative bacterialesions

Esophagus Persistent nausea Endoscopy Candida spp.Dysphagia Biopsy and cultures HSVRetrosternal burning Cytomegalovirus (CMV)

Liver/gallbladder/pancreas Right upper quadrant pain CT scan Gram-negative bacterianausea/vomiting Ultrasound Enterococcus↑ alkaline phosphatase Consider adverse effects of Anaerobic bacteria↑ transaminases concomitant medications Candida spp.(aminotransferases)↑ amylase/lipase↑ bilirubin

Colon/Intestines Crampy abdominal pain CT scan Salmonella spp.Loose, watery, diarrhea Surgical consultation Shigella spp.Bloody diarrhea Enteric stool cultures Giardia lamblia

Clostridium difficile toxin Candida spp.CMV

Contd

Occasionally, dysesthesias may also be noted.Unfortunately, many patients may have sinusi-tis without significant signs or symptoms, and –particularly in bone marrow transplant patientsor patients with acute leukemia – the diagnosismay often be delayed. Owing to the unreliabil-ity of routine sinus radiographs, most experts

now consider limited CT scans as the radi-ographic test of choice. Air fluid levels, mucosalthickening, or bone erosion may be noted aswell. ENT consultation is frequently required inorder to obtain material via aspiration or biopsyof the sinuses for appropriate stains andcultures to establish microbiologic diagnosis.


Table 8.1 History, examination and evaluation in the neutropenic cancer patient – contd

Site of infection Signs/symptoms Diagnostic evaluation Organisms to consider

Anaerobic bacteriaClostridium septicumStrongyloides stercoralisEnteric Gram-negativesPs. aeruginosa

Perirectal/groin Pain, induration Surgical consultation Gram-negative bacteriaGYN consultation Anaerobic bacteriaEndoscopy Enterococcus

Candida spp.

Respiratory tract Cough CXR, CT chest Gram-negative bacteriaSputum production Sputum stains and cultures S. aureusDyspnea Bronchoscopy Haemophilus influenzaePleuritic chest pain Bronchoalveolar lavage S. pneumoniae

Lung biopsy Seasonal viruses (respiratorysyncytial virus, influenza)Aspergillus spp.Pneumocystis cariniiLegionella spp.Candida spp.MycobacteriaCMV

Vascular access device Entry site erythema, Stains/cultures Staphylococcus epidermidis(VAD) tenderness Quantitative cultures through S. aureus

VAD and peripheral blood CorynebacteriumAcinetobacterPs. aeruginosaStenotrophomonasmaltophiliaBacillus spp.Candida spp.Non-tuberculous mycobacteria

Etiologic agents include Gram-positive bacteria,especially Streptococcus pneumoniae andStaphylococcus aureus, and Gram-negative bacte-ria; mixed infections are also common, includ-ing anaerobic organisms. Although lesscommon in solid tumor patients, Aspergillusspp., the Zygomycetes, and other molds may bethe cause of sinus infections in patients whohave undergone bone marrow transplantationor those who have acute leukemia.14

Mouth and oropharynx

The mouth and oropharynx are common sites ofinfection in neutropenic cancer patients.Mucositis due to chemotherapy disrupts thenormal protective mechanism in the mouth andoropharynx, and creates a portal of entry intothe bloodstream for potential pathogens.Unfortunately, pain and difficulty in swallow-ing are non-specific symptoms of mucositis, andare unreliable in determining which patientshave oropharyngeal infections. Periodontalinfections are common causes of fever in neu-tropenic cancer patients, and dental oncologyconsultation may be indicated for patients withpoor dentition.15 Oral lesions may become colo-nized with Gram-negative bacilli and cause bac-teremia. Alpha-hemolytic streptococci may alsoenter the bloodstream via the disrupted mucosalmembrane of patients with significant mucosi-tis.16 Mouth anaerobes and Candida spp., whichare often of low pathogenic potential, may causebloodstream infections and febrile episodes inneutropenic cancer patients with significant oralmucositis. Herpes simplex virus (HSV) may bereactivated in patients with mucositis, and maycause more extensive mucosal damage and pro-long healing time.17

Esophagus

Mucositis may also include the mucosal surfaceof the esophagus, and the presence of retroster-nal burning pain and dysphagia should prompt

the clinician to consider an esophageal sourcefor the patient’s febrile episode. Endoscopicevaluation may be warranted, with appropriateplatelet support, to obtain material for diagno-sis if the patient is thrombocytopenic. HSV andcytomegalovirus (CMV) should be consideredas potential pathogens in the appropriatepatient setting.18 Patients with oropharyngealcandidiasis may have esophageal candidiasis aswell. Although the classic symptoms of ret-rosternal burning and odynophagia may sug-gest an esophageal cause for the patient’sfebrile episode, more commonly chronic nauseaand/or vomiting may be the only symptoms ofesophageal infection.

Liver, gallbladder, and pancreas

Infections of the hepatobiliary tract are usuallysuspected because patients present withabdominal pain or right upper quadrant pain,or are found to have elevations in their alkalinephosphatase, bilirubin, transaminases, or amy-lase/lipase. The classic triad of right upperquadrant pain, fever, and jaundice, suggestingcolangitis, may be present in patients withhepatobiliary tract cancers who have eitherinternal or external drainage devices. Initialevaluation for infections in these areas shouldinclude ultrasound, CT, or magnetic resonanceimaging (MRI). Surgical consultation may bewarranted, and, for patients who have internalor external biliary stents, these may need endo-scopic evaluation to ensure that they are notobstructed. During the course of the patient’smanagement, it may be necessary to exchangeinfected stents. In addition to Gram-negativeenteric organisms, enterococci (includingvancomycin-resistant enterococci), and Candidaspp. are important pathogens to consider.19 Theclinician should also recognize that risingtransaminases, bilirubin, and alkaline phos-phatase in the setting of fever in the neu-tropenic cancer patient may not be due to aninfection, but may be due to an adverse effect ofone of the patient’s concomitant medications.


Therefore, re-evaluation of the patient’s medica-tion profile is often warranted. Patients withacute leukemia or patients who have undergonebone marrow transplantation who had pro-longed episodes of neutropenia may be predis-posed to hepatosplenic candidiasis, which ofteninitially presents as fever without an obviousfocus of infection or as an elevated alkaline phos-phatase.20 CT scans may show multiple smalllesions in the liver or spleen, which can be biop-sied to obtain a definitive diagnosis.

Intestines and colon

The presence of crampy abdominal pain anddiarrhea, with or without gastrointestinal bleed-ing, suggest the possibility of gastrointestinalmucositis as the cause of the patient’s infection.Enteric pathogens that cause febrile episodes innon-neutropenic hosts should be considered,and stool cultured for Giardia, Salmonella,Shigella, and Cryptosporidium. Stool should alsobe tested for Clostridium difficile toxin, even inthe absence of prior antibiotic therapy. Rarely,Strongyloides may be the cause of infection, andcan lead to intestinal obstruction andperitonitis.21 Typhilitis, or the so-called neu-tropenic enterocolitis, is suggested by fever,diarrhea, and abdominal pain. This syndromeusually occurs in patients with acute leukemiawho are neutropenic and have had intensivecytotoxic chemotherapy, and is caused byenteric Gram-negatives, including Pseudomonasaeruginosa.22 Other organisms in the differentialdiagnosis would include, as previously men-tioned, C. difficile colitis, CMV, and graft-versus-host disease. Endoscopy may be occasionallywarranted to obtain appropriate material forpathologic examination and culture. CT scansmay be useful in making the diagnosis and candemonstrate cecal wall thickening, local intra-mural hemorrhage, and edema of the ileum orparts of the colon.22 For patients with massivediarrhea, other supportive care measures suchas hydration, electrolyte replacement, and totalparenteral nutrition may be indicated.

Perirectum and groin

The patient may complain of painful defecationor may have a history of problems with rectalfissures or hemorrhoids. Examination of theperirectum and groin areas may reveal discreteerythema, induration, or fluctuance. Althoughroutine internal rectal exams are usually notindicated because of concerns about inducingbacteremia, gentle perirectal examinationsshould be performed in all neutropenic febrilecancer patients. Perirectal infections may beassociated with bacteremia, particularly withGram-negative organisms, and up to 10% ofthese patients may present with shock.23 Forlarge perirectal infections, particularly thosethat are fluctuant, surgical consultation fordrainage may be warranted. Cultures demon-strate that these infections are usually polymi-crobial, and are due to a mixture ofGram-negative enteric organisms, anaerobes,and enterococci. Candida infections are alsocommon. Perirectal infections are more com-mon in patients with acute leukemia, particu-larly monocytic and myelomonocytic leukemia,compared with patients with solid tumors.23

In addition to appropriate antibiotics andsurgical drainage, analgesics and stool softenersare indicated along with local therapy (e.g. sitzbaths).

Skin, soft tissues, and wounds

In an immunocompetent host, skin infectionsare easily diagnosed owing to localized ery-thema pain, tenderness, and the typical signsand symptoms due to local inflammatory reac-tions. Unfortunately, in neutropenic cancerpatients, many of these signs and symptoms areblunted. Skin infections in neutropenic cancerpatients may be a sign of a localized processsuch as an infection around a VAD, or may bepart of a hematogenous process due to bac-teremia. Elting et al24 reviewed the relationshipbetween the size of soft tissue lesions and out-comes in 163 cases of non-bacteremic soft tissue


infections in neutropenic cancer patients. Theresponse rate to initial antibiotic therapy was43% in patients whose soft tissue lesions meas-ured more than 5 cm, compared with 87%among those patients with smaller lesions(p < 0.0001). Soft tissue infections of any sizewith central necrosis and those more than 5 cmin size in the setting of bacteremia have alsobeen found to be associated with a poor clinicaloutcome.

Organisms associated with soft tissue andwound infections include S. aureus, coagulase-negative staphylococci, Gram-negative bacteria(Ps. aeruginosa and Stenotrophomonas mal-tophilia), and mixed Gram-negative/anaerobicinfections. Varicella zoster virus and HSV arealso common causes of skin infections in neu-tropenic cancer patients. Initial evaluation ofskin infections in neutropenic cancer patientsshould include aspiration and punch biopsiesof infected sites, with appropriate stains andcultures for bacteria, fungi, and atypicalmycobacteria.

Respiratory tract

Pulmonary infections are relatively frequentamong neutropenic cancer patients. Thesepatients may present with cough, dyspnea, andsputum production; however, one-third of neu-tropenic patients with pneumonia will have nosigns of rales or other symptoms indicating arespiratory tract infection. Owing to the lack ofan inflammatory response, a chest radiographis often normal during initial evaluation offebrile neutropenic patients who present withpneumonia.6 Patients with neutropenic pneu-monia may present with mental status changes,hypoxemia, and significant dyspnea. Cancerpatients with T-cell defects who are neu-tropenic may have a dry cough, and conversa-tional dyspnea, which suggests Pneumocystiscarinii pneumonia. Pneumonias that present atthe onset of the febrile neutropenic episode aretypically due to Gram-negative bacteria such asPs. aeruginosa, Klebsiella spp., and other

Enterobacteriaceae.25 Seasonal respiratory virusessuch as respiratory syncytial virus, influenzavirus, and CMV, as well as Legionella spp., mayalso be important pathogens in neutropeniccancer patients.26–29 The initial evaluationshould include blood cultures and examinationof the sputum for bacteria, fungi, and mycobac-teria. CT scans of the chest should be con-sidered when patients present with focal ornodular lesions, and bronchoalveolar lavageshould be considered for those patients whoinitially present with interstitial infiltrates.30

Additional laboratory tests such as fungal anti-gen assays may be warranted in certain situ-ations.

Urinary tract

Patients with urinary tract infections (UTIs)typically present with complaints of dysuria,urinary frequency, nocturia, and occasionallyhematuria. Unfortunately, because of the lackof inflammatory response in neutropenic cancerpatients, many of these symptoms are absent.In the absence of instrumentation of the urinarytract, or a prior history of frequent UTIs, theseinfections are a relatively infrequent cause offever in the neutropenic cancer patient.Nevertheless, routine Gram stain and urine cul-tures are warranted. Owing to the lack ofpyuria, routine urinalysis is generally con-sidered to be a low-yield test. Obviously, themost common causes of UTIs in neutropeniccancer patients would include Escherichia coliand other Gram-negative enterics; however,enterococci also need to be considered, sincemany of the initial empiric antibiotic regimensdo not provide adequate enterococcal coverage.

RISK ASSESSMENT

Although the value of risk assessment in the careof patients with fever and neutropenia is wellestablished, ‘risk’ has meaning only when theoutcome of interest is specified. Risk assessment


implies arranging patients along a spectrum ofrisk – but a risk of what? In various fever andneutropenia contexts, risk has been defined asthe likelihood of developing clinical infection,1

serious bacterial or fungal infection,31–33 a con-dition indicating medical instability, whether ornot infection-related,34,35 an incomplete clinicalresponse to an initial antibiotic regimen,36 or aneffective but slower resolution of infection.37

The varied definition of risk does not indicatemuddled research, but rather the flexibility ofthe risk assessment methodology: the outcomecan and should vary, depending on the clinicalquestion. For example, when patients are beingconsidered for outpatient treatment settingswhere medical surveillance is reduced and thusdetection of new problems potentially delayed,any evidence of medical instability may be per-tinent, while, when antibiotic drug regimensare being compared, prompt resolution of infec-tion is the appropriate outcome of interest. Thepurpose of risk assessment at the time of theinitial clinical evaluation is to substratify thisheterogeneous population into rational groupsbased upon clinically meaningful outcomes.

The Talcott clinical prediction rule

Talcott and colleagues34 were the first todevelop a formal clinical prediction rule basedupon statistical methods to stratify this hetero-geneous patient population. They asked animportant clinical question: are there easilyidentifiable factors that, when present on dayone of the initial febrile episode, predictadverse outcomes during the remainder of theepisode? Their goal was to identify a low-riskpatient group defined by the absence of identi-fiable medical instability during the febrile neu-tropenic episode, using information availablewithin 24 hours of presentation. To define‘medical instability’, they identified events orconditions that they labeled ‘major medicalcomplications’. The investigators specified anumber of these that required close medicalattention or intervention, such as systemic

hypotension, prolonged or heavy bleeding, anew cardiac dysrhythmia, altered mentalstatus, or any other condition that requiredobservation or intervention. These states werenot confined to those resulting directly or indi-rectly from severe infections, althoughinfection-related problems predominated. Inorder to assess risk more comprehensively, theyincluded serious complications that wererelated to the patient’s underlying malignancyor comorbid condition. A patient who is med-ically stable from the standpoint of infection butdevelops serious acute gastrointestinal bleedingduring a febrile neutropenic episode would bea poor candidate for outpatient management.

In a retrospective study of 261 episodes offever and neutropenia over 12 months, 22%resulted in one or more major medical compli-cation. This one-in-five risk of a serious medicalevent empirically justified the standard policyof keeping all patients with fever and neutrope-nia in the hospital. However, most of these seri-ous medical complications occurred in patientswho were identified as having one of thefollowing risk factors: those who were inpa-tients at the time at which fever and neutrope-nia developed (group I), evidence of anothersignificant comorbid condition that indepen-dently justified hospitalization within 24 hoursof presentation with fever and neutropenia(group II); or evidence of uncontrolled cancer(group III), defined as either acute leukemia notin documented complete remission or anothercancer that had progressed either clinically orradiologically during the most recent evaluablechemotherapy regimen. These patients (groupsI–III) had a complication rate of 36%, and 20%died (Table 8.2). The remaining patients with-out any one of these indicators of high risk(group IV) appeared to have a very low risk ofserious medical complications (2%), and noneof them died. The low-risk group was large:70% of patients who developed fever and neu-tropenia as outpatients. When indicators ofthese high-risk groups were put into a multiplelogistic regression model with the occurrence ofany serious medical complication as the depen-


dent variable, they were so dominant (all fac-tors significant p < 0.0001) that other factorsconventionally associated with high risk, suchas the diagnosis of acute leukemia (p < 0.0001),greater age (p < 0.0001), and more severe neu-tropenia (p < 0.0001), remained only marginallysignificant.34

To validate this prediction rule, these investi-gators prospectively evaluated the high-risk cri-teria in an additional population, including notonly patients at the original site, the Dana-Farber Cancer Institute in Boston, but alsocancer patients at the Miriam Hospital inProvidence, Rhode Island. In this confirmatorystudy, clinical reviewers without access toinformation on the patients’ clinical course afterthe initial 24-hour evaluation period assignedpatients to one of the four risk groups. Otherreviewers blinded to the initial 24-hour perioddetermined whether or not a major medicalcomplication subsequently occurred. Theseresults confirmed the initial model, althoughthe risk difference between the high-risk groupsand the remaining low-risk patients decreased.Patients in groups I–III were more likely thanthose in group IV to have any medical compli-cation, multiple medical complications, anddeath (25% versus 5%, 13% versus 0%, and 9%

versus 0%, respectively).35 Of the five group IVpatients with complications, one appeared tobe a documentation error, one had transienthypotension within minutes after the 24-hourobservation period, and three developed unam-biguous medical complications seven or moredays after admission, an apparently adequateperiod for detection by an appropriate programof medical follow-up. The combined data foroutcomes of the derivation set and validationset of this clinical prediction rule are shown inTable 8.3.

As a result of this work, clinicians had accessto a clinical decision rule allowing them toidentify a large group of low-risk patientswithin 24 hours of presentation with fever andneutropenia. This low-risk group (Talcott groupIV) of patients appeared to have a low enoughrisk of medical instability to make them appro-priate candidates for clinical trials of programsof outpatient medical management, wheremedical surveillance could only be episodicrather than the continued observation permit-ted by inpatient care. These criteria were usedas entry criteria for a pilot study and a subse-quent randomized trial of home intravenousantibiotic therapy by these investigators.38 Someconcerns were raised that the Talcott clinical


Table 8.2 Outcomes of patient groups: Talcott derivation study34

Percentage Number of SeriousPatient group of total patients complications Death

Inpatients (group I) 39 101 34 (34%) 23 (23%)

Outpatients with concurrent 8 22 12 (55%) 3 (14%)comorbidity (group II)

Outpatients with uncontrolled 10 26 8 (31%) 4 (15%)cancer (group III)

Low-risk outpatients (group IV) 43 112 2 (2%) 0 (0%)

All patients 100 261 56 (21%) 30 (11%)

prediction rule was derived and validated withpatients mostly from a single center.39 In addi-tion, sensitivity, specificity, and misclassifica-tion rates for the Talcott model were neverreported.

The MASCC Risk Index

The Multinational Association for SupportiveCare in Cancer (MASCC) Study Section onInfections developed an internationally vali-dated scoring system to identify low-riskpatients in a prospective cohort study.40

Between December 1994 and November 1997,this group gathered information on 1351patients at 20 institutions in 15 countries. Afterexcluding patients with undocumented fever,neutropenia, or prior cytotoxic chemotherapy,1139 patients were analyzed. The derivationand validation sets were developed by randomallocation of participating institutions ratherthan patients, greatly enhancing the generaliz-ability of the results compared with commonlyused statistical approaches, such as bootstrap-ping.41 Using a multiple logistic regressionanalysis, the authors identified eight statisti-cally significant risk factors, including baseline

demographic characteristics (age < 60 years,solid tumor), past medical history (no chronicobstructive pulmonary disease, no prior fungalinfection, outpatient status at the time of pre-sentation with fever and neutropenia), clinicianassessment of overall patient status (no symp-toms or mild symptoms or moderate symptomsindicating overall burden of illness) (Figure8.1), and the absence of particular acute patho-logical states (no hypotension and no dehydra-tion) (Table 8.4).

Based on the coefficient of each risk factor inthe multiple regression derivation model, ascoring system was developed, allowing usersof the index to vary the threshold for designat-ing patients at low risk (Table 8.5). Increasingscores indicated lower levels of risk. Using theoptimal cutoff of 21 points, the group were ableto identify 80% of patients who subsequentlyproved not to have a complication (sensitivity0.80), with an overall complication rate amonglow-risk patients of 6%. However, raising thethreshold one point to 22 resulted in signifi-cantly fewer patients identified as at low risk,with the sensitivity falling from 80% to 57%,although associated with a lower complicationrate of 3% among low-risk patients. In contrast,the Talcott clinical prediction rule, designed to


Table 8.3 Outcomes of patient groups: Talcott combined derivation and validation studies34,35

Percentage Number of SeriousPatient group of total patients complications Death

Inpatients (group I) 52 369 128 (37) 48 (13)

Outpatients with concurrent 9 65 26 (40) 8 (12)comorbidity (group II)

Outpatients with uncontrolled 8 55 14 (25) 8 (15)cancer (group III)

Low-risk outpatients (group IV) 31 216 7 (3) 0 (0)

All patients 100 705 175 (25) 64 (9)


No signs orsymptoms

Mild signs orsymptoms

Moderate signsor symptoms

Severe signsor symptoms

Moribund

Estimate the burden of illness, considering all comorbid conditions

How sick is the patient at presentation?

Figure 8.1 Visual analog score used in the MASCC Risk Index to measure burden of illness. No or mildsymptoms corresponds to 5 points in the scoring system.

Table 8.4 Multiple logistic regression model based on 746 episodes of fever and neutropenia40

Characteristic Coefficient OR (95% CI)a p value

Burden of illness:No or mild symptoms 2.1 8.2 (4.2–16.4) <0.001Moderate symptoms 1.3 3.7 (2.2–6.3) <0.001

No hypotension 2.0 7.6 (2.9–19.9) <0.001No chronic obstructive pulmonary disease 1.7 5.3 (1.9–15.5) <0.002Solid tumor or no previous fungal infection 1.6 5.1 (2.0–12.9) <0.001No dehydration 1.3 3.8 (1.9–7.7) <0.001Outpatient status 1.2 3.5 (2.0–6.0) <0.001

a OR, odds ratio; 95% CI, 95% confidence interval.

Table 8.5 MASCC Risk Index scoring system40

Characteristic Weight

Burden of illness: no or mild symptoms 5No hypotension 5No chronic obstructive pulmonary disease 4Solid tumor or no previous fungal infection 4No dehydration 3Burden of illness: moderate symptoms 3Outpatient status 3Age < 60 years 2

Note: Points attributed to the variable ‘burden of illness’ are not cumulative. The maximum theoretical score is therefore 26.

be conservative in designating patients at lowrisk, identified 32% of patients without subse-quent complications and a 4% complication ratein patients designated as at low risk. The rela-tionship between sensitivity and specificity inthe derivation set of 756 patients for theMASCC Risk Index in comparison with theTalcott model can be demonstrated in graphicformat (Figure 8.2).

In the validation set, a MASCC Risk Indexscore of 21 points identified 71% of patientswithout subsequent complications, with a 9%complication rate among designated low-riskpatients. Using a threshold of 22 points, the sen-sitivity dropped to 47%, and the complicationrate fell to 6%. The Talcott rule, which excludedall inpatients, identified only 30% of thosepatients with no complications as at low risk,with a 7% complication rate. Of interest, whenthe Talcott criterion of outpatient status wasadded to the MASCC scale using the authors’preferred threshold of 21, the sensitivity fell to46% of all patients without complications, whilethe complication rate decreased to 6%. A com-parison of the performance of the Talcott clini-cal prediction rule with the MASCC Risk Indexfor the validation set at a score of 21 is shown inFigure 8.3.

The development of the MASCC Risk Indexrepresented an important advance in the gen-eralizability of clinical prediction rules.Developed and tested in a broad range of clini-cal settings internationally, the study sharplyreduced concerns regarding limited geographicapplicability raised by the northeastern USAorigins of the clinical prediction rule by Talcottand colleagues. The study also demonstrated thevalidity of the Talcott criteria in an internationalsetting. The MASCC Risk Index increased thegroup of patients at low risk, expanding thepotential patient population for less aggressivetreatments, such as oral antibiotic regimens oroutpatient management. The large difference inperformance of the MASCC Risk Index associ-ated with small changes in the numerical cutoffraises some concern about its robustness.However, the Index provides a flexible, validclinical prediction rule developed and validatedin a very broad range of clinical circumstances,and variation in individual beliefs about theproper threshold is made possible by thenumerical form of its results. While its complex-ity may limit its use in routine clinical practice,it could easily be incorporated into prospectiveclinical trials. Currently, the Index is being usedin two clinical trials based in Brussels.42


1.0

0.8

0.6

0.4

0.2

0

Rat

e

8 14 17 19 20 21 22 23 24

Score

Sensitivity, MASCC Risk Index

Specificity, MASCC Risk Index

Sensitivity, Talcott

Specificity, Talcott

Figure 8.2 Sensitivity versus specificity for the MASCC Risk Index compared with the Talcott model (derivationset N � 756).

Risk assessment has been an integral part inthe management of cancer patients with feverand neutropenia since its initial development.Through the years, there has been a steadyincrease in the methodological sophistication ofrisk-assessment studies, as well as in our formalunderstanding of how to integrate the processinto clinical care. Risk assessment has providedpowerful management tools for clinicians, andwill likely continue to do so in the future.

REFERENCES


2. Middleman EL, Watanabe A, Kaizer H, BodeyGP, Antibiotic combinations for infections inneutropenic patients. Cancer 1972; 30: 573–9.

3. Bodey GP, Feld R, Burgess MA, �-lactam antibi-otics alone or in combination with gentamicinfor therapy of Gram-negative bacillary infectionsin neutropenic patients. Am J Med Sci 1976; 271:179–86.

4. Fainstein V, Bodey GP, Elting L et al, A random-ized study of ceftazidime compared to cef-tazidime and tobramycin for the treatment of

infections in cancer patients. J AntimicrobChemother 1983; 12: 101–10.

5. Winston DJ, Ho WG, Bruckner DA et al,Controlled trials of double beta-lactam therapywith cefoperazone plus piperacillin in febrilegranulocytopenic patients. Am J Med 1988;85(Suppl 1A): 21–30.

6. Flaherty JP, Waitley D, Edlin B et al, Multicenter,randomized trial of ciprofloxacin plus azlocillinversus ceftazidime plus amikacin for empirictreatment of febrile neutropenic patients. Am JMed 1989; 87(Suppl 5a): 278S–82S.

7. Rolston KVI, Berkey P, Bodey GP et al, A com-parison of imipenem to ceftazidime with orwithout amikacin as empiric therapy in febrileneutropenic patients. Arch Intern Med 1992; 152:283–91.

8. Schimpff SC, Empiric antibiotic therapy for gran-ulocytopenic cancer patients. Am J Med 1986;80(Suppl 5c): 13–20.

9. Caya JG, Farmer SG, Ritch PS et al, Clostridialsepticemia complicating the course of leukemia.Cancer 1986; 57: 2045–8.

10. Weinstein MP, Current blood culture methodsand systems: clinical concepts, technology, andinterpretation of results. Clin Infect Dis 1996; 23:40–6.

11. Donowitz GR, Harman C, Pope T, Stewart M, Therole of the chest roentgenogram in febrile neu-tropenic patients. Arch Intern Med 1991; 151: 701–4.


1.0

0.8

0.6

0.4

0.2

0

Rat

e

Score 21

Sensitivity

Specificity

Misclassification

Talcott

Figure 8.3 The sensitivity, specificity, and misclassification rates for the MASCC Risk Index at a proposedscore of 21 points compared with the Talcott clinical prediction rule (validation set N � 383).

12. Barloon TJ, Galvin JR, Mori M et al, High-resolution ultrafast chest CT in the clinical man-agement of febrile bone marrow transplantpatients with normal or nonspecific chestroentgenograms. Chest 1991; 99: 928–33.

13. Rubenstein EB, Kim YJ, Legha R, Rolston KVI onbehalf of the MASCC Infectious Disease StudySection, Documented infections in febrile neu-tropenic patients presenting with apparent feverof undetermined origin: a wolf in sheep’s cloth-ing. Supp Care Cancer 2000; 8: 242 (Abst 5).

14. Goering P, Berlinger NT, Weisdorf DJ,Aggressive combined modality treatment of pro-gressive sinonasal fungal infections in immuno-compromised patients. Am J Med 1988; 85:619–23.

15. Peterson DE, Sonis ST, Oral complications ofcancer chemotherapy: present status and futurestudies. Cancer Treat Rep 1982; 66: 1251–6.

16. Elting LS, Bodey GP, Keefe BH, Septicemia andshock syndrome due to viridans streptococci: acase–control study of predisposing factors. ClinInfect Dis 1992; 14: 1201–7.

17. Schubert MM, Peterson DE, Lloid ME, Oral com-plications. In: Hematopoietic Cell Transplantation,2nd edn (Thomas ED, Blume KG, Forman SJ,eds). Oxford: Blackwell Science, 1999: 751–63.

18. Crumpacker CS, Cytomegalovirus. In: Principlesand Practice of Infectious Diseases, 5th edn.(Mandell GL, Bennett J, Dolin R, eds).Philadelphia: Churchill Livingstone, 2000:1586–9.

19. Rolston KVI, Dholakia N, Rodriguez S,Rubenstein EB, Nature and outcome of febrileepisodes in patients with pancreatic and hepato-biliary cancer. Supp Care Cancer 1995; 3: 414–17.

20. Thaler M, Pastakia B, Shawker TH et al, Hepaticcandidiasis in cancer patients: The evolving pic-ture of the syndrome. Ann Intern Med 1988; 108:88–100.

21. Powell RW, Moss JP, Nagar D et al,Strongyloidiasis in immunosuppressed hosts.Presentation as massive lower gastrointestinalbleeding. Arch Intern Med 1980; 140: 1061–3.

22. Gomez L, Martino R, Rolston KVI, Neutropenicenterocolitis: spectrum of the disease and com-parison of definite and possible cases. Clin InfectDis 1998; 27: 695–9.

23. Rolston KVI, Bodey GP, Diagnosis and manage-ment of perianal and perirectal infection in thegranulocytopenic patient. In: Current Clinical

Topics in Infectious Diseases, Vol 13 (RemingtonJS, Swartz MN, eds). Boston: Blackwell Scientific,1993: 164–71.


25. Keating MJ, Bodey GP, Valdivieso M, RodriguezV, A randomized comparative trial of threeaminoglycosides – comparison of continuousinfusions of gentamicin, amikacin and sisomicincombined with carbenicillin in the treatment ofinfections in neutropenic patients with malig-nancies. Medicine 1979; 58: 159–70.

26. Whimbey E, Bodey GP, Viral pneumonia in theimmunocompromised adult with neoplastic dis-ease: the role of common community respiratoryviruses. Semin Respir Infect 1992; 7: 122–31.

27. Wendt CH, Weisforf DJ, Jordan MC et al,Parainfluenza virus respiratory infection afterbone marrow transplantation. N Engl J Med 1992;326: 921–6.

28. Emanuel D, Cunningham I, Jules-Elysee K et al,Cytomegalovirus pneumonia after bone marrowtransplantation successfully treated with thecombination of ganciclovir and high-dose intra-venous immune globulin. Ann Intern Med 1988;109: 777–82.

29. Ampel NM, Wing EJ, Legionellosis in the com-promised host. In: Clinical Approach to Infection inthe Compromised Host, 2nd edn (Rubin R, YoungLS, eds). New York: Plenum, 1988: 305–19.

30. National Comprehensive Cancer Network(Febrile Neutropenia Guidelines PanelMembers), NCCN Practice Guidelines for Feverand Neutropenia. NCCN Proceedings. Oncology1999; 13(5A): 197–257.

31. Pizzo PA, Robichaud KJ, Gill FA, Witebsky FG,Empiric antibiotic and antifungal therapy forcancer patients with prolonged fever and granu-locytopenia. Am J Med 1982; 72: 101–11.

32. Pizzo PA, Robichaud KJ, Gill FA et al, Durationof empiric antibiotic therapy in granulocytopenicpatients with cancer. Am J Med 1979; 67: 194–200.

33. Pizzo PA, Infectious complications in the childwith cancer. I. Pathophysiology of the compro-mised host and the initial evaluation and man-agement of the febrile cancer patient. J Pediatr1981; 98: 341–54.

34. Talcott JA, Finberg R, Mayer RJ, Goldman L, The


medical course of cancer patients with fever andneutropenia. Clinical identification of a low-risksubgroup at presentation. Arch Intern Med 1988;148: 2561–8.

35. Talcott JA, Siegel RD, Finberg R, Goldman L,Risk assessment in cancer patients with feverand neutropenia: a prospective, two-center vali-dation of a prediction rule. J Clin Oncol 1992; 10:316–22.

36. Pizzo PA, Hathorn JW, Hiemenz J et al, A ran-domized trial comparing ceftazidime alone withcombination antibiotic therapy in cancer patientswith fever and neutropenia. N Engl J Med 1986;315: 552–8.


38. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low-risk cancer patients withfever and neutropenia: a pilot study of 30

patients based on a validated prediction rule. JClin Oncol 1994; 12: 107–14.

39. Rubenstein EB, Rolston KVI, Outpatient treat-ment of febrile neutropenic patients with cancer.Eur J Cancer 1995; 31A: 2–4.

40. Klastersky J, Paesmans M, Rubenstein EB et alfor the Study Section on Infections of theMultinational Association for Supportive Care inCancer, The Multinational Association forSupportive Care in Cancer Risk Index: a multi-national scoring system for identifying low-riskfebrile neutropenic cancer patients. J Clin Oncol2000; 18: 3038–51.

41. Justice AC, Covinsky KE, Berlin JA, Assessingthe generalizability of prognostic information.Ann Intern Med 1999; 130: 515–24.

42. Klastersky J, Chami J, Frankard J et al,Prospective validation study of the MASCC clin-ical prediction rule for identification of low-riskpatients with febrile neutropenia and for thera-peutic decision. Proc Am Soc Clin Oncol 2000; 19:440a (Abst 1726).


9Risk-adjusted management of the febrileneutropenic cancer patientEdward B Rubenstein, Kenneth VI Rolston

INTRODUCTION

Chapter 4 discussed the evolution of clinicalprediction rules or statistically derived modelsfor determining risk during the febrile neu-tropenic episode. Since the original paper byBodey et al1 describing the relationship betweenthe circulating neutrophil count and the inci-dence of infections in patients with acuteleukemia, clinicians have used clinical trialsmethodology to improve treatment for thesepatients, and have made observations aboutfactors that influence the risk of good or badoutcomes during the febrile episode. A compar-ison of the characteristics of the two methodsfor determining risk stratification is shown inTable 9.1. In this chapter, we shall discuss sev-eral of these factors and provide an overview ofthe clinical-trials-based methods for risk-adjusted therapy in patients with fever andneutropenia.

OUTCOME AS A FUNCTION OF THE INITIALAND CHANGING NEUTROPHIL COUNT

As early as 1966, the fatality rate due to severeinfection in patients with acute leukemia was

noted to be linked to the change in neutrophilcount during the first week of infection.1 Table9.2 clearly demonstrates this relationship. Thefatality rate was highest (80%) in patients whoinitially started with absolute neutrophil counts(ANC) � 100/mm3 that did not change duringthe first week of infection, compared withthose patients who started out with ANC� 1000/mm3 that then rose to 1000/mm3

(27%). Many clinical trials have reported thatresponse rates to antibiotic regimens arestrongly influenced by the trend in the neu-trophil count during the febrile episode.2–5 In arandomized trial of 520 evaluable febrile neu-tropenic episodes treated with carbenicillinplus gentamicin, amikacin or sisomicin, theoverall response rate was higher (85%) whenthe neutrophil count rose, compared with whenit remained stable or decreased (59%)(p � 0.001).2 Thirteen years later, using moremodern antibiotics, Rolston and colleagues5

confirmed the influence of a rising neutrophilcount on response rates in 750 febrile neu-tropenic episodes in 567 patients treated withimipenem or ceftazidime with or withoutamikacin. In this study, the overall responserate was 73% if the initial neutrophil count rose,compared with 43% if it decreased or remained

unchanged (p � 0.00001). The response rate inpatients who were initially profoundly neu-tropenic (ANC � 100/mm3), but recovered fromneutropenia was 67%, compared with only 32%in patients who remained profoundly neu-tropenic (p � 0.0001). Clearly, bone marrowrecovery is a very important factor that influ-ences outcome during the febrile neutropenicepisode. Unfortunately it cannot be predicted atthe time of initial evaluation and is of little usein helping to risk-stratify patients. Delayedbone marrow recovery might be anticipated incertain patient subsets (e.g. those who havereceived multiple cycles of myelosuppressivechemotherapy, those with known bone marrowmetastases, or those who have received radia-tion therapy to the pelvis, spine, or long bones).

HOW DOES DURATION OF NEUTROPENIAINFLUENCE RISK?

In 1988, Rubin et al6 published a study from theUS National Cancer Institute (NCI) examining

the influence of the duration of neutropenia onthe response to empiric antimicrobial therapyand other important clinical outcomes inpatients with fever of undetermined origin(FUO). Patients with less than 7 days of neu-tropenia had response rates to initial antimicro-bial therapy of 95%, compared with only 32% inpatients with more than 14 days of neutropenia(p � 0.001), whereas patients with intermediatedurations of neutropenia of between 7 and 14days had response rates of 79%. These out-comes are shown in Table 9.3. At greatest riskare patients with acute leukemia and recipientsof high-dose chemotherapy with stem cell orbone marrow transplantation, because theduration of severe neutropenia often exceeds 15days. In contrast, most patients with solidtumors have neutropenia lasting less than 7–10days and are at much lower risk. Unfortunately,using duration of neutropenia as a decision ruleat the onset of the febrile episode precludes itsuse in identification/management of the low-risk patient: determining the duration of neu-tropenia is not possible until the neutropenia


Table 9.1 Risk assessment in febrile neutropenia

Clinical prediction rule(s) Clinical trial(s)

Methods • Statistical (logistic regression) • Pilot study – feasibility• Training set (derivation) • Phase II – reproducibility• Validation set • Phase III – compare with standard of care

Endpoints • Serious medical complication(s) • Response rate(s)• Death • Adverse events

Advantages • Large population – ‘all comers’ • Clinicians familiar with methods• Representative of true population • Well-defined eligibility and methodology

Limitations • Still needs testing in clinical trials • Generalizibility of results• Prognostic factors derived from secondary

analyses

has resolved. Pizzo and colleagues7 refined thecriterion from ‘duration of neutropenia’ to‘expected duration of neutropenia’. Presently,clinicians cannot accurately predict theexpected duration of neutropenia at the onset ofthe febrile episode, although patients with solidtumors receiving conventional-dose chemother-apy are likely to have neutropenia lasting lessthan 7 days in the majority of cycles afterchemotherapy. Observational studies are beingconducted to add this important variable tocurrent risk models.

PROGNOSTIC FACTORS IN PATIENTS WITHBACTEREMIA

Elting and colleagues8 conducted an extensiveanalysis of 909 episodes of bacteremia in 799febrile neutropenic cancer patients based upon10 consecutive published randomized clinicaltrials of empiric antibiotic therapy3,5,9–16 con-ducted at the University of Texas MD AndersonCancer Center between 1980 and 1993. Thesetrials in adults were prospectively conducted

and had similar eligibility criteria. Five out-comes were reported for each episode: responseto the initial antibiotic regimen, ultimate out-come of the infection, time to defervescence,duration of antibiotic therapy, and survival. Inan initial analysis, bacteremic patients weresubclassified according to the presence andamount of tissue involvement (simple versuscomplex) at the onset of the infectious episode.Table 9.4 shows the classification scheme usedto distinguish simple from complex bac-teremias. Twenty-one prognostic factors werestudied, including demographic variables (ageand gender), underlying malignancy/treatment(cancer diagnosis and bone marrow transplan-tation), clinical features at presentation/duringtherapy (shock, initial levels of serum crea-tinine, albumin, bilirubin, and serum trans-aminases, and neutrophil count/trend),infection-related variables (associated site ofinfection, size of soft-tissue lesions, necrosis,organism, and susceptibility to antibiotics), andtreatment factors (initial antibiotic regimen,number of antibiotics, vancomycin use, andschedule of antibiotics).

RISK-ADJUSTED MANAGEMENT OF THE FEBRILE NEUTROPENIC CANCER PATIENT 169

Table 9.2 Fatality rate of severe infections related to change in granulocyte level during the firstweek of infectiona

Granulocyte level Episodes

Initial ANC/mm3 Change Total number Percentage fatal

�100 None 15 80

�1000 None or fall 44 59

�1000 Rise, but still �1000 15 40

�1000 Rise to 1000 26 27

1000 Rise 44 32

a Adapted from Bodey GP, Buckley M, Sathe YS, Freireich EJ, Quantitative relationships between circulating leukocytes andinfection in patients with acute leukemia. Ann Intern Med 1966; 64: 328–40.

Results of the logistic regression for thisanalysis are shown in Figure 9.1 and Table 9.5for the initial response (the point at which theinitial antibiotic regimen was discontinued ormodified) and the ultimate outcome for theepisode. Certain factors, such as age, leukemia,prior bone marrow transplant, hematologicmalignancy versus solid tumor, serum albumin,presence or absence of shock, and complex bac-teremia, are either known or highly suspectedearly after the onset of the febrile episode, andcan be used to stratify patients into high-risksubsets. The prognostic significance of complexinfection associated with bacteremia on sur-

vival is demonstrated in Figure 9.2. At 21 days,20% of patients with complex bacteremias weredead, compared with only 5% of patients withsimple bacteremias (p � 0.0001). Other keyfindings from this study included the observa-tion that the median time to defervescence forpatients with simple bacteremias was half thatobserved for patients with complex bacteremias(2.5 days versus 5.3 days; p � 0.0001). As notedpreviously, response rates were strongly influ-enced by the trend in the neutrophil count dur-ing the febrile episode. Patients with complexbacteremias who were profoundly neutropenic(ANC � 100/mm3) at the onset of their febrile


Table 9.3 Clinical outcomes of patients with fever of unknown origin according to duration ofneutropeniaa

Low-risk Moderate-risk High-risk neutropenia: neutropenia: neutropenia:�7 days 7–14 days 14 days (331 patients) (166 patients) (93 patients)

Time to defervescence (days):Median 2 4 5Range 1–7 4–14 1–30

Recurrent fever 2 (0.6%) 2 (0.6%) 35 (38%)

Success without 315 (95%) 131 (79%) 30 (32%)modification of initialempiric antibioticregimen

Success with 14 (4%) 32 (19%) 60 (65%)modification ofinitial regimen

Death 2 (1%) 3 (2%) 3 (3%)

a Adapted from Rubin M, Hathorn JW, Pizzo PA, Controversies in the management of febrile neutropenic cancer patients.Cancer Invest 1988; 6: 167–84.

episode and did not experience neutrophilrecovery had a 63% response rate compared to86% for those patients whose neutrophils roseto ANC 1000/mm3 (p � 0.04). Profoundlyneutropenic patients with simple bacteremiashad a much higher response rate to antibiotics(94% versus 70%, p � 0.0001) compared topatients with complex bacteremias.

Based upon these and other studies, clinicalcriteria can be used to stratify patients into high-, moderate-, and low-risk strata shortly after theonset of the febrile neutropenic episode. Theseclinical criteria are reviewed in Table 9.6. Thesecriteria in one combination or other have beenused to select patients for risk-adjusted clinicaltrials of antibiotic therapy as described in subse-quent sections of this chapter.

TREATMENT OF HIGH-RISK PATIENTS

There is uniform agreement that high-risk neu-tropenic patients (see Table 9.6) need to betreated using standard, hospital-based, par-

enteral, broad-spectrum, empiric antibiotictherapy for the entire febrile episode.17 The vari-ous benefits of this approach, and the manytreatment options, including combination ther-apy and monotherapy, are discussed in detailin Chapters 1, 6, 7, and 14. There is also generalagreement that many patients do not fall intothe high-risk category, and that althoughimprovements in the methods for identifyingsuch patients accurately at the onset of a febrileepisode still need to be made, alternative treat-ment strategies for moderate- and low-riskpatients might be associated with substantialadvantages and need to be explored.18

ORAL ANTIBIOTIC THERAPY – HISTORICALPERSPECTIVES

The feasibility of oral therapy administered inthe hospital was first demonstrated twodecades ago with the use of trimethoprim/sul-famethoxazole in febrile neutropenic patients.19

This experience was not limited to ‘low-risk’


Table 9.4 A classification scheme for bacteremic febrile neutropenic cancer patientsa

Concomitant site of infection

Simple bacteremia Complex bacteremia

• Bacteriuria • Major organ:• Otitis Lungs• Pharyngitis Liver/spleen• Soft tissue infection �5 cm Kidneys

ColonBones/jointsVeins/heart/meninges

• Soft tissue infection, wound, or cellulitis 5 cm• Soft tissue infection any size with necrosis

a Adapted from Elting LS, Rubenstein EB, Rolston KVI, Bodey GP. Outcomes of bacteremia in patients with cancer andneutropenia: observations from two decades of epidemiological and clinical trials. Clin Infect Dis 1997; 25: 247–59.


WBC recovery

Vancomycin � Gram-positive bacteremia

Two antibiotics

Complex bacteremia

Shock

Resistant organism

Leukemia

Albumin�3.5 g/dl

�-streptococcal bacteremia

Other Gram-positive organisms

Initial outcome

Complex bacteremia

Shock

Pseudomonas spp.

Clostridium spp.

Albumin�3.5 g/dl

ANC recovery

Ultimate outcome

Protective Increased risk of failure

0.1 0.2 0.5 1 2 5 10 20 40

0.1 0.2 0.5 1 2 5 10 20 40Odds ratio and 95% confidence limits

Figure 9.1 Results of logistic regression analysis of factors predicting the outcome of initial antibiotic therapyfor and the ultimate outcome of infection in neutropenic patients with cancer and bacteremia who were enrolledin 10 clinical trials.8 An odds ratio of greater than 1 indicates a greater risk of therapeutic failure; a value ofless than 1 suggests a protective effect.

patients, and, although moderately successful,the most significant drawback of this approachwas the lack of activity of trimethoprim/sul-famethoxazole against Pseudomonas aeruginosa,a frequent and aggressive pathogen in neu-tropenic patients. This was overcome with thedevelopment of the newer synthetic quinolones(ciprofloxacin and ofloxacin), and oral therapybecame a potential option even for patientswith severe neutropenia. Early experience withoral ciprofloxacin at the MD Anderson CancerCenter was associated with an overall response

rate of 85%.20 More recently Malik et al21 com-pared several standard parenteral regimens inuse at the Aga Khan University Hospital,Karachi, Pakistan (amikacin plus piperacillin,carbenicillin, or cloxacillin) with oral ofloxacin(400 mg bid) in a prospective randomized trialof hospitalized neutropenic patients with feverwho were able to tolerate oral therapy. Theresponse rates to the original regimen (53%)were identical for oral ofloxacin and parenteraltherapy. The overall response rate (includingresponse after modification of the original regi-

men) was 73% for patients treated with par-enteral therapy and 77% (not statisticallysignificant) for patients treated with oralofloxacin. These data were encouraging, sincethese studies were not limited to low-riskpatients, creating the expectation that betterresponse rates might be seen in low-risk patients.

THE EMERGENCE OF OUTPATIENT THERAPYFOR LOW-RISK PATIENTS

The original impetus for exploring outpatienttherapy was provided by the availability ofcomputerized small-volume infusion pumps,which were being used to deliver outpatientchemotherapy. These had been tailored for out-patient parenteral antibiotic therapy of variousinfections in the general medical population(e.g. endocarditis, osteomyelitis, and diabeticsoft-tissue infections). Furthermore, it was pos-tulated that outpatient therapy for febrile neu-tropenic patients, if proven to be safe andeffective, might improve the quality of life ofcancer patients and their caregivers/family by

limiting the amount of time patients spent inthe confines of the hospital. Some investigatorsalso theorized that the risk of nosocomial infec-tions and potential iatrogenic complicationsmight outweigh the benefits of inpatienttherapy for low-risk patients. Another com-pelling reason to pursue this line of clinicalinvestigation was the promise of reducing thecost of healthcare for this common complicationof chemotherapy, with the principal premisebeing that oral outpatient therapy would be theleast expensive method, if such therapy couldbe administered safely. In the late 1980s, thebasic assumption made by the majority of clini-cians caring for febrile neutropenic patients wasthat the hospital was the safest place to treatsuch patients, and that one must demonstratethat other settings are ‘as safe as’ the hospitalbefore they are accepted as part of the standardof care.

Several pieces of information have recentlycome to light indicating that this assumptionis probably erroneous. Data presented at the4th Decennial International Conferenceon Nosocomial and Healthcare-Associated


Days of survival after initiation of therapy

Perc

enta

ge o

f pa

tient

s w

ho s

urvi

ved

Simple

Complex

100

90

80

70

0 7 14 21 28 35 42

Figure 9.2 Durationof survival inneutropenic patientswith cancer andsimple bacteremiasand in neutropenicpatients with cancerand complexbacteremias whowere enrolled in 10clinical trials.8 Theobserved difference isstatistically significantat p � 0.001.

Infections (Atlanta, GA, March 2000) hasdemonstrated that each year, approximatelytwo million patients in the USA acquire infec-tions while hospitalized for other conditions.These infections account for 88 000 deaths andcost more than 4.6 billion dollars. Additionally,at least 70% of these healthcare-associatedinfections diagnosed in hospitals are caused bybacteria that are resistant to at least one antimi-crobial agent generally used for the treatmentof such infections, and an increasing proportionof hospital-acquired isolates are multidrug-resistant.22 These data also demonstrate thatalthough similar infections occur at other sites

of healthcare delivery, such as nursing homes,dialysis centers, outpatient clinics, and patients’homes, they are much less frequent in a home-care setting than in a hospital or long-term caresetting (1% versus 5%). Early discharge fromthe hospital to an outpatient clinic or homecaresetting might, therefore, substantially reducethe frequency of resistant healthcare-associatedinfections, particularly in patients who areotherwise at very low risk of developing com-plications that require hospital-based monitor-ing.

Another disturbing document is the reportentitled To Err is Human, produced by the


Table 9.5 Prognostic factors for bacteremia in neutropenic patients with cancer who were enrolled in10 clinical trialsa

Factor No. of Initial p value Ultimate p valueepisodes response response

rate (%) rate (%)

Complex bacteremiab 138 38 73Simple bacteremiab 771 74 �0.0001 94 �0.0001Shock 34 35 41No shock 875 70 �0.001 93 �0.001Resistant organism 62 50 83Susceptible organism 748 70 �0.001 92 �0.02Leukemia 516 63 89Other hematologic malignancies 152 74 �0.0002 94 �0.07Solid tumor 241 77 93BMTc 189 59 88No BMT 720 71 �0.002 92 �0.11Age �50 years 423 73 92Age �50 years 486 64 �0.006 90 �0.33Albumin level �3.5 g/dl 711 66 90Albumin level �3.5 g/dl 198 77 �0.005 96 �0.01

a Adapted from Elting LS, Rubenstein EB, Rolston KVI, Bodey GP, Outcomes of bacteremia in patients with cancer andneutropenia: Observations from two decades of epidemiological and clinical trials. Clin Infect Dis 1997; 25: 247–59.b See Table 9.4.c Bone marrow transplantation.

Institute of Medicine.23 This report focuses onseveral studies conducted across the USA, andpoints out that the frequency of adverse eventsin US hospitals ranged between 2.9% and 3.7%of hospitalizations, and that between 8.8% and13.6% of these events led to deaths.24,25

Furthermore, over half of these adverse eventsresulted from medical errors that could havebeen prevented. When extrapolated to the more

than 33.6 million admissions to US hospitals in1997, these studies imply that between 44 000and 98 000 Americans die each year as a resultof medical errors. Although medical errorsoccur in all healthcare settings, the report alsopoints out that four out of five such eventsoccur in the hospital, with the rest occurring inphysicians’ offices, other non-hospital settings,or patients’ homes. These data again suggest


Table 9.6 Clinical risk stratification and guidelines for risk-adjusted therapy of febrile neutropenicpatients

Risk group Patient characteristics Treatment options

High-risk Hematologic malignancy Traditional, empiric broad-Allogeneic bone marrow transplantation spectrum, parenteral antibioticsSevere and prolonged neutropenia (14 days) for duration of febrile episode.Significant comorbidity or poor performance status Consider colony-stimulatingPresentation with shock, hypoalbuminemia, or factorscomplex infectionSlow response to initial therapya

Moderate-risk Solid tumor → intensive chemotherapy → Initial, parenteral, in-hospitalautologous bone marrow or hematopoetic stem cell therapy, followed by earlytransplantation discharge on parenteral or oralModerate duration of neutropenia (7–14 days) regimenClinically/hemodynamically stable, with minimalcomorbidityEarly response to initial therapya

Low-risk Solid tumor Outpatient therapy (parenteral,Conventional chemotherapy sequential, or oral)Short duration of neutropenia (�7 days)No comorbidityFever of unknown origin or simpleinfectionClinically and hemodynamically stable

a Although these responses are typical in high/moderate-risk patients, they cannot be used on day 1 to assign a particularpatient to either risk group, since response to antibiotics is often not determined until 72–96 hours after the initiation ofantibiotic therapy.

that the hospital is not necessarily the safestplace to deliver healthcare, and provide furtherimpetus for the evaluation of non-hospital-based settings for healthcare delivery. Insist-ence upon clinical trials that comparehospital-based treatment with that delivered innon-hospital-based settings before acceptingthe latter as a standard of care seems counter-intuitive, since such trials would expose a sub-stantial number of low-risk patients to thehazards of hospitalization. Fortunately, asignificant amount of progress has been madein the development and evaluation of altern-ative strategies, including the route(s) of anti-biotic administration and the setting(s) in whichtherapy is delivered. These are discussedbelow.

EARLY DISCHARGE AFTER INITIALHOSPITALIZATION

Using their prediction model to select low-riskpatients, Talcott et al26 conducted a pilot studyevaluating early discharge on parenteral antibi-otics after an initial 48-hour hospitalizationperiod. Patients with significant infections(pneumonia, bacteremia, urinary tract infec-tion), and those aged 65 years or more wereexcluded, even if they were predicted to below-risk. The initial hospital-based regimensincluded mezlocillin plus gentamicin, ormonotherapy with ceftazidime. After 2 days ofthis therapy, stable patients were discharged toreceive the same parenteral regimen at home,with daily home follow-up by a nurse. Patientswere readmitted if fever persisted or if compli-cations arose. Of the 30 patients treated in thismanner, 16 (53%) responded to the original reg-imen. Five patients were readmitted for persis-tent fever, and four developed seriouscomplications such as renal failure, hypoten-sion, and bacterial and fungal superinfection.There was a documented improvement in thepatients’ quality of life and a 44% reduction indaily medical charges for patients receivinghome antibiotics. Despite these favorable out-

comes and the fact that there were no deathsduring this trial, the high rate of readmission(30%) and alteration of the initial regimenraised some doubts about the practical applica-tion of Talcott’s prediction model. Perhaps theinclusion of patients with acute leukemiaand/or persistent neutropenia of more than 7days (5 patients (17%) had neutropenia of 13–36days duration) accounts for the disappointingresults of this pilot study. A newer predictionmodel has been developed by the MultinationalAssociation for Supportive Care in Cancer(MASCC), which is an improvement overTalcott’s model and has a lower misclassifica-tion rate.27 In this model, one of the factors pre-dictive of low risk is the presence of a solidtumor. Models to better predict the duration ofsevere neutropenia are also being developed,and should further enhance our risk-assessmentcapabilities. Until these models have beendeveloped and validated, it might be prudentto exclude patients with hematologic malig-nancies from those considered low-risk. Mostpatients who are stable enough to be dis-charged will probably have their parenteralregimen changed to an oral one, and parenteralhome antibiotic therapy will probably be limited to stable patients who are unable to tol-erate oral therapy because of factors such asmucositis.

HOSPITAL-BASED ORAL ANTIBIOTICS FORLOW-RISK PATIENTS

Two recently published prospective random-ized trials compared oral antibiotic therapywith standard parenteral regimens in hospital-ized, low-risk, febrile neutropenic patients.28,29

In the trial conducted by the InternationalAntimicrobial Therapy Cooperative Group ofthe European Organization for Research andTreatment of Cancer, equivalence wasdemonstrated in the comparison of intravenousceftriaxone plus amikacin (84% success rate)and oral ciprofloxacin plus amoxicillin/clavu-lanate (86% success rate). The frequency of


adverse events including death (related orunrelated to infection) was also similar for bothregimens.28

In the trial from the NCI and allied institu-tions, the oral regimen of ciprofloxacin plusamoxicillin/clavulanate and the intravenousregimen of ceftazidime monotherapy were alsoassociated with similar success rates (71% and67% respectively). There was a higher rate ofintolerance of the oral regimen (16%), in com-parison with the intravenous regimen (1%).There were no deaths in either arm of thisstudy. Patients with hematologic malignancieswere eligible for both of these trials, andalthough the mean duration of neutropeniaafter the onset of infection was approximately 4days, several patients had prolonged neutrope-nia (14–18 days), which may have contributedto the failure of the initial regimen and to intol-erance of the oral regimen, and led to the modi-fication of therapy in some patients.

The equivalence of oral and parenteral ther-apy demonstrated in these and other smallertrials may have significant implications for themanagement of neutropenic patients withfever, particularly in countries with limitedresources. In the USA, and other countries withsimilar reimbursement and legal systems, hos-pital-based oral antibiotic therapy will probablynot enjoy widespread use. In reality, mostpatients who are able to tolerate oral therapyand are clinically stable can probably bedischarged from the hospital and treated asoutpatients. A small number may require hos-pitalization for medical reasons not related totheir febrile episode, or might live alone and beunable to adequately care for themselves. Suchpatients might benefit from hospital-based oralantibiotic therapy.

EARLY DISCHARGE ON ORAL THERAPY

The availability of expanded-spectrum oralquinolones has also enabled clinicians to switchfrom parenteral regimens in the hospital to oralregimens upon discharge, for moderate-risk

patients after an initial period of stabilization inthe hospital.18 This strategy is known as sequen-tial, early-switch, or stepdown therapy. Severalclinical trials have successfully demonstratedthe utility of this approach. In a multicenterrandomized trial, comparing ceftazidime plusamikacin with ciprofloxacin plus azlocillin forempiric therapy of febrile neutropenia, earlyconversion to orally administered ciprofloxacinwas compared with continuation of the par-enteral regimen in patients showing initialresponse.30 Conversion to orally administeredciprofloxacin was possible for 65% of eligiblestudy patients after a mean of 6 days of par-enteral therapy, resulting in cost savings andshortened hospital stays. In another trial con-ducted at the NCI, febrile neutropenic patientswho defervesced within 72 hours of receiving aparenteral regimen (imipenem or ceftazidime)were randomized either to continue parenteralantibiotics or complete therapy with oralciprofloxacin.31 Twenty-four of 27 evaluableepisodes (89%) in patients randomized toreceive parenteral therapy, and 22 of 29episodes (76%) in patients switched to oral ther-apy, were successfully managed without anyfurther changes in antibiotic therapy or read-mission to the hospital. A number of otherstudies, somewhat limited in size, suggest thatthere is a role for the strategy of switching fromparenteral to oral antibiotics, enabling earlydischarge in selected febrile neutropenicpatients.32,33 Although many low-risk patientscan be managed safely with initial outpatienttreatment, sequential therapy with early dis-charge might offer a more comfortable manage-ment plan for some, particularly those withnausea or mucositis, which might limit oralintake at the onset.

OUTPATIENT ORAL ANTIBIOTIC THERAPY

Several trials conducted over the past decadehave demonstrated the efficacy of outpatientantibiotic therapy for febrile neutropenicpatients. In a study conducted in Pakistan,


Malik and colleagues34 supplied oral ofloxacin(400 mg bid) for self-administration to low-riskneutropenic patients (patients with non-hema-tologic malignancies and an expected durationof neutropenia of less than 1 week) who wereunable to afford hospitalization or lived too faraway from the oncology center when theydeveloped their febrile episode. Of the 111febrile episodes treated in this manner, 92 (83%)responded without hospitalization, and theoverall response rate to antibiotic therapy was97%. In a subsequent study, these same investi-gators compared inpatient and outpatient ther-apy with oral ofloxacin in low-risk febrileneutropenic patients.35 Overall, 78% of inpa-tients and 77% of outpatients responded to theinitial regimens and required no modificationof therapy. The mortality rate was 2% amonginpatients and 4% among outpatients. Both ofthese studies demonstrated relatively highresponse rates, most probably because theywere limited to low-risk patients. However, amortality rate of 4% in the outpatient settingwas of concern, and the question of whetherhospitalization could have prevented any ofthese deaths was raised. Monotherapy withquinolones such as ofloxacin, as used in thesetwo studies, is not endorsed in guidelines pub-lished by either the Infectious Disease Society ofAmerica or the National ComprehensiveCancer Network.17,36

Two studies among adult cancer patients andone in the pediatric population, comparing out-patient oral and parenteral antibiotic regimensin low-risk febrile neutropenic patients, havebeen conducted at the MD Anderson CancerCenter. In the first adult study, patients wererandomized to either a parenteral regimen(aztreonam 2 g q8h, plus clindamycin 600 mgq8h) or an oral regimen (ciprofloxacin 750 mgq8h, plus clindamycin 600 mg q8h) uponbecoming febrile.37 In this trial, 83 episodeswere evaluated: 40 in patients receiving the oralregimen and 43 in patients receiving the par-enteral regimen. The parenteral regimen wasassociated with a response rate of 95%, com-pared with 88% for the oral regimen (p � 0.19),

with a combined response rate of 92% for ‘out-patient therapy’. There were no infection-related complications such as septic shock andno infection-related deaths in this trial. Renaltoxicity, however, was documented in 10% ofpatients randomized to the oral arm. This wasprobably the result of multiple factors, includingpatients’ age and state of hydration, the adminis-tration of other nephrotoxic drugs (cisplatin),and possibly the high dose of ciprofloxacin.38

Consequently, in the second adult trial, the oralarm was modified (ciprofloxacin 500 mg q8h,plus amoxicillin/clavulanate 500 mg q8h), withthe parenteral arm being the same as in the firststudy.39 Of the 179 episodes that were evaluable,91 received parenteral and 88 received oral ther-apy. The response rate for the parenteral regi-men was 87% and that for the oral regimen 90%.Neither regimen was associated with any majortoxicity, and no patients developed septic shockor died as a result of their infection.

In the pediatric study, low-risk patients (agerange 2–16 years) were randomized to receiveeither oral ciprofloxacin (12.5 mg/kg q12h) orparenteral ceftazidime (50 mg/kg q8h) in theoutpatient setting, after receiving one dose ofparenteral ceftazidime during the initial evalu-ation/randomization period.40 Out of the 73episodes, 63 (86%) were successfully managedon an outpatient basis. The response rate for theparenteral arm (31 of 33 episodes) was slightlybetter than that for the oral arm (32 of 40episodes), but this difference was not statisticallysignificant. Of the 10 patients who were hospital-ized, 4 had prolonged fever and 3 developedemesis. Protracted neutropenia was linked withthe need for hospitalization. There were nodeaths, intensive care unit admissions or trans-fers, or other serious complications during thistrial.

These, and other smaller studies (Table 9.7),provide evidence that with careful patientselection, appropriate antimicrobial therapy,and adequate monitoring of patients, outpatienttherapy (both parenteral and oral) for low-riskfebrile neutropenic patients is safe and effect-ive.


RISK-ADJUSTED MANAGEMENT: GENERALISSUES

The above discussion has summarized much ofthe progress that has occurred regarding risk-adjusted management of febrile neutropenic

patients. This progress is the result of (a) anincreased understanding of ‘febrile neutrope-nia’, (b) technological advance in vascularaccess, infusion therapy, and outpatient moni-toring, (c) the availability of newer, morepotent, oral antimicrobial agents, and (d) the


Table 9.7 Selected trials of outpatient therapy in low-risk, febrile neutropenic patients

Authors Patient population and Treatment regimen Response rate to

nature of study initial regimen (%)

Malik et al34 Non-randomized trial; ofloxacin 83

111 episodes; all adults 400 mg p.o., bid

Malik et al35 Randomized trial of Inpatient: ofloxacin Inpatient: 78

inpatient versus outpatient 400 mg p.o., bid

oral therapy; 169 Outpatient: ofloxacin outpatient: 77

episodes; all adults 400 mg p.o., bid

Rubenstein et al37 Randomized trial of i.v. aztreonam 2 g q8h i.v.: 95

outpatient parenteral and plus

oral regimens; 83 i.v. clindamycin 600 mg q8h

episodes; all adults versus

p.o. ciprofloxacin 750 mg q8H p.o.: 88

plus

p.o. clindamycin 600 mg q8H

Rolston and Rubenstein39 Randomized trial of i.v. aztreonam 2 g q8h i.v.: 87

outpatient parenteral and plus

oral regimens; 179 i.v. clindamycin 600 mg q8h

episodes; all adults versus

p.o. ciprofloxacin 500 mg q8h p.o.: 90

plus

p.o. amoxicillin/clavulanate 500 mg q8h i.v.: 94

Mullen et al40 Randomized trial of i.v. ceftazidime 50 mg/kg q8h

parenteral and oral versus

regimens; 75 episodes; p.o. ciprofloxacin 12.5 mg/kg q12h p.o.: 80

all pediatric

current climate in the healthcare industry,which has provided much of the impetus forevaluating non-traditional methods and sites ofdelivery of care. It must be remembered thatmost of the trials and strategies discussedabove have been conducted and developed inlarge tertiary-care hospitals or comprehensivecancer centers, with a particular interest andsubstantial experience in caring for cancerpatients. The ability to create and maintain aninfrastructure that can handle these manage-ment strategies and the logistics involved,sometimes in a large number of patients, is crit-ical to the success of such programs. This infra-structure includes individuals from variousdisciplines – physicians, nurses, pharmacists,vascular access and infusion therapy teams,home healthcare personnel, and the patientsand their caregivers, all acting in concert toensure that the best possible care is delivered asefficiently and safely as possible. Some institu-tions caring for cancer patients may not havethe ability to create or maintain such an infra-structure, nor a group of healthcare providerswith sufficient interest or expertise to providesuch care. In such institutions, standard hos-pital-based care should continue to be pro-vided, until they acquire the ability to sustain arisk-adjusted therapeutic program.

Other issues that are vital to the success ofrisk-adjusted therapy (outpatient therapy in

particular) are listed in Table 9.8. It is importantto select appropriate empiric regimens, notmerely convenient ones, taking into considera-tion local microbiology and susceptibility/resis-tance patterns. At the MD Anderson CancerCenter, infections caused by Gram-negativebacilli, including Ps. aeruginosa, are docu-mented, even in low-risk patients.37,39 Usingonce-a-day ceftriaxone (a regimen that has beenused for empiric therapy in febrile neutropenicpatients, and is convenient) in such a settingwould not be appropriate as the initial empiricregimen.41 Careful patient selection, taking intoconsideration not only clinical/statistical risk-assessment criteria, but also issues such aspatient comfort and compliance, the availabilityof a caregiver at home, the availability of reli-able transportation (automobile) and communi-cation (telephone), and relative distance of thepatients residence from the hospital (we chose a30-mile radius, based on our local traffic pat-terns), are all important, and have an impact onthe overall success of these new strategies.Frequent monitoring of patients for response,lack of response, the development of medicalcomplications, toxicity, and to ensure com-pliance is also critical, and cannot be over-stressed. All these issues need to be worked outin advance in order to ensure a successful out-patient treatment program.


Table 9.8 Requirements for a successful program of risk-adjusted therapy

• Institutional support for an adequate infrastructure• Dedicated team of healthcare providers• Local epidemiologic/susceptibility-resistance data• Selection of appropriate (not merely convenient) antimicrobial regimens• Motivated, compliant patients and family or other support personnel• Adequate transportation and communication• Adequate monitoring of non-hospitalized patients• Access 24 hours a day to management team and ambulatory care facility (Emergency Department)

ADVANTAGES AND DISADVANTAGES OFRISK-ADJUSTED THERAPY

Standard, hospital-based therapy of the febrileneutropenic patient has been an extremely suc-cessful strategy. It has had a significant andpositive impact on the overall survival of cancerpatients, particularly on those who are proneto developing medical complications. Thisapproach, however, is expensive, consumesvaluable resources, and is not necessary or evenbeneficial for all febrile neutropenic patients.Risk-adjusted therapy is the new wave, and ifimplemented successfully, is associated withsubstantial benefits. These are outlined in Table9.9. Several studies have demonstrated thepositive economic benefits of early dischargeand/or outpatient oral antibiotic therapy, com-pared to hospital based parenteral therapy.26,37

There is also compelling data that hospitaliza-tion significantly increases the risk of acquiringinfections with multidrug-resistant pathogens,and exposes patients to a number of other haz-ards of hospitalization, which can probably beavoided by early discharge and/or outpatientmanagement.22,23 Studies have also documentedsignificant improvements in the quality of lifeof patients receiving risk-adjusted therapy, andincreased convenience for their caregivers aswell.37,39 This aspect does not get much press intoday’s financially-focused health care environ-ment where economics dictate many decisions.It is, however, an extremely importantconsideration in the care of cancer patients, andcannot be ignored any longer. We have hadpatients who have begged to be treated on ouroutpatient protocols/pathways, so that theymay be able to eat home-cooked meals, sleep intheir own beds, and smell the proverbial roses.Even high-risk, terminally ill, and dyingpatients have expressed the desire to spendtheir last few days at home. We as cliniciansentrusted with the overall wellbeing of thesepatients (including those who are terminally ill)need to respond to their wishes and needs.Risk-adjusted management is a step in the rightdirection.


Table 9.9 Advantages and disadvantages ofrisk-based therapy

Advantages• Avoidance of iatrogenic and other hazards of

hospitalization• Reduced rate of ‘healthcare associated’

infections• Lower cost of care• Enhanced quality of life (patients)• Increased convenience (family)• More efficient resource utilization

Disadvantages• Potential for serious complications in an

unsupervised setting• Potential for non-compliance• Need to maintain an (expensive?)

infrastructure

There are some potential disadvantages ofrisk-adjusted therapy. Low-risk does not mean‘no risk’, and the potential for developing com-plications such as septic shock or severe hemor-rhage in a relatively unsupervised environmentdoes exist. Careful patient selection and closemonitoring of patients generally prevents thedevelopment of such events, or enables one tomanage them promptly, should they occur.Some patients might be non-compliant, particu-larly those on oral therapy. Maintaining ade-quate venous access may occasionally become aproblem that can be difficult to address in anambulatory or home setting. Finally, patientsmight develop a false sense of security regard-ing their febrile episode, since it did not requirehopsitalization or parenteral therapy, andmight ignore early signs and symptoms of pro-gressive infection or other complicationsbecause of a perceived trivialization of theirillness. It is imperative that patients be givenspecific instructions regarding follow-up

monitoring, and be told to seek immediatemedical attention at the earliest sign of compli-cations. It has been our experience from morethan a decade of administering risk-adjustedtherapy that cancer patients are generally verycompliant, follow instructions meticulously,and do ‘whatever it takes’ to stay out of hos-pital. Consequently, we have seldom encoun-tered the problems listed above.

FUTURE CONSIDERATIONS

In recent years, much attention has beenfocused on devising risk-assessment and treat-ment strategies for low-risk patients.18 Futureconsiderations for this subset of patientsinclude further refinements to increase theaccuracy of risk-assessment models, andimprovements in therapeutic regimens asnewer antimicrobial agents become available.Currently, most moderate- and high-riskpatients are managed with standard, hospital-based therapy. It is in this group of patients thatnewer evaluation and management strategiesneed to be developed in order to fully embracethe concept of risk-adjusted management of thefebrile neutropenic patient.

EVALUATION AND MANAGEMENT OFMODERATE- AND HIGH-RISK PATIENTS

Our current systems for stratifying risk are notperfect, and misclassifications do occur.27,42,43

High-risk patients, as we currently identifythem, are a very heterogeneous population. Inthe MASCC Risk-Index derivation set, of the205 patients identified initially as high-risk, 79(39%) experienced a serious medical complica-tion including 29 deaths (14%). This means thatthe majority – 126 patients (61%) – initiallyidentified as being at high risk had resolution oftheir febrile episode without any problems.Errors of sensitivity lead to misclassification oflow-risk patients as being high-risk, therebyexposing them to the hazards of hospitalization,

including iatrogenic complications and nosoco-mial infections with resistant microorgan-isms.22,23 Errors of specificity, on the other hand,result in a more dangerous misclassification,namely high-risk patients being labelled as low-risk. Such patients could initially receive theircare in the outpatient setting and potentiallydevelop serious complications without ade-quate monitoring and supervision. These errorsin classification need to be minimized and moreaccurate risk assessment strategies need to bedeveloped. Meanwhile, Elting and colleagueshave suggested using a new outcome – the timeto clinical response – to facilitate implementa-tion of early-discharge strategies, therebyimproving safety and quality of care for hospi-talized patients initially classified as high-risk.

TIME TO CLINICAL RESPONSE

A traditional endpoint when comparing treat-ment regimens has been the overall response orsuccess rate at the end of therapy. Elting andcolleagues44 developed a working definition oftime to clinical response using pooled datafrom six prospective clinical trials of imipenem-based or ceftazidime based regimens conductedat the MD Anderson Cancer Center. The sensi-tivity, specificity, and predictive values ofobjective (temperature response) and subjectivevalues (self-report of improvement) were com-pared, as were time to clinical response, days ofhospitalization, and cost. An early-dischargestrategy based upon the time to clinicalresponse was generated and retrospectivelyapplied to 488 episodes of fever in 466 patients.The characteristics of these comparisons areshown in Table 9.10. A combination of defer-vescence for 24 hours and patient-reportedsubjective improvement resulted in a sensitivityof 93% and a specificity of 76%, with no deathsoccurring if an early-discharge strategy hadbeen implemented using this model. Theimipenem-based regimens had a quicker timeto response (5 days versus 7 days) when com-pared with the ceftazidime-based regimens


Tabl

e 9.1

0D

evel

opm

ent

of a

tim

e to

res

pons

e de

finit

ion:

pre

dict

ive

valu

es o

f su

bjec

tive

and

obj

ecti

ve c

linic

al c

rite

riaa

Clin

ical

Crite

rion

to

Pre

dict

True

-pos

itiv

eTr

ue-n

egat

ive

Fals

e-po

siti

veFa

lse-

nega

tive

Sel

ecte

d fo

rR

espo

nse

to d

isch

arge

Dea

ths

resp

onse

bra

tec

rate

dra

tee

(%)

rate

f(%

)st

ep-d

own

regi

men

(po

siti

ve

(sen

siti

vity

) (%

)(s

peci

ficit

y) (

%)

care

pred

icti

ve v

alue

)

One

tem

pera

ture

�37.8

°C9

98

28

21

474 (97%

)339 (72%

)29 (6%

)

One

tem

pera

ture

�37.8

°C p

lus

96

68

13

12

375 (77%

)328 (87%

)9 (2%

)

subj

ectiv

e re

spon

se

Thre

e co

nsec

utiv

e te

mpe

ratu

res

97

34

23

17

429 (88%

)332 (77%

)29 (7%

)

�37.8

°CTh

ree

cons

ecut

ive

tem

pera

ture

s94

72

11

15

364 (75%

)323 (89%

)9 (2%

)

�37.8

°C p

lus

subj

ectiv

e

resp

onse

24 h

ours

tem

pera

ture

�37.8

°C9

64

819

16

405 (83%

)329 (81%

)18 (4%

)

24 h

ours

tem

pera

ture

�37.8

°C9

37

610

16

356 (73%

)321 (90%

)0 (0%

)

plus

sub

ject

ive

resp

onse

Sub

ject

ive

resp

onse

onl

y96

65

13

12

380 (78%

)329 (87%

)9 (2%

)

aAd

apte

d fr

om E

lting

LS

, R

uben

stei

n EB

, R

olst

on K

VI e

t al

. Ti

me

to c

linic

al r

espo

nse:

an

outc

ome

of a

ntib

iotic

the

rapy

of fe

brile

neu

trop

enia

with

impl

icat

ions

for

qual

ity a

nd c

ost

of c

are.

J C

lin O

ncol

2000; 1

8:3699–7

06.

bPr

edic

tive

valu

e fo

r re

spon

se t

o th

e an

tibio

tic r

egim

en b

eing

rec

eive

d w

hen

the

clin

ical

crit

erio

n w

as m

et.

cPe

rcen

tage

of tr

ue a

ntib

iotic

res

pons

es p

rope

rly c

lass

ified

as

resp

onse

s.dPe

rcen

tage

of tr

ue a

ntib

iotic

fai

lure

s pr

oper

ly c

lass

ified

as

failu

res.

ePe

rcen

tage

of pa

tient

s w

ho w

ould

be

sele

cted

for

ste

p-do

wn

care

and

wou

ld fai

l.fPe

rcen

tage

of pa

tient

s w

ho w

ould

not

be

sele

cted

for

ste

p-do

wn

care

but

wou

ld n

ever

thel

ess

resp

ond

to t

hera

py.

(p � 0.003), providing another criterion forcomparing regimens that might have been con-sidered equivalent if only the endpoint of over-all response rate had been used. Ifprospectively validated, this outcome measurecould be used to identify appropriate regimensand implement early-discharge strategies,resulting in the advantages outlined in Table9.9.

USE OF HEMATOPOIETIC GROWTH FACTORS

The hematopoietic growth factors G-CSF andGM-CSF (granulocyte and granulocyte–macrophage colony-stimulating factors) shortenthe duration of neutropenia, and their role inthe primary or secondary prevention of fever inneutropenic patients has been reasonably wellclarified.45 Their role as adjuncts to antimicro-bial therapy in neutropenic patients with feverhas been difficult to define (see Chapter 13).Studies evaluating the therapeutic role of G-CSF and GM-CSF have been hampered by thelack of risk stratification.16,46–52 This has led totheir use principally in patients with shock,pneumonia, and other complex infections, andtheir benefit has been difficult to demonstrate inthis poor-prognosis group. A recent subsetanalysis from the MASCC Risk-Index studydemonstrated higher response rates to the ini-tial antibiotic regimen (86% versus 72%;p � 0.016), and fewer complications, includingdeath (4.2% versus 13.4%; p � 0.031) in neu-tropenic solid tumor patients receiving prophy-lactic hematopoietic growth factors, in whomthese factors were continued after the develop-ment of fever.53 This benefit was most markedin patients with sarcoma, a subgroup that has ahigh proportion of low-risk neutropenicpatients, and least apparent in patients withlung cancer, a subgroup with a low proportionof high-risk patients.37,39 Factors that accountedfor complications in the logistic regressionmodel included: an interaction betweenANC � 100/mm3 at onset of fever and

neutropenia and failure to respond to the initialregimen (p � 0.001), severe burden of illness ormoribund appearance at time of presentation(p � 0.001), uncontrolled cancer (p � 0.001), ageover 60 years, (p � 0.004), and presence of acomplex infection (p � 0.010). In this regressionmodel, prophylactic use of CSFs appeared tobe protective (p � 0.095). In contrast, Garcia-Carbonero and colleagues54 comparedantibiotics with or without G-CSF in ‘high-risk’solid tumor patients with fever and neutrope-nia, using days of hospitalization as their pri-mary endpoint. The criteria used to determine‘high-risk’, which failed to account for theimpact of microbiologically documented infec-tions on the time to defervescence, and therequirement that patients remain hospitalizeduntil afebrile for 2 days were significant limita-tions of this study. Garcia-Carbonero et al con-cluded that patients receiving G-CSF had ashorter duration of subsequent neutropenia,antibiotic therapy, and hospital stay. Using theendpoint of time to clinical response (adjustedfor microbiologically documented infections)might have produced different (more accurate)results. Determining the optimal role of thehematopoietic growth factors, particularly inhigh-risk patients, is a complex task and willrequire studies using prospective observationalcohort designs, and risk-stratified clinical trialsusing validated risk-assessment criteria.

REFERENCES


2. Keating MJ, Bodey GP, Valdivieso M, RodriguezV, A randomized comparative trial of threeaminoglycosides – comparison of continuousinfusions of gentamicin, amikacin, and sisomicincombined with carbenicillin in the treatment ofinfections in neutropenic patients with malig-nancies. Medicine 1979; 58: 159–70.

3. Fainstein V, Bodey GP, Elting L et al, A random-ized study of ceftazidime compared to cef-


tazidime and tobramycin for the treatment ofinfections in cancer patients. J AntimicrobChemother 1983; 12(Suppl A): 101–10.

4. Winston DJ, Ho WG, Bruckner DA et al,Controlled trials of double beta-lactam therapywith cefoperazone plus piperacillin in febrilegranulocytopenic patients. Am J Med 1988;85(Suppl 1A): 21–30.

5. Rolston KVI, Berkey P, Bodey GP et al, A com-parison of imipenem to ceftazidime with orwithout amikacin as empiric therapy in febrileneutropenic patients. Arch Intern Med 1992; 152:283–91.

6. Rubin M, Hathorn JW, Pizzo PA, Controversiesin the management of febrile neutropenic cancerpatients. Cancer Invest 1988; 6: 167–84.

7. Pizzo PA, Infectious complications in the childwith cancer. I. Pathophysiology of the compro-mised host and the initial evaluation and man-agement of the febrile cancer patient. J Pediatr1981; 98: 341–54.


9. Fainstein V, Bodey GP, Bolivar R et al,Moxalactam plus ticarcillin or tobramycin fortreatment of febrile episodes in neutropeniccancer patients. Arch Intern Med 1984; 144:1766–70.

10. Jones PG, Rolston KVI, Fainstein V et al,Aztreonam therapy in neutropenic patients withcancer. Am J Med 1986; 81: 243–48.

11. Anaissie EJ, Fainstein V, Bodey GP et al,Randomized trial of beta-lactam regimens infebrile neutropenic cancer patients. Am J Med1988; 84: 581–9.

12. Bodey GP, Fainstein V, Elting LS et al, β-lactamregimens for the febrile neutropenic patient.Cancer 1990; 65: 9–16.

13. Bodey G, Reuben A, Elting L et al, Comparisonof two schedules of cefoperazone plus aztre-onam in the treatment of neutropenic patientswith fever. Eur J Clin Microbiol Infect Dis 1991; 10:551–8.

14. Bodey GP, Elting LS, Narro J et al, An open trialof cefoperazone plus sulbactam for the treatmentof fever in cancer patients. J AntimicrobChemother 1993; 32: 141–52.

15. Raad II, Whimbey II, Rolston KVI et al, A com-

parison of aztreonam plus vancomycin andimipenem plus vancomycin as initial therapy forfebrile neutropenic cancer patients. Cancer 1996;77: 1386–94.

16. Anaissie EJ, Vartivarian S, Bodey GP et al,Randomized comparison between antibioticsalone and antibiotics plus granulocyte–macrophage colony stimulating factor (Escherichiacoli-derived) in cancer patients with fever andneutropenia. Am J Med 1996; 100: 17–23.

17. Hughes WT, Armstrong D, Bodey GP et al, Fromthe Infectious Diseases Society of America: 1997guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

18. Rolston KVI, New trends in patient manage-ment: risk-based therapy for febrile patients withneutropenia. Clin Infect Dis 1999; 29: 515–21.

19. Bodey GP, Grose WE, Keating MJ, Use oftrimethoprim/sulfamethoxazole for treatment ofinfections in patients with cancer. Rev Infect Dis1982; 4: 579–85.

20. Haron E, Rolston KVI, Cunningham C et al, Oralciprofloxacin therapy for infection in cancerpatients. J Antimicrob Chemother 1989; 24: 955–62.

21. Malik IA, Abbas Z, Karim M, Randomized com-parison of oral ofloxacin alone with combinationof parenteral antibiotics in neutropenic febrilepatients. Lancet 1992; 339: 1092–6.

22. Gerberding JL, Preventing Antimicrobial-ResistantHealthcare Infections: Beyond 2000. ClinicalUpdates in Infectious Diseases, NationalFoundation for Infectious Diseases, Vol V, Issue2, August 2000.

23. Kohn L, Corrigan J, Donaldson M (eds),Committee on Quality of Health Care inAmerica, To Err is Human: Building a Safer HealthSystem. Institute of Medicine Report. Washington,DC: National Academy Press, 2000.

24. Brennan TA, Leape LL, Laird NM et al,Incidence of adverse events and negligence inhospitalized patients: results of the HarvardMedical Practice Study I. N Engl J Med 1991; 324:370–6.

25. Hospital Statistics. Chicago: American HospitalAssociation, 1999.

26. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low risk cancer patients withfever and neutropenia: a pilot study of 30patients based on a validated prediction rule. JClin Oncol 1994; 12: 107–14.


27. Klastersky J, Paesmans M, Rubenstein EB et alfor the Study Section on Infections ofMultinational Association for Supportive Care inCancer, The MASCC Risk-Index: a multinationalscoring system to predict low-risk febrile neu-tropenic cancer patients. J Clin Oncol 2000; 18:3038–51.

28. Kern WV, Cometta A, DeBock R et al, Oral ver-sus intravenous empirical antimicrobial therapyfor fever in patients with granulocytopenia whoare receiving cancer chemotherapy. InternationalAntimicrobial Therapy Cooperative Group ofthe European Organization for Research andTreatment of Cancer. N Engl J Med 1999; 341:312–18.


30. Flaherty JP, Waitley D, Eldin B et al, Multicenterrandomized trial of ciprofloxacin plus Azlocillinversus ceftazidime plus amikacin for empirictreatment of febrile neutropenic patients. Am JMed 1989; 87(Suppl 5A): 278S–82S.

31. Rolston KVI, Rubenstein EB, Freifeld A, Earlyempiric antibiotic therapy for febrile neutropenicpatients at low risk. Infect Dis Clin North Am1996; 10: 223–37.

32. Lav RC, King SM, Richardson SE, Early dis-charge of pediatric febrile neutropenic cancerpatients by substitution of oral for intravenousantibiotics. Pediatr Hematol Oncol 1994; 11:417–21.

33. Bash RO, Katz JA, Cash JV et al, Safety andcost–effectiveness of early hospital discharge oflower risk children with cancer admitted forfever and neutropenia. Cancer 1994; 19: 522–7.

34. Malik IA, Khan WA, Aziz A et al, Self-adminis-tered antibiotic therapy for chemotherapy-induced low-risk febrile neutropenia in patientswith non-hematologic neoplasms. Clin Infect Dis1994; 19: 522–7.

35. Malik IA, Khan WA, Karim M et al, Feasibility ofoutpatient management of fever in cancerpatients with low-risk neutropenia: results of aprospective randomized trial. Am J Med 1995; 98:224–31.

36. National Comprehensive Cancer Network(Febrile Neutropenia Guidelines PanelMembers), NCCN Practice Guidelines for Fever

and Neutropenia. NCCN Proceedings. Oncology1999; 13(5A): 197–257.


38. Rolston KVI, Rubenstein EB, Ciprofloxacinnephrotoxicity. Arch Intern Med 1993; 153:2705–6.

39. Rolston KVI, Rubenstein EB, Elting LS et al,Ambulatory management of febrile episodes inlow risk neutropenic patients. In: Program andabstracts of the 35th Interscience Conference onAntimicrobial Agents and Chemotherapy, SanFrancisco. Washington, DC: American Society forMicrobiology, 1995: Abst 2235.

40. Mullen CA, Petropoulos D, Robert WM et al,Outpatient treatment of fever and neutropeniafor low risk pediatric cancer patients. Cancer1999; 86: 126–34.

41. Rolston KVI, Tarrand JJ, Pseudomonas aeruginosa– still a frequent pathogen in patients withcancer: 11 year experience from a comprehensivecancer center. Clin Infect Dis 1999; 29: 463–4.

42. Talcott JA, Finberg R, Mayer RJ, Goldman L, Themedical course of cancer patients with fever andneutropenia. Clinical identification of a low-risksubgroup at presentation. Arch Intern Med 1988;148: 2561–8.



45. Ozer H, Armitage JO, Bennett CL et al, 2000update of recommendations for the use ofhematopoietic colony-stimulating factors: evid-ence-based clinical practice guidelines. J ClinOncol 2000; 18: 3558–85.

46. Maher DW, Lieschke GJ, Green M et al,Filgrastim in patients with chemotherapy-induced febrile neutropenia: a double-blind,placebo-controlled trial. Ann Intern Med 1994;121: 492–501.

47. Mayordomo JI, Rivera F, Diaz-Puente MT et al,Improving treatment of chemotherapy-induced


neutropenic fever by administration of colony-stimulating factors. J Natl Cancer Inst 1995; 87:803–8.

48. Riikonen P, Saarinen UM, Makipernaa A et al,Recombinant human granulocyte–macrophagecolony-stimulating factor in the treatment offebrile neutropenia: a double-blind placebo-con-trolled study in children. Pediatr Infect Dis J 1994;13: 197–202.

49. Biesma B, de Vries EG, Willemse PH et al,Efficacy and tolerability of recombinant humangranulocyte–macrophage colony-stimulatingfactor in patients with chemotherapy–relatedleukopenia and fever. Eur J Cancer 1990; 26:932–6.

50. Vellenga E, Uyll-de Groot CA, de Wit R et al,Randomized placebo-controlled trial of granulo-cyte–macrophage colony-stimulating factor inpatients with chemotherapy-related febrile neu-tropenia. J Clin Oncol 1996; 14: 619–27.

51. Ravaud A, Chevreau C, Bonichon F et al,

Granulocyte–macrophage colony-stimulatingfactor (GM-CSF) in patients with neutropenicfever is potent after low-risk but not after high-risk neutropenic chemotherapy regimens: resultsof a randomized phase III trial. J Clin Oncol 1998;16: 2930–6.

52. Mitchell PLR, Morland BJ, Dick G et al, Clinicalbenefits and cost savings of interventional G-CSFtherapy in patients with febrile neutropeniafollowing chemotherapy. Blood 1995; 86(Suppl 1):500a.

53. Kim YJ, Rubenstein EB, Rolston KVI et al,Colony stimulating factors may reduce compli-cations and death in solid tumor patients withfever and neutropenia. Proc Am Soc Clin Oncol2000; 19: 612a (Abst 2411).

54. García-Carbonero R, Mayordomo JI, TornamiraMV et al, Granulocyte colony-stimulating factorin the treatment of high-risk febrile neutropenia:a multicenter randomized trial. J Natl Cancer Inst2001; 93: 31–8.


10Evaluation and management of vascularaccess device infections in febrileneutropenic cancer patientsClaude Afif, Issam Raad

EPIDEMIOLOGY OF VASCULAR ACCESSDEVICE INFECTIONS

Vascular access devices (VADs) are indispens-able tools in the management of cancerpatients.1 It is estimated that more than five mil-lion central venous catheters (CVCs) are used inthe USA every year.2 Of those, one million long-term CVCs (with duration of placement ofgreater than 30 days) are inserted in cancerpatients. Four types of long-term CVCs areused in cancer patients: tunneled CVCs(Hickman/Broviac/Groshong), and implantableports, non-tunneled subclavian CVC’s, andperipherally inserted central catheters (PICCs).

In contrast to the many advantages that theyoffer (namely, easy vascular access, andincreased ability to administer large volumes offluids, medications, blood products, and par-enteral nutrition), VADs are hampered by theoccurrence of serious complications, includinglocal site infection, thrombophlebitis (septic andnon-septic), endocarditis, and catheter-relatedbloodstream infection (CRBSI). Short-termperipherally inserted devices such as peripheralvenous catheters and peripheral arterialcatheters, as well as midline catheters, are asso-

ciated with lower risk of infections comparedwith CVCs. However, CVCs, including PICCsand short- and long-term vascular devices, aremore often associated with CRBSI.

A review by Press et al3 of 17 studies andanother review by Decker and Edwards4 of 21studies involving cancer patients reported theincidence of long-term tunneled CVC infectionto be approximately 1.4 per 1000 catheter-days.Howell et al5 included 26 studies of 3948 tun-neled catheters in 3478 adult cancer patients,and reported a CVC infection rate of 1.9 per1000 catheter-days. The rate of non-tunneledCVC-related infection (including PICC lines)was 1.4 per 1000 catheter-days.6 The directimplications of such infectious complicationsare an increased mortality and morbidity, withan extension of the hospital stay.7

PATHOGENESIS

Source of infection

Within 24 hours post-insertion, most VADs willbe colonized with microorganisms (Figure10.1). Electron microscopy studies have shown

that all VADs are colonized – even those withnegative cultures.8 Microorganisms causingVAD infection can originate from four potentialsources: the skin insertion site, the hub,hematogenous seeding from a distant focus,and infusate contamination.

The external surface colonization of a short-term indwelling vascular device usually occurswithin the first 10 days following its insertion,and results from the migration of skin floraalong the intracutaneous segment to reach thedistal intravascular segment, resulting inCRBSI.8

In contrast, luminal colonization occurs inlong-term devices (after 30 days of insertion),and is the result of hub contamination follow-ing frequent manipulation, which leads tomicrobial migration along the internal surfaceof the catheter.8,9

Hematogenous seeding is an unusual sourcefor VAD colonization. However, many infec-tions by Candida spp. are thought to originate inthe gastrointestinal tract, with secondary can-didemia leading to VAD colonization.10 Finally,parenteral nutrition solutions or lipid emul-sions promote the growth of bacteria andfungi,8 such as Candida parapsilosis11 andMalassezia furfur,10 resulting in CRBSI.

Cofactors for adhesions of organisms

Microorganisms can be found on VAD surfacesin a sessile form embedded in a biofilm12 or in afree-floating, disseminated form over the VADsurface. The adherence of microorganisms to acatheter surface depends on the interaction ofthe host, the microbial factors, and the cathetermaterial in reaction to the foreign nature of thecatheter.

The host reacts by enhancing the formationof a thrombin sheath by inducing the deposi-tion of adhesin proteins such as fibrinogen,fibronectin, laminin, and thrombospondin, thusallowing the adherence of microorganisms suchas Staphylococcus aureus, S. epidermidis, andCandida albicans.13–15


Sources of contamination1. Skin insertion site2. Hub contamination3. Contaminated infusate (rare)4. Hematogenous seeding (rare)

AttachmentHydrophobicity of organisms and catheters,surface irregularities and charge differences

Adherence1. Organisms binding to protein

adhesins (e.g. fibrin and fibronectin)2. Organisms producing microbial biofilm or

glycocalyx

MultiplicationColonizing organisms multiply on cathetersurfaces to reach a threshold. Factors enhancinggrowth and multiplication are parenteralnutrition/lipid emulsions, blood products, andprolonged catheterization

InvasionFree-floating organisms colonizing the intravascularsegment will ultimately invade the bloodstream

Figure 10.1 Flow diagram depicting thepathogenesis of catheter-related bloodstreaminfection.

Additionally, bacteria can promote their ownadherence mechanism. S. aureus and C. albicansare coagulase-producing organisms able toinduce thrombogenesis. Other microorganisms,

such as C. parapsilosis and S. epidermidis, pro-duce a biofilm, known as ‘glycocalyx’, whichacts as a barrier against phagocytosis andopsonization.16,17

Finally, the attachment of microorganisms tothe catheter surface depends on the intrinsicproperties of the catheter polymers, includinghydrophobicity, surface irregularities, andcharge differences. Furthermore, microorgan-isms adhere to polyvinyl chloride and polyeth-ylene surfaces better than to polyurethane orTeflon polymers.18,19

Microbiology

Although all VADs are colonized with microor-ganisms, only in a few cases will progression toCRBSI occur. In fact, infection depends onwhether the organisms on the catheter surface,particularly those in the free-floating phase,exceed a certain quantitative threshold.

The microbiology of VAD-associated blood-stream infections is predominantly representedby skin flora such as S. epidermidis, S. aureus,and occasionally Bacillus spp.20–23 The Gram-negatives are usually acquired from the hos-

pital environment and from the hands of med-ical personnel, and include Pseudomonas aerugi-nosa, Acinetobacter spp., and Stenotrophomonasmaltophilia. In addition to the Candida spp. thatmay occur in special settings, some reportshave described the emergence of new fungisuch as Malassezia furfur and Rhodotorula spp.Infection control surveillance at the MDAnderson Cancer Center detected 640 cases ofCRBSI occurring between 1990 and 1996, withthe following frequency: 25% caused bycoagulase-negative staphylococci, 25% by S.aureus, 14% by Gram-negative bacilli, and 15%by other organisms such as enterococci, Bacillusspp., and Corynebacterium.

EVALUATION AND DIAGNOSIS

The diagnosis of catheter-related infection isdifficult, and relies mostly on the isolation ofthe same organism from simultaneous bloodcultures (from the CVC and a peripheral vein)or the isolation of the same organism from acatheter culture and from a peripheral bloodculture with clinical signs and symptoms ofinfection (Table 10.1).24

VASCULAR ACCESS DEVICE INFECTIONS 191

Table 10.1 Diagnostic techniques for catheter infections

I. Semiquantitative or quantitative catheter culture techniques (with regular venipuncture bloodcultures)• Roll-plate catheter culture method (semiquantitative)• Sonication catheter culture method (quantitative)• Vortexing catheter culture method (quantitative)• Centrifuging catheter culture method (quantitative)

II. Simultaneous blood cultures (from catheter and peripheral vein)• Simultaneous quantitative blood cultures• Differential time to positivity

III. Rapid techniques (experimental – require further validation)• Staining of catheters• Endoluminal brush techniques• Acridine orange leukocyte cytospin test

Catheter culture technique

This method requires the removal of thecatheter, and thus is retrospective. It has littleimpact on the clinical decision to remove orretain the VAD.

Semiquantitative culture of the catheter tipThis consists of culturing the catheter tip usingthe roll-plate technique.22 A culture growth of15 or more colony-forming units (CFU) of agiven organism reflects catheter colonization. Agrowth of less than 15 CFU indicates cathetercontamination. Despite a specificity of 76%, thistechnique is limited by the isolation of theorganisms only from the external surface of thecatheter, which is a major limitation for thelong-term CVC.

Quantitative culture of catheter segmentsThis consists of culturing both the external andinternal surfaces of a catheter segment (usuallythe catheter tip and/or the subcutaneous tun-neled segment) by sonication and vortexing,thus releasing the microorganisms embeddedin the biofilm from the internal and externalsurfaces of the CVC.21 A cutoff value of 103 orgreater is indicative of catheter colonization.This technique was found to be highly sensitive(93%) and specific (94%), and is of better diag-nostic yield than the roll-plate technique forlong-term CVCs.25,26

Simultaneous blood culture methods

In febrile neutropenic cancer patients, the CVCis often removed prior to making the diagnosisof CRBSI. This results in unnecessary removalof the CVC.6 Catheter cultures require that thecatheter be removed then cultured in order tomake the diagnosis.25,26 Paired blood culturesfrom the CVC and peripheral vein may helpmake the diagnosis prior to catheter removal,and hence many CVCs may be saved.

Paired quantitative blood culturesSimultaneous quantitative blood cultures aredrawn through the catheter and a peripheralvein. A ratio of 5–10 : 1 or greater of bacterialgrowth from the VAD relative to the peripheralvein is consistent with the VAD being thesource of the bloodstream infection.27

Differential time to positivityThe time to detection of growth in a culture isclosely related to the inoculum size of themicroorganism. Simultaneous non-quantitative(regular) blood cultures are drawn from a CVCand peripheral vein. If the blood culture drawnfrom the CVC becomes positive at least twohours before the simultaneously drawn periph-eral blood culture, then this is highly suggestiveof CRBSI.28,29

PREVENTION

In order to develop useful strategies to preventVAD-associated bloodstream infections, oneshould take into consideration the several fac-tors (namely, host factors, catheter factors, andmicrobial factors) that interact to ultimatelycause the release of the microorganisms thatcolonize the device surface into the blood,resulting in the infectious complication.

Neutropenia as a host factor might predis-pose to catheter infections. However, since it isassociated with thrombocytopenia (which is aprotective factor) in patients with hematologicmalignancy, it has not been consistently shownto be a risk factor to CRBSI.

In a study conducted by Howell and col-leagues5 of patients with long-term indwellingtunneled CVCs who were followed for a total of12 410 catheter-days, neutropenia of less than500 neutrophils per mm3 of blood was the onlyindependent risk factor for catheter-relatedinfection (p � 0.018). Catheter infections weresignificantly more likely to occur during thefirst week of neutropenia than during theremaining neutropenic days. However, in astudy conducted at our center,6 neutropenia,


bone marrow transplantation, use of high-dosesteroids, or infusion of vesicant chemotherapyagents through the CVC did not predisposepatients to catheter infection. The only statisti-cally significant risk factor for catheter infectionwas hematologic malignancy (acute lym-phoblastic leukemia or acute myeloidleukemia).

The same was observed by Groeger and col-leagues30 in a study of patients with leukemia,who had shorter infection-free periods com-pared with patients with lymphoma ormyeloma (p � 0.02), but neutropenia was notevaluated as a risk factor except at the time ofcatheter insertion. In addition, patients withsolid tumors who had catheters had longerinfection-free periods than those who hadhematologic diseases (p � 0.005).30 Independentof neutropenia, patients with hematologicmalignancies may be at higher risk of infectionbecause of excessive manipulation of cathetersresulting from the high frequency of bloodtransfusions and blood withdrawals donethrough the catheter.

Several preventive measures have beenshown to decrease the rate of CRBSI (Table 10.2).

Maximum sterile barrier

In a large, prospective, randomized study doneat the MD Anderson Cancer Center, the use ofmaximal sterile barrier precautions (handwashing, wearing sterile gloves, a mask, agown, and a cap, and using a large drape) at thetime of insertion of the VAD31 was shown to

decrease the risk of long-term catheter infectionsixfold. Furthermore, the occurrence of theinfectious complication may be delayed by anaverage of two months compared with a weekwhen no sterile barriers were used. Given thesignificant decrease in CRBSI, the use of suchprecautions during insertion was found to behighly cost-effective.31 Currently, the USCenters for Disease Control and Preventionhighly recommend the use of maximal sterilebarrier precautions during all CVC insertions.24

Infusion therapy team

The insertion and maintenance of a CVC by anexperienced infusion therapy team not onlyhelps decrease the rate of infections but alsoprolongs the duration of placement of non-tunneled catheters.32,33 At the MD AndersonCancer Center, the mean duration of placementof long-term non-tunneled, non-cuffed siliconecatheters (PICC lines and non-tunneled siliconesubclavians) was 109 days and the infection ratewas 1.4/1000 catheter-days.6 These figures aresimilar to those reported for tunneled Hickmancatheters,3–5 which is attributed in part to thepresence of a skilled infusion therapy team atthe MD Anderson Cancer Center.

Tunneling and ports

The surgically implantable devices and thecompletely implanted subcutaneous ports aredesigned to prevent the migration of skin flora


Table 10.2 Prevention of catheter infections

Aseptic insertion/maintenance Novel/biotechnology

• Maximal sterile barrier • Antimicrobial/anticoagulant fluid solution• Infusion therapy team • Antimicrobial impregnation of catheters

• Silver iontophoresis

along the intercutaneous segment of the device,and hence to decrease the risk of CRBSI. Twoprospective randomized studies evaluated theeffect of catheter tunneling on catheter-relatedinfections.34,35 One study evaluated long-termCVCs (mostly silicone catheters) placed inimmunocompromised patients.34 The risks ofcatheter-related bacteremia associated with tun-neled and non-tunneled CVCs were 2% and 5%,respectively. The difference was not significant,and was most likely due to the relatively smallnumber of patients in each group (107 and 105patients in each group). In another studyinvolving short-term polyurethane cathetersplaced in the internal jugular veins of criticallyill patients, tunneled CVCs were associatedwith a significantly lower rate of catheter-related bacteremia than non-tunneled CVCs.Therefore, tunneling of CVCs may decrease therisk of CRBSI. But is the additional cost of tun-neling (the cost of insertion is $3000–4000 perdevice) justified by this risk reduction?

Ports seem to be associated with a lowerCRBSI rate than tunneled CVCs. A study byMirro and colleagues,36 involving 120 Hickmancatheters, 146 Broviac catheters, and 93implantable ports in children with malignancy,showed that when all causes of catheter failurewere considered (e.g. infection, obstruction,and dislodgment), indwelling ports had a sig-nificantly longer duration of use than did per-cutaneous Hickman or Broviac catheters(p � 0.0009). In a prospective, observationalstudy conducted on 1630 long-term venouscatheters (including 788 percutaneous cathetersand 680 ports), Groeger and colleagues30 foundthat the incidence of infection per device perday was 12 times greater with externalizedcatheters than with ports.

Selection of the VAD placement site

The risk of infection varies according to the siteof insertion of any vascular device. In general,lower-extremity insertion sites are associatedwith a higher risk of infection, mainly because

of exposure to enteric flora. Additionally, VADsinserted into the subclavian veins carry a lowerrisk for infections than those inserted in thejugular veins, because of the proximity of thelatter to oropharyngeal secretions.24

Routine exchange of vascular catheters overa guidewire

The routine exchange of a CVC over aguidewire at fixed intervals of time is notrecommended.24 Use of a guidewire should belimited to replacing a malfunctioning non-tunneled catheter, converting an existingcatheter, and determining the source of thebloodstream infection, allowing culture of theexchanged catheter.

Antimicrobial/anticoagulant flush solutions

The use of antimicrobial or anticoagulant flushsolutions consists of flushing the lumen of thecatheter with a combination of antimicrobialand antithrombotic agents. In several stud-ies,37–39 flushing of tunneled CVCs with a solu-tion of heparin and vancomycin decreased thefrequency of catheter-related bloodstreaminfection caused by Gram-positive organisms.However, this method is limited to the preven-tion of intraluminal colonization, and mayselect for the emergence of drug-resistantmicroorganisms such as vancomycin-resistantenterococci.

A new flush solution consisting of a mixtureof EDTA and minocycline has recently beendeveloped. This combination was found tohave an in vitro activity against a broad spec-trum of microorganisms, including Gram-positive and Gram-negative bacteria and someCandida spp., and was highly efficacious in pre-venting the recurrence of CRBSI in high-riskpatients.40 The efficacy of this combination indecreasing colonization and thrombotic occlu-sion associated with long-term CVCs used inhemodialysis patients has recently been


demonstrated in a prospective randomizedstudy.41

Antiseptic and antimicrobial impregnation ofcatheters

Coating catheters with antiseptics and antimi-crobials is one of the most promising methodsfor preventing CRBSI. However, this techniquehas only been applied to short-term poly-urethane CVCs, and has been used mostly incritically ill patients.

Short-term catheters coated with chlorhexi-dine and silver sulfadiazine on the external sur-face were twofold less likely to becomecolonized and fourfold less likely to producebacteremia when compared with uncoatedcatheters.42 However, these short-term cathetersare not effective if the catheter dwell timeexceeds two weeks. In a recent prospective ran-domized study involving leukemia patients, inwhom the average duration of catheter place-ment was three weeks, the use of antisepticcatheters impregnated with chlorhexidine andsilver sulfadiazine failed to decrease the rate ofCRBSI.43

In a prospective, randomized multicentertrial that included cancer patients, coating boththe internal and external surfaces of a catheterwith a combination of minocycline andrifampin was shown to be highly effective inpreventing CRBSI.44 Furthermore, when com-pared with antiseptic catheters coated withchlorhexidine and silver sulfadiazine, thesecatheters coated with minocycline and rifampinwere 12-fold less likely to be associated withCRBSI.45 Coating vascular devices with antimi-crobials has proven to be highly cost-effective,and catheters impregnated with minocyclineand rifampin were found to have more pro-longed anti-infective effect for up to six weeks.7

Currently, CVCs impregnated with minocy-cline and rifampin are being used in critically illcancer patients and bone marrow transplantpatients (pre-engraftment) at the MD AndersonCancer Center. In critically ill cancer patients,

these catheters were shown to decrease the riskof nosocomial bacteremia, and have resulted incost savings of more than one million dollars.46

Novel methods of impregnating long-termsilicone catheters with antimicrobials have beendeveloped, and ongoing trials are being con-ducted to demonstrate their efficacy in PICClines and non-tunneled silicone subclaviancatheters. It is possible that non-tunneled sili-cone subclavian catheters could serve as a cost-effective alternative to surgically implantablecatheters.

Silver iontophoretic catheters

Ionic silver has a broad spectrum of antimicro-bial activity.47 A silver-impregnated cuff thatcontains a biodegradable collagen with silverions was attached to short-term CVCs, and wasshown to reduce the incidence of CRBSI.48 Thehalf-life of a silver cuff is, however, short (fiveto seven days), and this is why it failed to offerprotection in long-term catheters.49 This can bepalliated by continuously generating silver ionsthrough an electric power source (a small bat-tery connected to the catheter by silver wires).This novel method, which was developed byBodey and colleagues, proved to have longantimicrobial durability in preventing cathetercolonization in vitro, and also in an animalmodel.50,51 However, the clinical efficacy andsafety of this device needs to be demonstratedthrough clinical trials.

MANAGEMENT AND TREATMENT

The plan of management of a CRBSI is complex,and involves the decision whether or not toremove the catheter. However, this decisionneeds to take into consideration several factors,including the degree of complexity of the infec-tion, the identification of the microbial etiology,availability of vascular access, and cost. A bal-anced approach to the management of the vas-cular catheter in febrile neutropenic patients


has been outlined in the National ComprehensiveCancer Network (NCCN) Practice Guidelinesfor Fever and Neutropenia.52 This approachrecommends the removal of the CVC in compli-cated cases with tunnel or pocket infection orwith persistence of the septicemia (bloodstreaminfection) for more than 48 hours on broad-spectrum systemic antimicrobial agents. Theremoval of the CVC is to be strongly consideredin patients with fungemias, rapidly growingmycobacteremias (Mycobacterium chelonae orM. fortuitum), and S. aureus, Ps. aeruginosa,Stenotrophomonas maltophilia, Corynebacteriumjeikeium, and Bacillus spp. bacteremias wherethe CVC is considered as the likely source. Amore detailed description of this approach isoutlined below.

CRBSI complications

Complicated CRBSIs are those associated withthe presence of a septic thrombosis or a deep-seated infection such as right-sided endocardi-tis, tunnel or port pocket infection, and/or thepersistence of a bacteremia 48 hours afterremoval of the vascular device despite ade-quate therapy. Most CRBSIs are non-complicated, and require therapy for 10–14days. However, in the presence of a complica-tion, the catheter should be removed, and ther-apy should be prolonged for at least four weeksin the setting of septic thrombosis or endocardi-tis.53

Microbial etiology (Table 10.3)

Coagulase-negative staphylococci (CoNS)CoNS are responsible for 25% of CBRSI, eventhough they are the most common cause ofcatheter colonization.8 The optimal duration oftherapy for CoNS is not yet defined; however, ifthe patient responds within 48–72 hours of ther-apy, a 7-day course of antimicrobials should beconsidered adequate.54 Catheter removal maynot be necessary; however, there is a 20%

chance of recurrence of the bacteremia if thecatheter is retained compared with 3% if it isremoved.55

Staphylococcus aureusS. aureus CRBSIs are associated with a high rateof complications, including septic throm-bophlebitis and deep-seated infections such asendocarditis, septic emboli, abscesses, andosteomyelitis.56 Retention of the catheter canlead to persistence of the bacteremia, to relapse,and to increased mortality.57

A short course of two weeks with parenteralantibiotics may be considered in uncomplicatedcases responding within 72 hours of therapy. Inthe case of occurrence of complicated S. aureusCRBSI, treatment should be continued for atleast four weeks.56,57

Candida speciesAll cases of catheter-related candidemia requiresystemic therapy because of the associationwith serious complications such as endoph-thalmitis, which may occur in up to 15% ofcases.58 Removal of the catheter is always pre-ferred. Fluconazole 400 mg daily is usually ade-quate except when C. glabrata or C. krusei aresuspected; then amphotericin B at doses of0.7 mg/kg should be used.59

Gram-positive bacilliRemoval of the catheter is suggested for CRBSIcaused by Gram-positive bacilli such as Bacillusspp. and coryneform bacteria.60,61 Vancomycinfor at least 7 days remains the therapy of choice.However, these suggestions are based on asmall number of cases derived from anecdotalreports. In some situations, where the infectionresponds rapidly to antimicrobial therapy,catheter retention is a consideration.

Gram-negative rodsThe common Gram-negative rods causingCRBSI are Ps. aeruginosa, Acinetobacter spp., andS. maltophilia. Failure to remove the catheter isassociated with higher rates of treatment failureand recurrence of the bacteremia.62,63 A treat-


ment course of 7–10 days should be consideredsatisfactory in non-complicated cases. Furtherstudies are underway at our institution to fur-ther delineate the appropriate management ofGram-negative bacillary bloodstream infection.

CONCLUSIONS

The long-term VADs used in cancer patients aretunneled catheters, non-tunneled subclaviancatheters, ports and PICC lines. The rate ofinfection for these catheters ranges from 1.4 to1.9 per 1000 catheter-days. The major route ofcatheter colonization and subsequent blood-

stream infection is luminal colonization origi-nating from hub contamination. The leadingorganisms causing CRBSI are coagulase-negative staphylococci, S. aureus, Gram-negative bacilli, such as Ps. aeruginosa and S.maltophilia, Candida spp. (C. albicans and C. para-psilosis), and Gram-positive bacilli, such asBacillus spp. and Corynebacterium spp. The diag-nosis of CRBSI without removing the catheterrequires simultaneous blood cultures drawnfrom the CVC and peripheral vein. If quantita-tive blood cultures are used, then a ratio of five-fold or greater of bacterial growth from theVAD relative to the peripheral vein is consis-tent with CRBSI. CRBSI is also highly suggested


Table 10.3 Treatment of catheter-infections according to microbial etiology

Catheter removal DurationOrganisms required? Antimicrobial (days)

Coagulase-negative No Vancomycin 7staphylococci (CoNS) Quinupristin/dalfoprisitin

Linezolid

Staphylococcus Yes MRSAa: like CoNS 10–14aureus MSSAa: antistaphylococcal

penicillins or cephalosporins

Gram-positive bacilli Yes Vancomycin 7–10

Gram-negative bacillib Yes Third-generation cephalosporin 7–10CarbapenemsQuinolones

Candida spp. Yes Fluconazole (for C. albicans and 14C. parapsilosis)

Amphotericin B (for C. kruseii andC. glabrata)

a MRSA, methicillin-resistant S. aureus; MSSA, methicillin-sensitive S. aureus.b For Pseudomonas aeruginosa bacteremia, a combination of a �-lactam and aminoglycoside is recommended.

if the blood culture drawn from the CVCbecomes positive at least two hours before thesimultaneously drawn peripheral blood cul-ture. Measures to prevent long-term catheter-related infections consist of maximal sterilebarrier precautions during insertion, insertionby a skilled infusion therapy team, the use ofports, and the use of an antimicrobial/anticoag-ulant flush solution. Antimicrobial impregna-tion of short-term catheters has been shown todecrease the frequency of CRBSI. This novelbiotechnology could be quite useful for long-term CVCs used in cancer patients, and couldrepresent a cost-effective alternative to surgi-cally implantable uncoated catheters. The man-agement of the vascular catheter in febrileneutropenic cancer patients includes removal ofthe CVC in situations with tunnel or pocketinfection or with persistence of the bloodstreaminfection for more than 48 hours on broad-spectrum antimicrobial agents. In addition,CRBSI caused by Candida spp., rapidly growingmycobacteria, S. aureus, Ps. aeruginosa, and S. maltophilia often require removal of thecatheter.

REFERENCES

1. Raad II, Bodey GP, Infection complications ofindwelling vascular catheters. Clin Infect Dis1992; 15: 197–208.

2. Maki DG, Infection Caused by IntravascularDevices: Pathogenesis Strategies for Prevention.London: Royal Society of Medicine, 1991.

3. Press OW, Ramsey PG, Larson EB et al, Hickmancatheter infections in patients with malignancies.Medicine 1984; 63: 189–200.

4. Decker MD, Edwards KM, Central venouscatheter infections. Pediatr Clin North Am 1988;35: 579–612.

5. Howell PB, Walters PE, Donowitz GR, Farr BM,Risk factors for infection of adult patients withcancer who have tunneled central venouscatheters. Cancer 1995; 75: 1367–74.

6. Raad I, Davis S, Becker M et al, Low infectionrate and long durability of nontunneled silasticcatheters: a safe and cost-effective alternative for

long-term venous access. Arch Intern Med 1993;153: 1791–6.

7. Saint S, Veenstra DL, Lipsky BA, The clinical andeconomic consequences of nosocomial centralvenous catheter related infection. Are antimicro-bial catheters useful? Infect Control HospEpidemiol 2000; 21: 375–80.

8. Raad II, Costerton W, Sabharwal U et al,Ultrastructural analysis of indwelling vascularcatheters: a quantitative relationship betweenluminal colonization and duration of placement.J Infect Dis 1993; 168: 400–7.

9. Linares J, Sitges-Serra A, Garau J et al,Pathogenesis of catheter sepsis: a prospectivestudy with quantitative and semiquantitativecultures of catheter hub and segments. J ClinMicrobiol 1985; 21: 357–60.

10. Maki DG, Pathogenesis, prevention and manage-ment of infections due to intravascular devicesused for infusion therapy. In: Infections Associatedwith Indwelling Medical Devices (Bisno AL,Waldvogel FA, eds). Washington, DC: AmericanSociety for Microbiology, 1989: 161–77.

11. Clarke DE, Raffin TA, Infectious complicationsof indwelling long term central venous catheters.Chest 1990: 97: 966–72.

12. Calwell DE, Korber DR, Lawrence JR, Imaging ofbacterial cells by fluorescence exclusion usingscanning confocal laser microscopy. J MicrobiolMeth 1992; 15: 249–61.

13. Hermann M, Vaudaux PE, Pittet D et al,Fibronectin, fibrinogen and laminin act as medi-ators of adherence of clinical staphylococcal iso-lates to foreign material. J Infect Dis 1988; 158:696–701.

14. Vaudaux P, Pittet D, Haeberli A et al, Host fac-tors selectively increase staphylococcal adher-ence on inserted catheters: a role for fibronectinand fibrinogen or fibrin. J Infect Dis 1989; 160:865–75.

15. Bouali A, Robert R, Tronchin G, Senet JM,Characterization of binding of human fibrinogento the surface of germ tubes and mycelium ofCandida albicans. J Gen Microbiol 1987; 133:545–51.

16. Christensen GD, Simpson WA, Younger JJ et al,Adherence of coagulase negative staphylococcito plastic tissue culture plates: a quantitativemodel for the adherence of staphylococci tomedical devices. J Clin Microbiol 1985; 22:996–1006.


17. Costerton JW, Irvin RT, Cheng KJ, The bacterialglycocalyx in nature and disease. Annu RevMicrobiol 1981; 35: 299–324.

18. Sheth NK, Franson TR, Rose HD et al,Colonization of bacteria on polyvinyl chlorideand Teflon intravascular catheters in hospital-ized patients. J Clin Microbiol 1983; 18: 1061–3.

19. Sherertz RJ, Carruth WA, Marosok RD et al,Contribution of vascular catheter material to thepathogenesis of infection: the enhanced risk ofsilicone in vivo. J Biomed Mater Res 1995; 29:635–45.

20. Raad II, Darouiche RO, Catheter related sep-ticemia: risk reduction. Infect Med 1996; 13:807–12; 815–16; 823.

21. Sherertz RJ, Raad II, Balani A et al, Three yearexperience with sonicated vascular catheter cul-tures in a clinical microbiology laboratory. J ClinMicrobiol 1990; 28: 76–82.

22. Maki DG, Weise CE, Sarofin HW, A semiquanti-tative culture method for identifying intra-venous catheter infection. N Engl J Med 1997;296: 1305–9.

23. Kiehn TE, Armstrong D, Changes in the spec-trum of organisms causing bacteremia andfungemia in immunocompromised patients dueto venous access devices. Dur J Clin MicrobiolInfect Dis 1990; 9: 869–72.

24. Pearson ML, Guidelines for prevention onintravascular device related infections. Part I:Intravascular device related infections: anoverview. Part II: Recommendations for the pre-vention of nosocomial intravascular devicerelated infections. Hospital Infection ControlPractices Advisory Committee. Am J InfectControl 1996; 24: 262–93.

25. Raad II, Sabbagh MF, Rand KH, Sherertz RJ,Quantitative tip culture methods and the diag-nosis of central venous catheter related infec-tions. Diagn Microbiol Infect Dis 1992; 15: 13–20.

26. Sherertz RJ, Heard SO, Raad II, Diagnosis oftriple-lumen catheter infection: a comparison ofroll plate, sonication, and flushing methodolo-gies. J Clin Microbiol 1997; 35: 641–6.

27. Capdevila JA, Planes AM, Palomar M et al,Value of differential quantitative blood culturesin the diagnosis of catheter related sepsis. Eur JClin Microbiol Infect Dis 1992; 11: 403–7.

28. Blot F, Schmidt E, Nitenberg G et al, Earlier posi-tivity of central venous versus peripheral bloodcultures is highly predictive of catheter related

sepsis. J Clin Microbiol 1998; 36: 105–9.29. Blot F, Nitenberg G, Chachaty E et al, Diagnosis

of catheter related bacteremia: a prospectivecomparison of the time to positivity of hub-blood versus peripheral-blood cultures. Lancet1999; 354: 1071–7.

30. Greoger JS, Lucas AB, Thaler HT et al, Infectiousmorbidity associated with long-term use ofvenous access devices in patients with cancer.Ann Intern Med 1993; 119: 1168–74.

31. Raad II, Hohn DC, Gilbreath BJ et al, Preventionof catheter related infections by using maximalsterile barrier precautions during insertion. InfectControl Hosp Epidemiol 1994; 15: 231–8.

32. Abi Said D, Raad I, Umphrey J et al, Infusiontherapy team and dressing changes of centralvenous catheters. Infect Control Hosp Epidemiol1999; 20: 101–5.

33. Tomford JW, Hershey CO, McLaren CE et al,Intravenous therapy team and peripheral venouscatheter associated complications: a prospectivecontrol study. Arch Intern Med 1984; 133: 1191–4.

34. Andrivet P, Bacquer A, Vu Ngoc C et al, Lack ofclinical benefit from subcutaneous tunnel insertionof central venous catheters in immunocompro-mised patients. Clin Infect Dis 1994; 18: 199–206.

35. Timset JF, Sebille V, Farkas JC et al, Effect of sub-cutaneous tunneling on internal jugular catheterrelated sepsis in critically ill patients: a prospec-tive randomized study. JAMA 1996; 276:1416–20.

36. Mirro J, Rao BN, Kumar M et al, A comparisonof placement techniques and complications ofexternalized catheters and implantable port usein children with cancer. J Pediatr Surg 1990; 25:122–4.

37. Schwartz C, Henrickson KJ, Roghmann K,Powell K, Prevention of bacteremia attributed toluminal colonization of tunneled central venouscatheters with vancomycin susceptible organ-isms. J Clin Oncol 1990; 8: 591–7.

38. Carratala J, Niubo J, Fernandez-Sevilla A et al,Randomized, double-blind trial of an antibiotic-lock technique for prevention of Gram-positivecentral venous catheter-related infection in neu-tropenic patients with cancer. Antimicrob AgentsChemother 1999; 43: 2200–4.

39. Henrickson KJ, Axtell RA, Hoover SM et al,Prevention of central venous catheter-relatedinfections and thrombotic events in immunocom-promised children by the use of vancomycin/


ciprofloxacin/heparin flush solution: a random-ized, multicenter, double-blind trial. J Clin Oncol2000; 18: 1269–78.

40. Raad I, Buzaid A, Rhyne J et al, Minocycline andEDTA for the prevention of recurrent vascularcatheter infections. Clin Infect Dis 1997; 25:149–51.

41. Bleyer A, Mason L, Raad I, Sherertz R, A ran-domized, double-blind trial comparing minocy-cline EDTA vs heparin as flush solutions forhemodialysis catheters. In: Program and Abstractsof the 4th Decennial Conference Program Committee,March 5–9, 2000, Atlanta, GA: 91 (Abstr P-S1-32).

42. Maki DG, Stolz SM, Wheeler S, Mermel LA,Prevention of central venous catheter relatedbloodstream infection by use of an antisepticimpregnated catheter: a randomized, controlledtrial. Ann Intern Med 1997; 127: 257–66.

43. Logghe C, Van Ossel C, D’Hoore W et al,Evaluation of chlorhexidine and silver–sulfadi-azine impregnated central venous catheters forthe prevention of bloodstream infection inleukemia patients: a randomized controlled trial.J Hosp Infect 1997; 37: 145–56.

44. Raad I, Darouiche R, Dupuis J et al, Centralvenous catheters coated with minocycline andrifampin for the prevention of catheter relatedcolonization and bloodstream infections. AnnIntern Med 1997; 127: 267–74.

45. Darouiche RO, Raad II, Heard SO et al, A com-parison of two antimicrobial impregnated centralvenous catheters. N Engl J Med 1999; 340: 1–8.

46. Raad I, Hackett B, Hanna H et al, Use of anti-biotic impregnated catheters associated withsignificant decrease in nosocomial bloodstreaminfections in critically ill cancer patients. In:Proceedings of the 10th Annual Meeting of theSociety for Healthcare Epidemiology, March 5–9,2000, Atlanta, GA.

47. Spadaro JA, Berger TJ, Barranco SD et al,Antibacterial effects of silver electrodes withweak direct current. Antimicrob Agents Chemother1974; 6: 637–42.

48. Maki DG, Cobb L, Garman JK et al, An attach-able silver impregnated cuff for prevention ofinfection with central venous catheters: aprospective randomized multicenter trial. Am JMed 1988; 85: 307–14.

49. Groeger JS, Lucas AB, Coit D et al, A prospectiverandomized evaluation of silver-impregnatedsubcutaneous cuffs for preventing tunneled

chronic venous access catheter infections incancer patients. Ann Surg 1993; 218: 206–10.

50. Raad I, Hachem R, Zermeno A et al, In vitroantimicrobial efficacy of silver iontophoreticcatheter. Biomaterials 1996; 17: 1055–9.

51. Raad I, Hachem R, Zermeno A et al, Silver ion-tophoretic catheter: prototype of long-term anti-infective vascular access device. J Infect Dis 1996;173: 495–8.

52. National Comprehensive Cancer Network,NCCN practice guidelines for fever and neu-tropenia. Oncology 1999; 13: 197–257.

53. Strinden WD, Helgerson RB, Maki DG, Candidaseptic thrombosis of the great central veins asso-ciated with central catheters. Ann Surg 1985; 202:653–8.

54. Hiemenz J, Skelton J, Pizzo PA, Perspective onthe management of catheter related infections incancer patients. Pediatr Infect Dis J 1986; 5: 6–11.

55. Raad I, Davis S, Khan A et al, Catheter removalaffects recurrence of catheter related coagulasenegative staphylococci bacteremia. Infect ControlHosp Epidemiol 1992; 13: 215–21.

56. Raad I, Narro J, Khan A et al, Serious complica-tions of vascular catheter related Staphylococcusaureus bacteremia in cancer patients. Eur J ClinMicrobiol Infect Dis 1992; 11: 675–82.

57. Raad II, Sabbagh MF, Optimal duration of ther-apy for catheter related Staphylococcus aureusbacteremia: a study of 55 cases and review. RevInfect Dis 1992; 14: 75–82.

58. Rose HD, Venous catheter associated can-didemia. Am J Med Sci 1978; 275: 265–9.

59. Rex JH, Walsh TJ, Sobel JD et al, Practice guide-lines for the treatment of candidiasis. Clin InfecDis 2000; 30: 662–78.

60. Saleh RH, Schorin MA, Bacillus spp. sepsis associ-ated with Hickman catheters in patients with neo-plastic diseases. Pediatr Infect Dis J 1987; 6: 851–6.

61. Young VM, Meyers WF, Moddy MR, SchimpffSC, The emergence of coryneform bacteria as acause of nosocomial infections in compromisedhosts. Am J Med 1981; 70: 646–50.

62. Elting LS, Bodey GP, Septicemia due toXanthomonas species and non aeruginosa Pseudo-monas species: increasing incidence of catheterrelated infections. Medicine 1990; 60: 196–206.

63. Benezra D, Kiehn TE, Gold JWM et al,Prospective study of infections in indwellingcentral venous catheters using quantitativeblood cultures. Am J Med 1988; 85: 495–8.


11Special considerations in children with feverand neutropeniaSarah W Alexander, Philip A Pizzo

INTRODUCTION

Fever and neutropenia is a common complica-tion of cancer therapy in both pediatric andadult patients. The general principles regardingthe care of these patients are the same irrespec-tive of age. However, there are some importantfactors that make the management of childrenwith fever and neutropenia unique. The pur-pose of this chapter is to review both the simil-arities and the important differences betweenchildren and adults with fever and neutropenia.

CANCER IN CHILDREN

Although cancer in children is a rare disease, itcontinues to be the cause of approximately 10%of deaths during childhood, and is the leadingcause of disease-related death in 3- to 14-year-olds.1 There are an estimated 12 000 new casesof cancer diagnosed each year in childrenbetween birth and 19 years in the USA, com-pared with approximately 1 000 000 new casesin the adult population.2,3

The types of cancer are different in childrencompared with adults (Table 11.1). Overall,brain tumors and leukemia account for approx-imately 50% of all pediatric oncology diag-noses.1 Within the pediatric group, there are

distinct patterns of diagnoses based on specificage group. In the first two years of life, the mostcommon tumors include neuroblastoma,Wilms’ tumor, retinoblastoma, primitive neu-roectodermal tumors, and hepatoblastoma.Acute lymphoblastic leukemia (ALL) has asharp incidence peak in the age range 2–4 years.Later in childhood, osteosarcoma, Ewing’s sar-coma, Hodgkin’s disease, and non-Hodgkin’slymphomas are all more common.4

In addition to the types of disease, theresponsiveness to therapy is, in general, morefavorable in children. Over the last threedecades, significant progress has been made inthe success rates in some types of pediatriccancer, most notably ALL, lymphomas, andsome soft tissue sarcomas. Overall, it is esti-mated that more than 65% of children withcancer are cured of their disease. There remain,however, some diagnoses in pediatric oncologywhere little therapeutic progress has been madeand the prognosis remains dismal – for exam-ple brainstem gliomas and rhabdoid tumors.

TREATMENT OF CANCER IN CHILDREN

The treatment of children with cancer involvesthe same three primary modalities as used inadult oncology care: chemotherapy, radiation,

and surgery. Chemotherapy is most oftenadministered in multiagent regimens. For somepediatric tumors, dose intensity, defined as theamount of drug administered over a givenperiod of time, is thought to be important. Ingeneral, children tolerate intensive therapy bet-ter than their adult counterparts, and often thedoses used are higher than they would be foradults.5

The ability of children to tolerate veryaggressive chemotherapeutic regimens com-pared with older individuals is at least in partdue to the fact that children have fewer comor-bid conditions. Sixty percent of all cancers inadults occur in patients more than 65 years old,many of whom are likely to have pre-existingconditions at the time of their cancer diagnosisthat pose additional clinical challenges.6,7 Forsome clinical events that are analogous in thepediatric and adult populations, such as febrileneutropenia in oncology patients, children con-

sistently have lower reported incidence of seri-ous morbidity and mortality.

COMMON INFECTIONS IN CHILDREN

Febrile illnesses are very common during nor-mal childhood. In the pre-school-age child, it isestimated that the average number of febrile ill-nesses per year ranges from 2 to 8.8 Childrenhave on average twice as many upper respira-tory tract infections per year as their adultcounterparts.9 This is largely due to the fact thatthe most common site for spread of viral ill-nesses is the school and group daycare setting,with secondary infections of siblings and adultsat home.10 Otitis media, a relatively rare infec-tion in adults, accounts for approximately one-third of all pediatrician office visits and isdiagnosed in 60–70% of all children by the ageof one year.11 Children are at risk for primary


Table 11.1 Predominant pediatric cancers by agea

Type of disease Percentage of disease type by age

<5 years 5–9 years 10–14 years 15–19 years

Leukemia 36.1 33.4 21.8 12.4Lymphoma 3.9 12.9 20.6 25.1Central nervous system tumors 16.6 27.7 19.6 9.5Neuroblastoma 14.3 2.7 1.2 0.5Retinoblastoma 6.3 0.5 0.1 0Wilms’ tumor 9.7 5.4 0.7 0.2Hepatoblastoma 2.2 0.4 0.6 0.6Malignant bone tumors 0.6 4.6 11.3 7.7Soft tissue sarcomas 5.6 7.5 9.1 8.0Germ cell tumors 3.3 2.0 5.3 13.9Carcinomas 0.9 2.5 8.9 20.9Other 0.5 0.4 0.8 1.2

a Adapted from Ries LAG, Smith MA, Gurney JG et al, Cancer Incidence and Survival Among Children and Adolescents: UnitedStates SEER Program 1975–1995. NIH Publication 99-4649, 1999.

infection with varicella zoster, with approxi-mately two-thirds of the cases occurring in chil-dren between the ages of five and nine years(prior to universal vaccination strategies).

Children with cancer are at risk for these‘regular’ childhood infections, in addition tobeing susceptible to other infectious pathogenssecondary to their immunocompromised state.In addition, ‘regular’ childhood infections cancause significant morbidity in compromisedchildren. For example, respiratory syncytialvirus (RSV) can cause severe pneumonia inpatients receiving chemotherapy, especially inthose undergoing bone marrow transplanta-tion.12,13 Otitis media can be recurrent orchronic, can be due to bacteria not usually asso-ciated with the disease, and can, in rare cases,progress to mastoiditis in the compromisedhost.14 Varicella zoster virus (VZV) andcytomegalovirus (CMV) are more likely tocause disease by primary infection in children,in contrast to reactivation disease in older indi-viduals. Primary varicella in immunocompro-mised hosts can cause severe encephalopathy,hepatitis, and pneumonia. In the 1970s, prior tothe use of acyclovir, this infection carried a 10%mortality rate in children with cancer.15

FEVER

In a prospective review of 1001 episodes offever in pediatric and young adult patients withcancer being treated at the US National CancerInstitute in the late 1970s and early 1980s,approximately one-half of all patients becamefebrile during their course of therapy.16 Of theseepisodes, 80% occurred when the patients wereneutropenic, while 20% occurred in patientswho were not neutropenic. Fevers in the non-neutropenic population were most oftenascribed to use of chemotherapy (especiallymethotrexate, cytarabine arabinoside, cyclo-phosphamide, and actinomycin D), to theunderlying malignancy, and to viral infections.

Fever in the setting of chemotherapy-induced neutropenia in the pediatric patient

with cancer remains a common problem. AtChildren’s Hospital, Boston over a one-yearperiod, 21% of 1107 admissions to the oncologyinpatient service were for children with feverand neutropenia.17 Of all patients receiving sys-temic chemotherapy during the period of obser-vation, 47% had at least one episode of feverand neutropenia.

ASSESSMENT OF THE CHILD PRESENTINGWITH FEVER AND NEUTROPENIA

The initial assessment of the child with feverand neutropenia is guided by the same generalprinciples as those used in adult oncologypatients. A careful history and meticulousphysical examination are imperative. Specialattention should be paid to areas at increasedrisk for infection in patients receiving cytotoxictherapy, including the oropharynx, perianalregion, central-line site if present, and any fociof recent invasive procedures.

Very young children usually cannot reporttheir own symptoms. The history often relies onobservations by caregivers. In addition, ininfants and toddlers, fewer ‘typical’ symptomsmay be recognizable. For example, the childwith mucositis may present with fussiness, dif-ficulty sleeping, or reduced oral intake.

Blood cultures should be obtained bothperipherally and from indwelling centralvenous catheters, if present. Other microbiolog-ical tests should be based on clinical suspicions,for example nasopharyngeal aspirate for respi-ratory viruses, stool assays for Clostridium diffi-cile, or skin scrapings for herpesviruses.Radiologic studies should be ordered based onspecific symptoms or physical findings.Screening chest radiographs should be con-sidered in all children with an anticipated pro-longed duration of febrile neutropenia (>7days) to provide a baseline study for futurecomparison. Serum chemistries, including anassessment of renal function, are usuallyobtained to ensure appropriate organ functionand antimicrobial dosing.

SPECIAL CONSIDERATIONS IN CHILDREN 203

It is sometimes necessary to modify diagnos-tic techniques on the basis of age. The volumeof blood that can be obtained for blood cultureoften is less in a neonate or small child. Youngchildren are often not able to provide urinesamples. Children often require sedation forradiologic tests that require being still for aperiod of time, for example computed tomogra-phy (CT) or magnetic resonance imaging (MRI)scans.18,19 More invasive diagnostic techniquesalso often require consideration of the child’ssize. In general, very small infants can undergobronchoscopy; however, transbronchial biop-sies are often difficult in the very young owingto limitations in the size of the equipment.20,21

Similarly, thorascopic lung biopsies are chal-lenging in young children because of the smallsize of the thorax, although continued miniatur-ization of the equipment may make this less ofa limitation in the future.22

THERAPY IN CHILDREN WITH FEBRILENEUTROPENIA

Initial antimicrobial coverage strategies havebeen studied extensively both in adult and inpediatric populations. Empiric therapy is basedon early antibiotic coverage for the most likelyinfecting organisms as well as those organismsthat have the potential of being rapidly lethal ifnot appropriately treated. There is no singlebest regimen for all patients with fever andneutropenia, with the choice being influencedby local patterns of infecting organisms andtheir resistance spectra, as well as considerationof toxicities, ease of administration, and cost.

The antibiotic regimens used in children donot, in general, differ from those used in adults.Either monotherapy or combination therapy,usually with a �-lactam and an aminoglycoside,is employed. Common antibiotics used in chil-dren with fever and neutropenia, together withdoses, are listed in Table 11.2.

Ceftazidime was the first agent to undergosignificant evaluation as monotherapy forempiric coverage of patients with fever and

neutropenia. Use of this agent has been shownto be as safe as that of combination regi-mens.23,25 Other agents that have been shown tobe efficacious and safe as monotherapy arecefepime and the carbapenems imipenem andmeropenem.26–30 The majority of the large stud-ies of empirical antibiotic therapy for fever andneutropenia with monotherapy have includedboth adult and pediatric patients.23,26,31

Vancomycin is not generally required for theinitial empiric therapy of children with febrileneutropenia.32–34 There are, however, certain cir-cumstances where its use should be considered.The 1997 recommendations from the InfectiousDiseases Society of North America in theirguidelines for use of antimicrobial agents inneutropenic patients with unexplained fever isthat vancomycin should ‘probably’ be used aspart of initial therapy in institutions with highrates of Gram-positive organisms leading to ful-minant infections (e.g. Streptococcus viridans), aswell as in those individuals with obviouscentral-line infections, those with significantmucositis or who are hypotensive at presenta-tion, and those with known colonization withGram-positive organisms resistant to the stan-dard empiric regimen.35

PROLONGED FEVER AND NEUTROPENIA

Children and adults with prolonged neutrope-nia are at high risk for serious infections, andrequire vigilant care.36 Persistent fever in theabsence of other findings in the neutropenichost is not in itself an indication for changingantibiotic therapy, except for the addition ofempiric antifungal therapy.37,38 Changes in clini-cal status or new microbiological data shouldbe used to guide antibiotic modifications.

Prolonged neutropenia puts the patient atsignificant risk for invasive fungal infection.The diagnosis of fungal infections in the neu-tropenic patient remains challenging, with feveroften being the only presenting clinical sign.Even in proven disseminated candidal disease,blood cultures often remain negative. Imaging



Table 11.2 Common antibiotics and doses used in children with fever and neutropeniaa

Antibiotic Dose Comments

Antibacterial agentsCeftazidime 30–50 mg/kg/dose i.v. q8h Broad-spectrum coverage, including

Pseudomonas aeruginosa

Cefipime 50 mg/kg/dose i.v. q8h Broader Gram-positive spectrum than ceftazidime

Imipenem/cilastin 15 mg/kg/dose i.v. q6h Cross-reactivity with 50% of patients withanaphylaxis to penicillin

Meropenem 20 mg/kg/dose i.v. q8h Less likely than imipenem to cause seizures

Piperacillin 75 mg/kg/dose i.v. q6h Should be combined with an aminoglycoside forcoverage of Ps. aeruginosa

Pipericilin– 75 mg/kg/dose i.v. q6h Insufficient data to date to use this agent astazobactam monotherapy

Gentamicin 2.5 mg/kg/dose i.v. q8h Levels should be monitored to prevent ototoxicityand nephrotoxicity

Amikacin 10 mg/kg/dose i.v. q8h Levels should be monitored

Vancomycin 15 mg/kg/dose i.v. q8h Is not needed for empiric therapy in mostpatients

Aztreonam 30 mg/kg/dose i.v. q6h Gram-negative coverage only

AntifungalsAmphotericin 0.5 mg/kg/dose i.v. q4h Higher doses may be needed for Aspergillus.

Significant nephrotoxicity

Lipid formulations 3–5 mg/kg/dose i.v. q24h Significantly less nephrotoxicity with equal(ABLC, efficacyAmBiosome)

Fluconazole 6–12 mg/kg/dose i.v. or p.o. Good coverage for Candida albicans; less goodfor other candidal species

AntiviralsAcyclovir 750–1500 mg/m2/day Dose for VZV is twice that for HSV, hydration

divided q8h should be ensured when giving high doses

Ganciclovir 5 mg/kg bid (for induction Granulocytopenia is the major dose limitingfor CMV) 5 mg/kg/day (for toxicitymaintenance)

Foscarnet 60–120 mg/kg/day divided Nephrotoxicity is the major dose-limitingq8h toxicity; renal function and electrolytes require

close monitoring

a Doses in children less than 28 days old may need to be modified.

with CT and MRI has become a routine part ofthe care of the persistently febrile neutropenicpatient to assess for evidence of hepatospleniccandidiasis at the time of resolution from neu-tropenia, although this infection has becomemuch less common in patients receiving anti-fungal prophylaxis. Subtle pulmonary findingscan be the first signs of invasive Aspergillusinfection, definitively diagnosed often only byopen lung biopsy.

The rational for empiric antifungal therapy isthe same as that for antibacterials – mainly todecrease infection related mortality by early ini-tiation of therapy. Two prospective randomizedtrials reported a benefit of empiric therapy withamphotericin B when started at either day 4 orday 7 of fever and neutropenia.36,39 Both of thesestudies included both adult and pediatricpatients.

Empiric therapy with amphotericin B hasbeen limited by significant renal toxicity.Several large prospective trials comparing lipo-somal preparations of amphotericin B withstandard amphotericin B for empiric therapy infebrile neutropenic patients have shown equalefficacy but less renal toxicity in those patientsreceiving the liposomal formulations.40–43 Thesestudies enrolled both adult and pediatricpatients. Empiric therapy with fluconazole hasalso been shown to be efficacious and less toxic.It should not, however, be used in thosepatients with clinical signs suggestive ofaspergillosis or in those who have received flu-conazole prophylaxis and are therefore athigher risk for infection with fluconazole-resistant candidal species.44,45

RISK STRATIFICATION IN CHILDREN WITHFEVER AND NEUTROPENIA

It has become increasingly clear that not allpatients with fever and neutropenia are atequal risk for serious infection.46–48 The identifi-cation of a subgroup at low risk for seriousinfection may allow for modifications ofempiric therapy, with a goal of less therapy-

related toxicity, an improved ‘quality of life’,and decreased cost. These modifications mayinclude the use of oral antibiotics as opposed tointravenous antibiotics, and care of the low-riskfebrile neutropenic patient in the outpatient set-ting.

Factors available at the time of presentationthat consistently appear to confer low risk are ashort duration of neutropenia (<7–10 days),being clinically well without significant comor-bid medical conditions or significant focal infec-tions, and having cancer that is notprogressive.49,50 These predictive factors havebeen evaluated in pediatric populations, withsimilar findings.17,51,52 In general, given thedecreased frequency of comorbid factors andthe overall better outcome,53 modification oftherapy for low-risk patients may have themost relevance for children with fever and neu-tropenia.

Based on risk stratification, there is nowevidence both in adults and in children that theuse of oral antibiotics in a subset of patientswith low-risk fever and neutropenia is safe andeffective.54–56 The oral regimen that was employedin the two large randomized studies55,56

was ciprofloxacin and amoxicillin/sulbactam.Outpatient therapy with intravenous and oralantibiotics in pediatric patients with low-riskfebrile neutropenia has also been shown to befeasible,57–59 as has early discontinuation of anti-biotic coverage in clinically well children withnegative blood cultures and evidence of earlybone marrow recovery.60–62

OUTCOMES IN CHILDREN WITH FEVER ANDNEUTROPENIA

Many of the large clinical trials evaluating dif-ferent therapies for fever and neutropenia havebeen conducted in combined pediatric andadult populations. In general, patient age hasnot been used as one of the criteria when evalu-ating subgroup outcomes. The largest body ofdata addressing the question of whether thereis a difference between pediatric and adult


patients with fever and neutropenia has comefrom a retrospective analysis of several largeEuropean trials.53 This study of 3080 patientswith fever and neutropenia compared the pedi-atric group, which comprised 25% of the studypopulation, with the adult group in terms ofinfection types and outcomes.

The rate of clinically documented infectionsin the pediatric group was lower, and con-sequently children had a higher incidence offever of unknown origin. Of the sites of infec-tion documented clinically, children were morelikely to have upper respiratory tract infections,compared with lower respiratory tract infec-tions in their adult counterparts.

The rate of bacteremia in the two groups wassimilar (22% versus 24%). The organisms caus-ing bacteremia were similar in children andadults. Gram-negative pathogens accounted for28% of the episodes in children, compared with30% of adults. Gram-positive infections caused64% of the bacteremias in children, comparedwith 57% in adults. Children were noted tohave a significantly higher rate of streptococcalinfections (29% versus 18%). Adults had ahigher incidence of polymicrobial infectionsthan children (12% versus 7%).

Overall mortality was lower in children (3%versus 10%), as was the incidence of deathrelated to infection (1% versus 4%, p � 0.001).In a separate study comparing pediatric andadult cancer patients with documented bac-teremia or fungemia, children had a lower over-all mortality as well as mortality related toinfection (7% versus 25%, p � 0.02; 4% versus8%, p � 0.03, respectively).63

SPECIAL CONSIDERATIONS IN ANTIBIOTICCHOICE IN CHILDREN

The US Food and Drug Administration (FDA)now permits the approval of drugs for use inchildren based on efficacy data gathered inadults, if the disease in children and adults issimilar, provided that pharmaceutical com-panies submit pharmacokinetic and toxicity

data from trials performed in an adequate num-ber of children. If the drug is likely to be usedin children, it is now mandatory for pharma-ceutical companies to provide such data in atimely fashion.64,65

There are, however, a substantial number ofantimicrobials used commonly in children withfebrile neutropenia (as is the case with manydrugs in use in pediatrics) that have not beenformally studied in children and are notFDA-approved. For example, ceftazidime hasbeen studied and is approved for children of allages; meropenem and imipenem are approvedfor those over the age of 3 months; and piperi-cillin, pipericillin–tazobactam, and cefepime areonly approved for use in those over 12 years ofage. For many of these agents, however, there issignificant clinical experience with use of thesedrugs in young children.

There are only a limited number of antimi-crobials that have been shown to have thepotential for significant and unique toxicity inchildren. Tetracyclines are avoided in childrenless than 8 years old owing to associated per-manent dental discoloration. The routine use offluoroquinolones in pediatrics has been limitedby the concern of a unique toxicity in youngindividuals leading to arthropathy. In experi-mental juvenile animals, exposure to fluoro-quinolones has been associated with a risk ofarthropathy expressed clinically as lamenessand associated with characteristic histologicfindings of blisters and erosions of articular car-tilage. This finding has been consistent for allfluoroquinolones tested.66 There is, however, agrowing body of evidence supporting the safetyof quinolone antibiotics in the pediatric popu-lation. Nalidixic acid, a non-fluorinatedquinolone, has been used in children fordecades. In animal studies, it causes the classiccartilage changes, but it has not been found tocause any arthropathy in children, includingthose treated for prolonged periods of time.67,68

In 1997, Hampel et al69 reviewed the worldwideexperience with ciprofloxacin in pediatricpatients based on its compassionate use, andconcluded that short courses of ciprofloxacin


appear to be safe. No cases of the experimen-tally induced cartilage damage have been con-firmed in humans.

SPECIAL CONSIDERATIONS IN ANTIBIOTICDOSING AND ADMINISTRATION INCHILDREN

Antibiotic dosing is based on the child’s weightor, less commonly, body surface area. In addi-tion, the differences in pharmacokinetics andpharmacodynamics between children and adultsneed to be taken into consideration, particularlyin the very young. The hepatic glucuronidationprocess, for example, is relatively immature dur-ing the first 2–3 months of life, thus decreasingthe metabolic clearance of many drugs. Renalexcretion reaches adult levels between 6 and 12months of life, owing to slow maturation ofglomerular filtration and tubular function, aswell as an increase in renal blood flow.70

Conversely, children may have more rapid clear-ance of some agents. For example, the pharma-cokinetics of atovaquone would suggest dosingof 30 mg/kg/day for those less than 3 monthsand greater than 24 months, but 45 mg/kg/dayin those from 3 to 24 months of age.71

Administration of oral antibiotics to childrenoften poses unique challenges. Children lessthan 5 years of age have difficulty swallowingtablets or capsules. Compliance with a regimenmay be limited if the agent is not palatable.Often there is no appropriate pediatric formula-tion available, and therefore provision of theagent requires an innovative pharmacist.

PROPHYLAXIS

The best infection prophylaxis in the care ofimmunocompromised patients (and others) isdiligent handwashing before and after contactwith patients. Teachers or daycare workersshould be made aware of the child’s immuno-compromised state and asked to notify the par-ents in case of an outbreak of a contagious

disease, such as varicella. Limiting social con-tacts for infection prevention is usually recom-mended for a period of time after allogeneicbone marrow transplantation.

Prophylactic antimicrobial regimens are usedextensively in patients receiving cytotoxicchemotherapy. Early on, it was noted that thepatient’s own enteric flora was often the culpritin documented infections during periods offever and neutropenia. Initially, it was shownthat non-absorbable antibiotics decreased therate of infectious complications; however, theseagents are in general unpalatable, and werelimited by difficulty with patient compliance,particularly in children.

The two agents that have received the mostattention as prophylactic therapy for bacterialinfections in neutropenic cancer patients have been trimethoprim–sulfamethoxazole(TMP–SMX) and the oral quinolones, primarilyciprofloxacin.72–74 Both agents have been shownto decrease documented infections, primarilybacteremia, in neutropenic patients; however, animpact on overall mortality has not been shown.The use of TMP–SMX is complicated by asignificant rate of allergic reactions and a risk ofbone marrow suppression, with the potential forprolongation of neutropenia. Fluoroquinolonesare in general well tolerated, but have been asso-ciated with a higher than expected rate of strep-tococcal bacteremia, leading some to suggest theaddition of penicillin or clindamycin to the pro-phylactic regimen. Newer quinolones with anexpanded spectrum and greater potency againstGram-positive organisms might be more effect-ive. The utility of any of these regimens is lim-ited by the potential for the emergence ofresistant pathogens.75,76 In children, the fluoro-quinoles as prophylaxis have been avoidedbecause of concerns of cartilage toxicity withprolonged exposure (see above).

USE OF GROWTH FACTORS

Growth factors, primarily granulocyte colony-stimulating factor (G-CSF, filgrastim) and gran-


ulocyte–macrophage colony-stimulating factor(GM-CSF, sargramostin), are used extensivelyin the care of pediatric cancer patients.77–79

These agents increase the number and phago-cytic function of polymorphonuclear cells in theperipheral blood.80–84 Given the association ofthe degree and duration of neutropenia withthe risk of serious infection, there was initiallygreat excitement about the potential impact thatgrowth factors might have in patients receivingmyelosuppressive chemotherapy. Guidelinesfor the use of these cytokines in patients havebeen developed by the American Society ofClinical Oncology (ASCO).85,86

Two strategies for growth factor use havereceived the greatest attention. The first is initi-ating therapy at the time of diagnosis withfever and neutropenia. Studies in both pediatricand adult patients have had variable results,with some investigators reporting a moderatedecrease in the number of days with fever, neu-tropenia, and antibiotic administration, as wellas a decrease in the length of hospitaliza-tion.87–91 No study has shown a reduction in therate of serious infection or infection-relatedmortality. Overall, there is no strong evidencesupporting the initiation of growth factors atthe time of fever and neutropenia.

The second strategy for growth factor usehas been one of primary prophylaxis, initiatingtherapy after completion of each cycle ofchemotherapy. Despite encouraging early clini-cal trials, multiple subsequent studies haveshown a decreased duration of neutropenia anda variable effect on the incidence of fever andneutropenia and of documented infection, butno discernible effect on infection-related mor-tality.79,92–95

The ASCO guidelines recommend primaryprophylaxis with growth factors when theexpected rate of fever and neutropenia is greaterthan 40%. This guideline is difficult to apply,particularly in pediatrics, where the data regard-ing incidence of febrile neutropenia for standardtreatment regimens has not been systematicallyevaluated. The patterns of growth factor use havebeen explored both in adults and in children.85,86

Pediatric oncologists are much more likely thantheir adult counterparts to use growth factors asprimary prophylaxis, often because of protocolrequirements, but also owing to the fact that mostchemotherapeutic regimens used for pediatricmalignancies will predictably result in periodswith marked neutropenia.

FUTURE DIRECTIONS IN THE CARE OFCHILDREN WITH CANCER AND FEBRILENEUTROPENIA

Significant progress has been made in the treat-ment of children with cancer. This has beenaccomplished by progress both in the use of theprimary therapeutic modalities of chemother-apy, surgery, and radiation therapy, and insupportive care measures.

The rapid and thorough assessment andappropriate management of children withchemotherapy-induced fever and neutropeniawill remain a common and important part ofthe care of children with cancer. Progress in thecare of these individuals will rely in part on theuse of risk-based stratification to allow for mod-ification of therapy based on individual risk fac-tors. Ongoing evaluation of the safety andefficacy, as well as cost and ease of administra-tion, of new antimicrobial agents will act tobroaden the therapeutic armamentarium, andwill hopefully increase the success rate anddecrease the burden of illness for childrentreated for cancer.

REFERENCES

1. Robison LL, General principles of the epidemiol-ogy of childhood cancer. In: Principles andPractice of Pediatric Oncology, 3rd edn (Pizzo PA,Poplack DG, ed). Philadelphia: Lippincott-Raven, 1997: 1–10.

2. Grovas A, Fremgen A, Rauck A et al, TheNational Cancer Data Base report on patterns ofchildhood cancers in the United States. Cancer1997; 80: 2321–32.


3. Ries LAG, Smith MA, Gurney JG et al, CancerIncidence and Survival among Children andAdolescents: United States SEER Program1975–1995. NIH Publication 99-4649, 1999.

4. Gurney JG, Severson RK, Davis S, Robison LL,Incidence of cancer in children in the UnitedStates. Sex-, race-, and 1-year age-specific ratesby histologic type. Cancer 1995; 75: 2186–95.

5. Balis FM, Holcenberg JS, Poplack DG, Generalprinciples of chemotherapy. In: Principles andPractice of Pediatric Oncology, 3rd edn (Pizzo PA,Poplack DG, eds). Philadelphia: Lippincott-Raven, 1997: 215–72.

6. Yancik R, Epidemiology of cancer in the elderly.Current status and projections for the future.Rays 1997; 22(1 Suppl): 3–9.

7. Yancik R, Wesley MN, Ries LA et al,Comorbidity and age as predictors of risk forearly mortality of male and female colon carci-noma patients: a population-based study. Cancer1998; 82: 2123–34.

8. Cherry J, The common cold. In: Textbook ofPediatric Infectious Disease, 4th edn (Feigin RCJ,Cherry T, ed). Philadelphia: Saunders, 1998:128–33.

9. Badger GF, Dingle JH, Feller AE, A study of ill-ness in a group of Cleveland families: Incidenceof the common respiratory diseases. Am J Hyg1953; 58: 31–40.

10. Hurwitz ES, Gunn WJ, Pinsky PF, SchonbergerLB, Risk of respiratory illness associated withday-care attendance: a nationwide study.Pediatrics 1991; 87: 62–9.

11. Teele DW, Klein JO, Rosner B, Epidemiology ofotitis media during the first seven years of life inchildren in greater Boston: a prospective, cohortstudy. J Infect Dis 1989; 160: 83–94.

12. Hall CB, Powell KR, MacDonald NE et al,Respiratory syncytial viral infection in childrenwith compromised immune function. N Engl JMed 1986; 315: 77–81.

13. Whimbey E, Couch RB, Englund JA et al,Respiratory syncytial virus pneumonia in hospi-talized adult patients with leukemia. Clin InfectDis 1995; 21: 376–9.

14. Berkow RL, Weisman SJ, Provisor AJ et al,Invasive aspergillosis of paranasal tissues in chil-dren with malignancies. J Pediatr 1983; 103: 49–53.

15. Feldman S, Hughes WT, Daniel CB, Varicella inchildren with cancer: seventy-seven cases.Pediatrics 1975; 56: 388–97.

16. Pizzo PA, Robichaud KJ, Wesley R, Commers JR,Fever in the pediatric and young adult patientwith cancer. A prospective study of 1001episodes. Medicine (Baltimore) 1982; 61: 153–65.

17. Alexander SW, Wade KC, Hibberd PL, ParsonsSK, Evaluation of risk prediction criteria forepisodes of febrile neutropenia in pediatriccancer patients. J Pediatr Hematol Oncol,Submitted for publication.

18. Hubbard AM, Markowitz RI, Kimmel B et al,Sedation for pediatric patients undergoing CTand MRI. J Comput Assist Tomogr 1992; 16: 3–6.

19. Krauss B, Green SM, Sedation and analgesia forprocedures in children. N Engl J Med 2000; 342:938–45.

20. Perez CR, Wood RE, Update on pediatric flexiblebronchoscopy. Pediatr Clin North Am 1994; 41:385–400.

21. Wood R, Pediatric Bronchoscopy. Chest Surg ClinN Am 1996; 6: 237–51.

22. Rothenberg SS, Thoracoscopy in infants and chil-dren. Semin Pediatr Surg 1998; 7: 194–201.

23. Pizzo PA, Hathorn JW, Hiemenz J et al, A ran-domized trial comparing ceftazidime alone withcombination antibiotic therapy in cancer patientswith fever and neutropenia. N Engl J Med 1986;315: 552–8.

24. Sanders JW, Powe NR, Moore RD, Ceftazidimemonotherapy for empiric treatment of febrileneutropenic patients: a meta-analysis. J Infect Dis1991; 164: 907–16.

25. De Pauw BE, Deresinski SC, Feld R et al, TheIntercontinental Antimicrobial Study Group,Ceftazidime compared with piperacillin andtobramycin for the empiric treatment of fever inneutropenic patients with cancer. A multicenterrandomized trial. Ann Intern Med 1994; 120:834–44.

26. Freifeld AG, Walsh T, Marshall D et al, Mono-therapy for fever and neutropenia in cancerpatients: a randomized comparison of ceftazidimeversus imipenem. J Clin Oncol 1995; 13: 165–76.

27. Lindblad R, Rodjer S, Adriansson M et al,Empiric monotherapy for febrile neutropenia – arandomized study comparing meropenem withceftazidime. Scand J Infect Dis 1998; 30: 237–43.

28. The Meropenem Study Group of Leuven,London and Nijmegen, Equivalent efficacies ofmeropenem and ceftazidime as empiricalmonotherapy of febrile neutropenic patients. JAntimicrob Chemother 1995; 36: 185–200.


29. Biron P, Fuhrmann C, Cure H et al, Cefepimeversus imipenem-cilastatin as empiricalmonotherapy in 400 febrile patients with shortduration neutropenia. CEMIC (Study Group ofInfectious Diseases in Cancer). J AntimicrobChemother 1998; 42: 511–18.

30. Wang FD, Liu CY, Hsu HC et al, A comparativestudy of cefepime versus ceftazidime as empirictherapy of febrile episodes in neutropenicpatients. Chemotherapy 1999; 45: 370–9.

31. Cometta A, Calandra T, Gaya H et al,Monotherapy with meropenem versus combina-tion therapy with ceftazidime plus amikacin asempiric therapy for fever in granulocytopenicpatients with cancer. The InternationalAntimicrobial Therapy Cooperative Group ofthe European Organization for Research andTreatment of Cancer and the Gruppo ItalianoMalattie Ematologiche Maligne dell’AdultoInfection Program. Antimicrob Agents Chemother1996; 40: 1108–15.

32. European Organization for Research andTreatment of Cancer (EORTC) InternationalAntimicrobial Therapy Cooperative Group,National Cancer Institute of Canada – ClinicalTrials Group, Vancomycin added to empiricalcombination antibiotic therapy for fever in gran-ulocytopenic cancer patients. J Infect Dis 1991;163: 951–8.

33. Rubin M, Hathorn JW, Marshall D et al, Gram-positive infections and the use of vancomycin in550 episodes of fever and neutropenia. AnnIntern Med 1988; 108: 30–5.

34. Ramphal R, Bolger M, Oblon DJ et al,Vancomycin is not an essential component of theinitial empiric treatment regimen for febrile neu-tropenic patients receiving ceftazidime: a ran-domized prospective study. Antimicrob AgentsChemother 1992; 36: 1062–7.


36. Pizzo PA, Robichaud KJ, Gill FA, Witebsky FG,Empiric antibiotic and antifungal therapy forcancer patients with prolonged fever and granu-locytopenia. Am J Med 1982; 72: 101–11.

37. Lee JW, Pizzo PA, Management of the cancerpatient with fever and prolonged neutropenia.Hematol Oncol Clin North Am 1993; 7: 937–60.

38. De Pauw BE, Raemaekers JM, Schattenberg T,Donnelly JP, Empirical and subsequent use ofantibacterial agents in the febrile neutropenicpatient. J Intern Med Suppl 1997; 740: 69–77.

39. EORTC International Antimicrobial TherapyCooperative Group, Empiric antifungal therapyin febrile granulocytopenic patients. Am J Med1989; 86(6 Pt 1): 668–72.

40. White MH, Bowden RA, Sandler ES et al,Randomized, double-blind clinical trial ofamphotericin B colloidal dispersion vs. ampho-tericin B in the empirical treatment of fever andneutropenia. Clin Infect Dis 1998; 27: 296–302.

41. Prentice HG, Hann IM, Herbrecht R et al, A ran-domized comparison of liposomal versus con-ventional amphotericin B for the treatment ofpyrexia of unknown origin in neutropenicpatients. Br J Haematol 1997; 98: 711–18.

42. Walsh T, Bodensteiner D, Hiemenz J et al, A ran-domized, double blind trial of AmBisome (lipo-somal amphotericin B) in the empirical treatmentof persistently febrile neutropenic patients. In:Program and Abstracts of the 37th InterscienceConference on Antimicrobial Agents and Chemo-therapy, Toronto. Washington, DC: AmericanSociety for Microbiology, 1997: Abstr LM-90.

43. Walsh TJ, Finberg RW, Arndt C et al, Liposomalamphotericin B for empirical therapy in patientswith persistent fever and neutropenia. NationalInstitute of Allergy and Infectious DiseasesMycoses Study Group. N Engl J Med 1999; 340:764–71.

44. Viscoli C, Castagnola E, Van Lint MT et al,Fluconazole versus amphotericin B as empiricalantifungal therapy of unexplained fever in gran-ulocytopenic cancer patients: a pragmatic, multi-centre, prospective and randomised clinical trial.Eur J Cancer 1996; 32A: 814–20.

45. Viscoli C, Castagnola E, Machetti M, Antifungaltreatment in patients with cancer. J Intern MedSuppl 1997; 740: 89–94.

46. Freifeld AG, Pizzo PA, The outpatient manage-ment of febrile neutropenia in cancer patients.Oncology (Huntingt) 1996; 10: 599–606; 611–12;discussion 615–16.

47. Buchanan GR, Approach to treatment of thefebrile cancer patient with low-risk neutropenia.Hematol Oncol Clin North Am 1993; 7: 919–35.

48. Elting LS, Rubenstein EB, Rolston KV, BodeyGP, Outcomes of bacteremia in patients withcancer and neutropenia: observations from two


decades of epidemiological and clinical trials.Clin Infect Dis 1997; 25: 247–59.

49. Talcott JA, Finberg R, Mayer RJ, Goldman L, Themedical course of cancer patients with fever andneutropenia. Clinical identification of a low-risksubgroup at presentation. Arch Intern Med 1988;148: 2561–8.


51. Lucas KG, Brown AE, Armstrong D et al, Theidentification of febrile, neutropenic childrenwith neoplastic disease at low risk for bac-teremia and complications of sepsis. Cancer 1996;77: 791–8.

52. Klaassen RJ, Goodman TR, Pham B, Doyle JJ,‘Low-risk’ prediction rule for pediatric oncologypatients presenting with fever and neutropenia. JClin Oncol 2000; 18: 1012–19.

53. Hann I, Viscoli C, Paesmans M et al, A compari-son of outcome from febrile neutropenicepisodes in children compared with adults:results from four EORTC studies. InternationalAntimicrobial Therapy Cooperative Group(IATCG) of the European Organization forResearch and Treatment of Cancer (EORTC). Br JHaematol 1997; 99: 580–8.

54. Malik IA, Abbas Z, Karim M, Randomised com-parison of oral ofloxacin alone with combinationof parenteral antibiotics in neutropenic febrilepatients. Lancet 1992; 339: 1092–6.


56. Kern WV, Cometta A, De Bock R et al, Oral ver-sus intravenous empirical antimicrobial therapyfor fever in patients with granulocytopenia whoare receiving cancer chemotherapy. InternationalAntimicrobial Therapy Cooperative Group ofthe European Organization for Research andTreatment of Cancer. N Engl J Med 1999; 341:312–18.

57. Mustafa MM, Aquino VM, Pappo A et al, A pilotstudy of outpatient management of febrile neu-tropenic children with cancer at low risk of bac-teremia. J Pediatr 1996; 128: 847–9.

58. Shemesh E, Yaniv I, Drucker M et al, Home

intravenous antibiotic treatment for febrileepisodes in immune-compromised pediatricpatients. Med Pediatr Oncol 1998; 30: 95–100.

59. Mullen CA, Petropoulos D, Roberts WM et al,Outpatient treatment of fever and neutropeniafor low risk pediatric cancer patients. Cancer1999; 86: 126–34.

60. Bash RO, Katz JA, Cash JV, Buchanan GR, Safetyand cost effectiveness of early hospital dischargeof lower risk children with cancer admitted forfever and neutropenia. Cancer 1994; 74: 189–96.

61. Aquino VM, Tkaczewski I, Buchanan GR, Earlydischarge of low-risk febrile neutropenic chil-dren and adolescents with cancer. Clin Infect Dis1997; 25: 74–8.

62. Wacker P, Halperin DS, Wyss M, Humbert J,Early hospital discharge of children with feverand neutropenia: a prospective study. J PediatrHematol Oncol 1997; 19: 208–11.

63. Krupova I, Kaiserova E, Foltinova A et al,Bacteremia and fungemia in pediatric versusadult cancer patients after chemotherapy: com-parison of etiology, risk factors and outcome. JChemother 1998; 10: 236–42.

64. Regulations requiring manufacturers to assessthe safety and effectiveness of new drugs andbiological products in pediatric patients-FDAFinal Rule. Fed Regist 1998; 63: 66631–72.

65. Miller JL, FDA to require pediatric studies fordrugs commonly used in children. Am J HealthSyst Pharm 1999; 56: 203.

66. Lietman PS, Fluoroquinolone toxicities. Anupdate. Drugs 1995; 49(Suppl 2): 159–63.

67. Schaad UB, Wedgwood-Krucko J, Nalidixic acidin children: retrospective matched controlledstudy for cartilage toxicity. Infection 1987; 15:165–8.

68. Nuutinen M, Turtinen J, Uhari M, Growth andjoint symptoms in children treated with nalidixicacid. Pediatr Infect Dis J 1994; 13: 798–800.

69. Hampel B, Hullmann R, Schmidt H,Ciprofloxacin in pediatrics: worldwide clinicalexperience based on compassionate use – safetyreport. Pediatr Infect Dis J 1997; 16: 127–9; discus-sion 160–2.

70. Reed MD, Besunder JB, Developmental pharma-cology: ontogenic basis of drug disposition.Pediatr Clin North Am 1989; 36: 1053–74.

71. Hughes W, Dorenbaum A, Yogev R et al, Phase Isafety and pharmacokinetics study ofmicronized atovaquone in human immunodefi-


ciency virus-infected infants and children.Pediatric AIDS Clinical Trials Group. AntimicrobAgents Chemother 1998; 42: 1315–18.

72. Gualtieri RJ, Donowitz GR, Kaiser DL et al,Double-blind randomized study of prophylactictrimethoprim/sulfamethoxazole in granulocy-topenic patients with hematologic malignancies.Am J Med 1983; 74: 934–40.

73. Kauffman CA, Liepman MK, Bergman AG,Mioduszewski J, Trimethoprim/sulfamethoxa-zole prophylaxis in neutropenic patients.Reduction of infections and effect on bacterialand fungal flora. Am J Med 1983; 74: 599–607.


75. Carratala J, Fernandez-Sevilla A, Tubau F et al,Emergence of quinolone-resistant Escherichia colibacteremia in neutropenic patients with cancerwho have received prophylactic norfloxacin. ClinInfect Dis 1995; 20: 557–60; discussion 561–3.

76. Donnelly JP, Is there a rationale for the use ofantimicrobial prophylaxis in neutropenicpatients? J Intern Med Suppl 1997; 740: 79–88.

77. Nemunaitis J, Rosenfeld CS, Ash R et al, PhaseIII randomized, double-blind placebo-controlledtrial of rhGM-CSF following allogeneic bonemarrow transplantation. Bone Marrow Transplant1995; 15: 949–54.

78. Piguet D, Chapuis B, Recombinant human gran-ulocyte–macrophage colony-stimulating factorin acquired or chemotherapy-induced neutrope-nia. An open clinical trial. Acta Oncol 1994; 33:639–43.

79. Pui C-H, Boyett JM, Hughes WT et al, Humangranulocyte colony-stimulating factor afterinduction chemotherapy in children with acutelymphoblastic leukemia. N Engl J Med 1997; 336:1781–7.

80. Wang JM, Chen ZG, Colella S et al, Chemotacticactivity of recombinant human granulocytecolony-stimulating factor. Blood 1988; 72:1456–60.

81. Welte K, Zeidler C, Reiter A et al, Differentialeffects of granulocyte–macrophage colony-stimulating factor and granulocyte colony-stimulating factor in children with severecongenital neutropenia. Blood 1990; 75: 1056–63.

82. Weisbart RH, Golde DW, Clark SC et al, Humangranulocyte–macrophage colony-stimulating

factor is a neutrophil activator. Nature 1985; 314:361–3.

83. Sieff CA, Emerson SG, Donahue RE et al,Human recombinant granulocyte–macrophagecolony-stimulating factor: a multilineagehematopoietin. Science 1985; 230: 1171–3.

84. Roilides E, Walsh TJ, Pizzo PA, Rubin M,Granulocyte colony-stimulating factor enhancesthe phagocytic and bactericidal activity of nor-mal and defective human neutrophils. J Infect Dis1991; 163: 5790–83.

85. American Society of Clinical Oncology,American Society of Clinical Oncology recom-mendations for the use of hematopoietic colony-stimulating factors: evidence-based, clinicalpractice guidelines. J Clin Oncol 1994; 12:2471–508.

86. American Society of Clinical Oncology, 1997update of recommendations for the use ofhematopoietic colony-stimulating factors:evidence-based, clinical practice guidelines. JClin Oncol 1997; 15: 3288.

87. Maher DW, Lieschke GJ, Green M et al,Filgrastim in patients with chemotherapy-induced febrile neutropenia. A double-blind,placebo-controlled trial. Ann Intern Med 1994;121: 492–501.

88. Mitchell PL, Morland B, Stevens MC et al,Granulocyte colony-stimulating factor in estab-lished febrile neutropenia: a randomized studyof pediatric patients. J Clin Oncol 1997; 15:1163–70.

89. Anaissie EJ, Vartivarian S, Bodey GP et al,Randomized comparison between antibioticsalone and antibiotics plus granulocyte–macrophage colony-stimulating factor(Escherichia coli-derived) in cancer patients withfever and neutropenia. Am J Med 1996; 100:17–23.

90. Riikonen P, Saarinen UM, Mäkipernaa A et al,Recombinant human granulocyte–macrophagecolony-stimulating factor in the treatment offebrile neutropenia: a double blind placebo-controlled study in children. Pediatr Infect Dis J1994; 13: 197–202.

91. Vellenga E, Uyl-de Groot CA, de Wit R et al,Randomized placebo-controlled trial of granulo-cyte–macrophage colony-stimulating factor inpatients with chemotherapy-related febrile neu-tropenia. J Clin Oncol 1996; 14: 619–27.

92. Pettengell R, Gurney H, Radford JA et al,


Granulocyte colony-stimulating factor to preventdose-limiting neutropenia in non-Hodgkin’slymphoma: a randomized controlled trial. Blood1992; 80: 1430–6.

93. Gerhartz HH, Engelhard M, Meusers P et al,Randomized, double-blind, placebo-controlled,phase III study of recombinant human granulo-cyte–macrophage colony-stimulating factor asadjunct to induction treatment of high-grademalignant non-Hodgkin’s lymphomas. Blood1993; 82: 2329–39.

94. Heil G, Chadid L, Hoelzer D et al, GM-CSF in adouble-blind randomized, placebo controlledtrial in therapy of adult patients with de novoacute myeloid leukemia (AML). Leukemia 1995; 9:3–9.

95. Ohno R, Tomonaga M, Ohshima T et al, A ran-domized controlled study of granulocyte colonystimulating factor after intensive induction andconsolidation therapy in patients with acutelymphoblastic leukemia. Int J Hematol 1993; 58:73–81.


12Prophylaxis of infectionsJ Peter Donnelly

INTRODUCTION

The treatment of cancer, particularly hemato-logic malignancies, entails using toxic drugcocktails that inflict considerable damage ontwo primary host defenses: the phagocytic lineof defense, primarily the neutrophils, and theintegument, i.e. the skin and the mucosa of theoral cavity and digestive tract. The importanceof neutropenia has been recognized for over 30years since the seminal publication of Bodeyand colleagues.1 More recently, it has becomeincreasingly clear that injury to the mucosalbarrier of the oral cavity and alimentary tractthat results from the same intensive chemother-apy is not simply an inevitable side-effect but isalso responsible for much of the morbidity thataccompanies myelotoxicity.2–4 In fact, patientsexposed to intensive radiotherapy and cyto-toxic chemotherapy are placed in double jeop-ardy, because concurrent mucositis andneutropenia leave them bereft of defensesagainst infection.

Prevention or cure

That prevention is better than cure is a long-held axiom of medicine, particularly wherevulnerable individuals are concerned and the

disease is preventable. Infections in neutropenicpatients can be devastating, often culminatingin the death of a patient who might otherwisehave achieved remission or even cure. When acause is identified, it is more often than not acommon rather than exotic microbial speciesthat is involved – frequently one that belongs tothe normally harmless commensal flora thatreside on body surfaces. This fact has motivatedmany in the past to resort to prophylactic regi-mens that seemed reasonable at the time buthad not been shown properly to work. Lookedat with hindsight, almost all the early trials ofprophylaxis fall short of modern standards, andadvocates of prophylaxis adopted regimensbased upon meager evidence. In fact, it is onlyvery recently that the spotlight of evidence-based medicine has been turned upon thefebrile neutropenic patient, and, not surpris-ingly, the claims made for antibacterial prophy-laxis and, in particular, antifungal prophylaxis,have been found wanting and based, at best, onlow-level evidence. Equally unsurprisingly, thecriticism leveled at prophylaxis by investigatorswho are considered deskbound by those facingthe challenge presented by febrile neutropeniaevery day has only served to fuel debate thathas often shed more heat than light. Meanwhilethere is probably no hematopoietic stem celltransplant center in the Western world that

does not give prophylaxis of one sort oranother, and very few centers dealing withleukemia who opt to do entirely without suchregimens. The purpose of this chapter is to tryto critically appraise prophylaxis and to exam-ine alternatives, if such exist. This cannot takeplace without first placing the issue in the con-text of neutropenia induced by chemotherapyused to treat malignant diseases and for theablative conditioning therapy for a hematopoi-etic stem cell transplant.

IMPAIRMENT OF HOST DEFENSES BYCANCER THERAPIES

Physical barriers

The physical barrier presented by the skin andthe mucosa of the respiratory, urinary, genital,and alimentary tracts forms first line of defenseagainst resident and transient microorganisms.Anatomically, the skin and these other organsform part of the external surface of the body,and comprise a continuum. Inevitably, thesesurfaces come into daily contact with micro-organisms, of which some are harmless, somebeneficial, and others detrimental. Evolutionhas ensured a formidable array of obstacles thata potential pathogen must overcome beforegaining entry into the host.

The outer body surface – the skinThe skin is a particularly effective barrier aslong as it remains intact, since it forms a barren,dry, hostile environment suited to a few Gram-positive bacteria, including Staphylococcus epi-dermidis, coryneforms, and certain yeasts.5 Thesalt in sweat, fatty acids released from the seba-ceous secretions, a low pH, the presence of bac-teriocins, and secretory IgA all help the healthyskin to resist colonization by foreign microor-ganisms, particularly the Gram-negative bacilli.

The inner body surfaces – the oral cavityFrom a microbial perspective, the oral cavitypresents an extremely complex environment

with varied topography, offering an enormousdiversity of ecological niches. Each surface pos-sesses its own unique consortium of commensalbacteria, from anaerobic bacteria such asFusobacterium spp. to the aerobes such asNeisseria spp. For instance, the oral viridansstreptococci are not uniformly distributed in theoral cavity. Rather, different species occupy dif-ferent niches, with Streptococcus sanguis andStreptococcus mitis biovar 1 predominating onthe buccal mucosa, Streptococcus mitis biovar 2residing on the dorsum of the tongue, andStreptococcus oralis being found almost exclu-sively in initial dental plaque.6 Besides beingpopulated by commensal flora, the oral cavity isalso the first port of call to most exogenous bac-teria. Fortunately, in health, professional andopportunistic pathogens do not normally colo-nize the oropharynx, since they are unable toovercome the formidable array of differentdefense mechanisms of antibacterial substancesand immunoglobulins secreted by the host andthe direct competition posed by the indigenousmicroflora.7 However, poor dental hygiene,peridontitis, diseased teeth, ill-health, and theloss of integrity of the oral cavity through drug-induced mucositis or by local infections such asherpes or candidiasis all increase the opportun-ities for certain potential pathogens to settleand establish an infective nidus. The Gram-negative bacilli, notably Pseudomonas aeruginosa,are adept at exploiting weaknesses, and areable to establish colonization of damaged tis-sue, where local infection can occur and evenlead to disseminated infection.

The inner body surfaces – the stomachThe extreme low pH of the normal stomachprovides an effective barrier to the transfer oforal bacteria to the intestinal tract. However,many patients treated with chemotherapy withor without irradiation experience nausea andvomiting and later gastric reflux, for whichantacids, H2 receptor antagonists, and proton-pump inhibitors are prescribed. Consequently,such patients are effectively achlorhydric andhence bereft of a barrier to microbial access.


This mechanism has been proposed to explainthe onset of severe infections due to S. mitis,8

and may also explain why these patients aremore at risk of developing Candida infections,including gastrointestinal candidiasis.9

The inner body surfaces – the intestinal tractIn health, the small intestine is virtually sterile,whereas the large intestine is colonized by avariety of different bacterial species, whichtogether account for almost half of the solidsfound in formed feces. Viable counts are esti-mated to be of the order of 1014 microorganismsper gram. It is a testament to evolution thatvery few species are capable of establishinginfection – usually certain Gram-negative bacilli(e.g. Escherichia coli) and the Candida yeasts –even in the most profoundly immunosup-pressed individual. The gut microflora aredominated by anaerobes, including Bacteroidesspp. and Clostridium spp., but it appears thatless-familiar Gram-positive non-sporing, lactic-acid-producing bacilli such as bifidobacteria areessential in maintaining a healthy commensalflora by providing the so-called ‘colonizationresistance’.10,11 A loss of this facility is markedby yeast overgrowth and the recovery of enter-ococci in almost pure culture from stool sam-ples. The loss of the normal bowel floraeffectively creates an ecological vacuum inwhich transient exogenous bacteria are nowable to gain a foothold and establish coloniza-tion (Figure 12.1).

Impact of chemotherapy and radiationtherapy on body surfaces

Injury to the skinChemotherapy and irradiation can radicallychange the microenvironment, since they causehair loss, dryness, and loss of sweat production.Infections can develop around the hair follicles,providing a potential nidus for systemic infec-tion. Abraded skin can also lead to local infec-tion, which can be a reservoir that promotesfurther spread to entry sites of intravenous

PROPHYLAXIS OF INFECTIONS 217

Infection

Potentialpathogen

GI tract

Translocation

Normalcommensalflora

Antibiotics

Selection

Systemicinfection

Chemotherapy

Injury

Figure 12.1 The origins of endogenous infection.

catheters. Needle punctures and vascularcatheters produce trauma, and provide a readymeans for microorganisms to gain direct accessto the bloodstream. See Chapter 10.

Effect of chemotherapy and irradiation on theoral cavityNeutropenic patients will experience varyingdegrees of mucosal damage, depending uponthe nature of their chemotherapy.12,13 Oralmucositis is not just a simple matter of failure ofdamaged cells to be replaced. Rather, it appearsto be complex biological response to chemother-apy and irradiation involving initial damage tothe endothelial and connective tissues, whichtriggers an initial inflammatory response.14

Briefly, mucositis proceeds in four phases –namely, inflammatory, epithelial, ulcerative andhealing phases. The cytokines tumor necrosisfactor � (TNF-�) and interleukin (IL)-1� appearto initiate and amplify the process and their pro-duction can be moderated by recombinantIL-11.15 Mucositis becomes clinically manifestduring the epithelial phase and infection duringthe ulcerative/bacteriological phase. Floridmucositis is initially characterized by dryness,then by edema, erythema, and pain. Ulcers then

occur singly or severally, and there may besignificant amounts of viscid mucus. Mucositisprogresses to a peak that is often accompaniedby fever, and then the signs and symptoms grad-ually subside as healing takes place.

The degree of mucositis and its onset andcourse vary, depending upon the constitutionof the myeloablative regimen(s) used to preparefor a hematopoietic stem cell (HSC) transplant.2

Similarly, the use of cytarabine (cytosine arabi-noside) in high doses and idarubicin have beenassociated with bacteremia due to oral viridansstreptococci.16,17 Oral mucositis not only pro-vides a portal of entry for these streptococci butalso for Stomatococcus mucilaginosus,18,19 Capno-cytophaga spp.,20–22 the anaerobes Fusobacteriumnecrophorum and Eubacterium spp., Leptotrichiabuccalis,21–23 as well as Candida albicans, Gram-negative bacilli, and possibly also S.epidermidis.25 Active herpes simplex infectioncan exacerbate existing mucositis,26 as well asincreasing the incidence of oral thrush.27

Effect of chemotherapy and irradiation on thegastric acid barrierDyspepsia is sufficiently commonplace forantacids such as H2 receptor antagonists, andeven the proton-pump inhibitors such asomeprazole, to be regularly prescribed. Theresulting lower gastric acidity inadvertentlydestroys the natural barrier, allowing bowelcolonization by oral commensals to ensue.When these bacteria are also only marginallysusceptible to antibiotics used for prophylaxis,their chances of establishing colonization areincreased. This might explain why only aminority of patients who develop viridansstreptococcal bacteremia progress towards sep-sis, the so-called ‘alpha-strep syndrome’, whichis associated with gut toxicity28 and the use ofH2 receptor antagonists.8,17 The sudden appear-ance in the bloodstream of overwhelming num-bers of streptococci from multiple sites ofmucosal damage may provide the mechanismfor inducing shock,29 since the cell wall materialof Gram-positive bacteria is capable of inducingTNF-� and other cytokines.30

Effect of chemotherapy and irradiation on theintestinal tractThe epithelia and connective tissue of the gutare damaged by chemotherapy and irradia-tion,15 and the mucosal barrier (at least that ofthe ileum) is compromised by the loss of the so-called ‘mucous blanket’, which can facilitateepithelial colonization.31 Even standardchemotherapy for acute myeloid leukemiainduces malabsorption and markedly increasesthe risk of candidiasis.32 Besides the damage tothe mucosa, chemotherapy and irradiationimpair gut function and lead to rapid alter-ations in permeability and increased absorptionof sugars.32–36 Perturbed gut function has beenshown to be one of the factors that, togetherwith antibiotic usage and colonization withCandida spp., predisposes patients withleukemia to invasive candidiasis, and alsoappears to be a risk factor for neutropenic ente-rocolitis.4,32 Impaired gut function and integritymay also facilitate translocation, particularly ofGram-negative bacilli such as Ps. aeruginosa,into the bloodstream of patients colonized withthe organism.37

Mucosal barrier injury is determined notonly by the nature of the chemotherapy andirradiation, but also by the accumulation of pro-inflammatory and other cytokines, the trans-location of the resident microflora and theirproducts across the mucosal barrier, exposureto antimicrobial agents that modulate themicroflora, and the origin of the HSC graft.4

Gut toxicity has also been shown to be respons-ible for reduced absorption of fluoro-quinolones,38,39 and has been implicated in theerratic bioavailabililty of the antifungal agentitraconazole.40,41 Finally, a dysfunctional gutwill have a marked effect upon the nutritionalstatus of the patient. Some cytotoxic drugs mayeven exert direct influence on oral and gut flora,since some of these agents have been shown topossess antibacterial activity, and even toenhance the effects of antimicrobial agents.42–47

The protracted diarrhea that often results fromtotal-body irradiation (TBI) leads to a lowermicrobial biomass, which may be more vulner-


able to antimicrobial agents. Aztreonam andimipenem are normally inactivated by feces,but in its absence may retain more of theiractivity.48,49 The ecology of the bowel flora willalso be altered markedly by diarrhea inducedby cytotoxic agents (e.g. cytarabine50), by graft-versus-host disease,51 and by TBI.52 In addition,prior exposure to antimicrobial agents thatinhibit the Gram-positive anaerobic flora of thebowel, such as the penicillins, rifamycin, clin-damycin, erythromycin, and vancomycin, willfurther alter the ecology considerably, since thisis associated with a measurable loss of ‘colo-nization resistance’.53

THE HOST AND THE MICROBIAL WORLD

Environment

Normal daily activity exposes us many times toexogenous microorganisms. Fortunately, thevast majority that are encountered are harm-less. Unless the individual is a carrier, profes-sional pathogens are acquired exclusively fromthe environment – either by direct contact withan infected individual or indirectly by inhalinginfected air, ingesting infected food and drink,or by coming into contact with infected objects.Bottled non-carbonated mineral water has beenidentified as a source of Pseudomonas spp. andStenotrophomonas maltophilia.54 Flowers (both cutand potted) are notorious sources of water- andsoil-borne Gram-negative bacilli, including Ps.aeruginosa. The global nature of floristry, withplants being imported from all over the world,may also increase the risk of infection by exoticspecies and by multiply resistant strains. Freshfruit and vegetables also arrive in supermarketsfrom all corners of the globe, and present a sim-ilar risk. Widespread travel may also open newavenues for acquiring infection, as can leisureand sports activities involving water. Even vis-iting a humble flower show where whirlpoolsare used can increase the risk of legionellosis.55

Increasing concern for the environment hasafforded molds such as Aspergillus fumigatus a

wider domain. This outstanding saprophyte isalready found almost everywhere, includingfireproofing material,56 silage,57 buildingdust,58–61 household dust,62 decaying mattersuch as dead leaves and old rotting furniture,63

compost heaps,64 biocontainers,65 potted plants,and foodstuffs such as ground pepper.66

Aspergillus and Penicillium prevail in theautumn and winter months.67 Other fungi, suchas the phycomycetes Mucor and Rhizopus, arealso ubiquitous saprophytes. Molds, includingAspergillus spp., Penicillium spp., Fusarium spp.,Absidia spp., and Cladosporium spp., are dis-persed in the air during the handling of grain.68

Hence, exposure to microorganisms is a con-stant factor of daily life in both normal hostsand patients receiving antineoplastic therapy.

Microbial diseases of the neutropenicpatient

There are in excess of 900 bacterial speciesknown to inhabit the body surfaces, but only avery few have been reported as causing infec-tion, even in the most immunosuppressedpatients. Infections in neutropenic patientsderive from two principal sources: the endoge-nous flora and the exogenous or environmentalflora. The resident commensal flora found onthe skin and mucosal surfaces contains poten-tial pathogens, but occasionally organisms canalso be acquired exogenously by ingestion ofcontaminated fluids and food or through directcontact. Such organisms may be transient, beingunable to establish a foothold and establish col-onization. Their opportunity to infect is there-fore limited unless there is repeated exposure,prolonged transit, and a ready-made portal ofentry available, such as an ulcer, abrasion, ordirect access via a catheter. Reverse barrier iso-lation, HEPA-filtered air, and a diet of lowmicrobial content all help reduce the chance ofacquiring potential pathogens exogenously. Asa consequence, most pathogens in neutropenicpatients arise from the resident flora inhabitingthe skin, the airways and the alimentary tract.


An extensive review of the epidemiology ofinfections in neutropenic patients is provided inChapter 4. Infections by much less commonbacteria have been attributed to catheter infec-tion by, for example, Tsukamurella pau-rometabolum (Gordona aurantiaca),69 to oralmucositis due to, for example, Capnocytophagaspp.,22 or to both routes (e.g. Stomatococcusmucilaginosus).19 The lesson again seems to bethat the presence of a portal of entry and apotential pathogen both conspire to cause infec-tion.

CHEMOPROPHYLAXIS

The debate about the utility or otherwise ofchemoprophylaxis continues to ebb and flow,although the emphasis has shifted from bacteriato fungi – probably because of prompt empiricantibiotic therapy successfully reducing mortal-ity through the years. The trend towards rely-ing more on evidence than eminence may havelowered the scientific appeal of prophylaxis,but not its intuitive attraction, since it hasbecome very much the norm. The estimation ofrisk and benefit is also in a continuous state offlux, and the concept of cost–effectiveness hastaken hold. In essence, choosing whether or notto give chemoprophylaxis depends uponanswering four basic questions (Table 12.1): Isthere an effective treatment for the infection? Isthe infection serious? Is prophylaxis effective inpreventing the infection? Does prophylaxishave few adverse effects?

Is there an effective treatment for theinfection?

When chemoprophylaxis was first attempted itwas aimed at reducing the morbidity and mor-tality that arose from infections caused byStaphylococcus aureus and the Gram-negativebacilli E. coli, Klebsiella pneumoniae, and Ps.aeruginosa, for which therapy was limited.However, nowadays, effective therapy is avail-

able for these infections if given to the patientpromptly after the onset of fever.70,71

Infections involving viridans streptococcialso have a high rate of cure using standardempiric regimens, even when there is evidenceof bacteremia. However, acute respiratory dis-tress syndrome associated with bacteremia dueto viridans streptococci, usually S. mitis, faresless well, but whether or not adding anotherantimicrobial agent such as penicillin is benefi-cial is uncertain. Instead, the syndrome may bebetter managed by complementing the empiricregimen with a short course of high-dose corti-costeroids.29

Similarly, effective therapy is available fortreating candidiasis, whether localized or dis-seminated.72–75 By contrast, although effectivetreatment is available for the most commonmold infection, invasive aspergillosis, inabilityto recognize the disease soon enough leads to adelay in starting treatment.76–82

Treatment of herpes simplex infection withacyclovir is highly effective, although resistantstrains have been encountered, resulting intherapeutic failure.83–85 Infections due tocytomegalovirus (CMV) pose a similar problemas does aspergillosis insofar as treating estab-lished disease such as CMV pneumonitis isineffective although, unlike aspergillosis, thedisease is invariably restricted to seropositiverecipients of an allogeneic HSC transplant.86

Whilst graft-versus-host disease (GVHD) playsthe greatest role in the development of CMVdisease, CMV viremia is the best predictor of itsdevelopment.87

Is the infection serious?

All infections that occur during neutropenia areregarded as serious, although those caused bycoagulase-negative staphylococci are generallyseen as indolent. Even when the attributablemortality is only marginal, few are prepared torun the risk of doing nothing and adopting await-and-see approach. Instead, empiric treat-ment is given on suspicion of infection and


Tabl

e 12.1

Pro

phyl

axis

or

not?

Infe

ctio

n du

e to

Gra

m-n

egat

ive

baci

llus

Her

pes

sim

plex

P. ca

rini

iC

andi

dasp

p.A

sper

gillu

ssp

p.

Is t

here

an

effe

ctiv

eYe

sYe

sYe

sYe

sYe

str

eatm

ent

for

the

infe

ctio

n?

Is it

a s

erio

us in

fect

ion?

Yes

No

Yes

Yes

Yes

Is t

he p

roph

ylax

is e

ffec

tive

Yes

Yes

Yes

Yes

No

in p

reve

ntin

g th

e in

fect

ion?

Doe

s th

e pr

ophy

laxi

s ha

veYe

saYe

sYe

sbYe

sN

oc

few

adv

erse

effec

ts?

Reg

imen

Cip

roflo

xaci

nAc

yclo

vir

Co-

trim

oxaz

ole

Fluc

onaz

ole

Non

e

aPr

ovid

ed a

ntib

iotic

res

ista

nce

is u

ncom

mon

.b

If sk

in r

ash

is c

onsi

dere

d un

impo

rtan

t.cIn

terf

eres

with

cyc

losp

orin

.

even complemented with other empiric agentssuch as a glycopeptide or amphotericin B.Hence it is a moot point whether or not infec-tion is serious in terms of increased morbidityand mortality, since clinicians act as though it isand this has become a standard of care.

Is prophylaxis effective in preventingbacterial infection?

Antimicrobial agents were first given to cancerpatients and those with hematologic malig-nancy in the early 1970s to try to reducethe infectious complications arising duringneutropenia.88–91 Non-absorbable regimens,particularly gentamicin plus vancomycin plusnystatin (GVN), were employed to sterilize thegut,92,93 but this proved futile – the compliancewas erratic and the risk of selecting resistantbacteria was actually higher. This wasexplained by the antibiotics destroying theanaerobic flora of the alimentary tract to suchan extent that much fewer exogenous Gram-negative bacilli were required to establish colo-nization than was the case under normalconditions.94 It therefore seemed more appro-priate to aim for partial or selective decontami-nation rather than attempt completesterilization of the body sites. This approachwas known by several acronyms, includingSDD (selective decontamination of the diges-tive tract), PAD (partial antimicrobial deconta-mination), and even by SAM (selectiveantimicrobial modulation).53 All the regimensemployed were targeted against the undesir-able Gram-negative bacilli found within theresident flora while preserving the microfloraresponsible for colonization resistance.However, it was only when Hughes and col-leagues95 reported that children given co-trimoxazole to prevent infection due toPneumocystis carinii also suffered fewerepisodes of bacterial infections that the demiseof total gut decontamination was assured andthe adopted standard approach became one ofselective oral antimicrobial prophylaxis.

Co-trimoxazoleCo-trimoxazole seems the ideal agent because itprevents infection with P. carinii, is effectiveagainst a wide range of respiratory pathogens,including Streptococcus pneumoniae andHaemophilus influenzae, and offers protectionagainst both S. aureus and the enteric Gram-negative bacilli.95 Small placebo-controlled trialsindicated that co-trimoxazole was beneficial asselective prophylaxis,96–102 as did comparativestudies.90,103–109 However, the risk of resistanceemerging and causing bacteremia was appar-ent,96,110 as was co-trimoxazole’s lack of activityagainst Ps. aeruginosa, necessitating the additionof colistin.108

The fluoroquinolonesThe introduction of the newer fluoroquinolones– norfloxacin,111 ciprofloxacin,112 ofloxacin,113

and pefloxacin114,115 – further expanded therange of agents available for prophylaxis.Moreover, compliance is better, side-effectsare lower with these drugs than with co-trimoxazole, and they do not seem to have anydeleterious effect on hematopoeisis.116 Theirspectrum of activity includes the most commoncauses of infection due to Gram-negative bacilli,S. aureus, and many of the coagulase-negativestaphylococci.117,118 The viridans streptococciand enterococci are only marginally suscepti-ble, if at all. Although only one of the manyprophylactic trials with the fluoroquinoloneshas been placebo-controlled,111 it is clear thatthey all provide better protection against infec-tion due to Gram-negative bacilli than does co-trimoxazole.113,114,119–127 However, among thedrugs studied so far, only ciprofloxacin offersthe most complete protection against Gram-negative bacilli, including Ps. aerugi-nosa.114,123,124,126

Norfloxacin was the first to be made avail-able, and was quickly followed byciprofloxacin, then pefloxacin and ofloxacin. Bythe time some of the newer fluoroquinolonesbecame available, practice patterns haddeveloped, and clinicians fell into two opposingcamps: those who employed a fluoroquinolone,


even though the evidence was far from conclu-sive that they were effective, and those whoabhorred the idea, fearing the emergence ofresistance. It is indeed surprising that there hasbeen no satisfactory placebo-controlled, double-blind trial of sufficient size to prove significantbenefit with enough power. The argumentsfor and against these drugs rest on a series ofsmall, mostly single-center, studies. Moreover,a variety of endpoints were employed, includ-ing the prevention of infection due to Gram-negative bacilli, infection per se, and even fever.A meta-analysis of 12 comparative and con-trolled studies is the best evidence available,and shows that whilst prophylaxis with a fluo-roquinolone is more effective than eitherplacebo or alternative regimens in preventinginfection due to Gram-negative bacilli,128 therewas no measurable benefit in terms of bac-teremia as a whole. There were more episodesof bacteremia due to Gram-positive cocci andno evidence of less mortality. Also, moreepisodes of fever of undetermined originoccurred with more than 80% of patients stillreceiving empiric therapy with broad-spectrumantibiotics.

Special cases – viridans streptococciRecognition that the oral viridans streptococciwere in the ascendancy prompted studies inwhich a fluoroquinolone was complementedby a penicillin,129–131 amoxycillin,132,133 vanco-mycin,134 or roxithromycin.135 Whilst there wasless bacteremia due to these streptococci, resis-tance developed against penicillin,128 and therewas no measurable impact on the incidence offever and other infective complications such aspneumonia and septic shock.135

Special cases – catheter-related infectionsIt might seem logical to attempt to preventthese infections by providing antibiotic cover-age during insertion or by instilling antibioticsthrough the lumen to apply an antibiotic block.A single bolus intravenous injection of 400 mgteicoplanin resulted in a lower incidence ofexit site and tunnel infections and catheter-

related Gram-positive bacteremia, particularlyamong patients who were already neutropenicwhen the Hickman catheter was inserted.136 Ashort course of three injections of 500 mg van-comycin perioperatively resulted in fewer ofthe central venous catheters in the vancomycinprophylaxis group becoming infected withGram-positive microorganisms than in the con-trol group.137 Giving vancomycin twice daily ata dose of 15 mg/kg from two days before HSCtransplant until resolution of neutropenia oruntil the first episode of fever prevented bac-teremia and focal infection, resulting in fewerdays with fever, and hence fewer days ofempiric antibiotic therapy.138 However, becausethe clinical course of coagulase-negativestaphylococcal infections is relatively benign,treatment with a glycopeptide is only war-ranted if there is a tunnel infection or coexistentthrombophlebitis.139,140 The enthusiasm for giv-ing vancomycin prophylactically has sincegiven way to a realization that there aresignificant hazards associated with the practicedue to the emergence of vancomycin resistanceamong the enterococci species (particularlyEnterococcus faecium) and the real fear that thetranposon responsible could cross over to moredangerous pathogens, such as S. aureus.141 Theconcern is such that most authorities advisespecifically against using vancomycin for pro-phylaxis even in neutropenic patients whomight benefit.142

Others have introduced antibiotic blocks toprevent contamination of the catheter lumenvia the hub.143 Vancomycin at the low concen-tration of 25 µg/ml in heparin has been used –apparently effectively.144 The justification forthis approach is that catheters can easilybecome infected with a nosocomial strain of S.epidermidis, since several clones survive on thesame hematology ward for quite some time bycolonizing patients’ skin.145 However, whetherthe putative benefit of using such blocks isgenerally found remains to be confirmed bystudies done on a much larger scale. The samecan be said of the claims made for impregnatedcatheters. If the patient’s skin is the principal


source of the staphylococcus and infectionsoriginate from the exit site, then coating theexternal surface of the catheter might providesome benefit.146,147 However, there are as yet toofew data to support this practice. In the absenceof good evidence, it seems more prudent tomanage catheters carefully, use them only asabsolutely necessary, remove them as soon asthey are no longer needed, and treat catheter-related infections with antibiotics when there isan obvious tunnel infection, evidence ofinfected thrombosis, or Gram-positive bac-teremia that persists for longer than 3–4 days;this should always be accompanied by swiftremoval of the device.

Does antibacterial prophylaxis have few side-effects?

Although bacteremia due to Gram-negativebacilli is reduced by prophylaxis, this results inmore unexplained fevers and more bacteremiasdue to Gram-positive cocci, which are eithernaturally resistant to the fluoroquinolones, suchas in the case of viridans streptococci, or haveapparently acquired resistance, as do manycoagulase-negative staphylococci, and this hasbeen observed during treatment withciprofloxacin.148 This observation is explicable,since ciprofloxacin is both excreted in the sweatand induces resistance amongst skin staphylo-cocci within a few days of exposure.149,150 Thesestaphylococci are commonly resistant totobramycin, co-trimoxazole, and methicillin,and may also be resistant to ciprofloxacin.151

Resistance among E. coli to the fluoro-quinolones norfloxacin,152 ofloxacin,153 andpefloxacin154 has also been observed. A recentstudy done in Taiwan may have inadvertentlyrevealed why, since 3 of 12 allogeneic HSCtransplant recipients given ciprofloxacin forprophylaxis developed bacteremia due to resis-tant E. coli during neutropenia, resulting in twodeaths from septic shock.155 Instead of using themore usual dose of 1000 or 1500 mg/day, theauthors opted for 500 mg/day, believing it to

be sufficient for prophylaxis. However, thedose may well have been too low to achieveadequate systemic levels for two reasons.Firstly, absorption of this drug is impairedfollowing chemotherapy, when mean peak con-centrations are reduced by half to 2.0 mg/l justaround the time neutropenia has reached itsnadir and mucositis its peak and shortly beforebacteremia occurs.39 Secondly, the fluoro-quinolones bind to feces,156 making even lessdrug available to inhibit the E. coli. The lowerdose of 500 mg/day would achieve even lowerpeak concentrations, reducing local and sys-temic efficacy. The impairment of drug absorp-tion might be explained by the fact that thistakes place in the upper part of the intestinaltract, i.e. duodenum and jejunum,157 which isalso where most damage occurs to the mucosalbarrier.33 Hence, apart from anything else, gutmucositis appears to result in altered drug dis-position. Lower absorption has also been notedfor ofloxacin.38

Bacteremia due to Gram-positive cocci suchas viridans streptococci might occur simply as aresult of plasma levels of the drug being lowerthan the minimum inhibitory concentration.117

Under these conditions, other multiply resistantGram-positive cocci, including staphylococciand enterococci, would also have a selectiveadvantage.123,148,151,158–161 Combine this selectivepressure with colonization of the mucosal sur-faces by resistant bacteria and severe mucositis,and all the necessary ingredients for infectionare present.

Generally, co-trimoxazole is well tolerated,but it induces a skin rash in 15% of patientsreceiving remission induction therapy, particu-larly when cytarabine is included.162 Moreover,the onset of skin rash frequently coincides withallergy to allopurinol and cytarabine, as well asoral mucositis, nausea, and diarrhea (e.g. afterremission induction of acute myeloidleukemia). This leads to less compliance andinterruption, if not to total discontinuation ofprophylaxis. Co-trimoxazole has also been asso-ciated with delayed hematopoiesis in bone mar-row transplant recipients.163 Fluoroquinolones


induce fewer side-effects, and although rashesdo occur, they do so much less frequently thanwith co-trimoxazole and seldom lead to prema-ture discontinuation. Side-effects are thereforeseldom the reason for stopping prophylaxis.Patients who discontinue taking medication forreasons other than the onset of fever or side-effects mostly do so simply because they areunable to swallow because of severe oralmucositis. Logically, the drug could have beencontinued parenterally when the oral route wasnot feasible, but this has never been donebecause such a move was perceived as therapyrather than prophylaxis. The fact that the dosesof both co-trimoxazole and the fluoro-quinolones were therapeutic was simply over-looked. Clearly, the psychological barrierimposed by the understanding of ‘prophylaxis’and ‘therapy’ was too great to overcome. Eventhe elegant study by Bow and colleagues164 thatshowed ciprofloxacin to be safe as an effectivestrategy to reduce the amount and duration of

empiric therapy directed against Gram-negative organisms in febrile neutropenicpatients did not attempt to switch to parenteraladministration to deliver the bioequivalentamount of drug. So, despite over a decade ofuse, not only is the correct dosage of fluoro-quinolones not yet known, but neither is theproper route of administration. Since there hasbeen no trial of fluoroquinolone that allowedcontinuation of the drug parenterally whendrug concentrations fell, we shall never knowwhether or not failure to maintain adequatelevels was the reason for breakthrough infec-tion. A more rational approach to prophylaxiswould, in fact, be to administer ciprofloxacinorally at a dose equivalent to the optimum ther-apeutic dose (e.g. 1500 mg/day), and to con-tinue prophylaxis parenterally should thepatient be unable to swallow the tablet or whenblood levels fall below a certain threshold(Figure 12.2). Oral treatment can then resumeonce mucositis subsides.


0.1

1

10

00 77 1414 2121 2828 3535 4242 4949 5656 6363DaysDays

39

40

41

Parenteral Oral

36

3738

Result DiagnosisProphylaxis given to

protect againstimminent infection or

death

Only complementprophylaxis whenthere is a problem

Microbiologically orclinically definedInfection?

Problem

Action Switch to parenteral when

• Compliance is poor• Blood levels are inadequate

Oral

Gra

nulo

cyte

cou

nt (

� 1

09/l

)Te

mpe

ratu

re (

°C)

Figure 12.2 Prophylaxis as the backbone for managing infections.

Should antibacterial prophylaxis be used?

Besides the problems of resistance, furtherdoubts have been cast upon the practice ofantibacterial prophylaxis in a recent review,165

which suggested that antibacterial prophylaxishas outstayed its welcome since timely admin-istration of empiric broad-spectrum antibioticshas been very successful in reducing mortalityattributed to Gram-negative infections. In addi-tion, there have been no controlled clinical trialsfulfilling the criteria required for firm evidence(high-power, prospective, randomized, blinded,multicenter studies), although there are stillreports being published that appear to showthat specific populations such as HSC trans-plant recipients do benefit.166 If put to a jury ofother than hematologists, a verdict of ‘notproven’ for prophylaxis would probably bereturned. But, despite the lack of evidence, pro-phylaxis with fluoroquinolones will still con-tinue to be given – certainly to recipients ofHSC transplants, and to other patients receiv-ing the sort of intensive chemotherapy thatinduces protracted neutropenia and injures themucosal barrier, if only because the fear of bac-teremia due to Gram-negative bacilli is beingassuaged. That being the case, it would still bedesirable to conduct a formal trial to help estab-lish the correct dose and means of administra-tion and identify those patients who wouldgain the most benefit, although it is doubtfulwhether or not a placebo would now be con-sidered ethical. However, this reservation ismore than compensated for by having the safe-guard of modern empiric regimens.

Is prophylaxis effective in preventing fungalinfection?

Prophylaxis of invasive fungal infectious dis-eases (IFIDs) has had a checkered history.Initially, studies were small, uncontrolled, donein single centers, and invariably inconclusive.Moreover, efficacy measures were indirect andsoft, relying on such entities as number of

febrile days, use of empirical amphotericin Band so on. Lately, there have been a few studiesthat did meet the rigorous requirements of arandomized, controlled trial, although theresults were not universally applicable. Forinstance, there are data that show fluconazoleeffective in preventing disseminated candidia-sis in recipients of allogeneic HSC trans-plant,167,168 and possibly in those being treatedfor acute myeloid leukemia,169,170 but there is no agreement about the optimum dose. InNorth America, 400 mg/day of fluconazole isused, whereas other investigators elsewherehave achieved similar results with only100–200 mg/day,171–176 calling into question theneed for the higher dose of fluconazole.Moreover, the drug offers no protection againstmold infections, including aspergillosis.

A recent meta-analysis showed so little bene-fit from prophylaxis that the authors concludedthere was insufficient evidence for its use.177

This certainly seemed to be the case in terms oflong-term survival, but doubts still remainabout the prevention of IFID per se. The studiesincluded in the meta-analysis also employeddifferent criteria for IFID. Thus, the fact remainsthat a study necessary to prove a clear benefit ofprophylaxis over placebo at reasonable cost hasyet to be done, although this now seems highlyunlikely, since few will feel comfortable offer-ing allogeneic HSC transplant recipients aplacebo.

The goal of prophylaxis must also be clearand explicit. Preventing death due to IFID isnot the same as preventing IFID itself. This isnot simply an academic issue. Death isregarded as a ‘hard’ fact, and, as such, is thepreferred endpoint for meta-analysis, healtheconomics, and by decision-makers. In contrast,clinicians are much more interested in prevent-ing morbidity, and thus, by implication, redu-cing mortality. Although all studies ofprophylaxis will take account of deaths, the pri-mary endpoint is the occurrence of IFID – hencethe importance of standard definitions, which,fortunately, we now have at our disposal.

Apart from selecting natively resistant non-


albicans Candida, including C. glabrata and C.krusei, prophylaxis with fluconazole might alsolead to the development of superinfection byAspergillus fumigatus.178

A further problem will be in obtaining agree-ment about how long such prophylaxis shouldbe maintained, since there are good reasons forcontinuing it for as long as the patient is at anyrisk of IFID, i.e. 1–2 years post HSC transplant,or even for life.179 Other problems will arise inchoosing the agent for study. The azole antifun-gal agents are the first choice, but the benefit offluconazole is limited to preventing IFID due toCandida spp.; confidence in itraconazole is stilllacking, and other potential alternatives are stillto complete phase III trials. A lipid form ofamphotericin B would be worth studying, butfor the inordinate acquisition costs and thereluctance to use the same drug for prophylaxisthat one would choose for treatment.

IFID – What’s in a name?Intuitively, prevention of fungal infectionseems the simplest approach, especially sincethe consequences of missing a case can be disas-trous – but it is not. Firstly, candidiasis andaspergillosis account for most IFIDs in allo-geneic HSC transplant recipients and in neu-tropenic patients in general,77,180,181 but these areradically different entities. Candidiasis almostinvariably develops in patients already colo-nized with the offending yeast,9,182,183 but thereare no internationally accepted criteria fordefining colonization and no standard methodsfor determining it, although there is someagreement that the same yeast should be recov-ered from the same site on two separate occa-sions or from at least two different sites on thesame occasion.9,182,183 The Candida spp., particu-larly C. albicans, form part of the normal resi-dent flora of the gastrointestinal tract, and areeasily detected in the oral cavity by culturingthe mucosal surfaces of the mouth or an oralgargle. The gut is less straightforward to sam-ple, since stools are not always available andthere is a general reluctance to take a rectalswab because of the risk of bleeding. There is

also evidence that catheter-related infection dueto Candida is preceded by colonization by thesame species around the exit site, particularlywhere Candida parapsilosis is concerned,184,185 butroutine surveillance swabs are not done.Differences in sampling alone account for muchof the different sensitivities for detecting colo-nization. Cultures may be done to obtain a pre-sent/absent result (qualitative), or they can beperformed in such a manner as to provide anestimate of numbers. Various different mediaare used, with some simply relying solely onstandard bacteriological media, whilst othersinclude one specially designed to recover fungi.A medium supplemented with an antibiotic tosuppress growth of bacteria to enhance detec-tion of yeast without allowing any discrimina-tion between the different species of yeast isalso inferior to the newer differential mediasuch as CHROMagar, which facilitate differen-tiation of yeasts reliably.186 Not surprisingly, notwo methods yield the same results and manywill not even bear comparison. It is therefore nosurprise that many centers simply do not makeany attempt to detect colonization, deeming it awaste of time and resources. Yet there are datato show a clear association between carriage ofyeast and subsequent infection, and, just asimportantly, the lack of colonization is highlypredictive of IFID due to Candida spp. beingunlikely.182,183,187–189

There have been many attempts at suppress-ing colonization by yeasts, including adminis-tering the polyenes by mouth as well as givingeach of the azole drugs available, but successvaries and is unpredictable.176,190,191 Both nys-tatin and amphotericin B have been givenorally to prevent fungal infection, althoughthere has never been a randomized controlledtrial. The practice seemed to have evolved fromseveral considerations. Nystatin was includedin the early decontamination regimens92,93,192 tosuppress the overgrowth by C. albicans thatoccurred as a result of altering the ecology ofthe gut microflora. The introduction of so-calledselective decontamination in Europe and theavailability of amphotericin B lozenges, tablets,


and suspension led to its inclusion for the samepurpose.53,100,120 Doses vary widely, and up to4 � 106 U/day nystatin is given, with variablesuccess and compliance. The suspension hasbeen most widely used, but tablets may bemore palatable and effective.193 Amphotericin Bmay be effective against candidiasis when atleast 200 mg/day is given orally, although atleast 8 times this amount appears necessary tosuppress gut colonization,194 and, once again,compliance is inconsistent.

Miconazole and clotrimazole suppress oralcolonization, and may both be more effectivethan placebo, but the data are sparse.183,199

Ketoconazole, fluconazole, and itraconazole areall effective in suppressing colonization, butthis is not translated into reducing infec-tion.177,194 By contrast, fluconazole not onlyreduces colonization and superficial infectionbut also lowers the risk of developing dissemi-nate candidiasis significantly.167,168,177,194,195

Similarly, itraconazole appears effective in sup-pressing yeast colonization196–198 IFID due toCandida spp., but not for that due to Aspergillusspp.77,171,199 Also, no study has shown an appre-ciable reduction in overall mortality, and therehas been little impact on fungal deaths.177,197

Moreover, because compliance will be variableduring mucositis and while the patient is suf-fering the side-effects of ablative treatment,both drugs may be given parenterally if treat-ment is to be continued, which, up untilrecently, limited the choice to fluconazolebecause of the lack of a parenteral form of itra-conazole.

In stark contrast to candidiasis, aspergillosisonly develops once the fungus has establisheditself in the airways, and Aspergillus spp. arenever normal residents of the upper respiratorytract. Thus, screening individuals at risk by tak-ing specimens for culture is doomed from thestart. Certain specific molds such as A. nigercan be detected beforehand from nasal secre-tions,200 and, as such, can identify those patientsat risk, but this seems to be of very limitedvalue and only to have been true in the contextof an outbreak. In theory, at least, it should be

possible to prevent the acquisition of moldssuch as Aspergillus spp. simply by supplyingHEPA-filtered air. Many centers also askpatients to inhale amphotericin B administeredin spray or nebulized form to destroy any sporesthat might have been inhaled or be lingering inthe airways. Whilst single centers that espousethe practice are convinced that it is effective,201–205

a large multicentre study failed to confirm this.206

Besides, even if this strategy were to work, itwould only help to suppress nosocomial IFID,and would have little, if any, influence on thecourse of disease established before admission.Ultimately, the best prevention againstaspergillosis in neutropenic patients is still thecontrol of the underlying disease, with subse-quent return of normal marrow function andresolution of neutropenia207 and control ofGVHD in allogeneic HSC transplant recipients.76

Special case – Pneumocystis cariniiThis erstwhile protozoon has now been reclassi-fied as a fungus. Infections seldom develop,and when they do occur, it is usually in patientswho are not protected by co-trimoxazole pro-phylaxis,208,209 such as following allogeneic HSCtransplant in a setting of chronic steroid use orGVHD. Co-trimoxazole 960 mg given oncedaily or three times a week or 960 mg bid twiceweekly is still the first choice for prophylaxis,with nebulized pentamidine providing a safer,though less effective alternative when there isintolerance.209,210 Dapsone is not considered suf-ficiently effective, since significantly higherrates of P. carinii pneumonitis have beenreported.211 When disease is apparent, co-trimoxazole given parenterally at the higherdose of 120 mg/kg/day in divided doses for 21days remains first choice for therapy, irrespec-tive of the severity.209

Does antifungal prophylaxis have few side-effects?

Unlike polyenes, all the azoles have the poten-tial drawback of selecting the resistant species


such as C. krusei and C. glabrata, and even, inthe long term, of inducing resistance in C. albi-cans.212 However, reality does not quite bear thisout, since in one study, recovery of C. glabratasteadily increased equally to around 30% inpatients whether or not fluconazole was given,whilst C. krusei were isolated exclusively frompatients given the drug.213 In another study of asimilar patient population, the reverse was true,with more C. glabrata being isolated frompatients given fluconazole, and C. krusei beingrecovered equally.214

The azole drugs are all relatively safe, andtreatment seldom needs to be stopped prema-turely because of actual side-effects. Howeverthe nausea and vomiting experienced bypatients following cytotoxic chemotherapy issufficiently troublesome that one in every fouror five patients stop taking itraconazolealtogether.199,215 This makes the argument all themore compelling for switching to parenteraltreatment – at least until all gastrointestinal tox-icities are resolved and normal intake isresumed. By contrast, patients seem to toleratefluconazole much better, and physicians arealready used to switching from oral to par-enteral therapy and back again.

Fluconazole has the least potential to interactwith other drugs, although interactions withcyclosporin have been reported when flucono-zole doses higher than 200 mg/day aregiven.216,217 This can be remedied by close moni-toring of cyclosporin concentrations and renalfunction.217 The increases in cyclosporin andtacrolimus concentrations seen are modestwhen these drugs are given orally, and may bea result of fluconazole inhibition of gut metabo-lism, resulting in greater absorption.218

By contrast ketoconazole, and to a lesserextent itraconazole are more potent inhibitorsof cytochrome P450 3A4 which metabolizedrugs such as cyclosporin and tacrolimus.Consequently a reduction in the dosage of theimmune suppressants is required and is evenseen by some as a beneficial interaction becauseit results in lower drug consumption and,hence, cost.219

Should antifungal prophylaxis be used?

Although most would wish for a broad-spectrum agent that can be given both orallyand parenterally, an ideal candidate has yet tobecome available. The presence of colonizationshould be used to select those patients at higherrisk of IFID due to Candida spp., especially sincethere has been no attempt to look at the poten-tial benefit of antifungal prophylaxis only inpatients who actually carry the yeast on theirmucosal surfaces. Similarly, the type of cyto-toxic chemotherapy used should be factoredinto a decision about whether or not to giveprophylaxis.

Given the low incidence of aspergillosis inthe recent studies of prophylaxis, an alternativeapproach seems to be warranted. The exampleof ganciclovir is instructive here, since the drugwas considered too toxic to be given to everyallogeneic HSC transplant recipient at risk ofCMV infection. Investigators decided to turn tothe laboratory, which had a test at its disposalfor detecting the pp65 antigen of CMV. This ledto the creative solution of screening for the anti-gen in patients at risk and only treating withganciclovir when there was a significant rise inantigen titers, which was assumed to indicateimminent infection. This pre-emptive approachmight prove worthwhile for patients at highrisk of developing pulmonary aspergillosis,since there is an enzyme-linked immunosorbentassay (ELISA) test for detecting galactomannanantigen; patients’ plasma can be monitored twoto three times weekly, and, if antigen isdetected, treatment could be started on theassumption that disease is imminent.82,220 A fur-ther refinement would be to take a high-resolution computed tomography scan of thelung, and only start treatment if this showedabnormalities consistent with an infectiveprocess. One strategy for managing aspergillo-sis would be to define the risk as shown inFigure 12.3 and then choose the type of treat-ment. When the risk is remote, prophylaxiswould only be indicated for special groups,such as allogeneic HSC transplant recipients.


During neutropenia, any patient at risk whohas developed fever that persists for 3–5 daysbut remains unexplained and is refractory toempiric antibacterial therapy would be con-sidered a candidate for empiric antifungaltherapy. At the other extreme, the few cases forwhich the diagnosis is proven by detecting fun-gus in the tissue obtained from the site of infec-tion would clearly be candidates for vigorousspecific therapy. Those patients who are at risk,and have some clinical manifestation of diseasesuch as a pulmonary infiltrate, would be con-sidered probable cases, provided that the labo-ratory has detected the presence of the fungusin blood or other body fluids and secretionsdirectly by culture or microscopy or indirectlyby detecting antigen or by polymerase chainreaction (PCR).

Is prophylaxis effective in preventing viralinfection?

Herpes simplexAcyclovir has been used routinely for HSCtransplant recipients for the two decades since itwas shown to be very effective in preventinglesions after reactivation of viral infection butless so in halting viral shedding.221 Reactivationtends to occur when pretransplantation herpessimplex IgG titers exceed 10 000 units using theELISA test. This method has been used by someto decide whether or not to initiate prophylaxisin patients being treated with conventional-dosechemotherapy for malignant diseases. In con-trast, prophylaxis forms an integral part of theprophylaxis schedule of HSC transplant recipi-ents.222 The discussion has focused more on forhow long treatment should be continued andwhether or not there is any measurable effect on


• At risk• Signs and symptoms• Mycology

• At risk• Febrile,

unresponsiveto antibiotics

-7 0 7 14 21 28 35 42 49 56 63-14Days after HSC transplant

Treatment

Diseaselikelihood

tissueTissue +

Pre-emptive

Probable disease

Specific

Proven

Empiric

Possible

Prophylaxis

Remote

CultureCulture +GMGM +

PCRPCR +

37383940

36

0.1

1.0

10.0 Cannot be excluded

• At risk• No signs• No symptoms

Gra

nulo

cyte

cou

nt(�

10

9/l

)Te

mpe

ratu

re(°

C)

Figure 12.3 Management of a patient at risk for aspergillosis.

the development of CMV disease.222 Opinionsare still divided, with some concluding categori-cally that acyclovir has no impact whatsoeveron the frequency of CMV infections,223 or on theincidence of CMV disease and CMV-relatedmortality – at least when ganciclovir is giveneither at engraftment or for CMV pp65 antigene-mia.224 Others claim the contrary, with a 20%survival advantage one year from transplant,provided that acyclovir is given intravenouslyin high doses (500 mg/m2 three times a day) 5days before transplant to 30 days after trans-plant, followed by 800 mg given orally fourtimes a day for 6 more months.225 There is alsosome dissent about general prophylaxis for HSCtransplant recipients, since restricting the drugto treating active infection would prove morecost-effective,226 and many centers administer itto every patient no matter what their serologicalstatus. Restricting prophylaxis to only thosewho are seropositive would clearly be justified,as might the application of a more stringent pol-icy, such as using a titer higher than 10 000 as athreshold to institute therapy. Few clinicianswould be happy nowadays to await infectionbefore acting, since this would be seen as lead-ing to unnecessary suffering, especially sinceacyclovir has so few side-effects.

It has also been suggested that herpes sim-plex infection might account for over 90% ofepisodes of otherwise unexplained persistentfever.227 The role that herpes plays in causingfever seems to be clear, since the number ofnon-fungal oral infections was reduced and theonset of fever was delayed by the use of acy-clovir prophylaxis, although the duration offever, use of antibacterial treatment, occurrenceof bacteremia, and need for systemic antifungaltherapy were not affected.228 An earlier studyshowed that oral prophylaxis was associatedwith a reduction of all microbiologicallydefined infections, although the drug was onlygiven during remission-induction therapy.229

CytomegalovirusAcyclovir given intravenously in high doseshas been advocated as prophylaxis, but the

issue is still controversial. A survey conductedby the European Group for Blood and MarrowTransplantation of 70 centers in 20 countriesshowed that prophylaxis was used in 59 centers(84%), with high doses of acyclovir beingemployed in 42 centers and ganciclovir in only7.230 Fifty four (77%) of the 70 centers whoresponded to the survey used prophylaxis.However, therapy was started early by 53 cen-ters (76%), mostly on the basis of detection ofviremia or CMV antigen in the blood, withCMV pneumonia being treated by using a com-bination of ganciclovir and intravenousimmunoglobulin in 64 (91%) centers. Pro-phylactic therapy with ganciclovir is generallygiven from the time of engraftment up to 3–4months post-transplantation to all patients atrisk of CMV disease, whilst the pre-emptiveapproach is reserved for those with evidentCMV infection. Each strategy has advantagesand disadvantages, and there is no evidence forthe superiority of one over the other, since theoverall survival is the same and the incidence ofdeath from CMV disease is similar.231

Fortunately, laboratories are now equippedwith the means of detecting active CMV infec-tion before disease becomes apparent, therebyallowing ganciclovir to be given pre-emptively,which has significantly decreased the incidenceof disease and mortality following allogeneicHSC transplantation.232 There are several assaysavailable for quantifying human CMV in theblood of immunocompromised patients, pro-viding the only reliable indication of the degreeof dissemination of CMV infection. These testsall detect CMV in peripheral blood leukocytesby culture, pp65 antigenemia, or quantitativePCR. The threshold values above whichCMV-related clinical symptoms are likely toappear have been estimated to be 10 or morefor viremia, 100 for antigenemia and 1000genome equivalents respectively.233 However,as with all diagnostic tests, each test differs inits ability to predict a positive or negative riskof developing CMV disease.234 Moreover, theunderlying prevalence of CMV disease will dif-fer from one study to the next, depending upon


the risk factors present and the degree of bias inselecting the population. For instance, T-cell-depleted stem cell grafts result in less GVHD,which, in turn, lowers the risk of CMV disease.Each strategy has its strengths and weaknesses,and there is no evidence that the overallsurvival differs or that the incidence of deathfrom CMV disease is different.235

Does antiviral prophylaxis have any side-effects?

Acyclovir has proven remarkably safe withresistance occurring rarely. In contrast, routineuse of ganciclovir is considered too toxic to bejustified for prophylaxis, except in very high-risk patients, since the risk of harm is con-sidered to outweigh the perceived benefit,mainly because of persistent neutropenia.87

Should antiviral prophylaxis be used?

There is more evidence in favor of giving acy-clovir as prophylaxis against herpes simplexinfection than there is for waiting until infectiondevelops to institute therapy. With CMV, quitethe opposite is true. Patients at risk should bemonitored for reactivation, which, when itoccurs, should provide the trigger for startingtreatment pre-emptively with ganciclovir.

CONCLUSIONS

Since not all neutropenic patients are the same,and some risk factors are already known, perhapsit is now time to apply this knowledge prospec-tively. For instance, we know that patients givenselective oral antimicrobial prophylaxis whodevelop mucositis and bacteremia due to Gram-positive cocci and who are colonized with Candidaare at higher risk of candidiasis than are otherpatients. We also know that cytotoxic regimensprone to induce damage to the gut mucosa alsoplace the patient at greater risk for developing the

same disease. Hence, treatment with cytarabine,colonization with C. albicans, and prophylaxiswith a fluoroquinolone could be used to selectcandidates for fluconazole or itraconazole pro-phylaxis. On the other hand, there seems no pointin attempting prophylaxis against Gram-negativeinfection when neutropenia is likely to be shorterthan 7 days or neutrophils are unlikely to dropbelow 0.5 � 109/l and mucositis is likely to bemild or absent, since co-trimoxazole and the fluo-roquinolones both require at least a week beforethe bacilli are effectively suppressed.235–237

Similarly, seronegativity for herpes simplexshould be used to preclude acyclovir prophylaxis,whilst allogeneic HSC transplant recipients are atgreater risk of aspergillosis if they have beengiven methotrexate for prophylaxis againstGVHD and have experienced other nosocomialinfections before the diagnosis of pneumonia.238

Prolonged neutropenia is also a major risk factorfor aspergillosis.239

From being as much a matter of faith as sci-ence, chemoprophylaxis is slowly evolving to amore rationale basis for its use as our under-standing of the prognostic factors for infectionduring neutropenia becomes more comprehen-sive and we start to apply the tools at our dis-posal to identify those most at risk. Theprinciples of evidence-based medicine are nowbeing adopted more readily to help us movefrom conviction to fact. Equally important, theHippocratic principle of at least doing no harmif one cannot do any good has acquired new lifeas we embrace the two sides of chemoprophy-laxis – namely a drug may still be effective, butnot good enough to outweigh the toxicities.Also, we now accept the need to have a betterestimate of the prevalence of an infectious dis-ease in our own particular patient group and todecide on the size of risk reduction that we con-sider important before deciding on whether toact on the evidence. The costs of implementingprophylaxis or not also have to be considered inthe broadest sense, since effective drugs tendalso to be expensive and should not be squan-dered, whilst losing a patient because of parsi-mony is usually a false economy, winning only


opprobrium. Equally, these decisions cannot betaken alone or in isolation, since they need toinvolve laboratories, pharmacists, and nurses,as well as clinical managers, more directly indecision making and in melding the evidencewith experience to achieve optimum results.

REFERENCES

1. Bodey GP, Buckley M, Sathe YS, Freirich EJ,Quantitative relationships between circulatingleucocytes and infection in patients with acuteleukaemia. Ann Intern Med 1966; 64: 328–40.

2. Wardley AM, Jayson GC, Swindell R et al,Prospective evaluation of oral mucositis inpatients receiving myeloablative conditioningregimens and haemopoietic progenitor rescue.Br J Haematol 2000; 110: 292–9.

3. Epstein JB, Schubert MM, Oral mucositis inmyelosuppressive cancer therapy. Oral SurgOral Med Oral Pathol Oral Radiol Endod 1999; 88:273–6.

4. Blijlevens NM, Donnelly JP, De Pauw BE,Mucosal barrier injury: biology, pathology, clin-ical counterparts and consequences of intensivetreatment for haematological malignancy: anoverview. Bone Marrow Transplant 2000; 25:1269–78.

5. Roth RR, James WD, Microbial ecology of theskin. Ann Rev Microbiol 1988; 42: 441–64.

6. Frandsen EVG, Pedrazzoli V, Kilian M, Ecologyof viridans streptococci in the oral cavity andpharynx. Oral Microbiol Immunol 1991; 6: 129–33.

7. Heimdahl A, Nord CE, Colonization of theoropharynx with pathogenic microorganisms –a potential risk factor for infection in compro-mised patients. Chemioterapia 1985; 4: 186–91.

8. Elting LS, Bodey GP, Keefe BH, Septicemia andshock syndrome due to viridans streptococci: acase-control study of predisposing factors. ClinInfect Dis 1992; 14: 1201–7.

9. Guiot HF, Fibbe WE, van ’t Wout JW, Risk fac-tors for fungal infection in patients with malig-nant hematologic disorders: implications forempirical therapy and prophylaxis. Clin InfectDis 1994; 18: 525–32.

10. Van der Waaij D, The ecology of the humanintestine and its consequences for overgrowthby pathogens such as Clostridium difficile. Ann

Rev Microbiol 1989; 43: 69–87.11. Vollaard EJ, Clasener HAL, Colonization resis-

tance. Antimicrobial Agents Chemother 1994; 38:409–14.

12. McGuire DB, Altomonte V, Peterson DE et al,Patterns of mucositis and pain in patientsreceiving preparative chemotherapy and bonemarrow transplantation. Oncol Nurs Forum1993; 20: 1493–502.

13. Sable CA, Donowitz GR, Infections in bonemarrow transplant recipients. Clin Infect Dis1994; 18: 273–81.

14. Sonis ST, Mucositis as a biological process: anew hypothesis for the development ofchemotherapy-induced stomatotoxicity. OralOncol 1998; 34: 39–43.

15. Sonis ST, Peterson RL, Edwards LJ et al,Defining mechanisms of action of interleukin-11on the progression of radiation-induced oralmucositis in hamsters. Oral Oncol 2000; 36:373–81.

16. Donnelly JP, Dompeling EC, Meis JF, de PauwBE, Bacteremia due to oral viridans streptococciin neutropenic patients with cancer: cytostaticsare a more important risk factor than antibacte-rial prophylaxis. Clin Infect Dis 1995; 20: 469–70.

17. Bochud PY, Calandra T, Francioli P, Bacteremiadue to viridans streptococci in neutropenicpatients: a review. Am J Med 1994; 97: 256–64.

18. Weers-Pothoff G, Nováková IRO, Donnelly JP,Muytjens HL, Bacteraemia caused byStomatococcus mucilaginosus in a granulocy-topenic patient with acute lymphocyticleukaemia. Neth J Med 1989; 35: 143–6.

19. McWhinney PHM, Kibbler CC, Gillespie SH etal, Stomatococcus mucilaginosus: an emergingpathogen in neutropenic patients. Clin Infect Dis1992; 14: 641–6.

20. Bilgrami S, Bergstrom SK, Peterson DE et al,Capnocytophaga bacteremia in a patient withHodgkin’s disease following bone marrowtransplantation: case report and review. ClinInfect Dis 1992; 14: 1045–9.

21. Baquero F, Fernandez J, Dronda F et al,Capnophilic and anaerobic bacteremia in neu-tropenic patients: an oral source. Rev Infect Dis1990; 12(Suppl 2): S157–60.

22. Shenep JL, Combination and single-agentempirical antibacterial therapy for febrile cancerpatients with neutropenia and mucositis. NCIMonographs 1990; 990: 117–22.


23. Reig M, Baquero F, Garcia-Campello M, Loza E,Leptotrichia buccalis bacteremia in neutropenicchildren. J Clin Microbiol 1985; 22: 320–1.

24. Vidal AM, Sarria JC, Kimbrough RC 3rd, KeungYK, Anaerobic bacteremia in a neutropenicpatient with oral mucositis. Am J Med Sci 2000;319: 189–90.

25. Wade JC Schimpff SC, Newman KA, WiernikPH, Staphylococcus epidermidis: an increasingcause of infection in patients with granulocy-topenia. Ann Intern Med 1982; 97: 503–8.

26. Seto BG, Kim M, Wolinsky L et al, Oral mucosi-tis in patients undergoing bone marrow trans-plantation. Oral Surg Oral Med Oral Pathol 1985;60: 493–7.

27. Beattie G, Whelan J, Cassidy J et al, Herpes sim-plex virus, Candida albicans and mouth ulcers inneutropenic patients with non-haematologicalmalignancy. Cancer Chemother Pharmacol 1989;25: 75–6.

28. van der Lelie H, van Ketel RJ, von dem BorneAE et al, Incidence and clinical epidemiology ofstreptococcal septicemia during treatment ofacute myeloid leukemia. Scand J Infect Dis 1991;23: 163–8.

29. Dompeling EC, Donnelly JP, Raemaekers JM,De Pauw BE, Pre-emptive administration ofcorticosteroids prevents the development ofARDS associated with Streptococcus mitis bac-teremia following chemotherapy with high-dose cytarabine. Ann Hematol 1994; 69: 69–71.

30. Bone RC, Gram-positive organisms and sepsis.Arch Intern Med 1994; 154: 26–34.

31. Walker RI, Brook I, Costerton JW et al, Possibleassociation of mucous blanket integrity withpostirradiation colonization resistance. RadiatRes 1985; 104: 346–57.

32. Bow EJ, Loewen R, Cheang MS, Schacter B,Invasive fungal disease in adults undergoingremission-induction therapy for acute myeloidleukemia: the pathogenetic role of theantileukemic regimen. Clin Infect Dis 1995; 21:361–9.

33. Bow EJ, Loewen R, Cheang MS et al, Cytotoxictherapy-induced D-xylose malabsorption andinvasive infection during remission-inductiontherapy for acute myeloid leukemia in adults. JClin Oncol 1997; 15: 2254–61.

34. Fegan C, Poynton CH, Whittaker JA, The gutmucosal barrier in bone marrow transplanta-tion. Bone Marrow Transplant 1990; 5: 373–7.

35. Johansson JE, Ekman T, Gastro-intestinal toxic-ity related to bone marrow transplantation: dis-ruption of the intestinal barrier precedes clinicalfindings. Bone Marrow Transplant 1997; 19:921–5.

36. Keefe DM, Cummins AG, Dale BM et al, Effectof high-dose chemotherapy on intestinal per-meability in humans. Clin Sci 1997; 92: 385–9.

37. Tancrede CH, Andremont AO, Bacterial trans-location and Gram-negative bacteremia inpatients with hematological malignancies. JInfect Dis 1985; 152: 99–103.

38. Brown NM, White LO, Blundell EL et al,Absorption of oral ofloxacin after cytotoxicchemotherapy for haematological malignancy. JAntimicrob Chemother 1993; 32: 117–22.

39. Johnson EJ, MacGowan AP, Potter MN et al,Reduced absorption of oral ciprofloxacin afterchemotherapy for haematological malignancy. JAntimicrob Chemother 1990; 25: 837–42.

40. Prentice AG, Warnock DW, Johnson SAN et al,Multiple dose pharmacokinetics of an oral solu-tion of itraconazole in patients receivingchemotherapy for acute myeloid leukaemia. JAntimicrob Chemother 1995; 36: 657–63.

41. Prentice AG, Warnock DW, Johnson SAN et al,Multiple dose pharmacokinetics of an oral solu-tion of itraconazole in autologous bone marrowtransplant recipients. J Antimicrob Chemother1994; 34: 247–52.

42. Bodet III CA, Jorgensen JH, Drutz DJ,Antibacterial activities of antineoplastic agents.Antimicrob Agents Chemother 1985; 28: 437–9.

43. Neuman M, The antimicrobial activity of non-antibiotics – interactions with antibiotics. ActaPathologica, Microbiologica et ImmunologicaScandinavia 1992; 100(Suppl 30): 15–23.

44. Jacobs JY, Michel J, Sacks T, Bactericidal effectof combinations of antimicrobial drugs andantineoplastic antibiotics against Staphylococcusaureus. Antimicrob Agents Chemother 1979; 15:580–6.

45. Michel J, Jacobs JY, Sacks T, Bactericidal effectof combinations of antimicrobial drugs andantineoplastic antibiotics against Gram-negative bacilli. Antimicrob Agents Chemother1979; 16: 761–6.

46. Moody MR, Morris MJ, Young VM et al, Effect oftwo cancer chemotherapeutic agents on theantibacterial activity of three antimicrobial agents.Antimicrob Agents Chemother 1978; 14: 737–42.


47. Bergström P, Grankvist K, Henriksson R,Interaction between antibiotics and antineoplas-tic drugs on antibacterial activity in vitro: estra-mustine phosphate sensitizes pneumococci toamikacin. Int J Oncol 1994; 4: 43–9.

48. Welling GW, Groen G, Inactivation of aztre-onam by faecal supernatants of healthy volun-teers as determined by HPLC. J AntimicrobChemother 1989; 24: 805–10.

49. Welling GW, Slootmaker-vandermeulen C,Jansen GJ, Inactivation of imipenem by faecalfractions from human volunteers and the effectof clavulanate and cilastatin. J AntimicrobChemother 1993; 31: 617–19.

50. Peters WG, Willemze R, Colly LP, Guiot HFL,Side effects of intermediate- and high-dosecytosine arabinoside in the treatment of refrac-tory or relapsed acute leukaemia and non-Hodgkin’s lymphoma. Neth J Med 1987; 30:64–74.

51. Guiot HF, Biemond J, Klasen E et al, Proteinloss during acute graft-versus-host disease:diagnostic and clinical significance. Eur JHaematol 1987; 38: 187–96.

52. Callum JL, Brandwein JM, Sutcliffe SB et al,Influence of total body irradiation on infectionsafter autologous bone marrow transplantation.Bone Marrow Transplant 1991; 8: 245–51.

53. Donnelly JP, Selective decontamination of thedigestive tract and its role in antimicrobial pro-phylaxis. J Antimicrob Chemother 1993; 31: 813–29.

54. Wilkinson FH, Kerr KG, Bottled water as asource of multi-resistant Stenotrophomonas andPseudomonas species for neutropenic patients.Eur J Cancer Care (Engl) 1998; 7: 12–14.

55. Hoepelman IM, Legionella epidemic in theNetherlands. Ned Tijdschr Geneeskd 1999; 143:1192–6 (in Dutch).

56. Aisner J, Schimpff SC, Bennett JE et al,Aspergillus infections in cancer patients.Association with fireproofing materials in anew hospital. JAMA 1976; 235: 411–12.

57. Bui AM, Germaud P, Normand De LaTranchade M, Touranchet A, Silage and allergicbronchopulmonary aspergillosis. Arch Mal ProfMed Trav 1994; 55: 335–7.

58. Ansorg R, van den Boom R, von Heinegg EH,Rath PM, Association between incidence ofAspergillus antigenemia and exposure to con-struction works at a hospital site. ZentralblBakteriol 1996; 284: 146–52.

59. Dewhurst AG, Cooper MJ, Khan SM et al,Invasive aspergillosis in immunosuppressedpatients: potential hazard of hospital buildingwork. BMJ 1990; 301: 802–4.

60. Streifel AJ, Lauer JL, Vesley D et al, Aspergillusfumigatus and other thermotolerant fungi gener-ated by hospital building demolition. ApplEnviron Microbiol 1983; 46: 375–8.

61. Hopkins CC, Weber DJ, Rubin RH, InvasiveAspergillus infection: possible non-ward com-mon source within the hospital environment. JHosp Infect 1989; 13: 19–25.

62. Korpi A, Pasanen AL, Pasanen P, Kalliokoski P,Microbial growth and metabolism in housedust. Int Biodeterioration and Biodegradation 1997;40: 19–27.

63. Streifel AJ, Stevens PP, Rhame FS, In-hospitalsource of airborne Penicillium species spores. JClin Microbiol 1987; 25: 1–4.

64. Marsh PB, Millner PD, Kla JM, A guide to therecent literature on aspergillosis as caused byAspergillus fumigatus, a fungus frequently foundin self-heating organic matter. Mycopathologia1979; 69: 67–81.

65. Reiss J, Molds in containers with biologicalwastes. Microbiol Res 1995; 150: 93–8.

66. Nolard N, Links between risks of aspergillosisand environmental contamination. Review ofthe literature. Pathol Biol Paris 1994; 42(7):706–10 (in French).

67. Beaumont F, Kauffman HF, van der Mark TH etal, Volumetric aerobiological survey of conidialfungi in the North-East Netherlands. I. Seasonalpatterns and the influence of metereologicalvariables. Allergy 1985; 40: 173–80.

68. Lappalainen S, Nikulin M, Berg S et al,Fusarium toxins and fungi associated with han-dling of grain on eight Finnish farms. AtmosEnviron 1996; 30: 3059–65.

69. Lai KK, A cancer patient with central venouscatheter-related sepsis caused by Tsukamurella-paurometabolum (Gordona-aurantiaca). Clin InfectDis 1993; 17: 285–7.


71. Pizzo PA, Current concepts: fever in immuno-compromised patients. N Engl J Med 1999; 341:893–900.


72. Goa KL, Barradell LB, Fluconazole: an updateof its pharmacodynamic and pharmacokineticproperties and therapeutic use in major superfi-cial and systemic mycoses in immunocompro-mised patients. Drugs 1995; 50: 658–90.

73. Voss A, de Pauw BE, High-dose fluconazoletherapy in patients with severe fungal infec-tions. Eur J Clin Microbiol Infect Dis 1999; 18:165–74.

74. Anaissie E, Bodey GP, Kantarjian H et al,Fluconazole therapy for chronic disseminatedcandidiasis in patients with leukemia and prioramphotericin B therapy. Am J Med 1991; 91:142–50.

75. Akova M, Akalin HE, Uzun O et al, Efficacy offluconazole in the treatment of upper gastroin-testinal candidiasis in neutropenic patients withcancer – factors influencing the outcome. ClinInfect Dis 1994; 18: 298–304.

76. Paterson DL, Singh N, Invasive aspergillosis intransplant recipients. Medicine (Baltimore) 1999;78: 123–38.

77. Warnock DW, Fungal infections in neutropenia:current problems and chemotherapeutic con-trol. J Antimicrob Chemother 1998; 41(Suppl D):95–105.

78. Severens JL, Donnelly JP, Meis J et al, Two strat-egies for managing invasive aspergillosis: adecision analysis. Clin Infect Dis 1997; 25:1148–54.

79. De Pauw BE, Practical modalities for preven-tion of fungal infections in cancer patients. Eur JClin Microbiol Infect Dis 1997; 16: 32–41.

80. Warnock DW, Fungal complications of trans-plantation: diagnosis, treatment and preven-tion. J Antimicrob Chemother 1995; 36: 73–90.

81. Winston DJ, Prophylaxis and treatment of infec-tion in the bone marrow transplant recipient.Curr Clin Top Infect Dis 1993; 13: 293–321.

82. Verweij PE, Donnelly JP, De Pauw BE, MeisJFGM, Prospects for the early diagnosis of inva-sive aspergillosis in the immunocompromisedhost. Rev Med Microbiol 1996; 7: 105–13.

83. Darville JM, Ley BE, Roome AP, Foot AB,Acyclovir-resistant herpes simplex virus infec-tions in a bone marrow transplant population.Bone Marrow Transplant 1998; 22: 587–9.

84. Engel JP, Englund JA, Fletcher CV, Hill EL,Treatment of resistant herpes simplex viruswith continuous-infusion acyclovir. JAMA 1990;263: 1662–4.

85. Snoeck R, Andrei G, Gerard M et al, Successfultreatment of progressive mucocutaneous infec-tion due to acyclovir- and foscarnet-resistantherpes simplex virus with (S)-1-(3-hydroxy-2 - p h o s p h o n y l m e t h o x y p r o p y l ) c y t o s i n e(HPMPC). Clin Infect Dis 1994; 18: 570–8.

86. Tsinontides AC, Bechtel TP, Cytomegalovirusprophylaxis and treatment following bone mar-row transplantation. Ann Pharmacother 1996; 30:1277–90.

87. Wood MJ, Viral infections in neutropenia – cur-rent problems and chemotherapeutic control. JAntimicrob Chemother 1998; 41(Suppl D): 81–93.

88. Schimpff SC, Infection prevention during pro-found granulocytopenia: new approaches to ali-mentary canal microbial suppression. AnnIntern Med 1980; 93: 358–61.

89. Van Der Waaij D, The colonization resistance ofthe digestive tract of man and animals.Zentralblatt für Bakteriologie 1979; Suppl 7:155–61.

90. Enno A, Catovsky D, Darrell J et al, Co-trimoxazole for prevention of infection in acuteleukaemia. Lancet 1978; ii: 395–7.

91. Guiot HF, Furth R, Partial antibiotic decontami-nation. BMJ 1977; i: 798–800.

92. Bender JF, Schimpff SC, Young VM et al, Roleof vancomycin as a component of oral nonab-sorbable antibiotics for microbial suppression inleukemic patient. Antimicrob Agents Chemother1979; 15: 455–60.

93. Levi JA, Vincent PC, Jennis F et al, Prophylacticoral antibiotics in the management of acuteleukaemia. Med J Austr 1973; 1: 1025–9.

94. Van Der Waaij D, Berghuis-De Vries JM,Lekkerkerk-Van Der Wees JEC, Colonizationresistance of the digestive tract of individualmice. J Hygiene 1971; 69: 404–11.

95. Hughes WT, Kuhn S, Chaudhary S et al,Successful chemoprophylaxis for Pneumocystiscarinii pneumonitis. N Engl J Med 1977; 297:1419–26.

96. The EORTC International AntimicrobialTherapy Project Group, Trimethoprim–sul-famethoxazole in the prevention of infection inneutropenic patients. J Infect Dis 1984; 150:372–9.

97. Gurwith MJ, Brunton JL, Lank BA et al, Aprospective controlled investigation of prophy-lactic trimethoprim–sulfamethoxazole in hospi-talized granulocytopenic patients. Am J Med


1979; 66: 248–56.98. Kauffman CA, Liepman MK, Bergman AG,

Mioduszewski J, Trimethoprim–sulphamethox-azole prophylaxis in neutropenic patients:reduction of infections and effect on bacterialand fungal flora. Am J Med 1983; 74: 599–607.

99. Henry SA, Armstrong D, Kempin S et al, Oraltrimethoprim/sulfamethoxazole in attempt toprevent infection after induction chemotherapyfor acute leukaemia. Am J Med 1984; 77: 663–6.

100. Dekker A, Rozenberg-Arska M, Sixma JJ,Verhoef J, Prevention of infection by trimetho-prim–sulfamethoxazole plus amphotericin B inpatients with acute nonlymphocytic leukaemia.Ann Intern Med 1981; 95: 555–9.

101. Weiser B, Lange M, Fialk MA et al, Prophylactictrimethoprim–sulfamethoxazole during consoli-dation chemotherapy for acute leukemia: a con-trolled trial. Ann Intern Med 1981; 95: 436–8.

102. Gualtieri RJ, Donowitz GR, Kaiser DL et al,Double-blind randomized study of prophylactictrimethoprim/sulfamethoxazole in granulocy-topenic patients with hematologic malig-nancies. Am J Med 1983; 74: 934–40.

103. Starke ID, Donnelly JP, Catovsky D et al, Co-trimoxazole alone for the prevention of bacter-ial infection in patients with acute leukaemia.Lancet 1982; i: 5–6.

104. Wade JC, Schimpff SC, Hargadon MT et al, Acomparison of trimethoprim–sulfamethoxazoleplus nystatin with gentamicin plus nystatin inthe prevention of infections in acute leukemia.N Engl J Med 1981; 304: 1057–62.

105. Bow EJ, Louie TJ, Riben PD et al, Randomizedcontrolled trial comparing trimethoprim/sul-famethoxazole and trimethoprim for infectionprophylaxis in hospitalized granulocytopenicpatients. Am J Med 1984; 76: 223–33.

106. Kurrle E, Dekker AW, Gaus W et al, Preventionof infection in acute leukaemia: a prospectiverandomized study of the efficacy of two differ-ent drug regimens for antimicrobial prophy-laxis. Infection 1986; 14: 226–32.

107. Malarme M, Meunier-Carpentier F, KlasterskyJ, Vancomycin plus gentamicin and cotrimoxa-zole for prevention of infections in neutropeniccancer patients (a comparative, placebo-controlled pilot study). Eur J Cancer Clin Oncol1981; 17: 1315–22.

108. Rozenberg-Arska M, Dekker A, Verhoef J,Colistin and trimethoprim–sulfamethoxazole

for the prevention of infection in patients withacute nonlymphocytic leukaemia: decrease inthe emergence of resistant bacteria. Infection1983; 11: 167–9.

109. Watson JG, Jameson B, Powles RL et al, Co-trimoxazole versus non-absorbable antibioticsin acute leukaemia. Lancet 1982; i: 6–7.

110. Wilson JM, Guiney DG, Failure of oraltrimethoprim–sulfamethoxazole prophylaxis inacute leukemia: isolation of resistant plasmidsfrom strains of enterobacteriaciae causing bac-teremia. N Eng J Med 1982; 306: 16–20.

111. Karp JE, Merz WG, Hendriksen C et al, Oralnorfloxacin for prevention of Gram-negativebacterial infections in patients with acuteleukemia and granulocytopenia: a randomized,double-blind, placebo-controlled trial. AnnIntern Med 1987; 106: 1–7.

112. Rozenberg-Arska M, Dekker AW, Prevention ofbacterial and fungal infections in granulocy-topenic patients. In: Antimicrobial Agents Annual(Verhoef P, ed). Amsterdam: Elsevier, 1987:471–81.

113. Arning M, Wolf HH, Aul C et al, Infection pro-phylaxis in neutropenic patients with acuteleukaemia – a randomized, comparative studywith ofloxacin, ciprofloxacin and cotrimoxazole/colistin. J Antimicrob Chemother 1990;26(Suppl D): 137–42.

114. D’Antonio D, Iacone A, Fioritoni G et al,Comparison of norfloxacin and pefloxacin inthe prophylaxis of bacterial infection in neu-tropenic cancer patients. Drugs Exp Clin Res1992; 18: 141–6.

115. Meunier F, Prevention of infections in neu-tropenic patients with pefloxacin. J AntimicrobChemother 1990; 26(Suppl B): 69–73.

116. De Pauw BE, De Witte T, Raemaekers JMM etal, Impact of ciprofloxacin and co-trimoxazoleon bone marrow growth as measured byCFU-GM and BFU-E assays. In: Ciprofloxacin.Microbiology–Pharmacokinetics, Clinical Experience– Proceedings of 6th Mediterranean Congress ofChemotherapy, Taormina, Italy, 1988: 94–7.

117. King A, Phillips I, The comparative in-vitroactivity of pefloxacin. J Antimicrob Chemother1986; 17(Suppl B): 1–10.

118. Reeves DS, The effect of quinolone antibacteri-als on the gastrointestinal flora compared withthat of other antibacterials. J AntimicrobChemother 1986 18(Suppl D): 89–102.


119. Dekker AW, Rozenberg-Arska M, Verhoef J,Infection prophylaxis in acute leukemia: a com-parison of ciprofloxacin with trimethoprim–sul-famethoxazole and colistin. Ann Intern Med1987; 106: 7–12.

120. Donnelly JP, Maschmeyer G, Daenen S,Selective oral antimicrobial prophylaxis for theprevention of infection in acute leukaemia –ciprofloxacin versus co-trimoxazole plus col-istin. The EORTC–Gnotobiotic Project Group.Eur J Cancer 1992; 28: 873–8.

121. Warren RE, Wimperis JZ, Baglin TP et al,Prevention of infection by ciprofloxacin in neu-tropenia. J Antimicrob Chemother 1990;26(Suppl F): 109–23.

122. Bow EJ, Louie TJ, Emerging role of quinolonesin the prevention of Gram-negative bacteremiain neutropenia cancer patients and in the treat-ment of enteric infections. Clin Invest Med 1989;12: 61–8.

123. Del Favero A, Menichetti F, The new fluori-nated quinolones for antimicrobial prophylaxisin neutropenic cancer patients. Eur J Cancer1993; 29A: 52–6.

124. The GIMEMA Infection Program, GruppoItaliano Malattie Ematologiche Malignedell’Adulto, Prevention of bacterial infection inneutropenic patients with hematologic malig-nancies. A randomized, multicenter trial com-paring norfloxacin with ciprofloxacin. AnnIntern Med 1991; 115: 7–12.

125. Brodsky AL, Minissale CJ, Melero MJ, AvalosJCS, Prophylactic use of fluoroquinolones inneutropenic patients. Medicina – Buenos Aires1993; 53: 401–7.

126. D’Antonio D, Piccolomini R, Iacone A et al,Comparison of ciprofloxacin, ofloxacin andpefloxacin for the prevention of the bacterialinfection in neutropenic patients with haemato-logical malignancies. J Antimicrob Chemother1994; 33: 837–44.

127. Jansen J, Cromer M, Akard L et al, Infectionprevention in severely myelosuppressedpatients – a comparison between ciprofloxacinand a regimen of selective antibiotic modula-tion of the intestinal flora. Am J Med 1994; 96:335–41.


129. Guiot HF, van der Meer JW, van den Broek PJet al, Prevention of viridans-group streptococcalsepticemia in oncohematologic patients: a con-trolled comparative study on the effect of peni-cillin G and cotrimoxazole. Ann Hematol 1992;64: 260–5.

130. Zinner SH, Calandra T, Meunier F et al,Reduction of fever and streptococcal bacteremiain granulocytopenic patients with cancer – atrial of oral penicillin V or placebo combinedwith pefloxacin. JAMA 1994; 272: 1183–9.

131. Broun ER, Wheat JL, Kneebone PH et al,Randomized trial of the addition of Gram-positive prophylaxis to standard antimicrobialprophylaxis for patients undergoing autologousbone marrow transplantation. Antimicrob AgentsChemother 1994; 38: 576–9.

132. Gluckman E, Roudet C, Hirsch I et al,Prophylaxis of bacterial infections after bonemarrow transplantation. A randomizedprospective study comparing oral broad-spectrum nonabsorbable antibiotics (van-comycin–tobramycin–colistin) to absorbableantibiotics (ofloxacin–amoxicillin). Chemotherapy1991; 1: 33–8.

133. Fanci R, Leoni F, Bosi A et al, Chemo-prophylaxis of bacterial infections in granulocy-topenic patients with ciprofloxacin vs.ciprofloxacin plus amoxicillin. J Chemother 1993;5: 119–23.

134. Archimbaud E, Guyotat D, Maupas J et al,Pefloxacin and vancomycin vs. gentamicin, col-istin sulphate and vancomycin for preventionof infections in granulocytopenic patients: arandomised double-blind study. Eur J Cancer1991; 27: 174–8.

135. Kern WV, Hay B, Kern P et al, A randomizedtrial of roxithromycin in patients with acuteleukemia and bone marrow transplant recipi-ents receiving fluoroquinolone prophylaxis.Antimicrob Agents Chemother 1994; 38: 465–72.

136. Lim SH, Smith MP, Machin SJ, Goldstone AH,A prospective randomized study of prophylac-tic teicoplanin to prevent early Hickmancatheter-related sepsis in patients receivingintensive chemotherapy for haematologicalmalignancies. Eur J Haematol Suppl 1993; 54:10–13.

137. Vassilomanolakis M, Plataniotis G, Koumakis Get al, Central venous catheter-related infectionsafter bone marrow transplantation in patients


with malignancies: a prospective study withshort-course vancomycin prophylaxis. BoneMarrow Transplant 1995; 15: 77–80.

138. Attal M, Schlaifer D, Rubie H et al, Preventionof Gram-positive infections after bone marrowtransplantation by systemic vancomycin: aprospective, randomized trial. J Clin Oncol 1991;9: 865–70.

139. Engelhard D, Elishoov H, Strauss N et al,Nosocomial coagulase negative staphylococcalinfections in bone marrow transplantationrecipients with central vein catheter: a 5 yearprospective study. Transplantation 1996; 61:430–4.

140. Dompeling EC, Donnelly JP, Deresinski SC etal, Early identification of neutropenic patientsat risk of Gram-positive bacteraemia and theimpact of empirical administration of van-comycin. Eur J Cancer 1996; 32A: 1332–9.

141. Edmond MB, Wenzel RP, Pasculle AW,Vancomycin-resistant Staphylococcus aureus:perspectives on measures needed for control.Ann Intern Med 1996; 124: 329–34.

142. Recommendations of the Hospital InfectionControl Practices Advisory Committee (HIC-PAC), Recommendations for preventing thespread of vancomycin resistance. Morb MortalWkly Rep 1995; 44(Rr-12): 1–13.

143. Kentos A, Struelens MJ, Thys JP, Antibiotic-locktechnique for the treatment of central venouscatheter infections. Clin Infect Dis 1996; 23:418–19.

144. Carratala J, Niubo J, Fernandez-Sevilla A et al,Randomized, double-blind trial of an antibiotic-lock technique for prevention of Gram-positivecentral venous catheter-related infection in neu-tropenic patients with cancer. Antimicrob AgentsChemother 1999; 43: 2200–4.

145. Nouwen JL, van Belkum A, de Marie S et al,Clonal expansion of Staphylococcus epidermidisstrains causing Hickman catheter-related infec-tions in a hemato-oncologic department. J ClinMicrobiol 1998; 36: 2696–702.

146. Maki DG, Stolz SM, Wheeler S, Mermel LA,Prevention of central venous catheter-relatedbloodstream infection by use of an antiseptic-impregnated catheter. A randomized, con-trolled trial. Ann Intern Med 1997; 127: 257–66.

147. Raad I, Darouiche R, Dupuis J et al, Centralvenous catheters coated with minocycline andrifampin for the prevention of catheter-related

colonization and bloodstream infections. A ran-domized, double-blind trial. The Texas MedicalCenter Catheter Study Group. Ann Intern Med1997; 127: 267–74.

148. Kotilainen P, Nikoskelainen J, Huovinen P,Emergence of ciprofloxacin-resistant coagulase-negative staphylococcal skin flora in immuno-compromised patients receiving ciprofloxacin. JInfect Dis 1990; 161: 41–4.

149. Høiby N, Johansen HK, Ciprofloxacin in sweatand antibiotic resistance. Lancet 1995; 346: 1235.

150. Høiby N, Jarløv JO, Kemp M et al, Excretion ofciprofloxacin in sweat and multiresistantStaphylococcus epidermidis. Lancet 1997; 349:167–9.

151. Hedin G, Hambraeus A, Multiply antibiotic-resistant Staphylococcus epidermidis in patients,staff and environment – a one-week survey in abone marrow transplant unit. J Hosp Infect 1991;17: 95–106.

152. Carratala J, Fernandez Sevilla A, Tubau F et al,Emergence of quinolone-resistant Escherichiacoli bacteremia in neutropenic patients withcancer who have received prophylactic nor-floxacin. Clin Infect Dis 1995; 20: 557–60.

153. Kern WV, Andriof E, Oethinger M et al,Emergence of fluoroquinolone-resistant Escheri-chia coli at a cancer center. Antimicrob AgentsChemother 1994; 38: 681–7.

154. Cometta A, Calandra T, Bille J, Glauser MP,Escherichia coli resistant to fluoroquinolones inpatients with cancer and neutropenia. N Engl JMed 1994; 330: 1240–1.

155. Yeh SP, Hsueh EJ, Yu MS et al, Oralciprofloxacin as antibacterial prophylaxis afterallogeneic bone marrow transplantation: a reap-praisal. Bone Marrow Transplant 1999; 24:1207–11.

156. Nord CE, Effect of quinolones on the humanintestinal microflora. Drugs 1995; 49(Suppl 2):81–5.

157. Staib AH, Beermann D, Harder S et al,Absorption differences of ciprofloxacin alongthe human gastrointestinal tract determinedusing a remote-control drug delivery device(HF-capsule). Am J Med 1989; 87: 66S–69S.

158. Giuliano M, Pantosti A, Gentile G et al, Effectson oral and intestinal microfloras of norfloxacinand pefloxacin for selective decontamination inbone marrow transplant patients. AntimicrobAgents Chemother 1989; 33: 1709–13.


159. Schaberg DR, Dillon WI, Terpenning MS et al,Increasing resistance of enterococci tociprofloxacin. Antimicrob Agents Chemother 1992;36: 2533–5.

160. Hillery SJ, Reisslevy EA, Increasing cipro-floxacin resistance in MRSA. Med J Austr 1993;158: 861.

161. Oppenheim BA, Hartley JW, Lee W, Burnie JP,Outbreak of coagulase-negative staphylococcusresistant to ciprofloxacin in a leukemia unit.BMJ 1989; 299: 294–7.

162. Verhagen C, Stalpers LJ, De Pauw BE, HaanenC, Drug-induced skin reactions in patients withacute non-lymphocytic leukemia. Eur J Haematol1987; 38: 225–30.

163. Schey SA, Kay HEM, Myelosuppression com-plicating co-trimoxazole prophylaxis after bonemarrow transplantation. Br J Haematol 1984; 56:179–80.

164. Bow EJ, Loewen R, Vaughan D, Reducedrequirement for antibiotic therapy targetingGram-negative organisms in febrile, neu-tropenic patients with cancer who are receivingantibacterial chemoprophylaxis with oralquinolones. Clin Infect Dis 1995; 20: 907–12.

165. Kerr KG, The prophylaxis of bacterial infectionsin neutropenic patients. J Antimicrob Chemother1999; 44: 587–91.

166. Engels EA, Ellis CA, Supran SE et al, Earlyinfection in bone marrow transplantation:quantitative study of clinical factors that affectrisk. Clin Infect Dis 1999; 28: 256–66.


168. Slavin MA, Osborne B, Adams R et al, Efficacyand safety of fluconazole prophylaxis for fungalinfections after marrow transplantation – aprospective, randomized, double-blind study. JInfect Dis 1995; 171: 1545–52.

169. Winston DJ, Chandrasekar PH, Lazarus HM etal, Fluconazole prophylaxis of fungal infectionsin patients with acute leukemia. Results of arandomized placebo-controlled, double-blind,multicenter trial. Ann Intern Med 1993; 118:495–503.

170. Rotstein C, Bow EJ, Laverdiere M et al,Randomized placebo-controlled trial of flucona-zole prophylaxis for neutropenic cancer

patients: benefit based on purpose and intensityof cytotoxic therapy. The Canadian FluconazoleProphylaxis Study Group. Clin Infect Dis 1999;28: 331–40.

171. Morgenstern GR, Prentice AG, Prentice HG etal, A randomized controlled trial of itracona-zole versus fluconazole for the prevention offungal infections in patients with haematologi-cal malignancies. UK Multicentre AntifungalProphylaxis Study Group. Br J Haematol 1999;105: 901–11.

172. Ehninger G, Schuler HK, Sarnow E,Fluconazole in the prophylaxis of fungal infec-tion after bone marrow transplantation. Mycoses1996; 39: 259–63.

173. Alangaden G, Chandrasekar PH, Bailey E et al,Antifungal prophylaxis with low-dose flucona-zole during bone marrow transplantation. BoneMarrow Transplant 1994; 14: 919–24.

174. Ellis ME, Clink H, Ernst P et al, Controlledstudy of fluconazole in the prevention of fungalinfections in neutropenic patients with haema-tological malignancies and bone marrow trans-plant recipients. Eur J Clin Microbiol Infect Dis1994; 13: 3–11.

175. Akiyama H, Mori S, Tanikawa S et al, Fluconazoleversus oral amphotericin-B in preventing fungalinfection in chemotherapy-induced neutropenicpatients with haematological malignancies.Mycoses 1993; 36: 373–8.

176. Menichetti F, Delfavero A, Martino P et al,Preventing fungal infection in neutropenicpatients with acute leukemia: fluconazole com-pared with oral amphotericin B. Ann Intern Med1994; 120: 913–18.

177. Gotzsche PC, Johansen HK, Meta analysis ofprophylactic or empirical antifungal treatmentversus placebo or no treatment in patients withcancer complicated by neutropenia. BMJ 1997;314: 1238–44.

178. Meis JF, Donnelly JP, Hoogkamp-Korstanje JA,De-Pauw BE, Aspergillus fumigatus pneumoniain neutropenic patients during therapy with flu-conazole for infection due to Candida species.Clin Infect Dis 1993; 16: 734–5.


180. Bodey G, Bueltmann B, Duguid W et al, Fungal


infections in cancer patients – an internationalautopsy survey. Eur J Clin Microbiol Infect Dis1992; 11: 99–109.

181. Denning DW, Marinus A, Cohen J et al, AnEORTC multicentre prospective survey of inva-sive aspergillosis in haematological patients:diagnosis and therapeutic outcome. EORTCInvasive Fungal Infections Cooperative Group.J Infect 1998; 37: 173–80.

182. Martino P, Girmenia C, Micozzi A et al,Prospective study of Candida colonization, useof empiric amphotericin B and development ofinvasive mycosis in neutropenic patients. Eur JClin Microbiol Infect Dis 1994; 13: 797–804.

183. Martino P, Girmenua C, Venditti M et al,Candida colonization and systemic infection inneutropenic patients. Cancer 1989; 64: 2030–4.

184. De Pauw BE, Raemaekers JMM, Schattenberg T,Donnelly JP, Empirical and subsequent use ofantibacterial agents in the febrile neutropenicpatient. J Intern Med 1997; 242: 69–77.

185. Weems JJ Jr, Candida parapsilosis: epidemiology,pathogenicity, clinical manifestations, andantimicrobial susceptibility. Clin Infect Dis 1992;14: 756–66.

186. Odds FC, Bernaerts R, CHROMagar candida, anew differential isolation medium for presump-tive identification of clinically important can-dida species. J Clin Microbiol 1994; 32: 1923–9.

187. Wenzel RP, Nosocomial candidemia: risk fac-tors and attributable mortality. Clin Infect Dis1995; 20: 1531–4.

188. Bow EJ, Invasive fungal infections in patientsreceiving intensive cytotoxic therapy for cancer.Br J Haematol 1998; 101(Suppl 1): 1–4.

189. Richet HM, Andremont A, Tancrede C et al,Risk factors for candidemia in patients withacute lymphocytic leukemia. Rev Infect Dis 1991;13: 211–15.

190. Johansen HK, Gotzsche PC, Problems in thedesign and reporting of trials of antifungalagents encountered during meta-analysis.JAMA 1999; 282: 1752–9.

191. De Gregorio M, Lee W, Ries C, Candida infec-tions in patients with acute leukemia: ineffec-tiveness of nystatin prophylaxis andrelationship between oropharyngeal and sys-temic candidiasis. Cancer 1982; 50: 2780–4.

192. Dietrich M, Rasche H, Rommel K, Hochapfel G,Antimicrobial therapy as a part of the deconta-mination procedures for patients with acute

leukemia. Eur J Cancer 1973; 9: 443–7.193. Schaferkorting M, Blechschmidt J, Korting HC,

Clinical use of oral nystatin in the prevention ofsystemic candidosis in patients at particularrisk. Mycoses 1996; 39: 329–39.

194. Denning DW, Donnelly JP, Hellreigel KP et al,Antifungal prophylaxis during neutropenia orallogeneic bone marrow transplantation: whatis the state of the art? Ad Hoc Working Group.Chemotherapy 1992; 1: 43–9.

195. Schuler US, Haag C, Prophylaxis of fungalfunctions. Mycoses 1997; 40(Suppl 2): 41–4.

196. Glasmacher A, Hahn C, Molitor E et al, Fungalsurveillance cultures during antifungal prophy-laxis with itraconazole in neutropenic patientswith acute leukemia. Mycoses 1999; 42: 395–402.

197. Gubbins PO, Bowman JL, Penzak SR,Antifungal prophylaxis to prevent invasivemycoses among bone marrow transplantationrecipients. Pharmacotherapy 1998; 18: 549–64.

198. Ninane J, Sluysmans T, Vermylen C et al,Itraconazole versus ketoconazole for theprophylaxis of fungal infection in neutropenicchildren: results of two consecutive nonran-domized studies. Pediatr Hematol Oncol 1989; 6:349–53.

199. Harousseau JL, Dekker AW, Stamatoullas-Bastard A et al, Itraconazole oral solution forprimary prophylaxis of fungal infections inpatients with hematological malignancy andprofound neutropenia: a randomized, double-blind, double-placebo, multicenter trial compar-ing itraconazole and amphotericin B. AntimicrobAgents Chemother 2000; 44: 1887–93.

200. Martino P, Raccah R, Gentile G et al, Aspergilluscolonization of the nose and pulmonaryaspergillosis in neutropenic patients: a retro-spective study. Haematologica 1989; 74: 263–5.

201. Egger T, Gratwohl A, Tichelli A et al,Comparison of fluconazole with oral polyenesin the prevention of fungal infections in neu-tropenic patients. A prospective, randomized,single-center study. Support Care Cancer 1995; 3:139–46.

202. De Laurenzi A, Matteocci A, Lanti A et al,Amphotericin B prophylaxis against invasivefungal infections in neutropenic patients: asingle center experience from 1980 to 1995.Infection 1996; 24: 361–6.

203. Erjavec Z, Woolthuis GM, de Vries-Hospers HGet al, Tolerance and efficacy of amphotericin B


inhalations for prevention of invasive pul-monary aspergillosis in haematologicalpatients. Eur J Clin Microbiol Infect Dis 1997; 16:364–8.

204. Conneally E, Caffeky MT, Daly PA, Nebulizedamphotericin B as prophylaxis against invasiveaspergillosis in granulocytopenic patients. BoneMarrow Transplant 1990; 5: 403–6.

205. Hertenstein B, Stefanic M, Novotny J et al, Lowincidence of invasive fungal infections afterbone marrow transplantation in patients receiv-ing amphotericin B inhalations during neu-tropenia. Ann Hematol 1994; 68: 21–6.

206. Schwartz S, Behre G, Heinemann V et al,Aerosolized amphotericin B inhalations as pro-phylaxis of invasive Aspergillus infections dur-ing prolonged neutropenia: results of aprospective randomized multicenter trial. Blood1999; 93: 3654–61.

207. Jeffery GM, Beard ME, Ikram RB et al,Intranasal amphotericin B reduces the fre-quency of invasive aspergillosis in neutropenicpatients. Am J Med 1991; 90: 685–92.

208. Hoyle C, Goldman JM, Life-threatening infec-tions occurring more than 3 months afterBMT. 18 UK Bone Marrow Transplant Teams.Bone Marrow Transplant 1994; 14: 247–52.

209. Miller RF, Le Noury J, Corbett EL et al,Pneumocystis carinii infection: current treatmentand prevention. J Antimicrob Chemother 1996;37(Suppl B): 33–53.

210. Vasconcelles MJ, Bernardo MV, King C et al,Aerosolized pentamidine as Pneumocystis pro-phylaxis after bone marrow transplantation isinferior to other regimens and is associatedwith decreased survival and an increased riskof other infections. Biol Blood Marrow Transplant2000; 6: 35–43.

211. Souza JP, Boeckh M, Gooley TA et al, Highrates of Pneumocystis carinii pneumonia in allo-geneic blood and marrow transplant recipientsreceiving dapsone prophylaxis. Clin Infect Dis1999; 29: 1467–71.

212. Edwards JE Jr, Bodey GP, Bowden RA et al,International Conference for the Developmentof a Consensus on the Management andPrevention of Severe Candidal Infections. ClinInfect Dis 1997; 25: 43–59.

213. Chandrasekar PH, Gatny CM, and the BoneMarrow Transplantation Team, The effect offluconazole prophylaxis on fungal colonization

in neutropenic cancer patients. J AntimicrobChemother 1994; 33: 309–18.

214. Laverdiere M, Rotstein C, Bow EJ et al, Impactof fluconazole prophylaxis on fungal coloniza-tion and infection rates in neutropenic patients.J Antimicrob Chemother 2000; 46: 1001–8.

215. Menichetti F, Del Favero A, Martino P et al,Itraconazole oral solution as prophylaxis forfungal infections in neutropenic patients withhematologic malignancies: a randomized,placebo-controlled, double-blind, multicentertrial. GIMEMA Infection Program. GruppoItaliano Malattie Ematologiche dell’ Adulto.Clin Infect Dis 1999; 28: 250–5.

216. Venkatakrishnan K, von Moltke LL, GreenblattDJ, Effects of the antifungal agents on oxidativedrug metabolism: clinical relevance. ClinPharmacokinet 2000; 38: 111–80.

217. Lopez-Gil JA, Fluconazole–cyclosporine inter-action: a dose-dependent effect? AnnPharmacother 1993; 27: 427–30.

218. Osowski CL, Dix SP, Lin LS et al, Evaluationof the drug interaction between intravenoushigh-dose fluconazole and cyclosporine ortacrolimus in bone marrow transplant patients.Transplantation 1996; 61: 1268–72.

219. Dresser GK, Spence JD, Bailey DG, Pharma-cokinetic-pharmacodynamic consequences andclinical relevance of cytochrome P450 3A4inhibition. Clin Pharmacokinet 2000; 38: 41–57.

220. Maertens J, Verhaegen J, Demuynck H et al,Autopsy-controlled prospective evaluation ofserial screening for circulating galactomannan bya sandwich enzyme-linked immunosorbent assayfor hematological patients at risk for invasiveaspergillosis. J Clin Microbiol 1999; 37: 3223–8.

221. Saral R, Burns WH, Laskin OL et al, Acyclovirprophylaxis of herpes-simplex-virus infections.N Engl J Med 1981; 305: 63–7.

222. Gluckman E, Lotsberg J, Devergie A et al,Prophylaxis of herpes infections after bone-marrow transplantation by oral acyclovir.Lancet 1983; ii: 706–8.

223. Ljungman P, Wilczek H, Gahrton G et al, Long-term acyclovir prophylaxis in bone marrowtransplant recipients and lymphocyte prolifera-tion responses to herpes virus antigens in vitro.Bone Marrow Transplant 1986; 1: 185–92.

224. Boeckh M, Gooley TA, Bowden RA, Effect ofhigh-dose acyclovir on survival in allogeneicmarrow transplant recipients who received


ganciclovir at engraftment or for cyto-megalovirus pp65 antigenemia. J Infect Dis 1998;178: 1153–7.

225. Prentice HG, Gluckman E, Powles RL et al,Long term survival in allogeneic bone marrowtransplant recipients following acyclovir pro-phylaxis for CMV infection. Bone MarrowTransplant 1997; 19: 129–33.

226. Sinnige LG, van der Meer JW, Gratama JW et al,Is acyclovir prophylaxis necessary after bonemarrow transplantation? Infection 1986; 14:122–4.

227. Baglin TP, Gray JJ, Marcus RE, Wreghitt TG,Antibiotic resistant fever associated with herpessimplex virus infection in neutropenic patientswith haematological malignancy. J Clin Pathol1989; 42: 1255–8.

228. Bergmann OJ, Mogensen SC, Ellermann-Eriksen S, Ellegaard J, Acyclovir prophylaxisand fever during remission-induction therapyof patients with acute myeloid leukemia: a ran-domized, double-blind, placebo-controlled trial.J Clin Oncol 1997; 15: 2269–74.

229. Lonnqvist B, Palmblad J, Ljungman P et al, Oralacyclovir as prophylaxis for bacterial infectionsduring induction therapy for acute leukemia inadults. The Leukemia Group of MiddleSweden. Support Care Cancer 1993; 1: 139–44.

230. Ljungman P, De Bock R, Cordonnier C et al,Practices for cytomegalovirus diagnosis, pro-phylaxis and treatment in allogeneic bone mar-row transplant recipients: a report from theWorking Party for Infectious Diseases of theEBMT. Bone Marrow Transplant 1993; 12:399–403.

231. Stocchi R, Ward KN, Fanin R et al, Managementof human cytomegalovirus infection and dis-ease after allogeneic bone marrow transplanta-tion. Haematologica 1999; 84: 71–9.

232. Dix SP, Wingard JR, Management of viral infec-

tions in bone marrow transplant recipients. ClinImmunother 1996; 6: 352–82.

233. Gerna G, Furione M, Baldanti F et al,Quantitation of human cytomegalovirus DNAin bone marrow transplant recipients. Br JHaematol 1995; 91: 674–83.

234. Hebart H, Kanz L, Jahn G, Einsele H,Management of cytomegalovirus infection aftersolid-organ or stem-cell transplantation.Current guidelines and future prospects. Drugs1998; 55: 59–72.

235. De Vries-Hospers HG, Sleijfer DT, Mulder NHet al, Bacteriological aspects of selective decont-amination of the digestive tract as a method ofinfection prevention in granulocytopenicpatients. Antimicrob Agents Chemother 1981; 19:813–20.

236. Bow EJ, Raynor E, Scott BA, Louie TJ, Selectivegut decontamination with nalidixic acid ortrimethoprim–sulfamethoxazole for infectionprophylaxis in neutropenic cancer patients:relationship of efficacy to antimicrobial spec-trum and timing of administration. AntimicrobAgents Chemother 1987; 31: 551–7.

237. Rozenberg-Arska M, Dekker AW, Verhoef J,Ciprofloxacin for selective decontamination ofthe alimentary tract in patients with acuteleukemia during remission induction treatment:the effect on fecal flora. J Infect Dis 1985; 152:104–7.

238. Pannuti C, Gingrich R, Pfaller MA et al,Nosocomial pneumonia in patients having bonemarrow transplant – attributable mortality andrisk factors. Cancer 1992; 69: 2653–62.

239. Gerson SL, Talbot GH, Hurwitz S et al,Prolonged granulocytopenia: the major risk fac-tor for invasive pulmonary aspergillosis inpatients with acute leukemia. Ann Intern Med1984; 100: 345–51.


13Cytokines and WBC transfusionsKatarzyna J Finiewicz, John R Wingard

INTRODUCTION

Neutropenia in cancer patients most oftenresults from cytotoxic chemotherapy or irradia-tion. Occasionally, a low granulocyte countmay also be caused by marrow replacementwith the tumor. Bone marrow failure states notrelated to a malignancy or its treatment accountfor only a minority of the cases of neutropeniaseen in clinical practice.

The association of neutropenia with infectionwas first recognized and reported in 1966 byBodey et al.1 The depth as well as the durationof neutropenia both contribute to the risk ofinfectious complications. The frequency ofinfection increases progressively as the neu-trophil count decreases below 1000/µl, reach-ing the highest level when the count falls below100/µl. The likelihood of neutropenic infectionalso correlates with the duration of neutrope-nia; patients in whom the process of recoverylasts longer than 10–14 days are at particularlyhigh risk for serious infectious complications.While most neutropenic infections are bacterial,after the first 10–14 days, the spectrum of infec-tious pathogens expands to encompass invasivefungi as well as other opportunistic organ-isms.2,3 Long-lasting neutropenia is associatedwith poor response rates to antibiotics and pro-longed hospital stay.4

Neutropenic fever after chemotherapy hascustomarily been treated with intravenousantibiotics. In severe infections not respondingto antibiotics, granulocyte transfusions haveoccasionally been employed. Despite remark-able progress attained in the treatment of neu-tropenic fever, achieved with the use of empiricbroad-spectrum antibiotics, morbidity fromneutropenic infections remains considerable. Aseffective as antibiotics are, they are ultimatelylimited by the potential for emergence of drugresistance. Thus, other measures that canreduce the use of antibiotics have been sought.Attempts to attenuate neutropenia led to thedevelopment of recombinant myeloid growthfactors. Two myeloid growth factors currentlyavailable for therapeutic use in the USA arerecombinant human granulocyte colony-stimulating factor and recombinant humangranulocyte–macrophage colony-stimulatingfactor. Both of these growth factors stimulateproliferation and maturation of myeloid pro-genitor cells. Both molecules predictablydecrease the depth and duration of neutropeniaafter myelotoxic cancer therapy.

In this chapter, we shall summarize the datafrom available clinical trials conducted to eval-uate the efficacy of hematopoietic growth fac-tors in the treatment of a neutropenic patientafter myelotoxic cancer therapy. We shall

present and discuss the indications forhematopoietic growth factors in light of thechanges in treatment of neutropenic fever thathave evolved over the last few years. The roleof granulocyte infusions as an alternativemethod of abating the effects of prolonged neu-tropenia will also be reviewed.

MYELOID GROWTH FACTORS

Hematopoiesis and hematopoietic growthfactors

Hematopoiesis is an orderly, continuousprocess by which primitive, multipotent pro-genitor cells give rise to mature hematopoieticcells.5,6 The net composition of the hematopoi-etic cell compartment is a result of continuousinterplay between pluripotent stem cells,maturing progenitors, bone marrow stroma,and hematopoietic growth factors. The interac-tions between the specific components of thehematopoietic cell compartment maintain astate of dynamic equilibrium, a state preciselyregulated to maintain the concentration ofmature blood cells in the circulation within anarrow homeostatic range. In response to a spe-cific stimulus, the hematopoietic system is capa-ble of rapid expansion of production of maturecells, up to 10-fold within several days. Thisoccurs mainly owing to an increased matura-tion of committed progenitors.

The compartment of the hematopoietic sys-tem predominantly engaged in the response toinfection is myelopoiesis. The baseline produc-tion of the neutrophils in the bone marrow isabout 1.0 � 1011/day in a healthy individual,which increases several-fold with infection.6,7

After myeloid precursors mature, they arereleased into peripheral blood, where they sur-vive approximately 6–10 hours. Half of the neu-trophils in peripheral blood freely circulate,while the other half remain in the microcircula-tion, adherent to vascular walls in the marginalpool. Neutrophils from the microcirculation canbe mobilized to circulate by stress or by med-

ications such as epinephrine (adrenaline) orcorticosteroids. However, neutrophilia inducedby growth factors occurs mainly through thestimulation of the myeloid progenitors in thebone marrow to increase production.

In the complex process of hematopoiesis, therole of hematopoietic growth factors is bothpermissive and instructive.8–12 Hematopoieticgrowth factors maintain the survival of theearly hematopoietic progenitor cells by pre-venting apoptosis. They also promote their proliferation, facilitate and direct theirdifferentiation, and activate effector functionsin mature cells. Although the commitment ofmultipotent hematopoietic progenitor cells to aspecific lineage is most likely random, furtherdevelopment depends on the instructionalinfluence of the microenvironment, in whichhematopoietic growth factors serve a crucialrole.

Hematopoietic growth factors are producedby a variety of stromal and hematopoietic cells.Most act locally, where they are produced, butsome have hormone-like activity. One cell maybe capable of producing several growth factors.The production of each growth factor is subjectto precise regulation by multiple autocrine andparacrine loops. Like hematopoietic cells,hematopoietic growth factors exist in hierarchy;some are multipotent and affect both early andlate progenitor cells, while others are lineage-specific. The pleotropic potential of many of thehematopoietic growth factors significantly lim-its their application as therapeutic agents, sincethe clinically desired effect may be accompa-nied by a spectrum of potentially deleteriouseffects. Hematopoietic growth factors interactwith progenitor cells through specific receptors.Each growth factor receptor is characterized byaffinity for several different growth factors,which can act in synergy.

Clonal culture techniques to grow differentmarrow progenitor colonies in vitro haveallowed the identification of hematopoieticgrowth factors with colony-stimulating activ-ities: granulocyte–macrophage colony-stimulating factor (GM-CSF), granulocyte


colony-stimulating factor (G-CSF), macrophagecolony-stimulating factor (M-CSF), and multi-colony stimulating factor (multi-CSF) orinterleukin-3 (IL-3).11 All four have been puri-fied and defined genetically. The first two, GM-CSF and G-CSF, have been manufacturedthrough recombinant engineering technologyand made available for clinical use. The lattertwo, M-CSF and IL-3, were tested in clinicaltrials, but were withdrawn from further devel-opment because of undesirable toxicities. As thenames indicate, GM-CSF and G-CSF preferen-tially stimulate growth and maturation of cellswith myeloid differentiation, and for thisreason will be referred to further as myeloidgrowth factors (MGFs).

The wide use of hematopoietic growth fac-tors in clinical practice raised theoretical con-cerns about potential deleterious influences onhematopoiesis, leading to stem cell exhaustionor ‘stem cell steal’. Stem cell steal can theoreti-cally result from preferential commitment ofstem cells to one specific lineage directed by thespecific growth factor, limiting differentiationto other lineages. Stem cell exhaustion refers tothe hypothetical possibility of stem cell deple-tion from repeated stimulation of marrow bygrowth factors. Neither of the two myeloidgrowth factors currently available in clinicalpractice appears to be able to recruit the stemcell directly, since they act on more committedmyeloid precursors. Based on observations ofpatients treated with MGFs for months andeven years, the concern that any of these twophenomena truly occur in humans does notseem substantiated, and MGFs seem quite safein this regard.

Biology of recombinant myeloid growthfactors

Human G-CSF is a glycoprotein of approxi-mately 20 kDa.13,14 G-CSF is encoded by a singlegene located on human chromosome 17q21–22.G-CSF is produced by activated macrophages,endothelial cells, and fibroblasts, as well as by

bone marrow stromal cells. The G-CSF receptoris a type I membrane protein belonging to thehematopoietic growth factor receptor family,but, unlike other receptors in this family(including the GM-CSF receptor), which func-tion as heterodimers or heterotrimers to bind totheir ligands, the receptor for G-CSF seems tofunction as a homodimer. Expression of theG-CSF receptor is specifically restricted to neu-trophilic progenitors, mature neutrophils, andvarious myeloid leukemia cells (Figure 13.1).After stimulation, the G-CSF receptor trans-duces the signals for both proliferation and dif-ferentiation.

G-CSF functions as the physiologic regula-tory factor for circulating granulocytes.Increased serum levels of G-CSF can bedetected in patients with neutropenia, fallingwith neutrophil recovery. Recombinant G-CSFelevates the level of circulating neutrophils in a dose-dependent fashion.15,16 Neutrophiliaresults from the shortening of the postmitoticmarrow transit time from approximately 6 daysto as short as 3 days with higher doses, withoutaffecting circulating neutrophil lifespan or thedistribution between the marginal and circulat-ing pools. Within 3–4 days after the start ofG-CSF, immature and committed progenitorcells appear in the circulation (Figure 13.1). Inaddition to its hematopoietic effect, G-CSFappears to modulate the function and microbi-cidal capacity of mature neutrophils (Table13.1).16–20

Human GM-CSF is a glycoprotein with amolecular mass of approximately 22 kDa and athree-dimensional structure composed of twopairs of antiparallel � helices.21,22 The humanGM-CSF gene is located on the long arm ofchromosome 5, near other cytokine genes.GM-CSF can be produced by a number of dif-ferent cells: T lymphocytes, macrophages,endothelial cells, fibroblasts, and stromal cells,as well as various malignant cells. GM-CSFreceptors are found on hematopoietic cells aswell as on various non-hematopoietic cells suchas trophoblasts, endothelial cells, oligodendro-cytes, and some malignant cells.23 GM-CSF

CYTOKINES AND WBC TRANSFUSIONS 247


Pluripotentstem cell

Lymphoidstem cell

CFU-GEMM

CFU-GMCFU-Meg

Platelets

BFU-E

Red bloodcell

Megakaryocyte

Neutrophil Monocyte

Macrophage

Dendritic cell

Eosinophil

CFU-Eo

Basophil

CFU-Bas

CFU-Blast

GM-CSF

GM-CSF and G-CSF

Figure 13.1 The biological effects of G-CSF and GM-CSFon myeloid progenitors are illustrated. The darkly shadedcells are those that are affected by G-CSF: differentiatedmyeloid precursors from the CFU-GM stage to the matureneutrophil. GM-CSF affects the same progenitors, but alsoearlier precursors (CFU-GEMM) and monocytoid,macrophage, and dendritic lineages (lightly shaded cells).

receptors have not been found on lymphocytes.The GM-CSF receptor is composed of two sub-units: a ligand-specific � chain that forms a low-affinity complex with GM-CSF and a � chain

that has no detectable binding to GM-CSF, but,together with the � chain, forms a high-affinityreceptor. The � chain is essential for high-affinity binding and signaling of IL-3 and IL-5.


Table 13.1 Comparison of biologic functions of G-CSF and GM-CSF

Target hematopoietic cell Biologic function G-CSF GM-CSF Therapeutic applications

Myeloid progenitor cells • Increased proliferation � � Mobilization of myeloid• Increased differentiation � � progenitor cells to

peripheral blood inpreparation for SCTa

Neutrophils • Maintenance of steady-state � � Control of infectionneutrophil numbers

• Increased antimicrobial � �

activity• Increased phagocytosis � �

• Increased chemotaxis � �

• Increased cytokine release � �

• Enhanced oxidative activity � �

Macrophages and • Increased intracellular � � Control of infectionmonocytes killing

• Increased phagocytosis � � Enhancement of immune• Increased APCb function; � � system response to

ADCCc infection and/or tumor• Increased cytokine release � �

• Enhanced oxidative activity � �

Dendritic cells • Increased production � � Enhancement of immune• Increased differentiation � � system response to• Increased APCb function � � infection and/or tumor

Structural cells • Induction of migration and � � Promotion of wound healing(endothelial cells, proliferationfibroblasts, • Increased adhesion � �

keratinocytes) molecule expression

a SCT, stem cell transplantation. b APC, antigen-presenting cell. c ADCC, antibody-dependent cellular cytotoxicity.

The recombinant GM-CSF molecule stimu-lates proliferation and maturation of a bipoten-tial neutrophil and macrophage progenitor,which leads to the release into peripheral bloodof monocytes and macrophages in addition toneutrophils (Figure 13.1).24–26 Treatment withGM-CSF also enhances the function of matureneutrophils, monocytes, and macrophages byincreasing antimicrobial activity, chemotaxis,and proinflammatory cytokine release. GM-CSFhas a significant effect on the antigen-presenting function of antigen-presenting cells(APC). Specifically, in vitro studies havedemonstrated that GM-CSF acts as the majorstimulatory cytokine for the production, differ-entiation, and viability of dendritic cells (Table13.1). Thus, GM-CSF appears to act not only asa stimulant of peripheral blood neutrophilrecovery but also as an enhancer of severalother components of the immune system’sresponse to infection and malignancy.27,28

Both GM-CSF and G-CSF have been tested toshorten neutropenia in chemotherapy-induced

aplasia and in bone marrow failure states fromother causes. MGFs have also been utilized inthe collection of stem cells in preparation forstem cell transplantation. The impact of MGFs,especially GM-CSF, on the function of maturemyeloid cells holds great potential for futureuse, but at the present time remains investiga-tional.

Parameters used to evaluate the benefits ofMGFs

The parameters that should be considered inthe assessment of the benefit from MGFs in aparticular clinical situation can be grouped intothree categories: (1) laboratory values, (2) clini-cal events, and (3) resource utilization (Table13.2).29 The accelerated recovery from neutrope-nia consistently seen with MGFs does notalways translate into less infectious morbidityand mortality. Therefore, the duration of neu-tropenia is not necessarily a useful surrogate


Table 13.2 Parameters considered in the assessment of the benefit from recombinant myeloid growthfactors

Laboratory values • Leukocyte count, neutrophil count• Functional assays of neutrophils and monocytes

Clinical events • Rates and severity of documented infection,fever, sepsis syndrome, infectiousmorbidity, and infectious mortality

• Disease control• Survival• Quality of life• Mucositis and other toxicities

Resource utilization • Antibiotics, antifungal agents and granulocytetransfusions

• Length of hospitalization• Diagnostic procedures• Cost

marker for clinically relevant outcomes. Themost important measure of the value of MGFsis their effect on clinical events, such as rate andseverity of infections, survival, and quality oflife (QoL).

Savings in resource utilization (e.g. antibioticutilization, diagnostic tests performed, andduration of hospitalization) are becoming morerelevant in an increasingly cost-consciousenvironment, and are now commonly used tojustify treatment with costly MGFs, even insituations where clinical gains are not apparent.However, pharmacoeconomic calculations maybe misleading. The lack of consistency in para-meters used in clinical decision making by vari-ous investigators limits the extrapolation of theinformation obtained from clinical trials andprevents valid comparative analyses. Thethreshold for initiating supportive treatmentsuch as antibiotics or blood product transfu-sions or the criteria used for discharge vary sig-nificantly from center to center. Practices usingmore stringent criteria may not appreciate thesame magnitude of savings as observed in clini-cal trials where much more conservativeapproaches were applied. Additionally, savingsin the field of supportive care are subject tochange over time. This is best illustrated by theshifting care of cancer patients from an inpa-tient to outpatient setting, which obviouslyundermines the significance of shortened hos-pital stay demonstrated by many studies.

Role of MGFs in the management of patientsundergoing chemotherapy

In 1991, recombinant G-CSF (filgrastim,Neupogen) became approved by the US Foodand Drug Administration (FDA) for clinicaluse, initially for protection from chemotherapy-induced neutropenia and then for shortening ofneutropenia and reduction of fever followingbone marrow transplantation (BMT). This wassoon followed by approval of GM-CSF (sar-gramostim, Leukine) for acceleration of neu-trophil recovery after autologous BMT in

patients with lymphoid malignancies (see Table13.3 for a summary of the available formula-tions of G-CSF and GM-CSF and theirapproved indications). Since then, both G-CSFand GM-CSF have gained wide acceptance inoncology clinical practice as an adjunct tochemotherapy in three clinical settings:

(1) primary prophylaxis, during the firstchemotherapy cycle to prevent anticipatedneutropenic infectious complications;

(2) secondary prophylaxis, during subsequentchemotherapy cycles, after documentedoccurrence of neutropenic fever during theprior cycle;

(3) therapeutic setting, for adjunctive treatmentof established neutropenia with or withoutfever.

The potential for broad application of thesecostly agents called for the need to define theclinical settings for their most appropriate use.A group of oncologists was convened by theAmerican Society of Clinical Oncology (ASCO)to review information obtained from clinicaltrials and comment on the appropriateness ofuse of MGFs in the treatment of neutropeniaand neutropenia-related complications. As aresult, in 1994, a set of guidelines for the use ofMGFs in particular clinical situations was pub-lished (the ASCO guidelines).30 In short, theASCO guidelines judged the use of MGF to besuitable: (1) in primary prophylaxis for patientswith an expected likelihood of neutropenicfever of over 40%; (2) in secondary prophylaxis,after a documented neutropenic fever duringthe prior chemotherapy cycle to reduce infec-tious complications and maintain chemother-apy dose intensity in subsequent cycles; and (3)after stem cell transplantation (SCT) to enhancehematopoietic recovery and to treat engraft-ment failure. The use of MGFs was judged tooffer only a marginal benefit in other clinicalsettings, and was not supported by publishedevidence under usual circumstances. Sinceguidelines are meant to be updated periodi-cally, on the basis of new evidence, the ASCOguidelines panel reviewed the interim literature


and in 1996 published an update that added arecommendation to use MGFs in patients over54 years of age with acute myeloid leukemia(AML) after completion of inductionchemotherapy.31 An extensive revision to theoriginal 1994 ASCO guidelines was recentlypublished, based upon a review of all the newliterature since 1994.32 A current summary ofthe 2000 ASCO Clinical Practice Guidelines forMGFs is given in the Appendix.

Owing to a paucity of data on MGF influenceon such clinical events as infectious morbidityand mortality, the ASCO panel used a numberof indirect measurements of clinical benefit inassessing trial results: effects on neutrophilcount, rates of neutropenic fever, antibiotic

therapy requirements, and need for hospitaliza-tion. These secondary endpoints were con-sidered valid surrogate measures if they werethought to be reflective of more meaningfulclinical outcomes, such as decreased infectiousmorbidity and mortality or an improvement inQoL. The same measures were also used toestimate the magnitude of economic gain fromMGFs. It should be noted that both improve-ment in QoL and economic gains can be meas-ured directly, by applying QoL scales and byperforming cost–benefit analyses. However, thedirect methods, although more accurate andreliable, were used in only a few randomizedtrials performed prior to 1994. They are beingincreasingly incorporated in the design of more


Table 13.3 Commercially available formulations of recombinant G-CSF and GM-CSF, and theirapproved usesa

Generic name Trade name Countries Indicationsb

G-CSF Filgrastim Neupogen Europe, USA, Canada, CIN, BMT, PBPC, SCN, ALAustralia

Gran Japan, Taiwan, Korea, CIN, BMT, AL, AA, SCN,China MDS, HIV

Lenograstim Neutrogin Japan, China CIN, BMT, SCN, MDS, AA

Granocyte Europe, Australia CIN, BMT

Nartograstim Neu-UP Japan CIN, BMT, SCN, AA

GM-CSF Molgramostim Leukomax Europe CIN, BMT, AIDS

Canada CIN, BMT

Sargramostim Leukine USA PBPC, BMT, AL

a Reprinted with permission from Armitage.27

b AA, aplastic anemia; AIDS, acquired immune deficiency syndrome; AL, acute leukemia; BMT, bone marrow transplantation;CIN, chemotherapy-induced neutropenia; HIV, human immunodeficiency virus; MDS, myelodysplastic syndrome; PBPC,peripheral blood progenitor cell transplantation; SCN, severe chronic neutropenia.

recent trials, which will hopefully provide moreprecise assessments.

Use of MGFs as primary prophylaxis inpatients with solid tumors and lymphomas

The majority of trials examining primary pro-phylaxis were aimed at demonstrating adecrease in the incidence and duration of neu-tropenia, leading to a reduction in the rate ofserious infections and improved survival. Fourstudies utilizing G-CSF and five studies utiliz-ing GM-CSF in primary prophylaxis had beenreported by 1994, and were considered in theASCO guidelines analysis (Tables 13.4 and13.5).30 Patient population in these trials washeterogenous, and included solid tumors aswell as non-Hodgkin’s lymphomas (NHL).

All four of the studies evaluating G-CSFdemonstrated a reduction in the severity andduration of neutropenia (Table 13.4).33–36 Inaddition, a decrease in the incidence of neu-tropenic fever by approximately 50% was seen.None of the studies showed a reduction in seri-ous morbidity or mortality from infectious com-plications. A decreased length of hospital stayand length of treatment with intravenousantibiotics was appreciated in only twostudies.33,34 One more recent study of G-CSF inprimary prophylaxis involved adults with lym-phoma.37 Similar to the results of the earlierinvestigations, there was a shortening of neu-tropenia and a decrease in neutropenic fever,but no other advantage.

Three of five trials of GM-CSF for primaryprophylaxis demonstrated a shortening of neu-tropenia (Table 13.5).38–42 The effect on neu-tropenic fever was inconsistent. A possibleexplanation for an apparent diminished efficacyof GM-CSF in comparison with G-CSF wasinclusion of patients with a lower risk of neu-tropenia in the GM-CSF trials, which madedetection of a positive effect more difficult.Alternatively, differences in biologic effectsbetween the two growth factors could alsoaccount for the observed different results.

Interestingly, the only randomized trial directlycomparing the two agents in primary prophy-laxis of chemotherapy-related neutropenic feverfound them to be equivalent in efficacy.43

The impact of MGFs on healthcare resourceutilization was examined using a risk-assessment model developed by Lyman et al toallow extrapolation of the data from the abovecited trials to a broad population of patientstreated with various chemotherapy regi-mens.44,45 In general, the use of prophylacticMGFs was found to be cost-effective only inpatients whose risk of neutropenic feverexceeded 40% over the course of chemotherapy.The model assumed that all patients with neu-tropenic fever would be hospitalized andtreated with intravenous antibiotics until reso-lution of neutropenia and all signs of infection.

However, one can justly question whethersuch an analysis is applicable today. Inresponse to the necessity to control health carecosts, many oncologic therapies, including sup-portive care, have been moved from an inpa-tient to outpatient setting, as noted earlier.46–49

The introduction of long-acting and oral broad-spectrum antibiotics, as well as the widespreadavailability of home care services, now allowmany low-risk patients with neutropenic feverto be treated on an outpatient basis. Suchchanges in practice necessitate re-examinationof the assumptions that formed the recommen-dations for the use of MGFs based on costanalyses performed years ago.

The decision analysis model employed bythe ASCO expert panel has another importantlimitation: the estimate of the risk of neu-tropenic fever has been based exclusively onthe myelotoxic potential of the chemotherapyregimen. Inclusion of additional factors thatcould help to predict prolonged, severe neu-tropenia in an individual patient was encour-aged, but left entirely to the judgement of thetreating physician. These risk factors were:bone marrow compromise from marrowinvolvement by the tumor, cumulative toxicityfrom prior chemotherapy and/or irradiation,and a history of severe neutropenia with prior


Tabl

e 13.4

Sum

mar

y of

ran

dom

ized

trial

s of

G-C

SF

in p

rim

ary

prop

hyla

xisa

Ref

eren

ceD

isea

seN

o. o

fTr

eatm

ent

Sho

rten

edC

linic

ally

rel

evan

t ev

ents

aR

esou

rce

utili

zati

onc

grou

papa

tien

tsar

ms

neut

rope

niac

Inci

denc

e of

neu

trop

enic

fev

erIn

fect

ious

Low

er r

ate

ofLo

wer

rat

e of

mor

talit

yan

tibi

otic

hosp

ital

izat

ion

Firs

t cy

cle

All

cycl

es

Firs

t cy

cle

All

cycl

esra

te (

%)

usag

e(%

)(%

)(%

)(%

)

Cra

wfo

rd e

t al

33

SC

LC199

G-C

SF

SS

28

d4

0d

3S

eS

e

Plac

ebo

57

77

3

Trill

et-L

enoi

r et

al3

4S

CLC

129

G-C

SF

SS

20

d2

6d

2S

SPl

aceb

o41

53

5

Pett

enge

ll et

al3

5N

HL

80

G-C

SF

NS

S—

23

d5

NS

NS

Plac

ebo

44

5

Geb

bia

et a

l36

Vario

us86

G-C

SF

—S

—1

2d

——

—Pl

aceb

o32

Zinz

ani e

t al

37

NH

L149

G-C

SF

—S

—5

d—

——

Con

trol

21

aM

odifi

ed a

nd u

pdat

ed w

ith p

erm

issi

on fro

m A

SC

O r

ecom

men

datio

ns.3

0

bS

CLC

, sm

all c

ell l

ung

carc

inom

a; N

HL,

non

-Hod

gkin

’s ly

mph

oma.

cS

, si

gnifi

cant

diff

eren

ce; N

S, no

sig

nific

ant

differ

ence

; —

, no

t as

sess

ed.

dS

igni

fican

tly d

iffer

ent.

e

Firs

t cy

cle

only

.

Tabl

e 13.5

Sum

mar

y of

ran

dom

ized

trial

s of

GM

-CSF

in p

rim

ary

prop

hyla

xisa

Ref

eren

ceD

isea

seN

o. o

fTr

eatm

ent

Sho

rten

edC

linic

ally

rel

evan

t ev

ents

cR

esou

rce

utili

zati

onc

grou

pbpa

tien

tsar

ms

neut

rope

niac

Inci

denc

e of

Infe

ctio

usLo

wer

rat

e of

Low

er r

ate

ofne

utro

peni

c fe

ver

mor

talit

yan

tibi

otic

usa

geho

spit

aliz

atio

nra

te (

%)

Kap

lan

et a

l38

HIV

-rela

ted

21

GM

-CS

F (E

. co

li)S

S1

0—

SN

HL

Con

trol

18

Ger

hart

z et

al3

9N

HL

125

dG

M-C

SF

(E. co

li)S

dS

d2

Sd

Sd

Con

trol

3

Ham

m e

t al

40

SC

LC148

GM

-CS

F (E

. co

li)S

eN

S—

NS

—C

ontr

ol

Nic

hols

et

al4

1G

erm

cel

l61

GM

-CS

F (E

. co

li)N

SN

S—

NS

—C

ontr

ol

Ham

m e

t al

42

SC

LC213

GM

-CS

F (y

east

)N

SI

—N

SN

SC

ontr

ol

aM

odifi

ed a

nd u

pdat

ed w

ith p

erm

issi

on fro

m A

SC

O r

ecom

men

datio

ns.3

0

bN

HL,

non

-Hod

gkin

’s ly

mph

oma;

SC

LC, sm

all c

ell l

ung

carc

inom

a.cS

, si

gnifi

cant

diff

eren

ce; N

S, no

sig

nific

ant

differ

ence

; I,

infe

rior;

—, no

t as

sess

ed.

dS

elec

ted

patie

nts

(inte

ntio

n-to

-trea

t da

ta n

ot a

vaila

ble)

.e

Firs

t cy

cle

only

.

chemotherapy regimens. More recent clinicalobservations demonstrated that the risk of infec-tious complications following chemotherapy isalso clearly influenced by a number of otherhost-related and disease-related factors, such asage, the presence of comorbidity, the type anddegree of control of the primary cancer, andwhether or not the fever occurred in an inpatientor outpatient setting.50–57 A better understandingof all risk factors could allow development of abetter risk-stratified approach.

Quality of life is often considered anotherway to appraise the value of MGF. An improve-ment in QoL is certainly an important goal inthe management of cancer patients; however, itis one that is not easy to measure. It has neverbeen convincingly demonstrated that shorten-ing the duration of a hospital stay represents anappreciable improvement in QoL. An answer tothis question should be directly evaluated infuture clinical studies.

Interestingly, an ASCO survey investigatingthe compliance of oncologists with the ASCOguidelines concluded that many physicians tendto use growth factors freely, regardless of themyelosuppressive potential of the chemotherapyregimen employed.58 The tendency to overusegrowth factors in primary prophylaxis may be inpart explained by a lack of appreciation of thefact that only a few of the chemotherapy regi-mens routinely used for treatment of solidtumors or lymphoma produce severe myelosup-pression with associated risk of febrile neutrope-nia exceeding 40% (see Table 3 in the ASCOrecommendations30). It should be noted thatnone of the chemotherapy regimens used in anyof the positive trials of primary prophylaxiscould be described as the standard of care for theneoplasms treated. Instead, the regimens that areconsidered the standard of care today are muchless myelosuppressive.

Use of MGFs in dose-intensive regimens

The effectiveness of dose intensification by com-bining multiple agents in non-cross-resistant

combinations, as outlined in the Goldie–Colemanhypothesis,59 has been established for suchtumors as leukemia, lymphoma, early-stagebreast cancer, and small cell lung cancer.Introduction of growth factors allowed the devel-opment of new dose-intensive and/or schedule-intensive chemotherapy regimens, with the hopethat they would lead to increased tumor responserates and improved survival. Many such intensi-fied regimens are currently being tested in clini-cal trials. The studies evaluating the role of MGFsin delivery of intensive chemotherapy can begrouped into two categories: (1) randomizedstudies between two identical intensive regimenswith or without MGF support,60–64 and (2) ran-domized studies between a new intensivechemotherapy regimen that could not be admin-istered without MGF support tested against theless intensive standard of care that usually doesnot require MGF support.65

Whether such a strategy will demonstrate abeneficial impact on disease control or long-termsurvival outcomes remains to be proven. Theearly results of clinical trials have been quite vari-able.60–64 Most studies failed to show an improvedsurvival. Interestingly, one recent trial of dose-intensive chemotherapy regimen with G-CSFsupport in patients with small cell lung cancerdemonstrated improved survival withoutimpairment in QoL.66 However, in light of thedisappointing experience with treating small celllung cancer by high dose chemotherapy withstem cell support, this finding is surprising andneeds confirmatory trials. Although MGF usagewas successful in alleviating neutropenia, thenon-hematologic toxicities became limiting withdose escalation. For example, a higher incidenceof dose-limiting thrombocytopenia offset the ben-efit of shorter duration of neutropenia in onetrial. Additionally, the preliminary analysis of atrial of dose-escalated chemotherapy regimen forHodgkin’s disease, BEACOPP, revealed anincreased rate of secondary malignancies inpatients on the dose-intensified treatment arm – afinding that raises valid concerns.65

Several trials testing the role of dose-intensified treatment regimens are underway.


Their results will be of utmost interest, particu-larly in two diseases with well-documenteddose–response effect: lymphoma and breastcancer.

Use of MGFs in secondary prophylaxis

There has been no large prospective random-ized trial to validate the use of MGFs for sec-ondary prophylaxis. A small study of GM-CSFpatients with lymphoma was inconclusive.67 The1994 ASCO recommendation for the use of MGFin this setting had been based on the data pro-vided by the study of G-CSF in primary prophy-laxis for patients receiving chemotherapy forsmall cell lung cancer.33 In that study, patientson the placebo arm who developed neutropenicfever during the first chemotherapy cycle werecrossed over to the G-CSF arm for the secondcycle. Treatment with G-CSF significantly short-ened the duration of neutropenia (2.5 days ver-sus 6 days) and prevented the recurrence offever following the second, full-dose cycle ofchemotherapy in 77% of patients. The impact onserious morbidity and mortality was notassessed. Interestingly, only 5% of patients onboth arms who did not have fever during thefirst cycle of chemotherapy developed neu-tropenic fever during the second cycle. Based onthis observation, the ASCO expert panel foundthe use of MGF in the setting of secondary pro-phylaxis justifiable, with the caveat, however,that an alternative maneuver, namelychemotherapy dose reduction, should be givenfirst consideration in the palliative setting.

The appropriateness of choosing MGF sup-port over chemotherapy dose reductionremains a matter of controversy. An ASCO pollof MGF usage conducted in 199758 found that indiseases with particularly high cure rates (suchas testicular cancer), oncologists tend to useMGF for secondary prophylaxis after the first-cycle neutropenia, even in the absence offever.58,68,69 This pattern of clinical care can beaccounted for by the practitioner’s belief thatchemotherapy dose reduction might negatively

affect the final outcome. The benefit from MGFsin maintaining dose intensity of standardchemotherapy regimens has not been proven(in part owing to their low myelosuppressivepotential), although subset analyses withinlarge trials do suggest that treatment withG-CSF may allow delivery of more optimaldoses in patients who otherwise could not tol-erate it.70–72

Use of MGFs for treatment of afebrile andfebrile neutropenia

The ASCO MGF survey revealed that mostoncologists prescribe MGFs for patients withestablished severe neutropenia with or withoutfever, despite the fact that the efficacy of MGFsin this context has never been established.58

Several randomized trials demonstrated thatG-CSF, GM-CSF, or both administered asadjuncts to antibiotics for febrile neutropeniashorten the duration of severe neutropenia, buthave no effect on mortality due to infections orresource utilization such as duration of hospitalstay or number of days of antibiotic therapy(Table 13.6).73–83 In a subset of these patientscharacterized by the presence of documentedserious infection, support with MGFs may intu-itively seem more beneficial and justifiable, butawaits validation by clinical trials.45,73

What about initiating treatment with MGFsduring severe neutropenia but before the devel-opment of fever? This approach, if found suc-cessful, would be of particular interest asprimary prophylaxis as it could limit the use ofMGFs only to patients at high risk for seriousinfection.84 Disappointingly, a randomized trialof G-CSF performed on outpatients with severechemotherapy-induced afebrile neutropeniafound that the time to neutrophil recovery wassignificantly shorter on the G-CSF arm, butthere was no effect on the rate of hospitaliza-tions, number of days in hospital, duration oftreatment with intravenous antibiotics, or num-ber of culture-positive infections.85 One expla-nation for the lack of effectiveness in this study


Tabl

e 13.6

Sum

mar

y of

ran

dom

ized

trial

s of

MG

Fs in

the

tre

atm

ent

of feb

rile

neu

trop

enia

Ref

eren

ceD

isea

seN

o. o

fM

GF

Neu

trop

enia

Clin

ical

ly r

elev

ant

even

tsb

Res

ourc

e ut

iliza

tion

b

grou

papa

tien

ts(m

edia

l dur

atio

n,Fe

ver

Infe

ctio

usA

ntib

ioti

c us

age

Hos

pita

lizat

ion

days

)b

(med

ian

dura

tion

,m

orta

lity

(med

ian

(med

ian

dura

tion

,da

ys)

rate

(%

)du

rati

on, da

ys)

days

)

Mah

er e

t al

73

Sol

id t

umor

s216

G-C

SF

3c

NS

NS

NS

8AL

LPl

aceb

o4

8Ly

mph

oma

May

ordo

mo

et a

l74

Sol

id t

umor

s60

G-C

SF

2c

NS

NS

5c

5c

Lym

phom

aPl

aceb

o3

77

May

ordo

mo

et a

l74

Sol

id t

umor

s61

GM

-CS

F (E

. co

li)2

cN

SN

S5

c5

c

Lym

phom

aPl

aceb

o3

79

Riik

onen

et

al7

5Pe

diat

ric A

LL58

dG

M-C

SF

(E. co

li)4

.5c

2c

0S

SS

olid

tum

ors

Plac

ebo

64

2

Anai

ssie

et

al7

6S

olid

tum

ors

107

GM

-CS

F (E

. co

li)N

S4

NS

NS

9Ly

mph

oma

Con

trol

41

0Le

ukem

ia

Bie

sma

et a

l77

Sol

id t

umor

s30

GM

-CS

FS

eN

S—

NS

—AL

LPl

aceb

oLy

mph

oma

Velle

nga

et a

l78

Sol

id t

umor

s134

GM

-CS

F (E

. co

li)3

31(?

)—

6Ly

mph

oma

Plac

ebo

43

07

Rav

aud

et a

l79

Sol

id t

umor

s68

GM

-CS

F (E

. co

li)3

c2

05

c6

c

Lym

phom

aPl

aceb

o4

21

67

Mitc

hell

et a

l80

,81

Sol

id t

umor

s186

dG

-CS

F3

c2

—5

c5

c

Pedi

atric

ALL

Plac

ebo

53

67

Lym

phom

a

aAL

L, a

cute

lym

phob

last

ic le

ukem

ia.

bS

, si

gnifi

cant

diff

eren

ce; N

S, no

sig

nific

ant

differ

ence

; —

, no

t an

alyz

ed.

cS

tatis

tical

ly s

igni

fican

t. d

Epis

odes

of ne

utro

peni

a. e

In n

on-B

MT

patie

nts.

may be the late start of G-CSF. It is thought thatin order to exert their beneficial effect, MGFshave to be initiated soon after completion ofchemotherapy, before the development of neu-tropenia, and continued through the period ofneutropenia.86

Role of MGFs in acute leukemia

Despite appreciable progress in supportivecare, infectious morbidity and mortality remaina major impediment in the treatment of acuteleukemia, particularly in older patients. Pre-existing immune deficiency caused by the dis-ease, as well as the profound and prolongedneutropenia associated with chemotherapy foracute leukemia, both contribute to the high riskof infectious complications in leukemia therapy.

MGFs have been assessed as an adjunct toinduction and to consolidation chemotherapyin acute leukemia. Several randomized studiesutilizing G-CSF or GM-CSF during inductionchemotherapy for AML have generated quiteconsistent results: a modest decrease in theduration of neutropenia was observed, with avariable effect on the incidence of serious infec-tions and resource utilization (Tables 13.7 and13.8).87–97 There has been no reproducibleimprovement in the clinically significant mea-sures of outcome: complete response (CR) rate,CR duration, or overall survival (OS). Only onestudy88 demonstrated a benefit in a subgroup ofpatients with persistent leukemia who receiveda second course of chemotherapy during neu-tropenia.98 One other study using G-CSFshowed an effect on CR rate,96,97 which did nottranslate, however, into a decrease in induction-related mortality or improved OS. These resultshave been summarized in several reviews.99–103

The majority of patients enrolled on these pro-tocols were over 60–65 years old.

Elderly patients’ tolerance of chemotherapyis especially poor. Patients older than 60 yearshave a risk of dying during the course of induc-tion chemotherapy for acute leukemiaapproaching 50%.104 The high mortality rate is

primarily due to uncontrolled infections devel-oping during neutropenia. Thus not surpris-ingly, most of the trials of MGFs administeredin conjunction with induction for AML targetedthis particular patient population (Tables 13.7and 13.8).

Two large randomized trials94,105 and onesequential cohort study106 evaluated the role of G-CSF during consolidation therapy forAML. Although there was no effect on overallsurvival, the reduction in the duration of neu-tropenia appeared more substantial: 5–6.5 dayswith consolidation versus 2–5 days with induc-tion. As opposed to induction therapy, mostpatients after consolidation can be followed onan outpatient basis. Thus, a decrease in the rateof even uncomplicated febrile neutropeniacould result in less hospitalization. None of thethree studies, however, was designed to assessthe impact of MGFs on clinical endpoints.

The role of MGFs as an adjunct to intensivechemotherapy for adult acute lymphoblasticleukemia (ALL) has been studied less exten-sively.107–110 One large trial107 demonstrated atrend towards a higher CR rate and fewerdeaths during remission induction, particularlyin older patients: the ultimate outcome, though,was not altered, and the leukemia-free survivaland OS were the same in the two arms.Interestingly, the time to completion of inten-sive chemotherapy was not shortened inpatients on the G-CSF arm, despite an acceler-ated recovery of the blood counts. This may bedue to the fact that infectious complicationswere not different between the two groups.There was also no apparent advantage to MGFsadministered in consolidation cycles. The inves-tigators felt that, based on the results of thistrial, G-CSF may be recommended in conjunc-tion with induction chemotherapy, but shouldnot be used routinely in the postremission treat-ment. The two other randomized trials con-firmed the finding of decreased neutropeniawith MGFs administered during inductionchemotherapy; one trial showed a significantreduction in the incidence of infections, whilethe other only a trend to fewer infections.108,109


Tabl

e 13.7

Sum

mar

y of

ran

dom

ized

trial

s of

GM

-CSF

wit

h in

itia

l ind

ucti

on t

hera

py for

pat

ient

s w

ith

AM

La

Tria

lN

o. o

fA

geb

Sta

rt o

f M

GF

inA

rms

Neu

trop

enia

Clin

ical

ly r

elev

ant

Res

ourc

e ut

iliza

tion

d

(coo

pera

tive

gro

up)

pati

ents

(yea

rs)

rela

tion

to

(med

ian

dura

tion

,ev

ents

chem

othe

rapy

(da

ys)

days

)In

duct

ion

CR

cA

ntib

ioti

cH

ospi

taliz

atio

nde

aths

(%

)(%

)us

age

Sto

ne e

t al

87

388

≥60 (69)

8G

M-C

SF

(E. co

li)1

5e

27

51

28

(CAL

BG

)Pl

aceb

o17

23

54

—30

Row

e et

al8

8124

55–7

0 (64)

11 (if

mar

row

apl

astic

)G

M-C

SF

(yea

st)

11

e6

60

——

(EC

OG

)Pl

aceb

o14

15

44

Witz

et

al8

9244

55–7

5 (67)

1G

M-C

SF

(E. co

li)2

4e

18

63

23

e3

0(G

OEL

AM)f

Plac

ebo

29

15

60

25

33

Zitt

oun

et a

l90

102

17–5

9 (43)

�1 o

r 8

GM

-CS

F (E

. co

li)2

16

43

e—

—(E

OR

TC/G

IMEM

A)f

Plac

ebo

24.5

87

7

Buc

hner

et

al9

163

16–7

5 (51)

�1

GM

-CS

F (E

. co

li)R

educ

ed5

74

——

Con

trol

58

2

Low

enbe

rg e

t al

92

253

<60 (42)

�1 o

r 8–9

GM

-CS

F (E

. co

li)2

4e

7 (al

l)77

—3

0(H

OVO

N/S

AKK

)fC

ontr

ol27

77

31

Low

enbe

rg e

t al

93

326

>60 (68)

�1 t

o ne

utro

phil

reco

very

GM

-CS

F (E

. co

li)2

3e

14

56

20

32

.5(E

OR

TC/H

OVO

N)f

Con

trol

25

13

55

16

32

aM

odifi

ed a

nd r

eprin

ted

with

per

mis

sion

fro

m S

chiff

er C

A, B

lood

1995; 8

8: 3675–8

5.9

9

bM

edia

n in

par

enth

eses

.cC

ompl

ete

resp

onse

rat

e. d

—, no

t as

sess

ed.

ep

<0

.05

.fS

tudi

es t

hat

also

eva

luat

ed p

rimin

g.

Tabl

e 13.8

Sum

mar

y of

ran

dom

ized

trial

s of

G-C

SF

wit

h in

itia

l ind

ucti

on t

hera

py for

pat

ient

s w

ith

AM

La

Tria

lN

o. o

fA

geb

Sta

rt o

f M

GF

inA

rms

Neu

trop

enia

Clin

ical

ly r

elev

ant

even

tsR

esou

rce

utili

zati

ond

(coo

pera

tive

gro

up)

pati

ents

(yea

rs)

rela

tion

to

(med

ian

dura

tion

, ch

emot

hera

pyda

ys)

Indu

ctio

nC

Rc

Ant

ibio

tic

Hos

pita

lizat

ion

(day

s)de

aths

(%

)(%

)us

age

Hei

l et

al9

4521

16–8

9 (54)

8G

-CS

F20

eN

A6

91

5e

23

e

(Am

gen

Mul

ti-In

stitu

tiona

l)Pl

aceb

o25

68

18.5

29

God

win

et

al9

5211

>55 (68)

11

G-C

SF

Red

uced

by

3–4

day

se2

04

12

22

9(S

WO

G)

(if m

arro

wPl

aceb

o19

50

26

29

hypo

cellu

lar)

Dom

bert

et

al9

6173

>65 (71)

8G

-CS

F21

e2

37

0e

——

(AM

L C

oope

rativ

e S

tudy

Plac

ebo

27

27

47

Gro

up: Fr

ance

/Bel

gium

)

aM

odifi

ed a

nd r

eprin

ted

with

per

mis

sion

fro

m S

chiff

er C

A, B

lood

1995; 8

8: 3675–8

5.9

9

bM

edia

n in

par

enth

eses

.cC

ompl

ete

resp

onse

rat

e. d

—, no

t as

sess

ed.

ep

<0

.05

.

No benefit from MGFs on resource utilizationwas seen. The impact on long-term survivalwas not reported in either of the two studies.

A few randomized studies looked at adjuvantG-CSF after induction chemotherapy in child-hood ALL.111–114 As with the results of the trialsin adult ALL, no beneficial effect on clinicalevents was demonstrated in this setting either.

In summary, the use of MGFs in the treat-ment of acute leukemia appears to be safe, but abeneficial effect on clinically important events,such as lower rates of serious neutropenic infec-tions, improved CR rate, or improved OS, hasnot been convincingly proven. Using GM-CSFand G-CSF with induction chemotherapy inelderly patients does deserve consideration,although supportive data in this regard do notappear very convincing.

Role of MGFs in high-dose chemotherapywith stem cell transplantation

The severity of neutropenia – depth as well asduration – is dependent on the dose intensity ofthe treatment regimen, and is particularly pro-found in the SCT population because of theintensive conditioning regimens employed.Recovery of hematopoiesis after SCT follows apredictable pattern of pancytopenia lasting 2–4weeks. A variable degree of other non-hematopoietic conditioning-regimen-inducedtoxicities affecting vital body organs usuallyoccurs during this period of time, which greatlycontributes to the risk of infections. Owing tothe relatively high morbidity and mortalityassociated with the early post-transplantperiod, necessitating intensive use of healthcare resources to treat these patients, the settingof SCT appears as a potentially attractive appli-cation for MGFs.

Twenty-two randomized controlled trialshave evaluated the effectiveness of MGFs infacilitating engraftment; 16 trials were con-ducted in autologous SCT, two in both autolo-gous and allogeneic SCT, and four exclusivelyin allogeneic SCT patients (Table 13.9).115–136

Disappointingly, the results, in general, weresimilar to the conclusions of studies investigat-ing the role of MGFs in non-transplant settings.All but three studies demonstrated statisticallysignificant shortening of time to granulocyterecovery by 3–13 days in the autologous trans-plant setting and by 2–6 days in the allogeneictransplant setting. However, the impact onother outcome parameters (especially infection)seems much less impressive. Only a few studiesshowed a reduction in infectious episodes ornumber of days with fever. No study showedany effect on the rate or severity of fungal infec-tions. Most importantly, no study demonstra-ted a reduced infectious mortality or improvedOS. A reduction in resource utilization and costwas demonstrated in 12 out of the 22 studies,mainly in terms of shortening of the hospitalstay. In only a few of the studies did the use ofMGFs also affect the use of intravenous antibi-otics or utilization of amphotericin B.

The magnitude of clinical benefit appearsmuch smaller in the allogeneic SCT setting.Among four studies of MGFs performed exclu-sively in allogeneic SCT recipients, two werenegative and two were positive. The two negat-ive studies showed no evidence of improve-ment in any of the parameters: laboratory,clinical, or resource utilization. In general,available data do not offer support for the rou-tine use of MGFs in the allogeneic transplantsetting.

It is not clear why a shortening of neutrope-nia after SCT has not translated into a greaterclinically measurable benefit. Certainly, the rateof serious infectious morbidity or mortalityafter autologous SCT, although much higherthan in a conventional chemotherapy setting, isstill so small that many of the clinical trials dis-cussed here were underpowered to documentthe difference, if it exists. In the setting of allo-geneic SCT, neutropenia is only one of thenumerous risk factors for infectious complica-tions, modification of which may not be suffi-cient to change the overall risk.

The remarkable efficacy of MGFs in mobiliz-ing stem cells from bone marrow into the circu-


Tabl

e 13.9

Effe

ct o

f G

-CSF

and

GM

-CSF

on e

ngra

ftm

ent

afte

r SC

T as

eva

luat

ed in

ran

dom

ized

con

trol

led

tria

ls

Ref

eren

ceN

o. o

fM

GF

Tran

spla

ntSte

mN

eutr

open

iac,

dC

linic

ally

rel

evan

t ev

ents

dR

esou

rce

utili

zati

ond

pati

ents

type

ace

llA

NC

AN

CFe

ver

Any

Bac

tere

mia

Hos

pita

lA

ntib

ioti

cU

se o

fso

urce

b

<100/µl

<500/µl

infe

ctio

nst

ayth

erap

yam

phot

eric

in B

Khw

aja

et a

l11

561

GM

-CS

FAu

toB

MN

SS

NS

NS

NS

NS

NS

NS

Gor

in e

t al

11

691

GM

-CS

FAu

toB

M—

SN

SN

SN

SS

NS

—G

ulat

i et

al1

17

24

GM

-CS

FAu

toB

M—

——

NS

NS

S—

—Li

nk e

t al

11

881

GM

-CS

FAu

toB

M—

SN

SS

NS

NS

NS

—N

emun

aitis

et

al1

19

128

GM

-CS

FAu

toB

MN

SS

NS

NS

NS

SS

NS

Sch

mitz

et

al1

20

54

G-C

SF

Auto

BM

—S

NS

NS

NS

NS

NS

—S

tahe

l et

al1

21

43

G-C

SF

Auto

BM

—S

SN

SN

SN

SN

S—

Adva

ni e

t al

12

269

GM

-CS

FAu

toPB

, B

M—

S—

SN

SN

S—

—Le

gros

et

al1

23

50

GM

-CS

FAu

toPB

—N

SS

NS

—N

SN

S—

Spi

tzer

et

al1

24

37

G&

GM

Auto

PBN

SS

—N

SN

SS

NS

—K

lum

pp e

t al

12

541

G-C

SF

Auto

BM

—S

NS

——

SS

—Li

nch

et a

l12

663

G-C

SF

Auto

BM

—S

——

—S

——

Bis

hop

et a

l12

75

4G

-CS

FAl

loPB

—S

——

—N

A—

—Le

e et

al1

28

24

G-C

SF

Auto

PB—

SN

S—

—S

NS

NS

McQ

uake

r et

al1

29

38

G-C

SF

Auto

PBN

SS

NS

——

SN

SS

Kaw

ano

et a

l13

063

G-C

SF

Auto

PB—

SN

S—

—N

A—

—O

jeda

et

al1

31

50

G-C

SF

Auto

PB—

S—

NS

—N

S—

—G

isse

lbre

cht

et a

l13

2315

G-C

SF

Allo

, Au

toPB

, B

M—

SS

—S

SN

S—

DeW

itte

et a

l13

35

7G

M-C

SF

Allo

BM

SN

SN

SN

SN

SN

SN

S—

Nem

unai

tis e

t al

13

41

09

GM

-CS

FAl

loB

MS

SN

SS

SS

NS

NS

Pow

les

et a

l13

54

0G

M-C

SF

Allo

BM

—N

SI

NS

NS

NS

I—

Linc

h et

al1

36

121

G-C

SF

Allo

, Au

toB

M—

SN

SN

SN

SS

NS

S

aAu

to, au

tolo

gous

; Al

lo, al

loge

neic

. b

BM

, bo

ne m

arro

w; PB

, pe

riphe

ral b

lood

. cAN

C, ab

solu

te n

eutr

ophi

l cou

nt.

dS

, p

<0.0

5; N

S, no

t si

gnifi

cant

ly d

iffer

ent;

—, no

t as

sess

ed; I,

MG

F gr

oup

infe

rior

to c

ontr

ol.

lation made peripheral blood stem cell trans-plants (PBSCT) possible, and, within a shortperiod of time, peripheral blood replaced bonemarrow as the major source of the stem cells forautologous SCT and is currently being evalu-ated in allogeneic SCT. The main advantage ofPBSCT over BMT is the ability to collect morestem cells, which facilitates faster engraftment.Many of the studies that demonstrated a benefitfrom MGFs after autologous SCT were per-formed before PBSCT became available.Whether the conclusions from these trials applyto PBSCT is debatable. An accelerated recoveryof peripheral blood counts and a reduction inthe antibiotic requirements with MGFs afterPBSCT have been confirmed in a few smallstudies.123,124,137–139 However, the same studiesplus a few small single-arm ones strongly sug-gest that the benefit of MGFs may be verysmall, and may even be negligible if the num-ber of the stem cells is optimal.128–131,138,140 Theonly published randomized trial of G-CSF afterallogeneic PBSCT showed faster recovery ofgranulocytes in this setting.141

Choosing a more cost-effective dosing sched-ule may be a method to decrease the cost oftransplant if one opts to use MGFs. Differentschedules of treatment with MGFs followingSCT have been evaluated. The dose of 5 µg/kg/day appears equivalent to 10 µg/kg/day.142

Furthermore, the results of several studies,143–152

including five randomized,143–147 strongly implythat the delayed initiation of treatment withMGF until day 5–7 does not negatively affectengraftment. Another way to possibly reconcilethe potential clinical benefit with pharmacoeco-nomic demands would be to adopt a modifiedrisk-stratified approach, which would call forMGF support only in patients who receive asuboptimal quality graft.

Dose and schedule of administration ofrecombinant MGFs

The MGF dose recommendations issued in theASCO guidelines were 5 µg/kg/day for G-CSF

and 250 µg/m2/day for GM-CSF.30 The initialapproval of MGF by the FDA for use in particu-lar clinical settings was accompanied by specificdose and schedule directions.153,154 The doses of G-CSF initially approved by the FDA were G-CSF 5 µg/kg/day in conventionalchemotherapy and 10 µg/kg/day in the trans-plant setting. In the clinical trials that followed,the issue of optimal dose was rarely raised.155,156

One exception was the setting of autologousSCT, where a randomized study demonstratedequal efficacy of G-CSF at 5 µg/kg/day and10 µg/kg/day.142 Although the issue ofincreased efficacy of the higher dose levels hasnot been addressed by randomized trials in thenon-transplant setting, circumstantial evidencesuggests this to be of no benefit. Interestingly,there is some information suggesting that lowerdoses of G-CSF, but not GM-CSF, may beequally efficacious.155–157 In lieu of this informa-tion, rounding the dose to the nearest vial sizehas been considered acceptable. Both subcuta-neous and intravenous routes of administrationare comparable, with the subcutaneous routepreferred by both physicians and patientsowing to cost and convenience.158–160

Initiation and stopping rules for MGFs rela-tive to chemotherapy have been a matter of con-troversy. Based on available data, it appears thatstarting MGFs 24–72 hours after completion ofchemotherapy may be optimal. Initiation ofMGFs prior to chemotherapy can lead to moreprofound neutropenia, and therefore should beavoided.161 Delayed start of therapy with MGFshas been employed by various centers todecrease utilization of these costly drugs. Onesmall phase II randomized study suggested thata delay in initiation of G-CSF from day 4 ofchemotherapy to day 6 was associated with asimilar pattern of hematologic recovery, but afurther delay to day 8 resulted in a less favor-able response.86 This contrasts with the autolo-gous SCT trials, where no advantage to an earlystart (vs delayed until day 5–7) was seen.

The optimal duration of therapy with MGFsis even more controversial. The FDA-approvedpackage circulars specify for G-CSF to be con-


tinued until the absolute neutrophil count(ANC) is > 10 000/µl in the conventionalchemotherapy setting and until the ANC is> 1000/µl for three consecutive days in thetransplant setting.153 The discontinuationinstruction for GM-CSF specifies recovery of the absolute neutrophil count to anANC > 20 000/µl.154 These recommendationsare certainly safe and effective, but such pro-longed courses of treatment appear not to benecessary for optimal effect. Indeed, many cen-ters have introduced early-stopping rules forMGFs with no deleterious effect on the outcome.

Toxicities with recombinant MGFs

In general, both G-CSF and GM-CSF are verywell tolerated at therapeutic doses. The side-effects seen in association with MGFs aresummarized in Table 13.10. The main toxicity ofMGFs is mild-to-moderate bone pain, usuallyeasily controllable by acetaminophen (paraceta-mol). Bone pain is practically the only signific-ant side-effect from G-CSF. The commonadverse reactions seen in patients treated withGM-CSF are those characteristic of proinflam-matory cytokine stimulation, such as fever,myalgia, and malaise.

The first dose of Escherichia coli-derivedGM-CSF can cause rash, pruritus, arthralgias,and even cardiovascular events such as tachy-cardia and hypotension. The administration ofsubsequent doses, though, is usually unevent-ful. A capillary leak syndrome has beenreported to occur in a dose-dependent fashion,but with dosages much higher than those rou-tinely used in clinical practice. These untowardside-effects, including the ‘first-dose effect’,have been described mainly with the E. coliproduct, and not with yeast GM-CSF.

The majority of clinical trials looking at theclinical activity of GM-CSF used the E. coli-derived formulation, which has a much lessfavorable toxicity profile in comparison withthe yeast-derived formulation. The toxicities ofthe only GM-CSF formulation commercially

available at present in the USA, i.e. a yeast-derived formulation, appear to be comparablewith those of G-CSF.43,162,163

GRANULOCYTE TRANSFUSIONS

The role of granulocyte transfusions (GTX) inrestoration of the host defense system inseverely neutropenic patients has been studiedfor more than 30 years (reviewed in references164–168). The initial trials performed in theearly 1970s yielded encouraging results. Theintuitive appeal of this approach gained manyproponents who argued that controlled trials ofthe seemingly obvious efficacy of GTX mightnot be necessary. However, reports of frequent,often severe, complications associated withGTX slowly dampened the initial enthusiasm.Additionally, the process of collection and stor-age of granulocytes proved very cumbersome.The required technology was not widely avail-able. As a consequence, GTX slowly fell out offavor as a therapeutic tool. This coincided intime with significant advances in the treatmentof neutropenic infections due to availability ofnew classes of antibiotics and MGFs.

However, in recent years, the emergence ofnew and more resistant pathogens, especiallyinvasive fungi and antibiotic-resistant bacteria,has led to renewed interest in this treatmentmodality. Additionally, advances in collectionmethods now allow the collection of muchlarger numbers of granulocytes, making thewhole procedure more feasible.

Methods of collection of white blood cellsfor GTX

The inability to collect a sufficient amount ofgranulocytes has been a major obstacle to theuse of GTX until the advent of MGFs in 1991.G-CSF administered to normal donors canincrease the number of circulating granulocytes10-fold and the number of circulating granulo-cyte progenitors 40-fold.15 This permits the


collection of a large number of neutrophils byroutine centrifugation leukapheresis. The aver-age yield of G-CSF-stimulated granulocytesrepresents a three- to fivefold increase (range3–7 � 1010) over that reported historicallyfollowing stimulation of donors with corticos-teroids alone.169,170 The progressive and sus-tained increases in precollection leukocytecount achieved with G-CSF result in greaternumbers of leukocytes collected during consec-utive leukapheresis days.171,172 GM-CSF appears

to be less effective in this regard, since it causesonly a twofold rise in precollection leukocytecount, without a continuing rise during succes-sive days of collection.171 In the majority of studies, G-CSF was administered at a dose of 5 µg/kg/day subcutaneously, 8–12 hoursbefore the first scheduled leukapheresis, andthen continued daily. After multiple doses ofG-CSF, neutrophilia was maintained. Therefore,the timing of subsequent leukaphereses inrespect to the G-CSF injections is not as critical.


Table 13.10 Side-effects associated with MGFsa

GM-CSF G-CSF

Non-dose-relatedFever Common RareBone pain Common (10%) Common (10%)Myalgia Common RareCatheter thrombosis Rare Not reportedSplenomegaly Rare Rare

Dose related (>32 µg/kg/day)Effusion

Pericardial � �

Pleural � �

Ascites � �

Pulmonary emboli � �

Edema � �

Weight gain � �

Laboratoryb

Increased LDH � �

Increased LAP � �

Increased uric acid � �

Increased ALP � �

Increased eosinophils � �

a Reprinted with permission from Negrin RS, Clinical applications of hematopoietic growth factors. In: The Cytokine Handbook,3rd edn (Thomson A, ed). London: Academic Press, 1998.b LDH, lactate dehydrogenase; LAP, leukocyte alkaline phosphatase; ALP, alkaline phosphatase.

A daily dose schedule of 10 µg/kg/day and analternative-day schedule appear to be compara-ble.173 More recently, the combined use ofG-CSF and corticosteroids to increase yield andto maintain the function of collected neu-trophils was reported in this setting.174,175 Itremains to be seen which mobilization regi-mens will prove superior.

Filtration leukapheresis through nylon/woolfibers has been used in the past for collection ofgranulocytes for GTX. This procedure was laterdiscovered to cause significant changes in neu-trophil function,176–178 which were believed to be responsible for many serious reactionsobserved in both the donor and the recipient.179

Currently, granulocytes are collected exclu-sively by centrifugal separation.180

The quality of GTX product

At present, GTX is not an FDA-approved treat-ment modality, and optimal granulocyte con-centrate specifications have not been defined orofficially formulated. It is generally acceptedthat a minimum of 1 � 1010 cells should be pre-sent in each unit of a concentrate.180 In vitroassays to test granulocyte function in a concen-trate would be desirable for quality assurance,but are not routinely performed.

Both G-CSF and corticosteroids included inthe mobilization regimen are known to alterneutrophil function. The clinical importance ofthe reported enhanced phagocytic and bacteri-cidal activity of granulocytes exposed to G-CSFhas not been studied extensively, but appearsrelatively minor.177,181 The decreased recoveryand prolonged circulatory half-life ofG-CSF-stimulated granulocytes have raised theconcern of impairment in cellular mobility.174,175

However, the ability of these neutrophils tomigrate into skin window chambers and intosites of inflammation and infection has beendocumented, reaffirming the lack of detrimen-tal effects of G-CSF/corticosteroids in thisregard.17,175–181 Information about the propertiesof stimulated granulocytes is still scant, and

hopefully will be studied further.Irradiation is generally performed as a

method to prevent transfusion-associated graft-versus-host-disease; whether this is necessaryremains unresolved. While this almost invari-ably fatal complication has been reported aftergranulocyte transfusion, it is rare.

Dose and side-effects of GTX

The optimal dose of granulocytes for the treat-ment of neutropenic infections is not known.Early trials using donors with CML suggestedthat a dose of 1 � 1010/m2 or more would bedesirable.182 Thus, a daily transfusion of a mini-mum of 1 � 1010 cells, which correspondsroughly to the product of one leukapheresisprocedure contained in one unit of the concen-trate, seems a logical target.183 The prolongedcirculation of G-CSF-stimulated transfusedgranulocytes suggests that an every-other-dayschedule may provide sufficient support.173,184 Itis thought that GTX should be continued untilresolution of infection or recovery of bloodcounts.

Acute toxicities such as fever, chills, andurticaria can occur during GTX. It appears thata slow rate of GTX infusion is crucial to avoidthese symptoms. Should they occur, furtherslowing down the infusion rate could alleviatethem. Routine premedication with antihista-mines and/or antipyretics might be helpful. Inthe event of a severe reaction, the infusionshould be stopped and the patient should betested for the presence of antileukocyte anti-bodies.

Respiratory distress and appearance of pul-monary infiltrates have also been seen in associ-ation with GTX. Severe pulmonary reactionsreported in the early 1980s were attributed tothe concomitant administration of GTX withamphotericin B.185,186 However, later reportsfailed to confirm the interaction between GTXand amphotericin B in causing lung injury.187,188

Indeed, in more recent studies, adverse reactionsto G-CSF-mobilized granulocyte transfusions


have been rare in general. Interestingly, therewas no correlation between the risk of unto-ward side-effects from GTX and the degree ofleukocyte antigen incompatibility between thedonor and the recipient,173–175 although,traditionally, GTX-related toxicities have beenthought to be more severe and more commonin alloimmunized recipients.189,190

Differences in the method of collection ofgranulocytes could explain the disconcordantresults between the older and newer studies.Many of the side-effects observed in the earlytrials might have been due to damage to granu-locytes procured through filtration leukaphere-sis. In vitro experiments demonstrated thatamphotericin B can trigger aggregation of neu-trophils injured by filtration leukapheresis.Whether the same interaction occurs with intactneutrophils collected through centrifugal sepa-ration is uncertain, but seems unlikely. None ofthe leukocyte antibodies detected in patientstreated on the newer protocols showed neu-trophil agglutination, a postulated prerequisitefor causing transfusion related acute lunginjury.175,189 Based on the reports of serious toxi-cities in early studies, it has become commonpractice to separate temporally the infusions ofamphotericin B and GTX by at least 8 hours.

Factors affecting the dynamics of transfusedgranulocytes

Transfusions of G-CSF-mobilized granulocytesinto neutropenic recipients produce incrementsin circulating granulocyte levels, measurablefor hours after transfusion – even up to 24hours.191 This was not observed when cortico-steroids alone were used for stimulation.171 Asustained increase in the granulocyte countafter G-CSF-mobilized GTX may be due pos-sibly to the large cell dose or possibly to infu-sion of early progenitors, with a longer half-life.Interestingly, though, a consistent relationshipbetween the cell dose delivered over a numberof days and the increments over a baseline neu-trophil count has not been found in a study

designed to address this question.173,191

The impact of alloimmunization of the recipi-ent on the effectiveness of treatment with GTXis a matter of controversy.192,193 In non-alloimmunized patients, granulocytes can betransfused safely. Granulocyte concentratesusually contain large numbers of erythrocytes,and for this reason transfusions should beABO-compatible. However, there is no evid-ence that ABO incompatibility affects the trans-fused granulocytes. Patients can develop twotypes of antibodies directed against two differ-ent classes of antigens on the surface of neu-trophils: HLA antibodies detectable bylymphocytotoxicity assays and granulocyte-specific antibodies detectable by leukoaggluti-nation or by indirect immunofluorescence.Alloimmunization is a known complication ofGTX in immunocompetent patients.192 How-ever, some of the newer studies suggest thatalloantibodies may be less of a problem in SCTrecipients. A few conflicting reports as to thesignificance of the alloantibodies in GTX recipi-ents have appeared in the recent literature. Thepresence of detectable antibodies prior to thestart of treatment or acquired antibodies duringtreatment with granulocyte infusions adverselyaffected neutrophil increments after transfusionof granulocytes and delayed engraftment in onerecent study.172

Evidence of efficacy of treatment with GTX:summary of clinical trials

Early clinical trials investigating granulocytetransfusions strongly suggested a considerablebenefit of this treatment modality. However,the findings were not consistent, and the clini-cal effects in individual patients were notalways obvious (see Table 13.11 for a summaryof the trials194–200). Most of the patients studiedhad Gram-negative bacterial septicemia.Significant benefits from GTX were seen onlywhen bone marrow recovery was delayed formore than 1 week (but did occur within 2–3weeks) and when leukocyte transfusions were


given for at least 4–7 days. There appears to bea dose-threshold effect, since the negative trialsused low doses (<1 � 1010). The positive trialsused higher doses (>1 � 1010). Two other factorsmight explain the negative results of some stud-ies: (1) good response to antibiotics in botharms, making the detection of the benefit, ifany, difficult in two studies; (2) not testing forleukocyte incompatibility.

Several encouraging case reports and pilotstudies utilizing G-CSF (�corticosteroid)-stim-ulated granulocyte infusions have beenreported in the past few years.173,174,201–204 The

trials suggest a benefit of G-CSF-stimulatedGTX in the treatment of bacteriemias and can-didemias. However, activity in the treatment ofinvasive mold infections such as aspergillushas not been consistently demonstrated.175

Unfortunately, most patients in these studieswere quite sick before starting treatment withGTX. Worsening of overall clinical status andprogression to multiorgan failure has beenobserved in many cases, despite evidence ofclearing the infection. Thus, very few patientsin these trials emerged as long-term survivors.This suggests that the GTX might have been


Table 13.11 Controlled studies of therapeutic granulocyte transfusion (GTX) in neutropenic patientsa

Reference Randomized Treatment No. of Survival Doseb HLA–WBCc Success of GTX

arms patients rate (%) (�1010 cells)

Higby et al194 Yes GTX 17 76 2.2 (FL) No–Yes Yes

Control 19 26

Vogler et al195 Yes GTX 17 59 2.7 (CL) Yes–Yes Yes

Control 13 15

Herzig et al196 Yes GTX 13 75 1.7 (FL) No–Yes Yes

Control 14 36 0.4 (CL)

Alavi et al197 Yes GTX 12 82 5.9 (FL) No–No Partial

Control 19 62

Winston et al198 Yes GTX 48 63 0.5 (CL) No–No No

Control 47 73

Graw et al199 No GTX 39 46 2.0 (FL) No–Yes Partial

Control 37 40 0.6 (CL)

Fortuny et al200 No GTX 17 78 0.4 (CL) No–Yes No

Control 22 80

a Modified and reprinted with permission from Strauss RG, J Pediatr Hematol Oncol 1999; 21: 475–8.165

b CL, centrifugation leukapheresis; FL, filtration leukapheresis.c HLA, human leukocyte antigen; WBC, white blood cell.

started too late. Improvements in early diag-nostic methods for fungal infections are neededto allow timely initiation of treatment withGTX.

In the face of a lack of randomized, con-trolled clinical trials, it is premature to draw adefinitive conclusion as to the effectiveness ofGTX collected from G-CSF-stimulated donorsin the prevention or treatment of severe neu-tropenic infections. The collection, storage, andreinfusion of granulocytes are both complexand costly. A randomized trial designed toassess the efficacy of G-CSF-stimulatd GTX inthe treatment of serious infections would be ofgreat interest.

CONCLUSIONS

It is apparent that recombinant growth factorshave a role in the management of febrile neu-tropenia. A hastened recovery from neutrope-nia has been consistently found. However, byand large, they have failed to fulfill all of theirearly promise. The observed decrease in depthand duration of neutropenia was expected toreduce the rate of infectious complications.While most clinicians believe that it does, thetrials have not convincingly demonstratedthis. Adjuvant treatment with MGFs was alsoanticipated to allow delivery of full and evenescalated doses of chemotherapy withoutdelays caused by prolonged neutropenia, and thus to improve its effectiveness.Disappointingly, only some decrease in thenumber of febrile episodes has been appreci-ated, but no reduction in the frequency ofsevere infections and no improved survivalfrom treatment with growth factors have beenconvincingly demonstrated. In light of the lackof convincing efficacy of MGFs in reducingserious morbidity or mortality, treatment withMGFs is often justified by economical or QoLconsiderations. Yet, even with these considera-tions, more information is needed. Clinicaltrials designed with these endpoints as pri-mary objectives could allow more precise and

reliable measurement of the magnitude of thisbenefit.

It is perplexing why shortening of neutrope-nia has not translated into clinically measurablebenefits. Certainly, rates of serious infectiousmorbidity or mortality in populations studiedwere often relatively small, and many of theclinical trials were underpowered to documentdifferences, if they exist. Another explanation isthat MGFs may not be able to prevent or affectthe duration of severe neutropenia of less than100/µl, the circumstance in which most neu-tropenic infections occur. Indeed, MGFs seemmost efficient in hastening the recovery of gran-ulocytes between 100/µl and 1000/µl, as seenin the SCT trials. It is also possible that alter-ations in neutrophil migratory functionreported by some investigators affect the capa-bility of mature granulocytes to reach infectedtissues, although this certainly remains dis-putable.

Much has been learned from the early clini-cal trials evaluating the efficacy of MGFs inparticular clinical settings. Identification ofpopulations of patients who could uniformlybenefit from adjuvant treatment with MGFsremains a challenge and would be of theutmost importance. The insights gained fromstudies identifying patients at high and lowrisk for severe neutropenic infection should beused in the design of future growth factortrials.

Treatment of neutropenic infections withGTX has become more feasible thanks to theavailability of MGFs. More importantly, therisks associated with these G-CSF-stimulatedGTX appear to be acceptable. The concernsregarding serious pulmonary reactions that lim-ited widespread application of this treatmentmodality in 1970s have been dissipated by themore recent studies of G-CSF-stimulatedGTX. Indeed, GTX now seem well tolerated byrecipients. However, large trials to establish theefficacy of this treatment modality in particularclinical settings are needed.


REFERENCES

1. Bodey GP, Buckley M, Sathe YS et al,Quantitative relationship between circulatingleukocytes and infection in patients with acuteleukemia. Ann Intern Med 1966; 64: 328–40.

2. Wingard JR, Santos, Saral R, Differencesbetween first and subsequent fevers duringprolonged neutropenia. Cancer 1987; 59: 844–9.

3. Hughes WT, Armstrong D, Bodey GP et al,Guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

4. Rubin M, Hathorn JW, Pizzo PA, Controversiesin the management of febrile neutropeniccancer patients. Cancer Invest 1988; 6: 167–84.

5. Babior BM, Golde DW, Production, distributionand fate of neutrophils. In: Williams’ Hematology(Beutler E, Lichtman MA, Coller BS, Kipps TJ,eds). New York: McGraw-Hill, 1995: 773–9.

6. Sieff CA, Williams DA, Hematopoiesis. In:Blood: Principles and Practice of Hematology(Handin RI, Stossel TP, Lux SE, eds).Philadelphia: JB Lippincott, 1995: 171–224.

7. Marsh JC, Boggs DR, Cartwright GE, WintrobeMM, Neutrophil kinetics in acute infection. JClin Invest 1967; 46: 1943–53.

8. Dancey JT, Deubelbeiss KA, Harket LA, FinchCA, Neutrophil kinetics in man. J Clin Invest1976; 58: 705–15.

9. Leary AG, Strauss LC, Civin KC, Ogawa M,Disparate differentiation in hemopoieticcolonies derived from human paired progeni-tors. Blood 1985; 66: 327–32.

10. Lemishka IR, Raulet DH, Mulligan RC,Developmental potential and dynamic behaviorof hematopoietic stem cells. Cell 1986; 45:917–24.

11. Metcalf D, Clonal Culture of Hemopoietic Cells.Amsterdam: Elsevier/North American Biomed-ical Press, 1984.

12. Shivdasani RA, Orkin SH, The transcriptionalcontrol of hematopoiesis. Blood 1996; 87:4025–39.

13. Demetri GD, Griffin JD, Granulocyte colony-stimulating factor and its receptor. Blood 1991;78: 2791–808.

14. Lieschke GJ, Burgess AW, Granulocyte colony-stimulating factor and granulocyte–macrophage stimulating factor. N Engl J Med1992; 327: 28–35; 99–106.

15. Price TH, Chatta GS, Dale DC, The effect ofrecombinant granulocyte colony-stimulatingfactor of neutrophil kinetics in normal youngand elderly humans. Blood 1996; 88: 335–40.

16. Welte K, Gabrilove J, Bronchud MH et al,Filgrastim (r-metHuG-CSF): the first 10 years.Blood 1996; 88: 1907–29.

17. Dale DC, Liles WC, Summer W, Nelson S,Granulocyte colony-stimulating factor: role andrelationship in infectious diseases. J Infect Dis1995; 172: 1061–75.

18. Liles WC, Huang JE, van Burik JH et al,Granulocyte colony-stimulating factor adminis-tered in vivo augments neutrophil-mediatedactivity against opportunistic fungal pathogens.J Clin Infect Dis 1997; 175: 1012–15.

19. Allen RC, Stevens P, Price TH et al, In vivoeffects of recombinant human granulocytecolony-stimulating factor on neutrophil oxida-tive functions in normal human volunteers. JInfect Dis 1997; 175: 1184–92.

20. Chatta GS, Price TH, Allen RC, Dale DC, Theeffects of in vivo recombinant human granulo-cyte colony-stimulating factor (rhG-CSF) onneutrophil response and peripheral bloodcolony forming cells in healthy young andelderly volunteers. Blood 1994; 84: 2923–9.

21. Aglietta M, Piacibello W, Sanavio F et al,Kinetics of human hematopoietic cells after in vivo administration of granulocyte–macrophage colony-stimulating factor. J ClinInvest 1989; 83: 551.

22. Neumunitis J, Granulocyte–macrophage colony-stimulating factor: a review from clinical development to clinical application. Transfusion1993; 33: 70–83.

23. Foulke RS, Marshall MH, Trotta PP, Von HoffDD, In vitro assessment of the effects of granu-locyte–macrophage colony-stimulating factoron primary human tumors and derived lines.Cancer Res 1990; 50: 6264–7.

24. Gasson JC, Molecular physiology of granulo-cyte–macrophage colony-stimulating factor.Blood 1991; 77: 1131–45.

25. Cannistra SA, Griffin JD, Regulation of the pro-duction and function of granulocytes andmonocytes. Semin Hematol 1988; 25: 173–88.

26. Hill ADK, Naama HA, Calvano SE, Daly JM,The effect of granulocyte–macrophage colony-stimulating factor on myeloid cells and its clini-cal applications. J Leucocyte Biol 1995; 58: 634.


27. Armitage JO, Emerging applications of recom-binant human granulocyte–macrophagecolony-stimulating factor. Blood 1998; 92:4491–508.

28. Root RK, Dale DC, Granulocyte colony-stimulating factor and granulocyte–macrophage colony-stimulating factor: compar-isons and potential for use in the treatment ofinfections in nonneutropenic patients. J InfectDis 1999; 179(Suppl 2): S342–52.

29. Wingard JR, Growth factors and otherimmunomodulators. In: Transplant Infections(Bowden R, Ljungman P, Paya C, eds).Philadelphia: Lippincott-Raven, 1998: 367–78.

30. American Society of Clinical OncologyRecommendations for the Use of HematopoieticColony-Stimulating Factors: evidence-based,clinical practice guidelines. J Clin Oncol 1994;12: 2471–508.

31. American Society of Clinical OncologyRecommendations for the Use of HematopoieticColony-Stimulating Factors: evidence-basedclinical practice guidelines. J Clin Oncol 1996;14: 1957–60.

32. Ozer H, Armitage JO, Bennett CL et al, 2000update of Recommendations for the Use ofHematopoietic Colony-Stimulating Factors:evidence-based clinical practice guidelines. JClin Oncol 2000; 18: 3558–85.

33. Crawford J, Ozer H, Stoller R et al, Reductionby granulocyte colony-stimulating factor offever and neutropenia induced by chemother-apy in patients with small-cell lung cancer (r-metHuG-CSF). N Engl J Med 1991; 325: 164–70.

34. Trillet-Lenoir V, Green J, Manegold C et al,Recombinant granulocyte colony-stimulatingfactor reduces the infectious complications ofcytotoxic chemotherapy. Eur J Cancer 1993; 33A:319–24.

35. Pettengell R, Gurney H, Radford JA et al,Granulocyte colony-stimulating factor to pre-vent dose-limiting neutropenia in non-Hodgkin’s lymphoma: a randomized controlledtrial. Blood 1992; 80: 1430–6.

36. Gebbia V, Testa A, Valenza R et al, A prospec-tive evaluation of the activity of humangranulocyte-colony stimulating factor on theprevention of chemotherapy-related neutrope-nia in patients with advanced carcinoma. JChemother 1993; 5: 186–90.

37. Zinzani PL, Pavone E, Storti S et al,

Randomized trial with or without colony-stimulating factor as adjunct to inductionVNCOP-B treatment in elderly high grade non-Hodgkin’s lymphoma. Blood 1997; 89: 3975–9.

38. Kaplan LD, Kahn JO, Crowe F et al, Clinicaland virologic effects of recombinant humangranulocyte–macrophage colony-stimulatingfactor in patients receiving chemotherapy forhuman immunodeficiency virus-associatednon-Hodgkin’s lymphoma: results of a random-ized trial. J Clin Oncol 1991; 9: 929–40.

39. Gerhartz HH, Engelhard M, Meusers P et al,Randomized double-blind, placebo-controlledphase III study of recombinant human granulo-cyte–macrophage colony-stimulating factor asadjunct to induction treatment of high-grademalignant non-Hodgkin’s lymphomas. Blood1993; 82: 2329–39.

40. Hamm JT, Schiller JH, Oken MM et al,Granulocyte–macrophage colony-stimulatingfactor (GM-CSF) in small cell carcinoma of thelung (SCCL): preliminary analysis of a random-ized controlled trial. Proc Am Soc Clin Oncol1991; 10: 255.

41. Nichols C, Bajorin D, Schmoll HJ et al, VIPchemotherapy with/without GM-CSF for poorrisk, relapsed, or refractory germ cell tumors(GCT): preliminary analysis of a randomizedcontrolled trial. Proc Am Soc Clin Oncol 1991; 10:167.

42. Hamm JT, Schiller J, Oken MM et al, Dose rang-ing study of recombinant human granulocyte-macrophage colony stimulating factor in smallcell lung cancer. Proc Am Soc Clin Oncol 1993;12: 335.

43. Beveridge RA, Miller JA, Kales AN et al, Acomparison of efficacy of sargramostim (yeast-derived RhuGM-CSF) and filgrastim (bacteria-derived RhuG-CSF) in the therapeutic setting of chemotherapy-induced myelosuppression.Cancer Invest 1998; 16: 366–73.

44. Lyman GH, Lyman CG, Sanderson RA et al,Decision analysis of hematopoietic growth fac-tor use in patients receiving cancer chemother-apy. J Natl Cancer Inst 1993; 85: 488–93.

45. Mayordomo JI, Rivera F, Diaz-Puente MT et al,Decision analysis of hematopoietic growth fac-tor use in patients receiving cancer chemother-apy. J Natl Cancer Inst 1993; 85: 1251–2.

46. Bash RO, Katz JA, Cash JV et al, Safety and costeffectiveness of early discharge of relatively low


risk children with cancer hospitalized for feverand neutropenia (F/N). Proc Am Soc Clin Oncol1992; 11: 381.

47. Rolston KVI, Rubenstein EB, Elting L et al,Ambulatory management of febrile episodes inlow-risk neutropenic patients. In: Program andAbstracts of the 35th Interscience Conference ofAntimicrobial Agents and Chemotherapy, SanFrancisco, CA, 1995: 333 (Abst 2235).

48. Kern WV, Cometta A, de Bock R et al, Oral ver-sus intravenous empirical antimicrobial therapyfor fever in patients with granulocytopenia whoare receiving cancer chemotherapy. N Engl JMed 1999; 341: 312–18.

49. Finberg RW, Talcott JA, Fever and neutropenia– how to use a new treatment strategy. N Engl JMed 1999; 341: 362–3.

50. Talcott JA, Finberg R, Mayer RJ, Goldman L,The medical course of cancer patients withfever and neutropenia. Arch Intern Med 1988;148: 2561–8.


52. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low-risk cancer patients withfever and neutropenia: a pilot study of 30patients based on a validated prediction rule. JClin Oncol 1994; 12: 107–14.

53. Rubenstein EB, Rilston K, Benjamin RS et al,Outpatient treatment of febrile episodes in low-risk neutropenic patients with cancer. Cancer1993; 71: 3640–6.

54. Lau RC, King SM, Richardson SE, Early dis-charge of pediatric febrile neutropenic cancerpatients by substitution of oral for intravenousantibiotics. Pediatr Hematol Oncol 1994; 11:417–21.

55. Rolston KVI, Rubenstein EB, Freifeld A, Earlyempiric antibiotic therapy for febrile neu-tropenic patients at low risk. Infect Dis ClinNorth Am 1996; 10: 223–7.


57. Klastersky J, Paesmans M, Rubenstein EB et al,The Multinational Association for Supportive

Care in Cancer Risk Index: a multinational scor-ing system for identifying low-risk febrile neu-tropenic cancer patients. J Clin Oncol 2000; 18:3038–51.

58. Bennett CL, Weeks JA, Somerfield MR et al, Useof hematopoietic colony-stimulating factors:comparison of the 1994 and 1997 AmericanSociety of Oncology Surveys regarding ASCOClinical Practice Guidelines. J Clin Oncol 1999;17: 3676–81.

59. Goldie JH, Coldman AJ, A mathematical modelfor relating the drug sensitivity of tumors totheir spontaneous mutation rate. Cancer Res1982; 42: 965–73.

60. Logothesis CJ, Finn LD, Smith T et al, EscalatedMVAC with and without recombinant granulo-cyte–macrophage colony-stimulating factor forthe clinical treatment of advanced malignanturothelial tumors: results of a randomized trial.J Clin Oncol 1995; 13: 2272.

61. Yau J, Neidhart J, Triozzi P et al, Randomizedplacebo-controlled trial of granulocyte–macrophage colony-stimulating factor supportfor dose intensive cyclophosphamide, etoposide,and cisplatin. Am J Hematol 1996; 51: 289.

62. Steward W, von Pawel J, Gatzenmeier U et al,Effects of granulocyte–macrophage colony-stimulating factor and dose intensification ofV-ICE chemotherapy in small cell lung cancer: aprospective randomized study of 300 patients. JClin Oncol 1998; 16: 642.

63. Bajetta E, DiBartolomeo M, Carnaghi C et al,FEP regimen in advanced gastric cancer, withand without low-dose GM-CSF: an Italian Trialin Medical Oncology study. Br J Cancer 1998; 77:1149.

64. Gisselbrecht C, Haioun C, Lepage E et al,Placebo controlled phase III study of lenogras-tim (glycosylated recombinant humangranulocyte-stimulating factor) in aggressivenon-Hodgkin’s lymphoma. Leuk Lymphoma1997; 25: 289–300.

65. Sieber M, Engert A, Diehl V, Treatment ofHodgkin’s disease: results and current conceptsof the German Hodgkin’s Lymphoma StudyGroup. Ann Oncol 2000; 11(Suppl 1): S81–5.

66. Thatcher N, Girling DJ, Hopwood P et al,Improving survival without reducing the qual-ity of life in small cell lung cancer patients byincreasing the dose intensity of chemotherapywith granulocyte colony-stimulating factor sup-


port: results of British Medical ResearchCouncil Multicenter Randomized Trial. J ClinOncol 2000; 18: 395–404.

67. Kaku K, Takhashi M, Moriyama Y et al,Recombinant human granulocyte–macrophagecolony-stimulating factor after chemotherapy inpatients with non-Hodgkin’s lymphoma. Aplacebo controlled double-blind phase III trial.Leuk Lymphoma 1993; 11: 229–38.

68. Bennett CL, Smith TJ, Weeks JC et al, Use ofhematopoietic colony-stimulating factors: theAmerican Society of Clinical Oncology Survey.The Health Services Research Committee of theAmerican Society of Clinical Oncology. J ClinOncol 1996; 14: 2511–20.

69. Swanson G, Bergstrom K, Stump E et al,Growth factor usage patterns and outcomes inthe community setting: collection through apractice-based computerized clinical informa-tion system. J Clin Oncol 2000; 18: 1764–770.

70. De Graaf H, Willemse PH, Bong SB et al, Doseintensity of standard adjuvant CMF with gran-ulocyte colony-stimulating factor for pre-menopausal patients with node-positive breastcancer. Oncology 1996; 53: 289–94.

71. Ribas A, Albanell J, Bellmund J et al, Frequentdose delays and growth factor requirementswith the sequential doxorubicin–CMF schedule.Acta Oncol 1997; 36: 701.

72. Silvestri F, Velisig M, Fanim R et al,Granulocyte colony-stimulating factor (G-CSF)allows the delivery of effective doses of CHOPand CVP regimens in non-Hodgkin’s lym-phoma. Leuk Lymphoma 1995; 16: 465.

73. Maher DW, Lieschke GJ, Green M et al,Filgrastim in patients with chemotherapy-induced febrile neutropenia: a double-blind,placebo-controlled trial. Ann Intern Med 1994;121: 492–501.

74. Mayordomo JI, Rivera F, Diaz-Puente MT et al,Improving treatment of chemotherapy-inducedneutropenic fever by administration of colony-stimulating factors. J Natl Cancer Inst 1995; 87:803–8.

75. Riikonen P, Saarinen UM, Makipernaa A et al,Recombinant human granulocyte–macrophagecolony-stimulating factor in the treatment offebrile neutropenia: a double-blind placebo-controlled study in children. Pediatr Infect Dis J1994; 13: 197–202.

76. Anaissie E, Vartivarian S, Bodey GP et al,

Randomized comparison between antibioticsalone and antibiotics plus granulocyte–macrophage colony-stimulating factor(Escherichia coli-derived) in cancer patients withfever and neutropenia. Am J Med 1996; 100:17–23.

77. Biesma B, de Vries EG, Willemse PH et al,Efficacy and tolerability of recombinant humangranulocyte–macrophage colony-stimulatingfactor in patients with chemotherapy-relatedleukopenia and fever. Eur J Cancer 1990; 26:932–6.

78. Vellenga E, Uyll-de Groot CA, de Wit R et al,Randomized placebo-controlled trial of granu-locyte–macrophage colony-stimulating factor inpatients with chemotherapy-related febrile neu-tropenia. J Clin Oncol 1996; 14: 619–27.

79. Ravaud A, Chevreau C, Bonichon F et al,Granulocyte–macrophage colony-stimulatingfactor (GM-CSF) in patients with neutropenicfever is potent after low-risk but not after high-risk neutropenic chemotherapy regimens:results of a randomized phase III trial. J ClinOncol 1998; 16: 2930–6.

80. Mitchell PLR, Morland BJ, Dick G et al, Clinicalbenefits and cost savings of interventionalG-CSF therapy in patients with febrile neu-tropenia following chemotherapy. Blood 1995;86(Suppl 1): 500a.

81. Mitchell PL, Morland B, Stevens MC et al,Granulocyte colony-stimulating factor in estab-lished febrile neutropenia: a randomized studyof pediatric patients. J Clin Oncol 1997; 15:1163–70.

82. Yoshida M, Karasawa M, Naruse T et al, Effectof granulocyte-colony stimulating factor onempiric therapy with flomoxef sodium andtobramycin in febrile neutropenic patients with hematological malignancies. Kan-etsuHematological Disease and Infection StudyGroup. Int J Hematol 1999; 69: 81–8.

83. Aviles A, Guzman R, Garcia E et al, Results of arandomized trial of colony-stimulating factor inpatients with infection and severe granulocy-topenia. Anticancer Drugs 1996; 7: 392–7.

84. Blay JY, Chauvin F, Le Cesne et al, Early lym-phopenia after cytotoxic chemotherapy as a riskfactor for febrile neutropenia. J Clin Oncol 1996;14: 636–43.

85. Hartmann LC, Tschetter LK, Habermann TM etal, Granulocyte colony-stimulating factor in


severe chemotherapy induced afebrile neu-tropenia. N Engl J Med 1997; 336: 1776–80.

86. Crawford J, Kreisman H, Garewal H et al, Theimpact of filgrastim schedule variation onhematopoietic recovery postchemotherapy. AnnOncol 1997; 8: 1117–24.

87. Stone RM, Berg DT, Stephen LG et al,Granulocyte–macrophage colony-stimulatingfactor after initial chemotherapy for elderlypatients with primary acute myelogenousleukemia. N Engl J Med 1995; 25: 1671–7.

88. Rowe JM, Andersen J, Mazza JJ et al, Phase IIIrandomized placebo-controlled study on granu-locyte–macrophage colony stimulating factor(GM-CSF) in adult patients (55–70 years) withacute myelogenous leukemia (AML). A studyof the Eastern Cooperative Oncology Group(ECOG). Blood 1995; 86: 257.

89. Witz F, Sadoun A, Perrin M-C et al, A placebo-controlled study of recombinant humanmacrophage colony-stimulating factor adminis-tered during and after induction treatment forde novo acute myelogenous leukemia in elderlypatients. Blood 1998; 91: 2722–30.

90. Zittoun R, Suciu S, Mandelli F et al,Granulocyte–macrophage colony-stimulatingfactor associated with induction treatment ofacute myelogenous leukemia: a randomizedtrial by the European Organization for theResearch and Treatment of Cancer LeukemiaCooperative Group. J Clin Oncol 1996; 14:2150–9.

91. Buchner T, Hiddemann W, Wormann B et al,GM-CSF multiple course priming and long-term administration in newly diagnosedAML. Hematologic and therapeutic effect. Blood1994; 84(Suppl 1): 27a.

92. Lowenberg B, Boogaerts MA, Daenen SMGJ etal, Value of different modalities of granulo-cyte–macrophage colony-stimulating factorapplied during or after induction therapy ofacute myeloid leukemia. J Clin Oncol 1997; 15:3496–506.

93. Lowenberg B, Suciu S, Archimbaud E et al, Useof recombinant granulocyte–macrophagecolony-stimulating factor during and afterremission induction chemotherapy in patientsaged 61 years and older with acute myeloidleukemia (AML): final report of AML-11, aphase III randomized study of the LeukemiaCooperative Group of European Organization

for the Research and Treatment of Cancer(EORTC–LCG) and the Dutch Belgian Hemato-Oncology Cooperative Group (HOVON). Blood1997; 90: 2952–61.

94. Heil G, Hoelzer D, Sanz MA et al, for theInternational Leukemia Study Group, A ran-domized, double-blind, placebo-controlled,phase III study of filgrastim in remission induc-tion and consolidation therapy for adults withde novo acute myeloid leukemia. Blood 1997; 90:4710–18.

95. Godwin JE, Kopecky KJ, Head DR et al, Adouble-blind placebo-controlled trial ofgranulocyte-colony stimulating factor in elderlypatients with previously untreated acutemyeloid leukemia: a Southwest OncologyGroup Study (9031). Blood 1998; 91: 3607–15.

96. Dombret H, Yver A, Chastang C et al, theCooperative Study Group, Increased frequencyof complete remission by lenograstim recombi-nant human granulocyte colony-stimulating fac-tor (rhG-CSF) administration after intensiveinduction chemotherapy in elderly patients withde novo acute myeloid leukemia (AML): finalresults of a randomized multicenter double-blindcontrolled study. N Engl J Med 1995; 332: 1678.

97. Dombert H, Chastang C, Fenaux P et al, A con-trolled study of recombinant human granulo-cyte colony-stimulating factor in elderlypatients after treatment for acute myelogenousleukemia. N Engl J Med 1995; 332: 1678–83.

98. Bennett CL, Stinson TJ, Tallman MS et al,Economic analysis of a randomized placebo-controlled phase III study of granulocytemacrophage colony stimulating factor in adultpatients (>55 to 70 years of age) with acutemyelogenous leukemia. Eastern CooperativeOncology Group (E1490). Ann Oncol 1999; 10:177–82.

99. Schiffer CA, Hematopoietic growth factors asadjunct to the treatment of acute myeloidleukemia. Blood 1995; 88: 3675–85.

100. Geller RB, Use of cytokines in the treatment ofacute myelocytic leukemia: a critical review. JClin Oncol 1996; 14: 1371–82.

101. Rowe JM, Liesveld JL, Hematopoietic growthfactors in acute leukemia. Leukemia 1997; 11:328–41.

102. Nemunaitis J, Use of cytokines in the treatmentof acute lymphoblastic leukemia. Leukemia 1997;11(Suppl 4): S36–7.


103. Estey E, Hematopoietic growth factors in thetreatment of acute leukemia. Curr Opin Oncol1998; 10: 23–30.

104. Hiddemann W, Kern W, Schoch C et al,Management of acute myeloid leukemia inelderly patients. J Clin Oncol 1999; 17: 3569–76.

105. Harousseau JL, Witz F, Lioure B et al,Granulocyte colony-stimulating factor afterintensive consolidation chemotherapy in acutemyeloid leukemia: results of a randomized trialof the Groupe Ouest–Est Leucémies AiguësMyéloblastiques (GOELAM). J Clin Oncol 2000;18: 780–7.

106. Moore JO, Dodge RK, Amrein PC et al,Granulocyte-colony stimulating factor (filgras-tim) accelerates granulocyte recovery followingintensive postremission chemotherapy for acutemyeloid leukemia with aziridinyl benzoquinone(AZO) and mitoxantrone: Cancer and LeukemiaGroup B Study 9022. Blood 1997; 89: 780–8.

107. Larson RA, Dodge RK, Linker CA et al, A ran-domized controlled trial of filgrastim duringremission induction and consolidationchemotherapy for adults with acute lym-phoblastic leukemia: CALGB Study 9111. Blood1998; 92: 1556–64.

108. Ottman OG, Hoelzer D, Gracien E et al,Simultaneous administration of granulocytecolony-stimulating factor (filgrastim) andinduction chemotherapy in acute lymphoblasticleukemia. A randomized phase III trial. Blood1995; 86: 444.

109. Gessler K, Koller E, Hubmann E et al,Granulocyte colony-stimulating factor as anadjunct to induction chemotherapy for adultacute lymphoblastic leukemia – a randomizedphase III study. Blood 1997; 90: 590–6.

110. Ohno R, Tomonaga M, Ohshima T et al, A ran-domized controlled study of granulocytecolony stimulating factor after intensive induc-tion and consolidation therapy in patients withacute lymphoblastic leukemia. Japan AdultLeukemia Study Group. Int J Hematol 1993; 58:73–81.

111. Pui CH, Boyett JM, Hughes WT et al, Humangranulocyte colony-stimulating factor afterinduction chemotherapy in children with lym-phoblastic leukemia. N Engl J Med 1997; 336:1781–7.

112. Laver J, Amylon M, Desai S et al, Randomizedtrial of r-metHu granulocyte colony-stimulating

factor in an intensive treatment for T-cellleukemia and advanced-stage lymphoblasticlymphoma of childhood: a Pediatric OncologyGroup pilot study. J Clin Oncol 1998; 16: 522–6.

113. Welte K, Reiter A, Mempel K et al, A random-ized phase III trial of the efficacy of granulocytecolony-stimulating factor in children with high-risk acute lymphoblastic leukemia. Blood 1997;87: 3143.

114. Michel G, Landman-Parker J, Auclerc M et al,Use of recombinant human granulocyte colony-stimulating factor to increase chemotherapydose-intensity; a randomized trial in very high-risk childhood acute lymphoblastic leukemia. JClin Oncol 2000; 18: 1517–24.

115. Khwaja A, Linch DC, Goldstone AH et al,rhG-CSF after bone marrow transplantation formalignant lymphoma: a BNLI double-blindplacebo-controlled trial. Br J Haematol 1992; 82:317.

116. Gorin NC, Coiffier B, Hayat M et al, rhGM-CSFafter high dose chemotherapy and ABMT withunpurged and purged marrow in non-Hodgkin’s lymphoma: a double-blind placebo-controlled trial. Blood 1992; 80: 1149.

117. Gulati SC, Bennett CL, GM-CSF as adjunct ther-apy in relapsed Hodgkin disease. Ann InternMed 1992; 116: 177.

118. Link H, Boogaerts MA, Carella AM et al, A con-trolled trial of rhGM-CSF after TBI, high-dosechemotherapy, and ABMT for ALL or malig-nant lymphoma. Blood 1992; 80: 2188.

119. Nemunaitis J, Rabinowe SN, Singer JW et al,rhGM-CSF after ABMT for lymphoid cancer. NEngl J Med 1991; 324: 1773.

120. Schmitz N, Dreger P, Zander AR et al, Resultsof a randomized controlled multicenter study ofrhG-CSF in patients with Hodgkin’s diseaseand non-Hodgkin’s lymphoma undergoingABMT. Bone Marrow Transplant 1995; 15: 261–6.

121. Stahel RA, Jost LM, Cerny T et al, Randomizedstudy of rhG-CSF after high-dose chemother-apy and ABMT for high-risk lymphoid malig-nancies. J Clin Oncol 1994; 12: 1931.

122. Advani R, Chao NJ, Horning SJ et al, GM-CSFas an adjunct to autologous hemopoietic stemcell transplantation for lymphoma. Ann InternMed 1992; 116: 183.

123. Legros M, Fleury J, Bay JO et al, rhGM-CSF vsplacebo following rh-GM-CSF-mobilized PBSCtransplantation. A phase III double-blind ran-


domized trial. Bone Marrow Transplant 1997; 19:209.

124. Spitzer G, Adkins DR, Spencer V et al,Randomized study of growth factors post PBPCtransplant: neutrophil recovery is improved withmodest clinical benefit. J Clin Oncol 1994; 12: 661.

125. Klumpp TR, Mangan KF, Goldberg SL et al,G-CSF accelerates neutrophil engraftmentfollowing PBSC transplantation: a prospective,randomized trial. J Clin Oncol 1995; 13: 1323.

126. Linch DC, Milliagan DW, Winfield DA et al,G-CSF significantly accelerates neutrophilrecovery after PBSC transplantation in lym-phoma patients and shortens the time in hos-pital: preliminary results of a randomized BNLItrial. Blood 1995; 86: 102a.

127. Bishop MR, Tarantolo SR, Geller RB et al, Arandomized, double-blind trial of filgrastim(granulocyte colony-stimulating factor) versusplacebo following allogeneic blood stem celltransplantation. Blood 2000; 96: 80–5.

128. Lee SM, Radford JA, Dobson L et al,Recombinant human granulocyte colony-stimu-lating factor (filgrastim) following high dosechemotherapy and peripheral blood progenitorcell rescue in high grade non-Hodgkin’s lym-phoma: clinical benefits at no extra cost. Br JCancer 1998; 77: 1294.

129. McQuaker IG, Hunter AE, Pacey S et al, Low-dose filgrastim significantly enhances neu-trophil recovery following autologousperipheral-blood stem-cell transplantation inpatients with lymphoproliferative disorders:evidence for clinical and economic benefit. JClin Oncol 1997; 15: 451.

130. Kawano Y, Takaue Y, Mimaya J et al, Marginalbenefit/disadvantage of granulocyte colony-stimulating factor therapy after autologousblood stem cell transplantation in children:results of a prospective randomized trial. Blood1998; 92: 4040–6.

131. Ojeda E, Garcia-Bustos J, Aguado MJ et al, Arandomized study of filgrastim (G-CSF) afterautologous peripheral blood transplantation.Blood 1998; 92(Suppl 1): 325b.

132. Gisselbrecht C, Prentice HG, Bacigalupo A et al,Placebo-controlled phase III trial of lenograstimin bone marrow transplantation. Lancet 1994;343: 696.

133. DeWitte T, Gratwohl A, Van Der Lely N et al,rhGM-CSF accelerates neutrophil and mono-

cyte recovery after allogeneic T-cell-depletedbone marrow transplantation. Blood 1992; 79:1359.

134. Nemunaitis J, Rosenfeld CS, Ash R et al, PhaseIII randomized, double-blind placebo-con-trolled trial of rhGM-CSF following allogeneicbone marrow transplantation. Bone MarrowTransplant 1995; 15: 949.

135. Powles R, Smith C, Milan S et al, rhGM-CSF inallogeneic bone-marrow transplantation forleukaemia: double-blind, placebo-controlledtrial. Lancet 1990; 336: 1417.

136. Linch DC, Scarffe H, Proctor S et al,Randomized vehicle-controlled dose-findingstudy of glycosylated rhG-CSF after bone mar-row transplantation. Bone Marrow Transplant1993; 11: 307.

137. Cortelazzo S, Viero P, Bellavita P et al, G-CSFfollowing PBPC transplant in non-Hodgkin’slymphoma. J Clin Oncol 1995; 13: 935.

138. Szilvassy SF, Hoffman R, Hematopoieticgrowth factors do not accelerate neutrophilrecovery after transplantation of optimallymobilized peripheral blood stem cells. Biol BloodMarrow Transplant 1996; 2: 2.

139. Brandwein JM, Callum J, Sutcliffe SB et al,Analysis of factors affecting hematopoieticrecovery after ABMT. Bone Marrow Transplant1990; 6: 291.

140. Bensinger WI, Longin K, Appelbaum F et al,Peripheral blood stem cells (PBSC) collectedafter recombinant granulocyte colony-stimulating factor (rhG-CSF): an analysis of fac-tors correlating with the tempo of engraftmentafter transplantation. Br J Haematol 1994; 87:825–31.

141. Bensinger W, Martin P, Clift R et al, A prospec-tive, randomized trial of peripheral blood stemcells (PBSC) or marrow (BM) for patientsundergoing allogeneic transplantation forhematologic malignancies. Blood 1999; 94: 369a(Abst 1637).

142. Stahel RA, Jost LM, Honegger H et al,Randomized trial showing efficacy of filgrastim5 µg/kg/day and 10 µg/kg/day following highdose chemotherapy and autologous bone mar-row transplantation in high risk lymphomas. JClin Oncol 1997; 15: 1730–5.

143. Faucher C, Le Corroller AG, Chabannon C et al,Administration of G-CSF can be delayed aftertransplantation of autologous G-CSF-primed


blood stem cells: a randomized study. BoneMarrow Transplant 1996; 17: 533–6.

144. Bence-Bruckler I, Bredeson C, Atkins H et al, Arandomized trial of granulocyte colony-stimulating factor (Neupogen) starting day 1 vsday 7 post-autologous stem cell transplantation.Bone Marrow Transplant 1998; 22: 965–9.

145. Ciernik IF, Schanz U, Gmur J, Delaying treat-ment with granulocyte colony-stimulating factor after allogeneic bone marrow transplan-tation for hematological malignancies: aprospective randomized trial. Bone MarrowTransplant 1999; 24: 147–51.

146. Hagglud H, Ringden O, Oman S et al, Aprospective randomized trial of filgrastim (r-metHuG-CSF) given at different times afterunrelated bone marrow transplantation. BoneMarrow Transplant 1999; 24: 831–6.

147. Torres Gomez A, Jimenez MA, Alvarez MA etal, Optimal timing of granulocyte colony-stimulating factor (G-CSF) administration afterbone marrow transplantation. A prospectiverandomized study. Ann Hematol 1995; 71: 65–70.

148. Sobrevilla-Calvo P, Cortes P, Solano P et al,Starting G-CSF on day �7 or on day 0 is equallyeffective in accelerating neutrophil recoveryafter autologous peripheral blood stem celltransplantation. J Clin Oncol 1996; 15: 272 (Abst724).

149. Vey N, Molnar S, Faucher C et al, Delayedadministration of granulocyte colony-stimulating factor after autologous bone mar-row transplantation: effect of granulocyterecovery. Bone Marrow Transplant 1994; 14: 779.

150. Khwaja A, Mills W, Leveridge K et al, Efficacyof a delayed granulocyte colony-stimulatingfactor after autologous bone marrow transplan-tation. Bone Marrow Transplant 1993; 11: 479–82.

151. Viret F, Molina L, Plantaz D, Impact of delayedstart (day �5) G-CSF after allogeneic bone mar-row transplantation: a pilot study. Blood 1994;84(Suppl 1): 63 (Abst 369).

152. Masaoka T, Takaku F, Kato S et al, rhG-CSF inallogeneic bone marrow transplantation. ExpHematol 1989; 17: 1047.

153. NEUPOGEN (filgrastim) prescribing informa-tion. Package Circular (P40047H 50M Rev 4-98).

154. LEUKINE (sargramostim) Package Circular(Rev 0230-01).

155. Ikeda K, Tasaka T, Sasaki K et al, Low-dosecontinuous subcutaneous infusion of gran-

ulocyte colony-stimulating factor forchemotherapy-induced neutropenia in acutemyelogenous leukemia and its pharmacokinet-ics. Leukemia 1994; 8: 1838–41.

156. Toner GC, Shapiro JD, Laidlaw CR et al, Low-dose versus standard-dose lenograstim prophy-laxis after chemotherapy: a randomized,crossover comparison. J Clin Oncol 1998; 16:3874–9.

157. Kubota M, Akiyama Y, Mikawa H et al,Comparative effect of 100 versus250 µg/m2/day of G-CSF in pediatric patientswith neutropenia induced by chemotherapy.Pediatr Hematol Oncol 1995; 12: 393–7.

158. Kessinger A, Bishop MR, Anderson JR et al,Comparison of subcutaneous and intravenousadministration of recombinant human granulo-cyte–macrophage colony-stimulating factor forperipheral blood stem cell mobilization. JHematother 1995; 4: 81–4.

159. Sturgill MG, Huhn RD, Drachtman RA et al,Pharmacokinetics of intravenous recombinanthuman colony-stimulating factor (rhG-CSF) inchildren receiving myelosuppressive cancerchemotherapy: clearance increases in relation toabsolute neutrophil count. Am J Hematol 1997;54: 124–30.

160. Sugiura M, Yamamoto K, Sawada Y, Iga T,Pharmacokinetic/pharmacoeconomic analysisof neutrophil proliferation induced by recombi-nant granulocyte colony-stimulating factor(rhG-CSF): comparison between intravenousand subcutaneous administration. Biol PharmBull 1997; 20: 684–9.

161. Broxmeyer HE, Benninger L, Patel S et al,Kinetic response of human marrow progenitorcells to in vivo treatment of patients with gran-ulocyte colony-stimulating factor is differentfrom the response with granulocyte–macrophage colony-stimulating factor. ExpHematol 1994; 22: 100–2.

162. Bennett C, Stinson T, Bhoopalam N et al, Adouble-blind, randomized trial of toxicity,resource use and costs for filgrastim and sar-gramostim. Proc Am Soc Clin Oncol 2000; 19:436a (Abst 1712).

163. Beveridge R, Miller J, Kales A et al,Randomized trial comparing the tolerability ofsargramostim (yeast-derived RhuGM-CSF) andfilgrastim (bacteria-derived RhuG-CSF) incancer patients receiving myelosuppressive


chemotherapy. Support Care Cancer 1995; 5:289–98.

164. Strauss RG, Therapeutic granulocyte transfu-sions in 1993. Blood 1993; 81: 1675–8.

165. Strauss RG, Rebirth of granulocyte transfusions:Should it involve pediatric oncology and trans-plant patients? J Pediatr Hematol Oncol 1999; 21:475–8.

166. Ditcher JP, The potential benefit of granulocytetransfusion therapy. Cancer Invest 1989; 7:457–62.

167. Klein HG, Strauss RG, Schiffer CA, Granulocytetransfusion therapy. Semin Hematol 1996; 33:359–68.

168. Dale DC, Liles WC, Return of granulocyte infu-sions. Curr Opin Pediatr 2000; 12: 18–22.

169. Casper C, Seger R, Burger J, Gmur J, Effectivestimulation of donors for granulocyte transfu-sions with recombinant methionyl granulocytecolony-stimulating factor. Blood 1993; 81:2866–71.

170. Bensinger WI, Price TH, Dale DC et al, Theeffects of daily recombinant human granulocytecolony-stimulating factor administration onnormal granulocyte donors undergoing leuka-pheresis. Blood 1993; 81: 1883.

171. Jendiroba DB, Lichtinger B, Freireich EJ, Theuse of hematopoietic growth factors for recruit-ment of leukocytes after transfusions. In:Clinical Applications of Cytokines and GrowthFactors (Wingard JR, Demetri GD, eds).Dordrecht: Kluwer, 1999.

172. Adkins DR, Goodnough LT, Shenoy S et al,Effect of leukocyte compatibility on neutrophilincrement after transfusion of granulocytecolony-stimulating factor-mobilized prophylac-tic granulocyte transfusions and on clinical out-comes after stem cell transplantation. Blood2000; 95: 3605–12.

173. Jendiroba D, Lichtiger B, Anaissie E et al,Evaluation and comparison of three mobiliza-tion methods for the collection of granulocytes.Transfusion 1998; 38: 722–8.

174. Price TH, Bowden RA, Boeckh M et al, Phase I/IItrial of neutrophil transfusions from donors stim-ulated with G-CSF and dexamethasone for treat-ment of patients with infections in hematopoieticstem cell transplantation. Blood 2000; 95: 3302–9.

175. Dale DC, Liles WC, Llewellyn A et al,Neutrophil transfusions: kinetics and functionof neutrophils mobilized with granulocyte

colony-stimulating factor (G-CSF) and dexam-ethasone. Transfusion 1998; 38: 713–21.

176. Djerassi I, Filtration leukapheresis for the sepa-ration and concentration of transfusableamounts of normal human granulocytes. J Med(Basel) 1970; 1: 358–64.

177. Ts’ao C, Ruder EA, Ultrastructural damage ofleukocytes procured by Leukopak: vulnerabil-ity of leukocytes to mechanical injury.Transfusion 1976; 16: 336–44.

178. Schiffer CA, Buchholz DH, Aisner J et al,Clinical experience with transfusion of granulo-cytes obtained by continuous flow filtrationleukapheresis. Am J Med 1975; 58: 373–81.

179. Hammerschmidt DE, Carddock PR, McCul-lough F et al, Complement activation and pul-monary leukostasis during nylon fiber filtrationleukapheresis. Blood 1978; 51: 721–30.

180. Loftus TJ, White RF, Huestis DW,Leukapheresis: increasing the granulocyte yieldwith the Fenwal CS-3000. J Clin Apheresis 1983;1: 109–14.

181. Adkins D, Goodgold H, Hendershott L et al,Indium-labeled white blood cells apheresedfrom donors receiving G-CSF localize to sites ofinflammation when infused into allogeneicbone marrow transplant recipients. BoneMarrow Transplant 1997; 19: 809–14.

182. Freii E III, Levin RH, Bodey GP et al, The natureand control of infections in patients with acuteleukemia. Cancer Res 1965; 25: 1511–15.

183. Boggs DR, Transfusion of neutrophils as pre-vention or treatment of infection in patientswith neutropenia. N Engl J Med 1974; 290: 1055.

184. Colotta F, Re F, Polentarutti N et al, Modulationof granulocyte survival and programmed celldeath by cytokines and bacterial products. Blood1992; 80: 2012–20.

185. Wright DG, Robichaud KJ, Pizzo PA et al,Lethal pulmonary reactions associated with thecombined use of amphotericin B and leukocytetransfusions. N Engl J Med 1981; 304: 1185–9.

186. Schroeder ML, Louie TJ, Pulmonary complica-tions in patients receiving granulocyte transfu-sions and amphotericin B. Can Med Assoc J 1984;130: 593–7.

187. Dana BW, Durie DG, White RF et al,Concomitant administration of granulocytetransfusions and amphotericin B in neutropenicpatients: absence of significant pulmonary toxi-city. Blood 1981; 57: 90–4.


188. Dutcher JP, Kendall J, Norris D et al,Granulocyte transfusion therapy and ampho-tericin B: adverse reactions? Am J Hematol 1989;31: 102–8.

189. Schiffer CA, Aisner J, Daly PA et al,Alloimmunization following prophylactic gran-ulocyte transfusion. Blood 1979; 54: 766–74.

190. Bux J, Becker F, Seeger W et al, Transfusion-related acute lung injury due to HLA-A2-spe-cific antibodies in recipient and NB1-specificantibodies in donor blood. Br J Haematol 1996;93: 707.

191. Adkins D, Spitzer G, Johnston M et al,Transfusions of granulocyte-colony-stimulatingfactor-mobilized granulocyte components toallogeneic transplant recipients: analysis ofkinetics and factors determining posttransfu-sion neutrophil and platelet counts. Transfusion1997; 37: 737–48.

192. Stroncek DF, Leonard K, Eiber G et al,Alloimmunization after granulocyte transfu-sions. Transfusion 1996; 36: 1009–15.

193. Stroncek DF, Neutrophil antibodies. Curr OpinHematol 1997; 4: 455–8.

194. Higby DJ, Yates JW, Henderson ES, Holland JF,Filtration leukapheresis for granulocyte transfu-sion therapy. N Engl J Med 1975; 292: 761–6.

195. Vogler WR, Winton EF, A controlled study ofthe efficacy of granulocyte transfusions inpatients with neutropenia. Am J Med 1977; 63:548–55.

196. Herzig RH, Herzig GP, Graw RG Jr et al,Successful granulocyte transfusion therapy forGram-negative septicemia. N Engl J Med 1977;296: 701–5.

197. Alavi JB, Root RK, Djerassi I et al, A random-ized clinical trial of granulocyte transfusions forinfection in acute leukemia. N Engl J Med 1977;296: 706–11.

198. Winston DJ, Ho WB, Gale RP, Therapeuticgranulocyte transfusions for documented infec-tions. A controlled trial in ninety-five infectiousgranulocytopenic episodes. Ann Intern Med1982; 97: 509–15.

199. Graw RG Jr, Herzig G, Perry S, Henderson ES,Normal granulocyte transfusion therapy. NEngl J Med 1972; 287: 367–76.

200. Fortuny IE, Bloomfield CD, Hadlock DC et al,Granulocyte transfusions: a controlled study inpatients with acute nonlymphocytic leukemia.Transfusion 1986; 15: 548–58.

201. Dignani M, Anaissie E, Hester J et al, Treatmentof neutropenia-related fungal infections withgranulocyte colony-stimulating factor-elicitedwhite blood cell transfusions: a pilot study.Leukemia 1997; 11: 1621–30.

202. Di Mario A, Sica S, Salutari P et al, Granulocytecolony-stimulating factor-primed leukocytetransfusions in Candida tropicalis fungemia inneutropenic patients. Haematologica 1997; 82:362–3.

203. Catalano L, Fontana R, Scarpato N et al,Combined treatment with amphotericin B andgranulocyte transfusion from G-CSF-stimulateddonors in an aplastic patient with invasiveaspergillosis undergoing bone marrow trans-plantation. Haematologica 1997; 82: 71–2.

204. Ozsahin H, von Planta M, Muller I et al,Successful treatment of invasive aspergillosis inchronic granulomatous disease by bone mar-row transplantation, granulocyte colony-stimulating factor-mobilized granulocytes, andliposomal amphotericin B. Blood 1998; 15:2719–24.

205. Beyer J, Schwella N, Aingsem J et al,Hematopoietic rescue after high-dosechemotherapy using autologous peripheral-blood progenitor cells or bone marrow: a ran-domized comparison. J Clin Oncol 1995; 13:1328–35.

206. Schmitz N, Linch DC, Dreger P et al,Randomized trial of filgrastim-mobilizedperipheral blood progenitor cell transplantationversus autologous bone marrow transplanta-tion in lymphoma patients. Lancet 1996; 347:353–7.

207. Nademanee A, Sniecinski I, Schmidt GM et al,High-dose chemotherapy followed by autolo-gous peripheral-blood stem-cell transplantationfor patients with Hodgkin’s disease and non-Hodgkin’s lymphoma using unprimed andgranulocyte colony-stimulating factor mobi-lized peripheral blood stem cells. J Clin Oncol1994; 12: 2176–86.

208. Korbling M, Przepiorka D, Huh YO et al,Allogeneic blood stem cell transplantation forrefractory leukemia and lymphoma: potentialadvantage of blood over marrow allografts.Blood 1995; 85: 1659–65.

209. Dreger P, Haferlach T, Eckstein V et al,G-CSF-mobilized peripheral blood progenitorcells for allogeneic transplantation: safety,


kinetics of mobilization, and composition of thegraft. Br J Haematol 1994; 87: 609–13.

210. Schmitz N, Dreger P, Suttorp M et al, Primarytransplantation of allogeneic peripheral bloodprogenitor cells mobilized by filgrastim (granu-locyte colony-stimulating factor). Blood 1995; 85:1666–72.

211. Bensinger WI, Weaver CH, Appelbaum FR etal, Transplantation of allogeneic peripheralblood stem cells mobilized by recombinanthuman granulocyte colony-stimulating factor.Blood 1995; 85: 1655–8.

212. Snowden JA, Biggs JC, Milliken ST et al, A ran-domized, blinded, placebo-controlled, doseescalation study of the tolerability and efficacyof filgrastim for hematopoietic stem cell mobi-lization in patients with severe rheumatoidarthritis. Bone Marrow Transplant 1998; 22:1035–41.

213. Ho AD, Young D, Maruyama M et al,Pluripotent and lineage-committed CD34� sub-sets in leukapheresis products mobilized byG-CSF, GM-CSF vs. a combination of both. ExpHematol 1996; 24: 1460–8.

214. Meisenberg B, Brehm T, Schmeckel A et al, Acombination of low-dose cyclophosphamideand colony-stimulating factors is more cost-effective than granulocyte-colony-factors alonein immobilizing peripheral blood stem and pro-genitor cells. Transfusion 1998; 38: 209–15.

215. Cesana C, Carlo-Stella C, Regazzi E et al,CD34� cells mobilized by cyclophosphamideand granulocyte colony stimulating factor(G-CSF) are functionally different from CD34�cells mobilized by G-CSF. Bone MarrowTransplant 1998; 21: 561–8.

216. Krishnan A, Bhatia S, Slovak ML et al,Predictors of therapy-related leukemia andmyelodysplasia following autologous trans-plantation for lymphoma: an assessment of riskfactors. Blood 2000; 95: 1588–93.

217. Weisdorf D, Miller J, Verfaillie C et al,Cytokine-primed bone marrow stem cells vs.peripheral blood stem cells for autologoustransplantation: a randomized comparison ofGM-CSF vs G-CSF. Biol Blood Marrow Transplant1997; 3: 217–23.


APPENDIX: SUMMARY OF GUIDELINES UPDATES*

1. Guidelines for primary prophylactic MGF administration

1996 Recommendation: Primary administration of MGFs was shown to reduce the incidence offebrile neutropenia (FN) by approximately 50% in the three major ran-domized trials in adults in which the incidence of FN was greater than40% in the control group. The value of primary MGF administration hasnot been clearly established in less myelosuppressive regimens, and thecost–benefit of primary versus secondary administration for the majorityof initial chemotherapy regimens is unproven. It is recommended thatprimary administration of MGFs be reserved for patients expected toexperience levels of FN that are at least comparable to or greater thanthose seen in control patients in these randomized trials, i.e. an expectedincidence of 40% or more. Thus, for previously untreated patients receiv-ing most chemotherapy regimens, primary administration of MGFs can-not be recommended.

2000 Recommendation: No change.

Special circumstances

1996 Recommendation: Clinicians may occasionally be faced with patients who might benefitfrom relatively non-myelosuppressive chemotherapy but who havepotential risk factors for FN or infection because of bone marrow com-promise or comorbidity. It is possible that primary MGF administrationmay be exceptionally warranted in patients at higher risk forchemotherapy-induced infectious complications, even though the datasupporting such use are not conclusive. Such risk factors might includethe following: pre-existing neutropenia due to disease, extensive priorchemotherapy, or previous irradiation to the pelvis or other areas con-taining large amounts of bone marrow; a history of recurrent FN whilereceiving earlier chemotherapy of similar or lesser dose intensity; or con-ditions potentially enhancing the risk of serious infection, e.g. poor per-formance status and more advanced cancer, decreased immunefunction, open wounds, or already-active tissue infections. This is notmeant to be an all-inclusive list; it is anticipated that, depending on theunique features of the clinical situation, there will be instances when theadministration of an MGF will be appropriate outside of uses recom-mended in other guidelines.



* References are cited as superior numbers, and are given in the reference list to the main text of this chapter.

2. Guidelines for secondary prophylactic MGF administration

1996 Recommendation: There is evidence that MGF administration can decrease the probabilityof FN in subsequent cycles of chemotherapy after a documented occur-rence in an earlier cycle. Even if FN has not occurred, the use of MGFsmay be considered if prolonged neutropenia is causing excessive dosereduction or a delay in chemotherapy. However, in the absence of clini-cal data supporting maintenance of chemotherapy dose intensity, physi-cians should consider chemotherapy dose reduction as an alternative tothe use of MGFs.

2000 Recommendation: In the setting of many tumors, exclusive of curable tumors (e.g. germ celltumors), dose reduction after an episode of severe neutropenia shouldbe considered as a primary therapeutic option. No published regimenshave demonstrated disease-free or overall survival benefits when thedose of chemotherapy was maintained and secondary prophylaxis wasinstituted. In the absence of clinical data or other compelling reasons tomaintain chemotherapy dose intensity, physicians should considerchemotherapy dose reduction after neutropenic fever or severe or pro-longed neutropenia after the previous cycle of treatment.

3. Guidelines for MGF therapy

A. Afebrile patients

1996 Recommendation: Data are inadequate with regard to whether patients with neutropeniabut no fever will benefit clinically from the initiation of an MGF at thetime neutropenia is diagnosed; intervention with an MGF in afebrileneutropenic patients is not recommended.

2000 Recommendation: Current evidence supports the recommendation that MGFs should notbe routinely used for patients with neutropenia who are afebrile. Thestrength of this recommendation has increased with the trial reported in1997.85

B. Febrile patients

1996 Recommendation: For the majority of patients with FN, the available data do not clearlysupport the routine initiation of MGFs as adjuncts to antibiotic therapy.However, certain FN patients may have prognostic factors that are pre-dictive of clinical deterioration, such as pneumonia, hypotension, multi-organ dysfunction (sepsis syndrome), or fungal infection. The use ofMGFs together with antibiotics may be reasonable in such high-riskpatients, even though the benefits of administration under these circum-stances have not been definitively proven.

2000 Recommendation: The collective results of the eight trials73–79,81 provide strong and consis-tent support for the recommendation that MGFs should not be routinely


used as adjunct therapy for the treatment of uncomplicated fever andneutropenia. Uncomplicated fever and neutropenia are defined as fol-lows: fever of 10 days or less in duration; no evidence of pneumonia, cel-lulitis, abscess, sinusitis, hypotension, multiorgan dysfunction, orinvasive fungal infection; and no uncontrolled malignancies. The eighttrials have consistently shown a decrease in the duration of neutropeniaof less than 500/µl, but clinical benefit has not consistently accompaniedthe decreased duration of neutropenia.

Certain patients with fever and neutropenia are at higher risk forinfection-associated complications and have prognostic factors that arepredictive of poor clinical outcome. The use of an MGF for such high-risk patients may be considered, but the benefits of an MGF in these cir-cumstances have not been proven. These factors include profound(absolute neutrophil count (ANC) < 100/µl) neutropenia, uncontrolledprimary disease, pneumonia, hypotension, multiorgan dysfunction (sep-sis syndrome), and invasive fungal infection. Age greater than 65 yearsand post-treatment lymphopenia may also be high-risk factors, but havenot been consistently confirmed by multicenter trials.

4. Guidelines for use of MGFs to increase chemotherapy dose intensity

1996 Recommendation: Outside of clinical research trials, there is little justification for the use ofMGFs to increase chemotherapy dose intensity. In settings in which clin-ical research demonstrates that dose-intensive therapy not requiringprogenitor cell support produces improvement in disease control, MGFsshould be used when these therapies are expected to produce significantrates of FN (e.g. in 40% or more of patients).

2000 Recommendation: In the absence of more trials demonstrating a favorable effect on overallsurvival, disease-free survival, quality of life, or toxicity, there is no justi-fication for the use of MGFs to increase chemotherapy dose intensity orschedule or both outside of a clinical trial. This application of MGF useremains the domain of appropriately designed clinical investigation.

5. Guidelines for use of MGFs as adjuncts to progenitor cell transplantation

1996 Recommendation: MGFs can successfully shorten the period of neutropenia and reduceinfectious complications in patients undergoing high-dose cytotoxictherapy with autologous bone marrow transplantation (BMT). MGFs areeffective in mobilizing autologous peripheral blood progenitor cells(PBPC) for transplantation, and autologous PBPC transplantation(PBPCT) has been shown to lead to earlier hematopoietic recovery thanautologous BMT.205,206 Trials have demonstrated the value of MGFadministration after high-dose chemotherapy and PBPCT.120,125,207

Available data suggest clinical benefits after allogeneic BMT, and routineprimary MGF administration in this setting seems warranted.134 MGFscan also be used to mobilize donor PBPC for allogeneic transplanta-tion.208–211 There may also be a role for MGFs in assisting in the recovery


of patients who experience delayed or inadequate neutrophil engraft-ment after PBPCT. MGFs can be routinely recommended as adjuncts toallogeneic and autologous PBPCT, both for mobilization of PBPC and as a means to speed hematopoietic reconstitution after BMT orPBPCT. Administration of a MGF in cases of engraftment failure is war-ranted.

2000 Recommendation: MGFs are recommended to help mobilize PBPC and after PBPC infu-sion. Mobilized PBPC have largely replaced bone-marrow-derived cellsfor use in autologous transplantation. Side-effects associated with mobi-lization and subsequent apheresis are usually limited, and includeconstitutional symptoms and a decrease in platelets and otherhematopoietic elements, especially after mobilization with combinationsof chemotherapeutic agents and an MGF. The optimal dose of MGFs andchemotherapaeutic agents is the subject of ongoing investigations, but ahigher (10 µg/kg/day) dose of G-CSF in the setting of mobilization mayyield a greater content of CD34� progenitor cells in the PBPC product, asdocumented in patients with hematologic malignancies and in patientswith rheumatoid arthritis.207,212 Although the optimal method of mobi-lization needs further investigation, especially in heavily pretreatedpatients, administration of G-CSF, either alone or in combination withGM-CSF, or after the use of chemotherapeutic agents, generates PBPC,leading to rapid hematopoietic recovery, shorter hospitalization, andpossibly reduced costs.206,213–215 Further investigations are necessary toassess the potential risks, especially that of secondary hematologicmalignancies associated with the use of combining chemotherapeuticagents and MGFs.216 The role of MGF-mobilized donor bone marrow inthe autologous transplant setting is also under assessment.217

6. Guidelines for use of MGFs in patients with acute leukemia and myelodysplastic syndromes

A. Acute myeloid leukemia (AML)

1996 Recommendation: Primary administration of an MGF can be used after completion ofinduction chemotherapy in patients 55 years of age or older. Althoughthere are fewer data, it is likely that the results showing shortening ofthe duration of neutropenia may apply to younger patients as well.MGFs given before and/or concurrently with chemotherapy for primingeffects still cannot be recommended outside of a clinical trial.

2000 Recommendation: MGF use can be considered in this setting if benefits in terms of possibleshortening of hospitalization outweigh the costs of MGF use. Severalstudies have shown that MGF administration can produce modestdecreases in the duration of neutropenia when begun shortly after com-pletion of the initial days of chemotherapy of the initial or repeat induc-tion. Beneficial effects on endpoints such as duration of hospitalizationand incidence of severe infections have been variable and modest,


although patients 55 years of age or older are most likely to benefit fromMGF use. No study has yet demonstrated significant improvement incomplete response rates or long-term outcome. Thus, while there seemsto be minimal risk associated with the use of MGFs in this situation, thechoice of whether or not to use the MGF is likely to be determined bycost considerations. In a nutshell, the cost of the cytokine must be bal-anced against any possible shortening of hospitalization associated withthe slightly more rapid marrow recovery, as, for example, in patients 55years of age or older. It is not known from the published data whetherthe MGFs significantly accelerate recovery to an ANC of 100–200/µl. Inmost patients, regenerating counts of this level are sufficient to protectagainst infection so as to permit safe discharge of patients from hospital.

There is no evidence that MGFs given either before or concurrently withchemotherapy for priming effects are of benefit, and their use in this fash-ion cannot be recommended outside the setting of a clinical trial.

There seems to be more profound shortening of the duration of neutrope-nia after consolidation chemotherapy for patients with AML in remission.Although the randomized studies did not address this issue, it is likely thatthis will be associated with decreased rates of hospitalization and possiblyshorter durations of hospitalization in such patients. No benefit has beendemonstrated in terms of prolongation of complete response duration oroverall survival; however, the available evidence indicate that MGFs can berecommended after the completion of consolidation chemotherapy.

B. Myelodysplastic syndromes (MDS)

1996 Recommendation: MGFs can increase the ANC in neutropenic patients with MDS. Data sup-porting the routine, long-term, continuous use of MGFs in these patientsare lacking. Intermittent administration of MGFs may be considered in asubset of patients with severe neutropenia and recurrent infection.


C. Acute lymphoblastic leukemia (ALL)*

2000 Recommendation: The data are sufficient to recommend G-CSF administration begun aftercompletion of the first few days of chemotherapy of the initial inductionor first post-remission course, thus shortening the duration of neutrope-nia of less than 1000/µl by approximately 1 week. Effects on the inci-dence and duration of hospitalization and the acquisition of seriousinfections are less consistent. Although there was a trend for improvedcomplete response rates in one large study,107 particularly in olderadults, there was no prolongation of disease-free or overall survival inany of the trials. G-CSF can be given together with continued corticos-teroid/antimetabolite therapy, which is a feature of many ALL regi-


* This topic is new to the guidelines in 2000.

mens, without evidence that such concurrent therapy prolongs themyelosuppressive effects of the chemotherapy. As in AML, it is notknown from the published data whether the MGFs significantly acceler-ate ANC recovery to 100–200/µl. In most patients, regenerating countsof this level are sufficient to protect against infection so as to permit safedischarge of patients from hospital. The use of G-CSF for children withALL was associated with small benefits in days of antibiotic use or in-hospital days, although a small amount of additional costs was incurred,after the costs of the MGFs were taken into consideration. Cost estimatesof MGFs for adults with ALL have not been reported.

D. Leukemia in relapse*

2000 Recommendation: The available data are not sufficient to recommend either for or against theuse of MGFs in patients with refractory or relapsed ALL. Few controlledstudies have evaluated MGFs in patients with relapsed or refractory acuteleukemia. The available data suggest a shortening of the duration of neu-tropenia, but are inadequate to allow comment on any effects on infectiouscomplications and, in particular, on whether there may be an adverse effecton response rates in some patients with myeloid malignancies because of astimulatory effect on leukemia growth in a situation in which there is lessof a guarantee that chemotherapy will produce sufficient cytoreduction.Therefore, there is no evidence that MGFs are of important benefit inpatients with refractory or relapsed myeloid leukemia, and they should beused judiciously or not at all in such patients.

7. Guidelines for use of MGFs in patients receiving concurrent chemotherapy and irradiation

1996 Recommendation: MGFs should be avoided in patients receiving concomitant chemother-apy and radiation therapy.

2000 Recommendation: MGFs should be avoided in patients receiving concomitant chemother-apy and radiation therapy, particularly involving the mediastinum. Inthe absence of chemotherapy, in patients receiving radiation therapyinvolving large fields, therapeutic use of MGFs may be considered ifprolonged delays secondary to neutropenia are expected.

8. Guidelines for use of MGFs in the pediatric population

1996 Recommendation: In the absence of conclusive pediatric data, the guidelines recommendedfor adults are generally applicable to the pediatric age group. However,optimal MGF doses have yet to be determined. Further clinical researchinto the use of these factors in support of chemotherapy and PBPCT inthe pediatric age group should be given high priority.



* This topic is new to the guidelines in 2000.

9. Guidelines for MGF dosing and route of administration

1996 Recommendation: In adults, the recommended MGF doses are 5 µg/kg/day for G-CSF (fil-grastim) and 250 µg/m2/day for GM-CSF (sargramostim). These agentscan be administered subcutaneously or intravenously as clinically indi-cated. MGF dose escalation is not advised. The available data suggestthat rounding the dose to the nearest vial size may enhance patient con-venience and reduce costs without clinical detriment.

2000 Recommendation: In adults, the recommended MGF doses are 5 µg/kg/day for G-CSF (fil-grastim) and 250 µg/m2/day for GM-CSF (sargramostim) for all clinicalsettings other than PBPC mobilization. In the setting of PBPC mobiliza-tion, if G-CSF is used, a dose of 10 µg/kg/day seems preferable. Outsideof this indication, MGF dose escalation is not advised. Rounding thedose to the nearest vial size is an appropriate strategy to maximize costbenefit. The preferred route of MGF administration is subcutaneous.

10. Guidelines for initiation and duration of MGF administration

1996 Recommendation: Existing clinical data suggest that starting G-CSF or GM-CSF between 24and 72 hours subsequent to chemotherapy may provide optimal neu-trophil recovery. Continuing the MGF until the occurrence of an ANC of10 000/µl after the neutrophil nadir, as specified in the G-CSF packageinsert, is known to be safe and effective. However, a shorter duration ofadministration that is sufficient to achieve clinically adequate neutrophilrecovery is a reasonable alternative, considering issues of patient conve-nience and cost.

2000 Recommendation: The optimal timing and duration of MGF administration are still underinvestigation. Starting MGFs up to 5 days after PBPC reinfusion is rea-sonable based on available clinical data.

11. Special commentary on comparative clinical activity of G-CSF and GM-CSF

1996 Recommendation: Guidelines about equivalency of the available recombinant preparationsof G-CSF and GM-CSF cannot be proposed, because there have been nolarge-scale, prospective, comparative trials evaluating relative MGF effi-cacy. The strength of evidence to support the use of G-CSF or GM-CSFvaries based on the specific indication for MGF administration, e.g. sup-port after BMT or use with non-transplantation chemotherapy regimens.The panel strongly encourages additional clinical investigation that willguide clinical application of these biologically distinct molecules byaddressing issues of comparative clinical activity, toxicity, andcost–effectiveness.



14Clinical practical guidelines in patients withfever and neutropeniaNatasha Leighl*, Ronald Feld

INTRODUCTION

Fever in the neutropenic cancer patient isamong the most serious complications relatedto chemotherapy, and is also among the mostcommon. Management of this complication canvary widely, relating to differing geographicpatterns of commonly infecting organisms andof antimicrobial resistance, as well as issues oftreatment availability and cost containment. Tohelp promote evidence-based quality improve-ments in the care of cancer patients with neu-tropenic fever, a number of societies havedeveloped clinical practice guidelines.

Practice guidelines are statements that aresystematically developed to assist practitionerand patient decisions about appropriate health-care for specific circumstances.1 Many factorsinfluence the use of guidelines, including easeof accessibility, endorsement by a recognizedand respected body, perceived quality of evid-ence contained within the guideline, and themethod of synthesis into recommendations.Also, perceptions of relevance of the guidelinesto local practice, and their usefulness and

* NL was supported in part by the Department of MedicalOncology, Princess Margaret Hospital, Toronto through theAventis Fellowship Award in Lung Cancer.

applicability, as well as practitioner attitudestowards guideline use in general and themethod of guidelines promotion (e.g. peerreview publication and criteria for reimburse-ment), all factor into whether guidelines areadopted in clinical practice.

Similar to critical appraisal of originalresearch, one should not adopt a practice guide-line at face value, but rather appraise themethodology of guidelines, their development,and their recommendations before incorporat-ing them into day-to-day clinical practice. Inaddition to a review of content, the methodol-ogy of guidelines for therapy of febrile neu-tropenic patients will be briefly reviewed.Guides for the clinician include a criticalappraisal instrument for clinical guidelinespublished by Cluzeau et al at St George’sHospital Medical School, London, UK (http://www.sghms.ac.uk/phs/hceu/form.htm),2 andthe popular ‘Users’ guide to the medical liter-ature’ series.3,4

ANTIMICROBIAL THERAPY IN NEUTROPENICPATIENTS WITH FEVER

Using an English-literature search throughMedline (up to June 2000), and the Agency forHealthcare and Policy Research Guideline

Clearinghouse,5 several guidelines for empirictherapy in febrile neutropenia have been identi-fied. These include the guidelines commis-sioned by the Infectious Diseases Society ofAmerica, initially developed in 1990 and subse-quently revised in 1997.6,7 The NationalComprehensive Cancer Network,8 the Fed-eration of French Cancer Centers,9 and theJapanese Infectious Disease Society10 have alsopublished guidelines in this area. The ItalianAssociation for Pediatric Hematology andOncology recently published clinical guidelinesfor empiric antimicrobial therapy of febrile neu-tropenia in the pediatric population.11 We havebeen unable to identify published evaluationsof these guidelines, but anticipate that thesemay be available in the future.

Guidelines from the Infectious DiseasesSociety of America

The Infectious Diseases Society of America(IDSA) sponsored the development of theseguidelines, involving infectious disease andoncology experts in both the adult and pediatricpopulations, adequately representing the keydisciplines in the area, and including one of theauthors of this chapter (RF). The initial guide-lines were published in 1990,6 with a revision in1997 to address several new issues that hademerged in the management of febrile neu-tropenia, namely the emergence of antibiotic-resistant bacteria, the use of colony-stimulatingfactors, and the issue of cost containment.7

While the methodology for identifying studieswas not explicit, the sources used as well as theindividual studies are clearly listed. In the earl-ier publication, the rating of strength of the rec-ommendation from the guidelines committeewas intertwined with the strength of the sup-porting evidence, making the rating systemoccasionally confusing. In the revised docu-ment, these are clearly separated, with both arating system for the strength of the recommen-dation made and a rating of the quality of evid-ence for that recommendation. Formulation of

the 1997 recommendations was based upon thelevel of evidence presented, and where ade-quate data were lacking, consensus of expertopinion was offered. The methodology for reso-lution of disagreements among the experts wasnot reviewed in the document. The guidelineswere externally reviewed through the peerreview process as well as by the PracticeGuidelines Committee, and were further sub-ject to approval by the IDSA Council.

In the 1997 IDSA guidelines, a consensus def-inition of febrile neutropenia is offered, namelya temperature of 38.3°C, or 38.0°C for an hour’sduration, in combination with an absolute neu-trophil count less than 500/µl, or less than1000/µl with an anticipated drop to below500/µl. The consensus for standard evaluationin neutropenic patients was described, withsome discussion of controversy surroundingsurveillance cultures for colonization, and therole of routine central catheter versus periph-eral blood cultures. Prompt initial therapy wasrecommended in all patients, with threeempiric options (see Figure 14.1). One approachincludes combination therapy with vancomycinand ceftazidime, particularly where empiricvancomycin is felt to be clinically appropriate.On the whole, less empiric use of vancomycin isrecommended in the revised guidelines com-pared with the previous ones, in an effort tohelp reduce the frequency of development ofvancomycin-resistant organisms. Indicationsfor empiric vancomycin use include obviouscatheter-related infections, in the setting ofintensive chemotherapy with mucosal damage,quinolone prophylaxis, colonization with �-lactam-resistant Streptococcus pneumoniae orwith methicillin-resistant Staphylococcus aureus,Gram-positive bacteria identified in a blood cul-ture prior to final identification and susceptibil-ity testing, and evidence of hemodynamicinstability. The other two options for empirictherapy include monotherapy (with cef-tazidime, imipenem, meropenem, or cefepime)and duotherapy (with an aminoglycoside andan anti-pseudomonal �-lactam). Duotherapywith two �-lactam agents is discussed but not


clearly endorsed – a shift from the 1990 recom-mendations, where it was presented as a clearalternative in initial empiric therapy.

In cases where there is patient defervescencewithin three days of treatment, it is recom-mended that clinically stable patients be con-sidered for a change to oral antibiotics, whilehigher-risk patients continue on initial therapy,or that treatment be tailored to microbiologicevidence of infection where appropriate. Inthose patients with ongoing fever, reevaluationis recommended on the fourth or fifth day oftherapy. If patients are clinically unchanged,options include continuing the same manage-

ment, or discontinuing vancomycin therapy. Ifa patient has deteriorated, a change in antibi-otics is warranted, incorporating coverage thatwas not started empirically. This includesamphotericin B institution if patients remainfebrile on the fifth to seventh day of therapy,with or without a change in the other antibi-otics administered (see Figure 14.2).

With respect to the duration of therapy,patients who are afebrile by the third day oftreatment, with granulocyte recovery, can beconsidered for a seven-day course of therapy.Additional recommendations for treatmentduration are outlined in Figure 14.3, and

CLINICAL PRACTICE GUIDELINES 291

Fever (38.3°C) and neutropenia (�500/mm3)

Evaluation

Is vancomycin needed?

Indications

Vancomycin�

ceftazidime

Monotherapy

No indications

Ceftazidime orimipenem*

Reassess after three days

Aminoglycoside**�

anti-pseudomonal �-lactam

DuotherapySevere mucositis, quinolone prophylaxis,colonized with methicillin-resistant Staphylococcus

aureus or penicillin/cephalosporin-resistantStreptococcus pneumoniae, obvious catheter-related

infection, hypotension

*Recent studies suggest that cefepime or meropenem may be as effective as ceftazidime or imipenem as monotherapy.**Avoid if patient is also receiving nephrotoxic, otoxic, or neuromuscular blocking agents; has renal or severe electroytedysfunction; or is suspected of having meningitis.

Figure 14.1 Guide to the initial management of the febrile neutropenic patients. (Reproduced with permissionfrom Hughes WT et al, J Infect Dis 1990; 161: 381–96.7)


Persistent fever during first 3 days of treatment: no etiology

Reassess patient on day 4–5

Add amphotericin Bwith or without

antibiotic change

Continue initialantibiotics

Changeantibiotics

If febrile throughdays 5–7 and resolution

of neutropenia is notimminent

If no change in patient,consider stopping

vancomycin

If progressive disease,add vancomycin plus

Gram-negativebacillary coverage

Figure 14.2 Treatment of patients who have persistent fever after three days of treatment and for whom theetiology of the fever is not found. (Reproduced with permission from Hughes WT et al, J Infect Dis 1990; 161:381–96.7)

Afebrile by day 3

ANC500/µlby day 7

Fever persists

Stop after7 days

ANC�500/µlby day 7

Low risk orclinically well

Stop whenafebrile

5–7 days

High risk orANC�500/µl,mucositis, or

clinically unstable

Continueantibiotics

Stop 4–5days after

ANC500/µl

Reassess

Continuefor 2 weeks

Reassess

ANC500/µl ANC�500/µl

Stop if noevidence

of disease,and stable

Figure 14.3 Duration of antibiotic therapy (ANC, absolute neutrophil count). (Reproduced with permission fromHughes WT et al, J Infect Dis 1990; 161: 381–96.7)

include discontinuation of therapy at four tofive days after neutrophil recovery. In persis-tently febrile patients without neutrophil recov-ery, it is suggested that all antimicrobials maybe discontinued after two weeks of therapyincluding amphotericin B, provided there is noevidence of clinical infection (includingcomputed-tomographic lung and abdominalscans to rule out systemic fungal infection).These patients must then be followed with closeobservation.

Antivirals are not routinely recommended bythe IDSA as part of initial empiric therapyunless clinically indicated, and for the useof hematopoietic colony-stimulating factors(CSFs), clinicians are referred to the AmericanSociety of Clinical Oncology guidelines for CSFuse.12–14 However, it was suggested that empiricCSF use might be indicated in patients withhypotension or sepsis-related multiorgan dys-function, as well as in those with pneumonia,systemic fungal infections, severe cellulitis orsinusitis, and an anticipated delay in marrowrecovery. Granulocyte transfusions were notrecommended. Except for prophylaxis ofPneumocystis carinii infection, these guidelinesdid not recommend routine antibiotic prophy-laxis. The panel reviewed the effectiveness oftrimethoprim–sulfamethoxazole (TMP–SMZ) orquinolone prophylaxis, which demonstrated areduction in infection rates but not in mortalitywith TMP–SMZ, and a reduction in Gram-negative bacillary infections but not in the rateof infections overall with quinolones. Despitethis evidence, the panel declined to recommendroutine antibiotic prophylaxis in neutropenicpatients because of the concern regarding theemergence of drug-resistant bacteria due toextensive antibiotic use. This represents a shiftfrom the 1990 guidelines, which had endorsedprophylaxis with TMP–SMZ in certain patients.Finally, the updated guidelines review severalcost-containment suggestions, including the useof lower-dose antibiotic therapy, steppingdown to outpatient oral therapy where pos-sible, and avoiding expensive agents (e.g. lipo-somal amphotericin B and CSFs) where they are

unnecessary. While initial empiric outpatienttherapy is not endorsed, it is highlighted as anarea of promise, and potential cost contain-ment.

With respect to the appraisal of guidelinecontent, the reasons for and objectives of guide-line development are clear, with a satisfactorydescription of the disease definition andpatients to whom the guidelines apply. It isclearly stated that the guidelines are to be usedin the context of appropriateness for the indi-vidual patient and setting, and should not beused blindly or against the better judgment orexperience of the clinician. A multiplicity oftreatment options is clearly stated, and the rec-ommendations are clearly presented for clini-cians, with an adequate description oftreatment benefits, as well as treatment toxici-ties. As mentioned, there is a section reviewingeconomic considerations, and most recommen-dations do appear supported by the benefits,toxicity, and costs of the interventions. Wherethis is not the case, the panel did highlight rea-sons for the discrepancy. The IDSA is presentlyrevising the guidelines again, and publicationof the revision is anticipated in 2001.

Guidelines from the National ComprehensiveCancer Network

The National Comprehensive Cancer Network(NCCN) has also drafted guidelines for febrileneutropenia management, using consensus ofexpert opinion and published in 1999.8 Thepanel of experts was drawn from member insti-tutions of the NCCN, with the co-chair havingalso been on the IDSA guidelines developmentpanel. The document states that the recommen-dations made are based on scientific publica-tions, peer-reviewed information formallypresented at meetings, and, where data arelacking, expert opinion. The rating system ofrecommendations is based on the degree ofconsensus achieved, ranging from uncontestedrecommendations to those that caused cleardisagreement among the panel experts. The


definition and evaluation of febrile neutropeniaoutlined in the document parallel those of theIDSA. There is also a review of the utility oftwo-site culturing for serum cultures, which isnot routinely recommended. Initial treatmentrecommendations again parallel those of theIDSA, with recommendations for monotherapy(using imipenem, meropenem, ceftazidime,cefepime), duotherapy with aminoglycosidesand anti-pseudomonal penicillins, and alsomention of the use of two �-lactam agents,although this was not universally supported,similar to the IDSA guidelines. Initial use ofvancomycin as part of combination therapy wasintroduced as an option, with similar clinicalindications to those of the IDSA. One clear dif-ference from the IDSA guidelines is the optionof duotherapy with ciprofloxacin and a urei-dopenicillin.

The guidelines are unique in the develop-ment of easy-to-use algorithms on site-directedtherapy. Various scenarios of clinical infectionin the neutropenic cancer patient are reviewed,including clinical signs, appropriate testingstrategies, evaluation of treatment response,and duration. There is also a focus on potentialetiologic agents and site-specific therapy,including recommendations for empiric antivi-ral and antifungal therapy tailored to specificclinical scenarios. While these algorithms arevery attractive to the clinician for their ease ofpractical application, the quality of the evidenceunderpinning the specific therapies in the algo-rithms is not always explicit in the document,making critical appraisal of that evidence diffi-cult.

The NCCN guidelines include a section onoutpatient therapy, with clear recommenda-tions to consider outpatient therapy in experi-enced centers, based upon the Talcott model ofrisk assessment and small, randomizedtrials.15–18 For the use of CSFs, conformation tothe ASCO guidelines is recommended.12–14 Withrespect to prophylaxis, there are several recom-mendations for specific use, in contrast to theIDSA guidelines. These recommendationsinclude the use of quinolone or TMP–SMZ pro-

phylaxis for severe neutropenia (<100 neu-trophils/µl for �7 days, but not for routineshort-term neutropenia. Antifungal and antivi-ral prophylaxis are also recommended in spe-cific settings, with a review of the existingevidence (namely fluconazole in marrow trans-plant patients, herpes simplex virus prophy-laxis in marrow transplant and acute leukemiapatients with re-induction). With respect to P.carinii prophylaxis, several groups are identi-fied as appropriate candidates. These includerecipients of allogeneic marrow transplantationand patients receiving therapy for acute lym-phoblastic leukemia, and mention is made ofconsideration of prophylaxis for patients on flu-darabine, those with high steroid use (i.e.>20 mg/day of prednisone), and those receiv-ing autologous peripheral blood stem cells.Expert opinion likely underpins considerationof prophylaxis of the latter group, although thisis not explicit. Additional suggestions includethe use of antibiotic-impregnated vascularcatheters, and antifungal treatment strategiesincluding multiple antifungal agents, andadjunctive therapy such as CSFs or even granu-locyte transfusions. Again, the overall speci-ficity of the recommendations lends itself wellto practical application in the clinical setting,but a clear rating of the supporting evidence forthese was not clearly stated in the documenttext.

Guidelines from the Federation of FrenchCancer Centers

Recommendations for the management of briefneutropenia from the Federation of FrenchCancer Centers were published in 1998, in abs-tract form in English.9 The practice guidelinewas developed by a multidisciplinary group ofexperts, with feedback from oncologists. Datausing literature search (Medline and CurrentContents) and personal reference lists wereused, with clearly defined outcomes for studyendpoints. Reviewers from 20 French cancercenters reviewed the evidence and formulated


recommendations. These recommendations arenot dissimilar to those of the IDSA. Theyinclude informing patients of the risks ofchemotherapy, observing afebrile neutropenicepisodes without antibiotic prophylaxis, andempiric antibiotic therapy for all febrile neu-tropenic episodes, which may include �-lactamand aminoglycoside combination therapy ormonotherapy with a broad-spectrum �-lactamexcept in the case of septic shock or respiratorydisease. Glycopeptides such as vancomycin canbe added if catheter-related or cutaneous infec-tion is obvious, as well as in the case of microbi-ologically documented infection caused byoxacillin-resistant Gram-positive bacteria, or inthe clinically deteriorating patient with persis-tent fever. The evidence supporting outpatienttreatment was felt to be insufficient to recom-mend it at that time, and participation in stud-ies to identify factors predicting low risk andassessing feasibility and safety of early dis-charge and outpatient therapy was recom-mended.

Guidelines from the Japanese InfectiousDisease Society

Available in abstract form in English, guide-lines developed by the Japanese InfectiousDisease Society were published in 2000, usingformal consensus methodology where indi-vidual panel members vote on each recommen-dation.10 The definition and evaluation aresimilar to those in the guidelines previouslydiscussed, as is empiric therapy, with a recom-mendation for either monotherapy with a car-bapenem or third-generation �-lactam orcombination of either with an aminoglycoside.Initial glycopeptide treatment is not mentionedin the abstract, but may well be discussed in thefull text. If additional therapy is required, theuse of glycopeptide or antifungal agents is rec-ommended. The abstract states that the guide-lines require study for confirmation, implyingfuture outcome assessment, including evalu-ation of response rates of different recom-

mended treatment regimens, and the potentialfor use in evaluation of newer agents for febrileneutropenia.

Guidelines from the Italian Association ofPediatric Hematology and Oncology

The aim of these guidelines was to address thetremendous variability in management offebrile neutropenia in children in Italian cen-ters, and the impact on costs and developmentof antibiotic resistance.11 A chairperson wasnominated to prepare a first draft based uponevidence, which was subsequently reviewed byexperts in the society. Through publication, thedocument also underwent external peer review.The rating system is similar to that used by theIDSA in 1990, linking recommendation accep-tance by experts with quality of evidence, andwithout a separate rating of the quality of sup-porting evidence. A similar definition of febrileneutropenia to that in the adult population isreached. As with other guidelines in this area,clinicians are cautioned to apply the guidelineswith awareness of their local patterns of infect-ing organisms and antimicrobial resistance.Standard evaluation without surveillance cul-tures is recommended, with a review of evid-ence for additional tests. A recommendationformulated from expert opinion for ongoingcultures daily or even twice daily was pre-sented – a contrast to the IDSA guidelines foradults. It is recommended that all patientsreceive inpatient therapy, with recognition thatoral or outpatient therapy is a feasible approachcurrently under investigation. The three recom-mended options for empiric therapy of febrileneutropenia in children are similar to those inadults, namely combinations of �-lactams andaminoglycosides, double �-lactam coverage,and monotherapy with a carbapenem or third-generation cephalosporin. Intravenous or oralmonotherapy with a quinolone is also men-tioned, but the supporting evidence is notclearly reviewed in the document. The pros andcons of all options, including cost, are


presented. The favored option, particularly inhigher-risk patients (i.e. those with prolongedduration of neutropenia), is combination ther-apy with a �-lactam and aminoglycoside.Overall, empiric use of vancomycin orteicoplanin is discouraged for fear of develop-ment of resistant organisms, and their use is notendorsed unless staphylococcal infection risk isclinically likely, in centers with high rates ofmethicillin resistance. A similar caution is maderegarding the development of resistant organ-isms through the use of single-agent car-bapenem therapy, and the authors concludethat this should not be used routinely as part ofinitial empiric treatment. Single daily dosage ofaminoglycosides, specifically amikacin giventhe burden of evidence evaluating this drug, issupported.

With respect to the duration of treatment,continuing antibiotics for four to seven daysafter defervescence is suggested, withconsideration of early discharge if patients areafebrile and at low risk. Patients at low riskinclude those with evidence of marrow recov-ery and controlled disease, and who haveaccess to proper monitoring at home. The use ofantifungal therapy was reviewed, and empiricantifungal therapy was discouraged, except inselected groups of patients. Planned outcomeevaluation of the guideline development andplans for revision were not specified in the text,but may be published in the future.

GUIDELINES FOR METHODOLOGY INCLINICAL TRIALS IN FEVER ANDNEUTROPENIA

While guidelines for practice are becomingincreasingly sophisticated in the evaluation oftreatments and comparison of different agents,a major problem remains in the lack of unifor-mity in available evidence, particularly in clini-cal trial endpoints. It is intuitively clear that asthe definition of failure of response to therapychanges, so will a study’s results and its conclu-sions about treatment efficacy. This was ele-

gantly demonstrated in a study conducted byElliott and Pater,19 in which three different mea-sures of outcome were used to define responsein testing new antibiotic regimens for febrileneutropenic episodes. Using two of the defini-tions, a significant difference betweenresponses to treatment regimens weredemonstrated, but using the third definition,only treatment equivalence was shown. Thislatter definition used death from infection asthe only definition of treatment failure, andsince patients could be rescued with treatmentmodifications as part of the study, the infection-related death rates are, not surprisingly, equiv-alent. Thus, changing the height of thebenchmark of the trial endpoint can change thesuccess rate in these trials.

Several definitions of treatment response intrials of febrile neutropenia exist, includingguidelines proposed by the Immuno-compromised Host Society (ICHS),20 theInternational Antimicrobial Therapy Coopera-tive Group of the European Organization forResearch and Treatment of Cancer (EORTC),21

Pizzo and co-workers,22,23 and others (see Table14.1).

An earlier attempt to develop a consensusguideline in this area resulted in adoption ofthe ICHS guidelines for response,24 and hasbeen extensively reviewed elsewhere.25,26

Despite the US Food and Drug Administrationrequirement that trials carried out to obtainapproval for antimicrobials in patients withfebrile neutropenia comply with these guide-lines, this has not resulted in the widespreadreport of these endpoints in the actual publica-tions of trial results. As the published resultsare what clinicians use to guide therapy, andcomprise the evidence on which practice guide-lines are commonly based, this is a key area ofconcern. It seems clear that in our paradigmshift in the clinical practice of medicine to amore evidence-based approach, the use of auniform guideline on how to conduct investiga-tion and establishment of that evidence must bepromoted. To that end, the ICHS guidelines formethodology of the design and conduct of clini-



Table 14.1 Criteria for response in studies of empiric therapy for neutropenic sepsisa

Pizzo criteria for response1. Evaluation: at 72 hours and end of treatment2. Success: survival of febrile episode until neutrophil recovery with no modification of original regimen3. Success with modification: survival of febrile episode but requiring addition of another agent or

complete change of regimen4. Failure: death from infection

EORTC criteria for response1. Evaluation: at 72–96 hours and end of treatment2. Success: all signs of infection resolved and infecting organisms eradicated without any modification

of primary regimen; must be maintained for 7 days after stopping initial antibiotics3. Failure: death from infection; bacteremia for more than 24 hours; breakthrough bacteremia; no

response to initial therapy; documented resistance of pathogen to initial therapy; modification ofregimen with shock, acute respiratory failure, disseminated intravascular coagulation

4. Non-evaluable for response: mixed or non-bacterial infection; non-infectious febrile episode;treatment stopped because of toxicity

5. Protocol violation: change of therapy from �-lactam, with susceptible Gram-positive infection insetting of persistent fever

ICHS criteria for responseMicrobiologically defined infection1. Success: eradication of signs and microbiologic evidence of infection on primary therapy; no

recurrence for 7 days after stopping initial antibiotics2. Initial response but regimen modified: success as above, but secondary infection arises requiring

addition of another agent3. Failure: death from infection; any change to the initial antibiotic regimen to eradicate infection

Clinically defined infection (i.e. no microbiologic isolate)1. Success: same as above without evidence of bacteriologic cure2. Initial response but regimen modified: same as above3. Failure: same as above

Unexplained fever1. Success: patient defervesces on initial regimen and recovers from neutropenia; no recurrence of

fever within 7 days of completing initial antibiotic regimen2. Initial response but regimen modified: patient develops new fever after initial defervescence,

requiring additional agent outside spectrum of initial therapy3. Non-response or failure: death from infection; any change to initial regimen for persistent fever

aAdapted with permission from Feld R, Support Cancer Care 1998; 6: 444–8.26

cal trials in febrile neutropenia are beingupdated, with the assistance of theMultinational Association for Supportive Carein Cancer (MASCC). In addition to the empha-sis on patients considered high-risk, they willalso include guidelines for trials of outpatienttherapy, in patients at lower risk, using theTalcott15 and/or MASCC criteria for this defini-tion.27 These guidelines are currently being pro-duced by both international societies, andideally would be endorsed by other societiessuch as ASCO and IDSA prior to publication,which is expected in 2001.

Predicting the future of guidelines is not anexact science. But it can be seen from thisreview that many local documents are based onone or two major guideline publications, andwe anticipate this will be the way of the future.While local guidelines assist practitioners inapplying minor regional differences to generalmanagement approaches – for example tailor-ing therapy guidelines to pathogens andantimicrobial resistance patterns in a specificgeographic area – we expect that only a fewmajor guidelines in each subject area will con-tinue to be updated. Through the maintenanceof these guidelines as current and relevant toclinical practice, we can continue to promotequality improvements in our practice patternsand in the supportive care of tomorrow’s cancerpatient.

REFERENCES

1. Grimsahw J, Eccles M, Clinical practice guide-lines. In Evidence Based Practice in Primary Care(Silagy C, Haines A, eds). Plymouth, UK:Latimer Trend, 1998: 110–22.

2. Cluzeau F, Littlejohns P, Grimshaw J, Feder G,Appraisal Instrument for Clinical Guidelines. London:St George’s Hospital Medical School, 1997.

3. Hayward RSA, Wilson MC, Tunis SR et al,Users’ guides to the medical literature VIII. Howto use clinical practice guidelines A. Are the rec-ommendations valid? JAMA 1995; 274: 570–4.

4. Wilson MC, Hayward RSA, Tunis SR et al,Users’ guides to the medical literature VIII. How

to use clinical practice guidelines B. What are therecommendations and will they help you in car-ing for your patients? JAMA 1995; 274: 1630–2.

5. ECRI (International Nonprofit Health ServicesResearch Agency), Collaborating Center of theWorld Health Organization (WHO), and EPC(Evidence-based Practice Center) of the USAgency for Healthcare Research and Quality(AHRQ). National Guidelines Clearinghouse,1997 (http://www.guidelines.gov)

6. Hughes WT, Armstrong D, Bodey GP et al,Guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever. JInfect Dis 1990; 161: 381–96.

7. Hughes WT, Armstrong D, Bodey GP et al, 1997Guidelines for the use of antimicrobial agents inneutropenic patients with unexplained fever.Clin Infect Dis 1997; 25: 551–73.

8. National Comprehensive Cancer Network,NCCN practice guidelines for fever and neu-tropenia. Oncology 1999; 13: 197–257.

9. Biron P, Guhrmann C, Escande MC et al,Standards, options, and recommendations forthe management of brief neutropenias.Federation Nationale des Centres de LutteContre le Cancer. Bull Cancer 1998; 85: 695–711.

10. Masaoka T, Febrile neutropenia – guideline inJapan. Gan To Kagaku Ryoho 2000; 27: 161–5.

11. Viscoli C, Castagnola E, Caniggia M et al, Italianguidelines for the management of infectiouscomplications in pediatric oncology: empiricalantimicrobial therapy of febrile neutropenia.Oncology 1998; 55: 489–500.

12. American Society of Clinical Oncology Ad HocColony-stimulating Factor Guidelines ExpertPanel, American Society of Clinical Oncology,Recommendations for the use of hematopoieticcolony-stimulating factors: evidence-based, clini-cal practice guidelines. J Clin Oncol 1994; 12:2471–508.

13. Ozer H, Miller LL, Schiffer CA et al, AmericanSociety of Clinical Oncology update of recom-mendations for the use of hematopoietic colony-stimulating factors: evidence-based clinicalpractice guidelines. J Clin Oncol 1996; 14:1957–60.

14. Ozer H, Armitage JO, Bennett CL et al, 2000update of recommendations for the use ofhematopoietic colony-stimulating factors:Evidence-based clinical practice guidelines. JClin Oncol 2000; 18: 3558–85.



16. Talcott JA, Whalen A, Clark J et al, Home anti-biotic therapy for low-risk cancer patients withfever and neutropenia: a pilot study of 30patients based on a validated prediction rule. JClin Oncol 1994; 12: 107–14.

17. Rubenstein EB, Rolston K, Benjamin RS et al,Outpatient treatment of febrile neutropenia inlow-risk neutropenic patients with cancer.Cancer 1993; 71: 3640–6.

18. Rolston KVI, Rubenstein EB, Eltin LS et al,Ambulatory management of febrile episodes inlow risk neutropenic patients. In: Programs andAbstracts of the 35th Interscience Conference onAntimicrobial Agents and Chemotherapy, SanFrancisco. Washington, DC: American Society ofMicrobiology, 1995: Abst 2235.

19. Elliott C, Pater JL, The effect of different mea-sures of outcome on the results of studies ofempiric antibiotic therapy in febrile neutropenicpatients. Clin Invest Med 1988; 11: 327–30.

20. Immunocompromised Host Society, Consensuspanel of the design analysis and reporting ofclinical trials on the empirical antibiotic manage-ment of the neutropenic patient. J Infect Dis 1990;161: 397–401.

21. Cometta A, Zinner S, de Bock R et al, TheInternational Antimicrobial Therapy CooperativeGroup of the European Organization of Researchand Treatment of Cancer, Piperacillin-tazobac-

tam plus amikacin versus ceftazidime plusamikacin as empiric therapy for fever in granu-locytopenic patients with cancer. AntimicrobAgent Chemother 1995; 39: 445–52.

22. Pizzo PA, Hathorn JW, Hiemen ZJ et al, A ran-domized trial comparing ceftazidime alone withcombination antibiotic therapy in cancer patientswith fever and neutropenia. N Engl J Med 1986;315: 552–8.

23. Freifeld AG, Walsh T, Marshall D et al,Monotherapy for fever and neutropenia incancer patients: a randomized comparison of cef-tazidime versus imipenem. J Clin Oncol 1995; 13:165–76.

24. Hughes W, Pizzo PA, Wade JC et al, Evaluationof new anti-infective drugs for the treatment offebrile episodes in neutropenic patients. ClinInfect Dis 1992; 15(Suppl 1): S206–15.

25. Feld R, Methodology for clinical trials in cancerpatients with febrile neutropenia: Do we have aconsensus? Support Care Cancer 1998; 6: 423–4.

26. Feld R, Criteria for response in patients in clini-cal trials of empiric antibiotic regimens forfebrile neutropenia: Is there agreement? SupportCare Cancer 1998; 6: 444–8.

27. Klastersky J, Paesmans M, Rubenstein EB et alfor the Study Section on Infections of theMultinational Association for Supportive Care inCancer, The Multinational Association forSupportive Care in Cancer Risk Index: a multi-national scoring system to predict low-riskfebrile neutropenic cancer patients. J Clin Oncol2000; 18: 3038–51.


15Healthcare economic concepts in fever andneutropeniaThomas D Szucs

I am finally summing up with an urgent appealto adopt this or some uniform system of publish-ing the statistical records of hospitals. If theywere obtained, they would show subscribers howtheir money was being spent, what amount wasreally being done with it, or whether the moneywas doing more mischief than good.

Florence Nightingale, 1863

WHY MEDICAL ECONOMICS?

In the past few years, the discipline of healtheconomics has experienced an extraordinaryboom within the healthcare sector. Researchersfrom a wide range of disciplines havedeveloped new techniques to evaluate theimpact of clinical care and medical technology.Clinicians, pharmacists, economists, epidemiol-ogists, and operations researchers have con-tributed to the new field of medical economicsto study how different approaches to patientcare influence the resources consumed in clini-cal medicine. Faced with the basic economicnotion that resources are limited and desires aswell as needs are infinite, health economists tryto find solutions to the problem of how theseresources can be allocated appropriately tomaximize the production of health. The com-mon denominator is the search for increased

efficiency and effectiveness of healthcare ser-vices and products. Among these efforts, sev-eral researchers from a wide range ofdisciplines have begun to study the economicimpact of medicines on healthcare provision onmicro- and macroeconomic levels. The newfield of pharmacoeconomics has grown at atremendous rate, supplying ample evidence ofthe economic benefits of modern therapeutics.

Not all medical interventions and proce-dures, however, have to be subject of an eco-nomic evaluation. A simple decision tool isdisplayed in Figure 15.1. This box takes the twomain outcomes of a potential evaluation intoconsideration: costs and clinical results or qual-ity of a medical intervention. In circumstances

Higher costs

Evaluate

Reject

ACCEPT(dominant)

Evaluate

Lower costs

Out

com

e

Figure 15.1 The four possible outcomes of aneconomic evaluation.

where a drug is costly and leads to betterresults, an economic evaluation is advisable.Where a drug is cheaper and yields better out-comes, the drug should be accepted. In theopposite case, where a more costly drug or pro-cedure produces less favorable outcomes, oneshould reject it. Economic evaluations may,however, assist in identifying those drugs thathave poorer outcomes but are also cheaper.

THE MAJOR COMPONENTS OF ANECONOMIC EVALUATION

All economic studies investigate the balancebetween inputs (the consumption of resources)and outcomes (improvements in the state ofhealth of individuals and/or society).

Inputs (costs)

Although the unit price of a drug or procedureis often a prime factor in decision making,economic outcomes research provides a morecomprehensive interpretation of cost. This isaccomplished by determining the overall cost ofa given diagnostic and therapeutic process fromthe initiation of diagnosis until a final outcomeis achieved. The various types of costs can begrouped under the following categories: directmedical costs, direct non-medical costs, andindirect costs.

Direct medical costsInterpretations of what belongs in each of thesecategories vary in the economic literature.Direct medical costs are defined as thoseresources used by the provider in the deliveryof medical care. As an example, direct medicalcosts for a hospital include:

• drugs;• laboratory tests;• medical supplies;• use of diagnostic equipment (e.g. magnetic

resonance imaging, CAT scans, X rays);• medical staff time for personnel such as

physicians, nurses, pharmacists, physicaltherapists, laboratory technicians, etc.;

• room and board: the cost of supplies,equipment, and personnel required for rou-tine patient-related services such as food,laundry, and housekeeping.

These are examples of costs that can be directlyrelated to the care of patients. Other costs ofoperating a hospital include plant maintenanceand repairs, utilities, telephone, accounting,legal fees, insurance, taxes, real estate costs, andinterest expenses. In general, most economicstudies do not factor general operating costsinto the dollar value assigned to the cost ofresources expended for a given medicine.

Looking down the list of direct medicalexpenses, it is easy to see why length of stay isan important cost factor to hospitals, especiallywhen payment is determined by diagnosis-related groups (DRGs). Costs such as room andboard are directly tied to the length of stay,regardless of the reason. The cost of laboratorytests, supplies, and medical staff time vary withthe medical condition being treated, but aremultiplied by the length of stay.

Since the introduction of DRGs, a commonselling strategy has been to emphasize how agiven technology such as a new antimicrobialdrug can help shorten hospital stays. Morerecently, economic studies began to collect andanalyze data linking specific diagnostic strat-egies to length of stay.

Length of stay in hospital settings translatesto the number of patient visits in managed caresettings. Although the specific items includedunder the category of direct medical costs willbe slightly different in managed care organiza-tions, the same principles of cost analysis apply.Drugs that achieve results quickly and pre-dictably not only benefit the patient, but alsobenefit the provider by reducing the number ofpatient visits. Every patient visit incursprovider resource costs that may not be reim-bursed by a third-party payer. Interventionsthat minimize patient visits are clearly cost sav-ings for the healthcare organization.


Direct non-medical costsEconomic literature generally defines directnon-medical costs as out-of-pocket expensespaid by the patient for items outside the health-care sector. This category includes such costs as:

• travel to and from the hospital, clinic, ordoctor’s office;

• travel and lodging for family members wholive elsewhere;

• domestic help or home nursing services;• insurance co-payments and premiums;• treatment not covered by third-party pay-

ers.

Although these costs are generally classified as‘non-medical’, to the patient they are real andoften substantial costs of medical care. Whatmakes them ‘non-medical’ is that they are notcosts incurred by the healthcare provider, andare somewhat difficult to measure. For example:

• A patient’s inability to afford competentfollow-up care at home may result in poorcompliance with drug therapies and even-tual treatment failure. This may lead toadditional hospital stays or office visits,which affect the provider’s bottom line.

• A patient’s inability to bear the unreim-bursed cost of medications may also lead topoor compliance and costly complications.

• High transportation costs may lead tomissed appointments for necessary follow-up visits, which can result in deteriorationof a patient’s medical condition andincreased treatment costs for the provider.

Even though the provider may not directlyincur these costs, they can be used in specificsituations by making the provider aware oftheir potential economic impact. It may also bepossible to use these costs to encourage payers(e.g. employers and insurance companies) todiscuss the use of a more cost-effective test withthe healthcare provider.

Indirect costsOne definition of indirect costs (also called‘intangible costs’ by some economic analysts) is

the overall economic impact of illness on thepatient’s life. These include:

• loss of earnings due to temporary, partial,or permanent disability;

• unpaid assistance by family members inproviding home healthcare;

• loss of income to family members who for-feit paid employment in order to remain athome and care for the patient.

Like direct non-medical costs, indirect costs arereal to the patient, but abstract to the provider –but may impact the provider’s direct medicalcosts. For example, patients who cannot earnincome may not be able to pay their bills –including medical bills. Economic hardshipmay result in poor compliance with drug thera-pies as patients reduce doses or fail to refill pre-scriptions in order to save money. The medicalprovider may have to bear the additional costsof managing complications. Economic hardshipmay also result in missed follow-up appoint-ments, leading to the same types of problemsfor providers as described previously withdirect non-medical costs.

Consequences and outcomes

Final states or outcomes can be negative (some-times referred to as the ‘five D’s’):

• death;• disability (the patient is permanently dis-

abled and unable to return to work orschool, perform household chores, etc.);

• discomfort (the patient is in a constant stateof moderate to severe pain);

• dissatisfaction (the patient is not satisfiedwith the course of treatment or servicesprovided);

• disease (the patient’s condition is not beingcontrolled, resulting in frequent relapses,rehospitalization, and expenditure of addi-tional resources).

There are also positive outcomes:

• the patient is cured;

HEALTHCARE ECONOMIC CONCEPTS IN FEVER AND NEUTROPENIA 303

• the patient is able to resume normal func-tions;

• there is an improved or satisfactory qualityof life;

• the patient’s medical condition is success-fully managed or stabilized by continueddrug therapy.

The use of outcomes research (see below)represents an important advance in medicaleconomic analysis because of the relationshipbetween the final state, or result, of diagnosisand therapy and overall cost–effectiveness. If itcan be demonstrated that a product or inter-vention will achieve cost-effective, positive out-comes, this will increase the chances of thetechnology being diffused within healthcare.

IMPORTANT ECONOMIC CONCEPTS

Average, marginal, and joint costs

Most decisions in healthcare are not concernedwith whether or not a service should be pro-vided, or whether or not a particular procedureshould be undertaken, but rather with howmuch of the service should be provided. Thatis, should existing levels of provision beexpanded or contracted? For example, shouldthe existing provision of daycare for peoplewith mental illness be expanded, and, if so, byhow much? What family planning servicesshould be made available? How many patientspresenting with head injuries should have com-puted tomograms? All these decisions requirethat attention should be focused on marginalcosts – that is, the change in total costs resultingfrom a marginal change in activity.

In the short run, there is often an importantdifference between the marginal cost of anactivity and its average cost, where the averagecost is defined as the total cost divided by thetotal number of units of output (Table 15.1). Anexample is provided by a study of thecost–effectiveness of antihyperlipemic treat-ment in the prevention of coronary heart dis-


Table 15.1 Total, average, and marginalcosts: an illustration

No. of Total Average Marginalpatients cost cost costtreated ($) ($) ($)

10 4000 400 020 5000 250 10030 6000 200 10040 6800 170 8050 7400 148 60

ease.1 In this study, the marginal cost was calcu-lated per life-year saved of continuing drugtreatment for successive periods of 5 additionalyears for patients between 35 and 70 years ofage. These indicated large differences betweenaverage and marginal cost–effectiveness, sincemarginal costs increased quite steeply after 55years of age.

Another context in which the distinctionbetween average and marginal costs is impor-tant is in relation to the duration of hospitalstay of inpatients. Many new procedures havereduced the amount of time necessary for apatient to remain in hospital, and thereby yieldcost savings. When valuing these savings,however, it is important to keep in mind thatusing average costs/day will generally over-state the savings, since the later days of astay usually cost less than the earlier ones. Itis the marginal cost/day that is the relevantmeasure.

Another problem of cost measurement arisesin connection with joint costs. Often a singleproduction process can result in multiple out-puts. For example, a single chemical analysis ofa blood sample can diagnose the presence ofmany diseases. How should the cost be alloc-ated to each diagnosis? Similarly, within a

hospital setting, there are many common ser-vices (such as medical records, radiology, oper-ating theatres, laundry, catering, and cleaning)that contribute to a number of specialities.Economic evaluation requires some method forallocating the joint costs of these services toindividual programs or procedures. There areseveral methods that may be used to do this.Most of them use some physical unit of utiliza-tion, such as the number of laboratory tests,hours of operating theatre use, or square metersof ward space, to apportion total laboratory,operating theatre, and ward cleaning costs.

Costs of capital

Investments in buildings and equipment thatyield a flow of services over a number of yearsgive rise to capital costs. Generally, investmentexpenditure will be undertaken at the begin-ning of a project, but the use of items of capitalequipment will generate annual capital costsover the lifetime of the asset.

These costs have two components: interestand depreciation. Interest costs should beincluded even if the asset was not acquiredwith borrowed money, because tying-upmoney in an item of capital equipment involvesan opportunity cost, namely, interest foregone.Depreciation costs arise because of the wearand tear that an asset receives through use andthe consequent reduction in the length of itsuseful life. Land, however, is a capital asset,which is not assumed to incur depreciationcosts.

Sometimes an item of capital expenditure isunique to a particular use, and has little or noalternative use value (opportunity cost). In suchcases, it is referred to as a sunk cost. A hospitalbuilding or an item of medical equipment may,for example, have considerable value in itsexisting use but little resale value. This can pro-vide a powerful case for continuing to use exist-ing assets instead of undertaking newinvestments, because, in an economic evalu-ation, sunk costs should not be included among

annual capital costs. In practice, this considera-tion is likely to be more important in the case ofmajor capital developments than of individualprocedures.

Adjusting for differences in timing:discounting

The current (operating) costs associated withmost procedures can be expected to extend overa number of years into the future, but their timeprofiles may differ. In the case of many preven-tive procedures, such as treatment for hyper-tension, costs will be incurred regularly over anumber of years. The alternative of no preven-tive treatment may well incur zero expenditurein the early years, but incur the costs of surgeryearlier than would otherwise have been thecase. Discounting offers a means of standardiz-ing different cost–time profiles so that totalcosts can be compared.

Discounting is based on the assumption thatcosts incurred in the immediate future are ofgreater importance than costs incurred in thedistant future. This is because earlier access tofinance would permit investment at a positiverate of interest, thereby yielding a larger sum inthe future (there is an opportunity cost) orbecause people and society attach more impor-tance to current opportunities than to futureones (positive time preference).

For these reasons, economic evaluationweights costs by a discount rate, according tothe year in which they accrue, before addingthem up and expressing total costs in present-value terms (values in the current year).

In essence, discounting is the reverse applica-tion of the more familiar compound interestformula – instead of sums being calculated for-wards, they are discounted backwards.Fortunately, the application of discounting doesnot require close familiarity with the formula,since many finance and accounting textbooksinclude discount tables. These indicate the pre-sent values of the dollar at different discountrates.


Table 15.2 shows how present discountedvalues will vary for selected combinations ofthe discount rate and the years in which thecosts accrue. Looking across the second row ofthe table, for example, shows that in the fifthyear, $1.00 will be worth 86 cents at a discountrate of 3%, but only 68 cents at a discount rateof 8%. The choice of the appropriate discountrate depends in part on national recommenda-tions. Given the sensitivity of valuations to thechoice of discount rate, however, and the factthat the ranking of different projects with dif-ferent time profiles could depend on the ratechosen, it is good practice to compute costs interms of a range of discount rates.

Inflation

Most programs that extend over several yearswill be affected by inflation. It is important,however, to distinguish between changes in thegeneral price level and changes in relativeprices. In the case of general inflation, there willbe no change in the relative cost of inputs (theiropportunity costs remain constant). As such, allfuture inputs can be valued at current pricesand discounted by a real (excluding inflation-ary effect) rate of interest.

If, however, some input prices are expectedto increase more than others, there will be rela-tive changes in their opportunity costs, andthese need to be taken into account. One way ofdoing this is to use the general rate of inflationas a benchmark and to adjust the future pricesof individual inputs – upwards or downwards– by an amount that reflects the differencebetween their rate of inflation and the generalrate. Thereafter, all costs should once again bediscounted by the same real rate of interest.

METHODS OF HEALTH ECONOMICEVALUATION

The most common methods employed byhealth economists are based on classical researchdesigns such as cost-minimization, cost–benefit,cost–effectiveness, and cost–utility analyses.2–4

(Table 15.3).

Cost-minimization analyses (CMA)

Cost-minimization analysis is concerned withcomparing the costs of different treatmentmodes that produce the same result. For exam-ple, this form of analysis could be used to com-


Table 15.2 Present value of 1$

Discount rate

Year 3% 4% 5% 6% 7% 8%

1 0.9709 0.9615 0.9524 0.9434 0.9346 0.92595 0.8626 0.8219 0.7835 0.7473 0.7130 0.6806

10 0.7441 0.6756 0.6139 0.5584 0.5083 0.463215 0.6419 0.5553 0.4810 0.4173 0.3624 0.315220 0.5537 0.4564 0.3769 0.3118 0.2584 0.214525 0.4776 0.3751 0.2953 0.2330 0.1842 0.1460

pare the cost of two programs that involveminor surgery for adults. Both have the sameoutcome in terms of the surgical procedure, butthe first program might require the patient tostay overnight at the hospital, while the secondmight be done through day surgery withoutrequiring hospitalization. Given these twoalternatives, the search would be for the leastcostly treatment. While we might be interestedin the extent to which daycare surgery shiftscosts from the institution to the patient, themain efficiency comparison would be on a costper surgical procedure basis.

As far as pharmaceuticals are concerned, thistype of study is used most frequently when anew drug is introduced into a therapeutic classthat includes close competitors and no measur-able therapeutic effect between them has beendocumented.

When the costs of two interventions arebeing compared, cost-minimization analysisoften assumes that they lead to identical healthoutcomes. Studies of this nature should reportevidence to support the contention that out-come differences are non-existent or trivial innature. In most cases, however, the issues aremore than that of cost alone. It is rarely the casethat two therapies having the same indicationproduce identical health outcomes in everyrespect.

Cost–benefit analyses (CBA)

As applied to healthcare, cost–benefit analysismeasures all costs and benefits of competingtherapies in terms of monetary units. Generally,a ratio of the discounted value of benefits to


Table 15.3 Overview of the main types of pharmacoeconomic evaluations

Type of study Intervention Consequences Measurement of Compares Assumescosts consequences alternatives equivalent

effectiveness

Cost–benefit Monetary value Monetary value Economic Not necessarily, Noanalysis of resources of outcomes although(CBA) consumed comparisons

are implicit

Cost– Monetary value Effects on health • Lives saved Yes Noeffectiveness of resources • Cases treatedanalysis consumed • Years of life saved(CEA)

Cost–utility Monetary value 1. Utility of health 1. Quality-adjusted Yes Noanalysis of resources 3. effects 3. life-years (QALYs)(CUA) consumed 2. Indirect costs 2. Economic

3. Subsequent 3. Economic3. use of 3. resources

costs (the present value of both) is calculatedfor each competing therapy. The ratios for eachof the competing therapies and for competingprograms (e.g. intensive care unit versus newdiagnostic equipment) can be readily com-pared.

Cost–benefit analysis has the shortcoming ofrequiring the assignment of a dollar value tolife and to health improvements, includingquality-of-life variables. This presents equalbenefit issues as well as substantial measure-ment problems. For these reasons cost–benefitanalysis has not been widely used in recentyears for evaluating drug therapies.

Cost–effectiveness analyses (CEA)

Cost–effectiveness studies measure changes inthe costs of all relevant treatment alternatives,but measure the differences in outcomes insome natural unit such as actual lives saved,years of lives saved, or children immunized.Cost–effectiveness analysis can also be appliedequally to cases where the outcome is in termsof quality of life. Cost–effectiveness analysis isuseful in comparing alternative therapies thathave the same outcome units (e.g. years of lifeexpectancy or lives saved) but where the treat-ments do not have the same effectiveness (i.e.one drug may lead to greater life expectancy).The measure compared is the cost of therapydivided by the units of effectiveness, and hencea lower number signifies a more cost-effectiveoutcome.

The most challenging endpoint of acost–effectiveness analysis is the cost per life-yeargained. This requires the researcher to calculatethe survival benefit of comparative strategies.This is sometimes not an easy task, specificallywhen the medical economist does not haveaccess to the raw data. When Kaplan–Meiersurvival curves are available, the life-yearsgained are represented by the area between thetwo curves. Without such survival curves, lifeexpectancy has to be estimated by using lifetables,5 epidemiological formulas (declining

exponential approximation of life expectation,DEALE),6,7 or modeling techniques.8

This type of study has the advantage that itdoes not require the conversion of health out-comes to monetary units, and thereby avoidsequal benefit and other difficult issues of thevaluation of benefits. It has the disadvantage ofnot permitting comparison across programs,which have different endpoints. In other words,a drug whose function is aimed at reducinginfant mortality rates cannot be compared witha drug designed to improve functional status ofsenior citizens. Moreover, it cannot compareoutcomes measured in clinical units withquality-of-life measures.

Cost–utility analyses (CUA)

Cost–utility analysis compares the added costsof therapy with the number of quality-adjustedlife-years (QALY) gained. The quality adjust-ment weight is a utility value, which can bemeasured as part of clinical trials or indepen-dently. The advantage of cost–utility analysis isthat therapies that produce different or multipleresults can be compared.

The QALY, which has been the standardmeasure of benefit thus far, is arrived at in eachcase by adjusting the length of time affectedthrough the health outcome by the utility value(on a scale of 0 to 1) of the resulting healthstatus. Many analysts are more comfortablewith this measure of the consequence of med-ical care than with the use of money as the mea-sure of benefits.

With respect to drug therapy, cost–utilityanalysis is an improvement over cost–effective-ness analysis because it can measure the effectsof multiple outcomes (such as the impact of lab-oratory tests on both morbidity and mortalityor the impact on both pain and physical func-tional status).

In contrast to classical quality-of-life mea-sures, utility assessments focus on health statepreference valuation. Some programs may havethe ability to save lives; the years of life gained


across the entire population that could behelped are often used as a measure of success ofthe program. However, this measure has beenrefined to recognize that even though the bene-fits may be in years of life saved, the physicalstate of affairs of the affected group may be lessthan optimum during this extended time. Asnoted previously, weights are attached to theextended time, with a value of 1 given to nor-mal function and a value less than 1 given to astate of less than normal activity or discomfort.A value of 0 would represent death (at timeswhere life is said to be worse than death, avalue less than zero is the minimum). Theseweights are termed utilities, and they are usedas the adjusting factor to multiply the years oflife extended to obtain the quality-adjusted life-year. Cost per QALY can be computed andcompared across alternative treatment scenar-ios9 (Table 15.4).

Utilities can be assessed either through directmeasurement approaches (e.g. standard gambleor time-tradeoff techniques) or through multi-attribute instruments (e.g. the Quality ofWellbeing Index,10 the Rosser Index,11 or theHealth Utility Index12). The direct methodsoriginate from research in the field of gametheory, as early as the 1950s.13 (Table 15.5 listsutility values exemplified by the case ofmetastatic breast cancer.)

In addition, the field of quality-of-life studieshas been increasingly integrated into economicstudies. Preferably, these analyses require aprospective study design and should be incor-porated into the early clinical developmentprocess. In cases where this is not feasible,sound retrospective studies may be performed,employing analytical tools such as meta-analysis, modeling, and decision analysis. Theconcept of costs plays an important role in theconduct of such studies, and exerts a greatimpact on the corresponding results. Theapproach used by health economists is to con-sider costs as opportunity costs; i.e. they definea cost to be the consumption of a resource thatcould otherwise be used for another purpose.Once the resource has been used, the opportun-

ity to use it for another purpose is lost. Thevalue of that resource is that of the next bestuse. In pharmacoeconomic studies, costs as wellas benefits are classified in three classes: direct,indirect, and intangible costs. It is important todetermine all costs and consequences withinthe relevant timeframe and to avoid omittingcosts that may not be readily available. Beforeconducting an economic analysis, the perspec-tive or point of view of the study must bedefined. The view may be on the level of soci-ety, the patient, payer, or provider.

OUTCOMES RESEARCH

Against this background, what is now calledoutcomes research or medical effectivenessresearch has a tremendously important role toplay.14,15 Several different activities fall under thegeneral rubric of outcomes research. The first isclarification of the clinical, functional, and eco-nomic impacts of individual technologies andpractice alternatives. Outcomes research alsoinvolves evaluating patients’ preferences andcombining these in conjunction with the results oftechnology assessments to define what might beconsidered optimal practice. Once optimal prac-tice has been defined, an additional type of out-comes research seeks to compare actual practicewith the standards of practice developed fromthe preceding evaluations. When actual and pre-ferred practices differ, outcomes researchersneed to evaluate why they differ, and thendevelop interventions that will shift actual prac-tice into closer compliance with what is thoughtto be optimal practice. Outcomes research willultimately provide the information that isneeded to make product pricing decisions,coverage decisions, payment decisions, acquisi-tion decisions, and usage decisions in a cost-conscious and value-driven environment. It willbe critical to providers’ and manufacturers’ sur-vival in an increasingly competitive environ-ment. Given this perspective, researchers haveclearly been performing outcomes and effective-ness research for a long time.



Table 15.4 League table for various medical interventions in oncology

Intervention Costs/life-year Ref

Routine carcinoembryonic antigen monitoring of colon cancer 031 000–6 600 000 32

Allogenous bone marrow transplantation for 421 000 33relapse of Hodgkin’s disease

Bone marrow transplantation and high (versus 129 179 34standard) chemotherapy for breast cancer

Allogeneic bone marrow transplantation in 115 800 34metastatic breast cancer

Chemotherapy of acute non-lymphocytic 080 300 35leukemia

Adjuvant CMF (cyclophosphamide, methotrexate, 044 000 365-fluorouracil) in 75-year-old women withbreast cancer

Postsurgical chemotherapy for 60-year-old women 022 105 36with breast cancer

Postsurgical chemotherapy for 018 107 36premenopausal women with breast cancer

Interferon-�2b in hairy cell leukemia 013 800 37

Paclitaxel as first-line chemotherapy of ovarian 006 400–11 400 38cancer (six European countries)

Adjuvant CMF in 45-year-old women 004 900 36with breast cancer

Tamoxifen in advanced breast cancer 004 810 39

What is it then that is new about outcomesand effectiveness research? First, researchersare now focusing on a broader spectrum of out-come measures than has historically been thecase. Newer measures include functional status,health-related quality of life, and costs. Second,they are now interested in measuring outcomesin everyday practice, not just in the settings inwhich randomized trials are typically con-ducted. Third, they are using new types of datasets, such as insurance claims and hospital dis-charge abstracts, to perform outcomes research.The use of such datasheets, as well as the use ofnon-randomized trial methodologies in out-comes research, is controversial. Researchers,however, simply are not going to be able to doa randomized trial on everything, and in manycircumstances it would not be cost-effective todo so. Finally, increased attention is being paidto the patients’ viewpoints in the outcomesresearch that is currently being performed.

HEALTH ECONOMIC ASPECTS OF FEBRILENEUTROPENIA

During the past decade, several excellent eco-nomic evaluations have been performed in the


Table 15.5 Utility values exemplified by thecase of metastatic breast cancer38

Health state Utility

Partial response 0.81Stable disease 0.62Before commencement of 2nd-line

therapy 0.59Partial response plus severe

neuropathy 0.53Progressive disease 0.41Sepsis 0.20Terminal disease 0.16

field of febrile neutropenia, with great empha-sis on antimicrobial therapy and colony growthfactors. These studies have clarified the costsand consequences of drug interventions and, inseveral instances, have significantly impactedformulary decisions. The conduct of economicanalyses in febrile neutropenia is quite easy,because the benefits of therapy can usually beseen within a relatively short time frame. Thus,future costs and consequences do not have tobe discounted to their present value.

What are the costs of treating patients withfebrile neutropenia?

There have been only few studies to dateassessing the costs of febrile neutropenia.Faulds et al16 noted that the costs of treatmentvaried widely among institutions. Chaplin17

summarized the financial implications of febrileneutropenia management, and concluded thatdetailed cost studies were urgently required.

In a retrospective study, Leese et al18 collecteddata over a 1-year period from the medicalrecords of patients admitted to a district generalhospital – either with febrile neutropenia or whodeveloped this complication whilst receivinginpatient chemotherapy. Costs were calculatedfor inpatient stay, drug treatment, and diagnos-tic tests. The median total costs for 46 episodesof febrile neutropenia were £2068.35 and themedian total cost per day was £139.41. Inpatientbed-days accounted for 57.8% of total costs, fol-lowed by drug treatment at 25.8% and diagnos-tic tests at 16.4%. The costs of blood productswere excluded, since they are frequently admin-istered irrespective of the neutropenia.

Which treatment setting is more cost-effective and preferred by patients andfamilies?

Mullen et al19 measured resource allocation inoutpatient management of fever and neutrope-nia in low-risk pediatric patients with cancer

and its impact on their families in a prospectiveclinical trial. Eligible patients received a singledose of intravenous antibiotics, and wereobserved for several hours in clinic. Patientswere randomly assigned to continue eitherintravenous or oral antibiotics, and were seendaily as outpatients. Charges were calculatedbased on the number of resources used andMedicare/Medicaid reimbursement schedules.A questionnaire was used to measure theimpact of outpatient treatment on the family. Atotal of 73 episodes of fever and neutropeniawere studied. The median duration of treat-ment was 4 days. Of the episodes, 86% weremanaged without hospitalization. The mediancalculated charge was $1840. The median calcu-lated charge for patients receiving oral antibi-otics was $1544, and was significantly less thanthe $2039 median charge for outpatients treatedwith intravenous antibiotics. The estimatedcharge for comparable inpatient treatment was$4503. Nearly all families preferred outpatientcare, and few reported a loss of work hours orincreased childcare expenses. The investigatorsconcluded that outpatient treatment of low-riskepisodes of fever and neutropenia is substan-tially less costly than inpatient care, and is pre-ferred by most families.

What has been learned by the expanded useof costly colony-stimulating factors from aneconomic point of view?

The colony-stimulating factors have been usedeffectively in a variety of clinical settings to pre-vent febrile neutropenia and to assist patientsreceiving dose-intensive chemotherapy with orwithout stem cell support. Several studies haveconfirmed the clinical efficacy of the colony-stimulating factors used prophylactically inboth solid tumors and hematologic malig-nancies. The cost of these agents, along withtheir large-scale clinical use, has prompted anumber of economic investigations. Economicanalyses based on measures of resource utiliza-tion derived from randomized clinical trials

have provided febrile neutropenia risk thresh-old estimates for the cost-saving use of prophy-lactic colony-stimulating factors. A number ofimportant studies concerning the clinical andeconomic impact of these agents have beenreported recently. These include a revised cost-minimization study based on improved febrileneutropenia cost information and a cost–effec-tiveness analysis in the adjuvant breast cancersetting based on a clinical prediction model toselect patients at high risk for neutropenic com-plications. Continuing clinical and economicevaluation along with updating of clinical prac-tice guidelines is needed owing to the rapidtechnological and clinical advances in this area.

Which type of antibacterial regimen is morecost-effective in the empiric management offebrile neutropenia?

There is evidence to suggest that single-agentbroad-spectrum antibacterials may be cost-effective alternatives to combination antibioticsfor the empiric management of febrile neu-tropenia in cancer patients. Dranitsaris et al20

compared the clinical effectiveness of cef-tazidime monotherapy with that of two combi-nation antibiotic regimens in cancer patientswith febrile neutropenia. The two comparatorregimens consisted of tobramycin pluspiperacillin, either with (CAP regimen) or with-out (AP regimen) cefazolin. They also per-formed a cost–effectiveness analysis of the threeregimens. Meta-analysis of randomized com-parative trials between the three therapygroups was performed to determine the aver-age overall response rate after 3–5 days of treat-ment. Seven clinical studies were selected foranalysis. The overall incidence of adverse drugreactions (ADRs) was determined using theresults of comparative and non-comparativestudies. A comparative cost-analytic model wasapplied from a hospital perspective. The costsof primary therapy, hospitalization, laboratorytests, routine patient care, and treating ADRswere calculated, as were future costs.


Monotherapy with ceftazidime was associatedwith an overall response rate of 63.5% andmean per-patient costs of C$12 000 to C$14 000.In comparison, the AP regimen was associatedwith an overall response rate of 58.8% andmean costs of C$13 000 to C$16 000 per patient.The overall response rate in patients receivingthe CAP regimen was 75.3%, and the mean costper patient was C$11 000 to C$12 000. Thus, theCAP regimen was the most cost-effective ther-apy from a hospital perspective.

What is the economic benefit of a step-downregimen in high-risk neutropenic patients?

In a recent study, Marra and co-workers21 deter-mined treatment outcomes and the economicimpact of a ciprofloxacin step-down programfor high-risk febrile neutropenic adults from thehospital’s perspective. In an unblinded, two-phase, single-center study, adult leukemia andstem cell transplant (high-risk) adults withfebrile neutropenia were studied. Twoapproaches were analyzed: a multidisciplinaryciprofloxacin step-down program involving areduction in parenteral ciprofloxacin dose from400 to 200 mg and conversion to oralciprofloxacin when criteria were met; and anapproach without the parenteral step-downprogram. In total, 46 sequential treatmentcourses were compared with a 42-treatmentcourse from 6-month periods in pre-intervention (P1) and post-intervention (P2)phases. Assessed parameters were clinical andmicrobiologic outcomes, ADRs, and directmedical resource use and costs (1998 Canadiandollars) for the episode of febrile neutropenia.A decision-analytic model was used to mapprobabilities and costs and to conduct sensitiv-ity analyses. To supplement standard statisticaltesting, 1000 bootstrap samples were created,and the mean cost difference was calculatedbetween phases for each sample. Patient demo-graphics, percentage intravenous-to-oral step-down, and duration of therapy were similarbetween phases. Clinical success (83% P1, 81%

P2), microbiologic eradication (15% P1, 24%P2), and possible ADRs (6% P1, 9% P2) did notdiffer. Intravenous-to-intravenous dose step-down occurred in 33% of P2 and no P1 treat-ment courses (p � 0.001). Resource use andcosts were similar between phases, although areduction was seen in the drug’s mean totalcost/day (C$58 P1, C$52 P2, p � 0.04). Therewas also a trend toward a decrease in meantotal treatment costs (C$4843 P1, C$3493 P2,p � 0.08). Of 1000 bootstrap samples, 99.8%showed a cost advantage for P2. The model wasrobust to sensitivity analyses. Finally, theauthors concluded that this intervention influ-enced the administration of ciprofloxacin with-out an apparent compromise of patientoutcomes and resulted in a reduction in totalcosts of treating febrile neutropenia.

USING ECONOMIC ANALYSES IN DECISIONMAKING

Just conducting pharmacoeconomic research isoften not enough. What has to be done is toincrease the impact of such evaluations. A lot ofeconomic data has already been compiled, butis not being used properly. So the future liesalso in using results and increasing the impactof those evaluations.22 One means, for example,is to involve decision makers in the planning ofsuch studies. In the past, manufacturers haveproduced data and tried to convince decisionmakers, instead of working together withdecision makers beforehand. It should alwaysbe noted that economic analysis and economicdata represent only part of the informationrequired for the decision process. The next stepis to make decision makers aware of the useful-ness of an economic evaluation. Whoever thedecision maker is, there will be no use, if theydon’t feel that they can make a better decisionbased on this economic evidence. It is extremelyimportant to have the data present before thedecision is being taken.

The next point is to make the study known inthe community by all means of publication and


communication, preferably through the chan-nels that reach decision makers. This meansthat methodologies have to be reviewed interms of their credibility, financial costs, andtime completion. Classifications such as thoseused in the Cochrane Collaboration may also besuitable for analyzing and rating the sources foreconomic evaluations.23 Table 15.6 lists a rankorder of methodologies for assessing credibil-ity. In those cases where economic studies can-not be combined with randomized controlledstudies, modeling techniques have to beemployed.24 However, also here good modelingpractice should be envisaged.25 Checklists havebeen developed in order to facilitate theappraisal of the quality of economic analysesand assist in minimizing possible bias.26,27 Thesecriteria are also being increasingly used in thepeer-review process by many biomedical jour-nals28 and discussed accordingly.29

In many countries, such as Canada andAustralia, economic appraisal is a prerequisitefor acceptance of a new pharmaceutical productto be considered reimbursable.30,31


Table 15.6 Rank order of methodologies interms of credibility, financial cost and time tocompletion

• Economic analysis integrated into arandomized controlled clinical trial

• Economic analysis integrated intocase–control or cohort observational study

• Model based on published randomizedclinical trial

• Model based on published observationalstudy

• Model based on expert opinion• Model with unclear or incomplete source of

data

FUTURE OUTLOOK

Medical economics will become one of the mostsignificant strategic success factors for health-care providers in an era of cost containment.The challenge will be not only to meet therequirements of government agencies and pay-ers who are increasingly asking for economicassessments of commercial products, but also toaddress the value of medical economics to clini-cians. In the future, it will certainly be neces-sary for clinicians to apply the tools ofeconomic analyses both in research and in prac-tice. Instead of waiting for policy analysts,third-party payers, or governmental agencies tohand down decisions about which services aredeemed worth the cost, physicians could even-tually become practicing clinical economists.Another approach is to explore ways in whichclinical decisions are influenced by as well asinfluencing the cost of care. Clinicians need tointegrate economic thinking into their decisionmaking if medical care is to be rational butnot rationed. Pharmaceutical and device-manufacturing companies can contribute signif-icantly to this process by expanding economicresearch on their products, by providing train-ing and know-how to medical professionals,and by encouraging customers to acknowledgethe validity of such research.

REFERENCES

1. Oster G, Epstein AM, Cost–effectiveness of anti-hyperlipemic therapy in the prevention of coro-nary heart disease. JAMA 1987; 258: 2381–7.

2. Luce BR, Elixhauser A, Standards for theSocioeconomic Evaluation of Health Care Services.Berlin: Springer-Verlag, 1990.

3. Drummond MF, O’Brien B, Stoddart GL,Torrance GW, Methods for the Evaluation of HealthCare Programmes, 2nd edn. Oxford: OxfordUniversity Press, 1997.

4. Siegel JE, Weinstein MC, Russell LB, Gold MR,for the Panel on Cost–Effectiveness in Healthand Medicine, Recommendations on reporting

cost–effectiveness analyses. JAMA 1996; 276:1339–41.

5. Metropolitan Life Insurance Company, BloodPressure: Insurance Experience and its Implications.New York: Metropolitan Life InsuranceCompany, 1961.

6. Beck, JR, Kassirer JP, Pauker SG, A convenientapproximation of life expectancy – the DEALE. I.Validation of the method. Am J Med 1982; 73:883–8

7. Beck JR, Kassirer JP, Pauker SG, A convenienceapproximation of life expectancy – theDEALE. II. Use in medical decision-making. AmJ Med 1982; 73: 889–97.

8. Sonnenberg FA, Beck JR, Markov models inmedical decision making: a practical guide. MedDecis Making 1993; 13: 322–38.

9. Drummond M, Cost–effectiveness league tables:more harm than good. Soc Sci Med 1993; 37:33–40.

10. Kaplan R, Bush J, Health-related quality of lifemeasurement for evaluation research and policyanalysis. Health Psychol 1982; 1: 61–80.

11. Rosser R, Kind P, A scale of valuations of statesof illness: Is there a social consensus? Int JEpidemiol 1978; 7: 347–58.

12. Feeny D, Furlong W, Boyle M, Torrance GW,Multi-attribute health classification systems:Health Utilities Index. PharmacoEconomics 1995;7: 490–502.

13. von Neumann J, Morgenstern O, Theory of Gamesand Economic Behavior, 3rd edn. New York:Wiley, 1953.

14. Epstein AM, The outcomes movement: Will itget us where we want to go? N Engl J Med 1990;323: 266–70.

15. Wennberg JE, Outcomes research, cost contain-ment, and the fear of health cure rationing. NEngl J Med 1990; 323: 1202–4.

16. Faulds D, Lewis N, Milne RJ, Recombinant gran-ulocyte colony-stimulating factor: pharmacoeco-nomic considerations in chemotherapy-inducedneutropenia. PharmacoEconomics 1992; 1: 231–49.

17. Chaplin S, Cancer therapy complication costly totreat. Hosp Doctor 1991; May 9: 308.

18. Leese B, Collin R, Clark DJ, The cost of treatingfebrile neutropenia in patients with malignantblood disorders. PharmacoEconomics 1994; 6:233–9.

19. Mullen CA, Petropoulos D, Roberts WM et al,Economic and resource utilization analysis of

outpatient management of fever and neutrope-nia in low-risk pediatric patients with cancer. JPediatr Hematol Oncol 1999; 21: 212–18.

20. Dranitsaris G, Tran TM, McGeer A, Narine L,Pharmacoeconomic analysis of empirical therapywith ceftazidime alone or combination antibi-otics for febrile neutropenia in cancer patients.PharmacoEconomics 1995; 7: 49–62.

21. Marra CA, Frighetto L, Quaia CB et al, A newciprofloxocin stepdown program in the treat-ment of high-risk febrile neutropenia: a clinicaland economic analysis. Pharmacotherapy 2000; 20:731–40.

22. Coyle D, Increasing the impact of economic eval-uations on health-care decision making.Discussion Paper 108, York University, York,1993.

23. Chalmers I, Haynes B, Reporting, updating, andcorrecting systematic reviews of the effects ofhealth care. BMJ 1994; 309: 862–5.

24. Hillner BE, role of decision analysis in relation toclinical trials and a US perspective of the Battellemodel. PharmacoEconomics 1996; 9(Suppl 2): 30–6.

25. Buxton MJ, Drummond MF, Hout BA et al,Modelling in economic evaluation. A first andlast resort? Health Econ 1997; 6: 217–27.

26. Drummond MF, Stoddard GL, Torrance GW,Methods for the Economic Evaluation of Health CareProgrammes. Oxford: Oxford University Press,1987.

27. Task Force on Principles for Economic Analysisof Health Care Principles, Economic analysis ofhealth care technology. A report on principles.Ann Intern Med 1995; 123: 61–70.

28. Drummond MF, Jefferson TO, Guidelines forauthors and peer reviewers of economic submis-sions to the BMJ. BMJ 1996; 313: 275–83.

29. Kassirer JP, Angell M, The journal’s policy oncost effectiveness analysis. N Engl J Med 1994;331: 669–70.

30. Commonwealth Department of Health, Housingand Community Services, Guidelines for thePharmaceutical Industry on the Preparation ofSubmissions to the Pharmaceutical Benefits AdvisoryCommittee. Canberra: Australian GovernmentPublishing Service, 1992.

31. Canadian Coordinating Office of HealthTechnology Assessment (CCOHTA), Guidelinesfor the Economic Evaluation of Pharmaceuticals.Ottawa: CCOHTA, 1994 (www.ccohta.com).

32. Kievit J, van de Velde CJH, Utility and cost of


carcinoembryonic antigen monitoring in coloncancer follow-up evaluation. A Markov analysis.Cancer 1990; 65: 2580–7.

33. Desch CE, Lasala MR, Smith TJ et al, The optimaltiming of autologous bone marrow transplanta-tion in Hodgkin’s disease patients following achemotherapy relapse. J Clin Oncol 1992; 10:200–9.

34. Hillner BE, Smith TJ, Desche CE, Efficacy andcost–effectiveness of autologous bone marrowtransplantation in metastatic breast cancer.JAMA 1992; 267: 2055–61.

35. Welch HG, Larson EB, Cost–effectiveness ofbone marrow transplantation in acute nonlym-phocytic leukemia. N Engl J Med 1989; 321:807–12.

36. Hillner BE, Smith TJ, Efficacy and cost–effective-ness adjuvant chemotherapy in women withnode-negative breast cancer. N Engl J Med 1991;324: 160–8.

37. Ozer H, Golomb HM, Zimmerman H et al,Cost–benefit analysis of interferon alpha 2b intreatment of hairy cell leukemia. J Natl CancerInst 1989; 81: 594–602.

38. Berger K, Szucs T, Cost–effectiveness analysis ofpaclitaxel and cisplatin versus cyclophos-phamide and cisplatin as first-line therapy inadvanced ovarian cancer. A European perspec-tive. Eur J Cancer 1998; 34: 1894–901.

39. Rees GJG, Cost–effectiveness in oncology. Lancet1985; ii: 1405–8.


16Fever in the neutropenic patient: Past lessonsand future prospectsPhilip A Pizzo

The association of fever, neutropenia, and riskof infection began with the seminal observa-tions of the individuals to whom this book isdedicated. It was Gerry Bodey who reported inthe mid-1960s that a drop in the neutrophilcount in cancer patients, especially when pro-found and protracted, is associated with aheightened risk for infection. When these infec-tions go unrecognized or are not treatedpromptly, morbidity and mortality are signific-ant. The first and perhaps most enduringimpact on the infection-related mortality associ-ated with fever and neutropenia emanatedfrom the studies of Jean Klastersky in Europeand Stephen Schimpff in the USA, whodemonstrated the role of early empiric anti-biotic therapy in the management of fever andneutropenia.

It is now more than three decades that haveunfolded since these initial discoveries, andmuch has changed in the diagnosis, manage-ment, prevention and outcome of the infectiouscomplications that occur in conjunction withfever and neutropenia. In this book, a numberof distinguished experts have reviewed andcodified the many advances and progress thathave been accomplished, and the challengesthat remain.

PAST LESSONS AND THE PROGRESSOBSERVED DURING THE PAST THREEDECADES

Although the approach to the diagnosis andmanagement of fever in the neutropenic patientemployed today is thematically consonant withprinciples generated more than a quarter of acentury ago, there have been considerablechanges as well. These include the patients atrisk for infection, the types of anticancer ther-apy they receive, the nature of the infectingorganism and the patterns of infectionsobserved, the antimicrobial armamentarium,the increasing availability of biological thera-pies, and evolving concepts regarding preven-tive strategies. Many of these changes havebeen discussed in detail in this book. In thischapter, I shall summarize some of these devel-opments and offer some perspective on howthey inform current management and futureresearch.

Changes in the patients at risk

When the principles surrounding empiricantimicrobial therapy for neutropenic patientswho became febrile were initially elucidated,they largely applied to individuals receiving

chemotherapy for acute leukemias, especiallyadults with acute myeloid leukemia. Thesepatients generally suffered the greatest periodsof neutropenia and were among the most inten-sively treated with chemotherapy.

Since the late 1960s and early 1970s, moreintensive chemotherapy regimens have beenadministered to adults and children with varioussolid tumors as well as those with leukemias andlymphomas. Although the duration of bone mar-row suppression, and hence the risk of infection,is greatest in individuals with underlying bonemarrow disease or those who have receivedmarrow-ablative regimens, such a risk is nowclearly extended to other patient populations.These include both adults and children withsolid tumors who are receiving chemotherapy(especially dose-intensive regimens with orwithout stem cell reconstitution), patients withprimary or secondary bone marrow failurestates, and individuals who may be treated withcytotoxic therapies for non-malignant processes.

Accordingly, with the exception of patientswho have transient neutropenias associatedwith antecedent viral infections, it can be gener-ally assumed that broad-spectrum antibiotictherapy should be promptly and empiricallyadministered to every individual who hasdeveloped a fever while neutropenic. This isespecially true for individuals who havereceived prior cytotoxic chemotherapy.

It should also be noted that, even in theabsence of fever, neutropenic patients withlocalizing symptoms that are compatible with apossible site of infection (e.g. abdominal pain)should be treated similarly to neutropenicpatients with fever. Indeed, certain organisms(e.g. Clostridium septicum) can cause devastatinginfections in neutropenic patients in the absenceof fever.

Changes in cancer treatment and theirimpact on outcome

Combination chemotherapy began in the mid1960s and early 1970s with the treatment of

acute leukemias and lymphomas (especiallyHodgkin’s disease). Multimodal therapy is nowpart of most therapeutic regimens, and, inmany cases, dose-intensive therapies are acomponent of the therapeutic regimen. At thesame time, based on the underlying disease andthe chemotherapy regimen, it is possible to cat-egorize patients as ‘low-risk’ or ‘high-risk’. Asdetailed in Chapters 8 and 9, this risk stratifica-tion has important implications for patientmanagement, including the type and route ofinitial empiric antimicrobial drug delivery, theneed for subsequent modifications of therapy,and even whether patients are treated in or outof hospital.

It has also become clear that while the depthand duration of neutropenia are perhaps themost important determinants for the risk ofinfection, both the underlying disease andchemotherapy regimen that is administered canimpact other host defenses. Notably, the impacton cellular immunity, measured by sustainedage-related depressions in CD4� cell number,underscores the multidimensional impact ofcancer therapy on innate and acquired hostdefenses.

Change in the organisms causing infectionand the patterns of disease

As discussed in Chapters 1, 6 and 7, the need forprompt empiric broad-spectrum antibiotic ther-apy in the 1960s–1980s was underscored by thedominance of Gram-negative organisms causingsevere infections in cancer patients. Most notablewere infections due to Pseudomonas aeruginosa,Escherichia coli, and Klebsiella spp. Although theseorganisms can still be responsible for seriousinfections, there has been a notable decline inGram-negative infections in most centers treat-ing cancer patients in the 1990s through the pre-sent. The reasons for this decline are not fullyunderstood, but it is also clear that Gram-negative bacteria remain a significant problem atselected institutions and, in particular, in devel-oping nations.


At the same time that Gram-negative organ-isms declined, Gram-positive isolates increased,especially with coagulase-negative staphylo-cocci. Although this is largely attributed to theremarkably increased use of indwelling sialasticcatheters in cancer patients, it is important tonote that some of the trends with Gram-positiveorganisms occurred prior to frequent catheteruse. In addition to the coagulase-negativestaphylococci, S. aureus, streptococci (perhapsmost notably some of the viridans streptococci),and enterococci have also emerged as signific-ant pathogens.

As also detailed in Chapters 1, 5, 6 and 7, thechoice of empiric antibiotic therapy and out-come of the patient can be linked to the infect-ing organism. However, because empiricantibiotic therapy is initiated with the onset offever and includes broad-spectrum regimens, itis also notable that the ability to diagnose eithera site of infection or microbiological etiology isnow the exception rather than the rule. Indeed,unlike the patterns observed in the 1960s and1970s, when a clinical or microbiological site ofinfection was diagnosed in two-thirds ofpatients presenting with fever and neutropenia,this now occurs in less than a third of patients.Of course this does not mean that these patientsare uninfected, but rather that early therapy hassuppressed or muted the ability to define a clin-ical or microbiological site or cause of the fever.Clearly, this also makes management decisionsmore challenging, since the initial therapeuticregimen, and its modification and duration, arenow guided by persistent or recurrent feverrather than clinical or microbiological end-points.

Although bacteria account for the majority ofthe initial fevers in neutropenic cancer patients,it has also become clear during the last threedecades that other classes of organisms areresponsible for primary or secondary fevers orinfections. Among these are the herpesvirusesas well as the respiratory viruses, fungi (whichcan be influenced by geography), and variousparasites, either due to reactivation or newacquisition.

Given the wide range of potential offendingorganisms, it has also become increasingly clearthat neutropenic cancer patients can have mul-tiple infectious complications during a periodof neutropenia, especially when the length ofneutropenia exceeds 10 days. These ‘high-risk’patients deserve scrupulous attention and care-ful management, as has been reviewed inChapters 1, 6 and 7.

Changes in the therapeutic armamentariumand impact on supportive care

In addition to changes in the patients at riskand the organisms responsible for infection, oneof the most remarkable changes during the lastthree decades has been in the antimicrobialagents available for supportive management.Although the limited spectrum of antibioticcoverage mandated that combination therapybe administered during the 1960s and 1970s,this began changing in the 1980s with the avail-ability of third-generation cephalosporins, andthen subsequently, the carbapenems and fluo-roquinolones. As discussed in Chapters 5 and 9,these newer agents have permitted newapproaches to empiric antibiotic management,including monotherapy and, in ‘low-risk’patients, the prospect of oral regimens and out-patient management.

Of course, advances in antibiotic therapyhave been coupled with the emergence of resis-tant microorganisms, making antibiotic utiliza-tion an area that requires considerable scrutinyand regulation. Inappropriate or unnecessaryantibiotic use, especially with drugs such asvancomycin, aminoglycosides, and fluoro-quinolones, has been associated with the emer-gence of drug resistance in a number of centersaround the world. This has further underscoredthat empiric therapy is not an excuse for unreg-ulated or indiscriminate drug regimens.Moreover, the routine use of antibiotics that areimportant parts of the therapeutic armamentar-ium (e.g. fluoroquinolones) should not beemployed as prophylaxis.

PAST LESSONS AND FUTURE PROSPECTS 319

In addition to advances in antibiotics,progress has also been made in newer antiviraland antifungal agents, driven in part because ofthe role that some of these organisms play inpatients with AIDS. Although there has beenprogress, there is still much work to be done inthis area, since there remain serious limitationsin the antiviral and antifungal therapeuticarmamentarium.

Over the past two decades, attention has alsoturned to the use of biologicals as either thera-peutic adjuncts or as a means to bolster orrestore the altered host defenses in cancerpatients. The initial forays into this area ofresearch addressed the role of leukocyte trans-fusions, passive and active immunization, andthe use of interferons. Although some of theseapproaches were grounded in logic, technicallimitations most often precluded their success.

Beginning in the late 1980s and extending tothe present, considerable attention has beengiven to the hematopoietic cytokines (granulo-cyte and granulocyte–macrophage colony-stimulating factors: G-CSF and GM-CSF) toshorten the duration of neutropenia in cancerpatients. As discussed in Chapter 13, a trulyevidence-based evaluation of these cytokineslimits their utilization to relatively specific clini-cal indications, focusing particularly on higherrisk patients.

In summary, it is increasingly clear that thesupportive management of fever in neutropenicpatient is closely linked to their risk status.Low-risk patients with short durations of neu-tropenia (i.e. <10 days) can be treated with sim-pler antimicrobial therapies, either parenterallyor orally, and with relatively little need foradditional antimicrobial additions or modifica-tions of the initial regimen. There is little needfor hematopoietic cytokines in these low-riskpatients. Moreover, therapy for many of thesepatients can be done in an ambulatory setting,potentially at home.

In contrast, high-risk patients, categorized byprolonged durations of neutropenia (>10 days),require inpatient management, are at risk formultiple secondary infections that require addi-

tions or modifications of their initial regimen,and may benefit biological response modifiersthat improve immune or hematologic recovery.

Changes in clinical trial design and relatedexpectations

Improving the management of fever and infec-tion in neutropenic cancer patients is closelydependent on evidence-based data emergingfrom appropriately conducted clinical studies.Because of the progress that has been made inreducing infection-related mortality, and sincea clinical or microbiological cause of the initialfever is frequently lacking, the conduct of clini-cal trials is methodologically challenging. Tooptimize the design and analysis of clinicaltrials in febrile neutropenic cancer patients, anumber of international societies including theInternational Immunocompromised HostSociety (ICHS), the Infectious Disease Society ofAmerica (IDSA), and the MultinationalAssociation for Supportive Care in Cancer(MASCC) have attempted to establish guide-lines that optimize clinical trial design and theability to compare the results of studies done indifferent settings and patient populations. Theissues and governing concerns have also beencovered in Chapters 2 and 14.

CURRENT PROBLEMS AND CHALLENGESFOR THE FUTURE

The past three decades have witnessed majorprogress in the supportive management ofcancer patients who develop fever and neu-tropenia. Morbidity and mortality have beendramatically reduced, and therapies are simplerand less toxic and more appropriately delin-eated according the patient’s risk status, dis-ease, age, and clinical setting. Despite thisprogress, however, numerous challenges andopportunities remain to be addressed and prob-lems solved.

Almost certainly as a result of the early initi-


ation of empiric therapy when neutropenicpatients become febrile, the ability to diagnosewhether a patient is truly infected, and, if so, atwhat site and with what organism(s), remains amajor challenge. Improved rapid non-culture-dependent diagnostic tests from accessiblebody sites or fluids are needed. These assaysneed to address bacteria, viruses, fungi, andselected parasites. Clearly, these measures needto be available to children as well as adults. Inaddition, improved imaging studies that couldhelp to localize occult sites of infection wouldbe a significant advance.

Improved risk stratification of patients is alsoimportant. This can be based on improved clini-cal measures of risk, but should also addressthe biological factors within the host that con-tribute to the risk of infection or their specificexpressions. For example, genetic predisposi-tion to colonization or infection could addressboth the host and the pathogen.

Refined assessment of host defense factorsare also important in determining whichpatients are at risk for specific infections. Thiscould include rapid determination of innate oracquired host immune factors that modulatethe risk for infection during neutropenia.Pharmacogenetic assessment may also help

determine which patients are more susceptibleto toxic effects from anticancer therapies.

Development of new antimicrobial agentsthat overcome resistance or that provide moreeffective therapy, especially for viruses andfungi, is also critically needed. With improvedantibacterial agents, opportunistic fungi and anumber of herpesviruses and respiratoryviruses have emerged as major pathogens.Clearly, an expanded and enhanced therapeuticarmamentarium is important. Improving hostdefenses with cellular or humoral biologicaltherapies is another goal.

Although treating or preventing infectionswith antimicrobial agents and/or biologicals isimportant when patients receive cytotoxicimmunosuppressive therapies, the greatestimpact will come with more selective and spe-cific anticancer therapies that do not damagethe host defense matrix. While this was a dreamjust years ago, the rapid progress in elucidatingthe molecular pathogenesis of cancer is leadingto the development of new agents that impactcancer cells without resulting in side-effectssuch as neutropenia. When such therapiesbecome available, the problems and challengesassociated with fever and neutropenia may wellbe overcome.

PAST LESSONS AND FUTURE PROSPECTS 321

Index

Key: autoPBSCT, autologous peripheral blood stem cell reinfusion; BSCT, blood stem cell transplan-tation; BMT, bone marrow transplantation; HSCT, hematopoietic stem cell transplantation; PBSC,peripheral blood stem cells; PBSCT, peripheral blood stem cell transplants.

abdominal infections 117Absidia spp. 219absolute monocyte counts 63absolute neutrophil counts 151Acinetobacter spp. 64, 95

in vascular access device colonization 191,196

acquired organisms 5–6, 112, 219Actinomyces 62acute abdomen 69acute disseminated candidiasis 73acute leukemia 58–9, 61, 168

and growth factors 259–62acute lymphoblastic leukemia 201, 259, 262acute myeloid/myelocytic leukemia 69, 115

and growth factors 252, 259–61, 285–6acute non-lymphocytic leukemia 2acyclovir 101, 133, 230–1, 232adaptive clinical trial designs 51adrenaline 246adverse events 175, 176–7Aeromonas hydrophila 112Agency for Policy Research Guideline

Clearinghouse 290airway obstruction 93allergic reactions 208alloBMT patients 74, 75allogeneic peripheral blood stem cells 60, 133alloimmunization 134, 268

alpha-hemolytic streptococci 155alpha-strep syndromes 120–1, 218Ambisome see liposomal amphotericin BAmerican Society of Clinical Oncology (ASCO)

on growth factorscomparative activity of 288in prophylaxis 253, 257use of 133, 209, 251–2, 264

on oncologist compliance 256amikacin 14, 15, 17aminoglycosides 12, 32amphotericin B 32, 70, 227–8, 291

candidemia 101, 196in children 206empiric therapy 139and granulocyte transfusions 267, 268in HSCT 131nephrotoxicity 142for persistent fever 19–20prophylaxis 132see also liposomal amphotericin B

anaerobes 112, 216, 217antibiotics active against 12–13in mouth infections 155

anaerobic infections 101analysis of clinical trials 48antacids 113, 216, 218anthracyclines 64antibiotic resistance see antimicrobial resistance

antibiotic-associated colitis 70antibiotic-resistant bacterial colonization 60antibiotic-resistant Enterobacteriaceae 66–7antibiotic-resistant organisms 94antibiotics

absorbtion 38against anaerobes 12–13broad-spectrum 12, 13–14, 136against fungal infections 5oral 171–3, 228, 295

and early discharge 176hospital-based 176–7low-risk patients 176–7outpatient 177–8

route 151, 312selection factors in HSCT 136therapy 6, 7–8, 50

appropriateness 11in children 205, 207–8duration 291, 292, 293, 296early initiation 8–10and growth factors 251oral 38, 291

anticoagulant flush solutions 194antifungal therapy

and aspergillosis 74guidelines 296

anti-infective therapy 27, 33–7antileukocyte antibodies 267antimicrobial agents

changes over time 319CNS penetration 39and indigenous flora 5protein binding 39–41

antimicrobial flush solutions 194antimicrobial hand rubs 129antimicrobial resistance 64, 65, 113

to �-lactams 67–8emergence of 319of fungi 228–9glycopeptides 99local patterns 115, 180multidrug 174, 181patterns 66, 128and peak concentrations 29and prophylaxis 224rates 67

and synergistic activity 8see also colonization resistance

antimicrobial therapydiscontinuation 122economic evaluations 311guidelines 289–96

antineoplastic therapy 49antipseudomonal PK:PD data 31antithrombolytic agents 194anti-virals 32aplastic anemia 60aplastic bone marrow 23area under the curve 29arthropathy 207ascending cholangitis 105ASCO see American Society of Clinical

Oncology aspergillosis 73–4, 228

and allogeneic transplantation 139–40in HSCT patients 141–2invasive 71management 229, 230recurrent 130, 131

Aspergillus spp. 131, 206, 219in HSCT 140infection in neutropenia 5prophylaxis against 228in sinuses/nasal passages 155A. flavus 74, 141A. fumigatus 74, 118, 219

in HSCT 141prophylaxis against 227

A. niger 74A. terreus 74

aspiration pneumonia 93aspiration sites 4atelectiasis 103autologous HSCT 130autologous peripheral blood stem cell

reinfusion (autoPBSCT) 59–60patients with varicella zoster 76

average costs 304axillae 3, 4azithromycin 28azoles 32aztreonam 115, 218

324 INDEX

Bacillus spp.glycopeptide resistance 99in HSCT 138in surgical wound infections 95in vascular access device colonization 191,

196bacteremia 62–8, 151, 171

in children 64, 207complex 170–1febrile episodes and 6, 7Gram-negative 7, 9, 11, 66, 67

and neutrophil count 10response to therapy 14, 15

and granulocyte transfusions 269and neutrophil count 2patients with solid tumors 98and perirectal infections 156prognosis 63–4, 169–71risk factors 63of unknown origin 4, 5

bacterial infectionsbloodstream, duration of 137diarrhea 135Gram-negative 155, 157

bacterial translocation 5, 60, 218bacteriuria 69Bacteroides spp. 217

B. fragilis 5Baltimore Cancer Research Center 7barriers, anatomic, disruption of 93Bayesian trial designs 47, 51B-cell recovery 126�-lactamase 67� -lactams 12, 32

antimicrobial resistance 67–8broad-spectrum 57

bias 46biliary tract obstruction 93binding 224biofilms 190, 191biological agents 320biopsies 135, 142, 157, 204Bipolaris 140blinding of trials 48, 51blood culture methods, simultaneous 192blood cultures 116, 120, 135

in children 203

blood culture studies 61blood stem cell transplantation (BSCT) 59–60bloodstream infections 137, 138, 139bone disorders 94, 126, 245bone marrow function 126bone marrow involvement in acute leukemia

58bone marrow recovery 167bone marrow stroma 246bone marrow suppression 208bone marrow toxicity 137bone marrow transplantation (BMT) 38

allogeneic 2and growth factors 251infectious complications 59–60recipients 74

and fungal infections 70and viral infection 76, 102

see also hematopoietic stem celltransplantation

bone pain 265bowel flora after cancer therapy 218brain involvement in invasive aspergillosis 74brain tumors in children 201breakpoints, drug activity 30breakthrough infections 38, 115

fungal 139breast cancer surgery 104, 105broad-spectrum antibiotics 12, 13–14, 136broad-spectrum �-lactams 57bronchoalveolar lavage 74, 142Burkholderia pickettii 66

cancer in children 201, 201–2cancer treatment, changes in 318cancer-mimicking infections 106Candida spp. 5, 71

C. albicans 140, 141, 218C. glabrata 72, 141, 196C. krusei 72, 141, 196C. lusitaniae 72C. parapsilosis 72, 141, 190, 191C. tropicalis 72, 140catheter-related infection 112colonization 131and CRBSIs 196prophylaxis against 227, 228

INDEX 325

Candida spp. – cont.sites of infection 155–6in surgical wound infections 95–6in vascular access device colonization 190

candidemia 190and granulocyte transfusions 269in patients with solid tumors 101

candidiasis 220, 227–8gastrointestinal 217hepatosplenic 156in HSCT patients 140–1invasive 218

capillary leak syndrome 265Capnocytophaga spp. 218, 220carbapenems 12caspofungin 142catheter culture techniques 192catheter-related infections 64, 69, 72, 93, 220

Candida 196, 227classification 137fungemia 72Gram-positive 121in hematologic patients 117in HSCT patients 137–9prophylaxis 223see also vascular access device infections

catheterscoated 195colonization 64damage to integument 3intravenous 112urinary catheters 93

cefazolin cost-effectiveness 312–13cefepime

for children 207empiric therapy 115

monotherapy 13, 204cefpirome 115ceftazidime 16, 290

for children 207cost–effectiveness 312–13early discharge strategy 182empiric therapy 115, 139

monotherapy 13, 204intravenous prophylaxis 131plus amikacin 14, 15and vancomycin-resistant enterococci 67

cellular immunity 318dysfunction 102, 126

cellulitis 94breast 104perianal/perirectal 70, 117

central venous catheters 189cephalosporins 12cerebrospinal fluid shunts 93chemoprophylaxis 220chemotherapy 60, 318

in acute leukemia 58in children 202impact on skin/mucosa 3, 93, 217–19intensive 38, 262–4for solid tumors 59

chemotherapy-induced neutropenia seeneutropenia

chemotherapy-induced vomiting 4chest examination 68

X-rays 116, 134, 152, 203childhood leukemias 59, 262children 201–9

antibiotics 205, 207–8with bacteremia 64, 207cancer in 201, 201–2with febrile neutropenia 203–7

assessment 203–4outcomes 206–7prolonged 204, 206therapy 204

with fever 203guidelines for treatment 295implantable ports 194leukemic 76, 117with pneumonia 68prophylaxis 222

Chlamydia pneumoniae 68chlorhexidine/silver sulfadiazine coating 1952-chlorodeoxyadenosine 59chronic disseminated candidiasis 73Chryseobacterium meningosepticum 66ciliary dysfunction 3, 4ciprofloxacin 207–8, 224, 313

oral 131, 172, 177circulating neutrophils 111, 246, 247Citrobacter spp. 64Cladosporium spp. 219

326 INDEX

clean-air ventilation 130clearance, renal 38

in children 208clinical guidelines 128–9clinical prediction rule 162clinical risk stratification 175clinical trials 48–53

analysis and reporting 48blinding 48, 51comparability of groups 49–50and cost–effectiveness studies 53designs 45, 46–7, 50–1, 320

alternative 48generality 49ineligibility 52methodology 296–8monitoring 47–8multiple entries and withdrawals 49, 52–3outcomes 50, 53risk-adjusted 171stratification 49–50validity 53

clinically documented infection 62–76, 137–9,151

Clostridium spp. 121, 217C. difficile 70, 98, 135, 156C. perfringens 112C. septicum 318

clotrimazole 228CNS drug concentrations 32, 39coagulase-negative staphylococci 20, 21, 105,

222, 319and bacteremia 64–5and CRBSIs 196in HSCT 128, 129and soft tissue infection 157

coagulase-producing organisms 190coated catheters 195Coccidiodomycetes 140collagen/silver ions 195collection of granulocytes 265–7, 270colonization 5, 112, 226–30

bowel 218fungal 60, 131, 140–1, 227, 229oral antibiotic therapy 228of vascular access devices 189–90

colonization resistance 5, 13, 217

colonizing organisms 11colony-stimulating factors see myeloid growth

factorscombination pharmacodynamics 28combination therapy

current recommendations 23, 290, 295for empiric regimens 41, 114patients with solid tumors 99and Ps. aeruginosa 101

commensal flora 219comorbid conditions 143, 202comparability of groups 49–50complicated secondary infections 57complications

CRBSIs 196of gynecological cancer 106with high-dose cytosine arabinoside 59of neutropenic enterocolitis 69and risk-adjusted therapy 181with streptococcal bacteremia 64with vascular access devices 189

computed tomography (CT) scans 116, 152for aspergillosis 229, 230chest 68in children 206of hepatobiliary tract 155, 156in HSCT 130, 134–5sinus/nasal passages 154

concentration–killing relationship 28concentration-dependent pharmacodynamics

28–9concentration-independent pharmacodynamics

29consequences 303–4corticosteroids 246Corynebacterium spp. 64, 95, 191, 196

C. jeikeium 128, 138C. parvum 112

coryneforms 216cost analysis principles 302cost of treatment 142

early discharge 176granulocyte transfusions 270growth factors 251, 262outpatient therapy 173tunneling 194

cost time profiles 305

INDEX 327

cost–benefit analyses 307–8of growth factors 252

cost–effectiveness 304herpes simplex treatment 231of impregnated catheters 195maximum sterile barrier 193

cost–effectiveness analyses 307, 308cost–effectiveness studies 53cost–utility analyses 307, 308–9cost-containment suggestions 293cost-minimization analyses 306–7, 312costs 137, 302–3

febrile neutropenia 311of medical interventions 310of methodologies 314site of care 311–12weighting 305see also under economic evaluation

costs per quality-adjusted life-year 309, 310co-trimoxazole 224

prophylaxis 222, 228see also trimethoprim–sulfoxamethazole

Cox proportional hazards regression 48creatinine clearance 38credibility of methodologies 314critical appraisal instrument 289Cryptococcus neoformans 118Cryptosporidium 156cutaneous lesions 98cyclodextran–itraconazle 132cyclosporin–itraconazole 132cytokine response 61, 249cytomegalovirus 75, 155, 232

in children 203in HSCT patients 141prophylaxis 230–2risk of transmission 134treatment efficacy 220

cytosine arabinoside 64

data-dependent randomized designs 51daunorubicin 59deaths

BSCT and autoPBSCT 60early 52from primary infection 57

decision analysis model 253

decision making 313–14decision tool 301decontamination of gastrointestinal tract 38,

112, 222defervescence, time to 62dental discoloration 207depreciation costs 305design of clinical trials 45, 46–7, 50–1, 320diagnosis

early 6in immunocompromised hosts 111

diagnosis-related groups 302diagnostic techniques 321

modification for children 204diarrhea 135, 218, 219differential time to positivity 192diffuse leptominingeal disease 104dimorphic fungi 140direct medical costs 302direct non-medical costs 303discounting/discount tables 305–6disseminated candidiasis 73, 130, 204

prophylaxis 226, 228distribution factors 32distribution of organisms 95–6DNAemia 61documented infection 62–76, 137–9, 151donor priming 134dosage adjustments 38dose-intensive therapies 318dosing, antimicrobial 27–32drug interactions 229drug penetration 38duotherapy 290–1, 294dyspepsia 218dysphagia 117

early deaths 52early discharge 176, 177, 182early-switch therapy 177economic analyses 313–14economic benefits of outpatient therapy

181economic evaluations

classifications 314components 302–3concepts 304–5

328 INDEX

economic hardship and non-compliance303

economic impact of medicines 252, 301economic outcomes 50ecthyma gangrenosum 65, 117elderly patients 259, 262elimination rate 32empiric therapy 7–8, 11, 41, 290, 295

antiviral 293bacteremia 169and causative organism 319in children 204, 206combination 12–13cost-effectiveness 312–13growth factors 293in hematologic patients 113–16high-risk patients 171and improved outcomes of FN 57initial 136–7lack of response 139modifications 118–22, 206and neutropenia duration 168outpatient 293and pathogen predominance 20patients with solid tumors 99and PK:PD ratios 31prompt 6, 290reassessment 139for risk-adjusted therapy 180

endoscopic evaluation 155, 156Enterobacter spp. 64, 128

E. cloacae 112Enterobacteriaceae 96

antibiotic-resistant 66–7enterococci 319

in HSCT 128, 129entry-site infections 137, 138EORTC see European Organisation for

Research and Treatment of Cancer epidemiology 58–60, 318–19epinephrine 246Epstein–Barr virus 75Escherichia coli 318

bacteremia 9, 11bacterial translocation 5in cellulitis 70fluoroquinone resistant 66

in HSCT 128synergy versus single agents 16, 18in urinary infections 157

Escherichia coli-derived GM-CSF 265esophagitis 4Eubacterium spp. 218European Group for Blood and Marrow

Transplantation 231European Organisation for Research and

Treatment of Cancer (EORTC) 5, 19, 98Antimicrobial Therapy Cooperative Group

176International Antimicrobial Therapy

Cooperative Group 58, 63, 64, 66, 67on treatment response 296, 297

trial 14on vancomycin 136

evaluation of neutropenic patients, standard290

evidence-based medicine 215examination, physical 96, 116, 152

of children 203exogenous microorganisms 219

FDA approval 264febrile episodes and bacteremia 6, 7febrile neutropenia 11–20, 290–1, 293

children with 203–7costs 311

Federation of French Cancer Centers 294–5fever

definition 151early onset in acute leukemia 58and growth factors 257incidence rates 57, 60persistent 121

treatment of 19–20, 292, 293fever of unknown origin 61

in children 207and neutropenia duration 168, 170patients with solid tumors 98, 99–100

filamentous fungi 68filtration leukapheresis damage 266–7

and amphotericin B 268first dose effect of GM-CSF 265fixed-sample-size randomized designs 50–1Fleming two-stage design 47

INDEX 329

fluconazole 229for candidemia 101, 196colonization suppression 228empiric therapy 206prophylaxis 132, 141, 226

flucytosine tissue concentrations 32fludarabine 59, 75fluid shifts 38fluoroquinolones 207, 222–3, 224

gut toxicity 218prophylaxis 66, 67, 131

and bacteremia 62and C. difficile colitis 70and streptococcal bacteremia 64

side-effects 224–5tissue concentrations 32

flush solutions 194follow-up monitoring 181–2Fourth Decennial International Conference on

Nosocomial and Healthcare-AssociatedInfections 173–4

Fred Hutchinson Cancer Research Center 70,128

aspergillosis trial 130–1use of growth factors 133–4

fungal colonization see under colonizationfungal infections 96

cancer-mimicking 106duration of 137evaluation/management 139–43in hematologic patients 117invasive 61, 70–4, 122, 139–40, 226

in children 204patients with solid tumors 101and persistent fever 19pneumonia 68rare/unusual organisms 72therapy 5, 226–8, 269–70

fungal spore counts 73–4fungal superinfections in acute leukemia 58fungemia 72, 72–3fungi 219Fusarium spp. 73, 96, 140

exogenous sources 219Fusobacterium necrophorum 218Fusobacterium spp. 62, 216

galactomannan detection 140, 142, 229ganciclovir 231, 232gastric acid barrier 218gastrointestinal damage 59gastrointestinal tract decontamination 38G-CSF see granulocyte colony-stimulating

factors Gehan two-stage design 47Gemella 62generality of trial results 49gentamicin 18Giardia 156glucose non-fermenting Gram-negative rods

65glycocalyx 190, 191glycopeptides

in empiric therapy 115, 121patients with solid tumors 99, 100risk factor for fungemia 72supplementation with 114

GM-CSF see granulocyte–macrophage colony-stimulating factors

Gordana aurantiaca 220graft-versus-host disease 38, 218

acute 135chronic 126, 141and cytomegalovirus 220and fluconazole 132and PBSC 133

Gram-negative bacilli/rods 5, 216, 217, 219and CRBSIs 196predominance 20prophylaxis 222

Gram-negative bacteremia 7, 66, 67and neutrophil count 10response to therapy 14, 15

Gram-negative bacteria 318and respiratory tract infection 157sinus/nasal passages 155and soft tissue infection 157therapy 100–1

Gram-positive bacteria 7, 319bacilli 196cocci 20, 128therapy 100

granulocyte colony-stimulating factors (G-CSF)23, 133, 142, 320

330 INDEX

biology of 247, 249–50FDA approval 251, 264formulations and uses 184, 208–9, 252and neutropenia 245as primary prophylaxis 253–4, 256production of 246–7

granulocyte increase with G-CSF 265–6granulocyte mobility impairment 267granulocyte transfusions 134, 245, 265–70, 320

dose and side-effects 267–8efficacy of treatment 268–70granulocyte dynamics 268product quality 267

granulocyte–macrophage colony-stimulatingfactors (GM-CSF) 23, 133, 142, 320

biology of 247–50FDA approval 251formulations and uses 184, 208–9, 252and neutropenia 245as primary prophylaxis 253, 255–6production of 246

granulocyte-specific antibodies 268granulocytopenia 16granuloma formation 73granulomatous tissue reaction 74granulopoiesis 122growth factors see myeloid growth factorsguidelines 320

antimicrobial therapy 289–96clinical practice 289–98risk-adjusted therapy 175on use of growth factors 209, 251–2

for acute lymphoblastic leukemia 286–7for acute myeloid leukemia 285–6with chemotherapy/irradiation 284,

287dosing and administration route 288initiation and duration 288for leukemia in relapse 287for myelodysplastic syndromes 286and progenitor cell transplantation

284–5prophylaxis 282, 283therapy 283–4

guidewires 194gut microflora 217gynecologic-cancer-related infections 105–6

Haemophilus influenzae 96, 222haemorrhages 111handwashing 129, 208hazards of hospitalization 174–6health economics 301–9

in decision making 313–14evaluation methods 306–9febrile neutropenia 311–13

health state preference value 308hematogenous seeding 190hematologic malignancies 106

and catheter-related infections 193hematopoiesis 246–7

after stem cell transplant 262hematopoietic colony-stimulating factors 122,

184, 246–7and neutropenia duration 112see also macrophage colony-stimulating

factors; myeloid growth factorshematopoietic stem cell transplantation (HSCT)

125allogeneic 231and neutropenia duration 168

hepatic abscesses 105hepatic glucuronidation 208hepatitis 102hepatobiliary infection 105heptosplenic candidiasis 73herpes simplex virus 75, 218

effect of prophylaxis 133in mouth infections 155prophylaxis 230–1, 232reactivation 75and soft tissue infection 157treatment efficacy 220

herpesvirus HHV-6 75, 76herpesvirus HHV-7 75, 76herpesvirus infections 7, 75high-dose chemotherapy 168high-dose cytosine arabinoside 7, 19, 59high-efficiency particulate air (HEPA) filters

113, 219, 228in HSCT 130

high-efficiency particulate air (HEPA) filtration74

high-risk patients 171, 181–2Histoplasma 140

INDEX 331

history, medical 96–8, 116, 152of children 203

HIV-related neutropenia 60HLA antibodies 268Hodgkin’s disease 59, 201

and growth factors 256–7home-administered IV therapy in hematologic

patients 118hospital-acquired organisms 6

resistance, multidrug 174, 181see also nosocomial pathogens

hospital-based therapy 171oral 176–7

hospitalizationearly discharge 176and risk 175, 176, 181

host defence barriers 5, 321impairment 215–19

host defense deficits 59, 60host factors 49, 74host response 73hub contamination 190hyalohyphomycosis 74

invasive 71, 72hypokalemia 114

idarubicin 59identification of causative pathogens 137IDSA see Infectious Diseases Society of

America IDSA clinical care guidelines 135imaging procedures

in children 204, 206of HSCT patients 134

imipenemearly discharge strategy 182empiric therapy 139

monotherapy 13, 204serum bactericidal activity 16–17

imipenem–cilastatin 115immunization history 97Immunocompromised Host Society

on clinical trials methodology 296, 298on treatment response 296, 297

immunocompromised patients 208immunodeficiency and fungal infections

72

immunoglobulinand cytomegalovirus 231for respiratory viruses 143

immunosuppressive therapy 60implantable ports 189impregnation of catheters 195incidence of febrile neutropenia in solid tumors

59indirect costs 303indwelling vascular access catheters 20,

152ineligibility 52infection control 129–31infection management 101infection sites see sites of infectioninfection-related factors 49infections

cancer-mimicking 106causitive organisms 4, 219–20, 318–19in children 202–3focal 68–9in hematologic patients 111–22in HSCT 126–8, 134–5incidence and neutrophil count 1, 2, 6neutropenia-associated 128patients with solid tumors

breast cellulitis 104clinical features/diagnosis 96–8gynecologic cancer 105–6hepatobiliary 105and Ommaya reservoirs 104pulmonary 103–4sites 94, 96, 97spectrum 94–5therapy 98–103

prevention 222–4seriousness 220–1sites 127treatment efficacy 220

Infectious Diseases Society of America (IDSA)128, 178

on management of febrile neutropenia290–1, 293

inflammatory response 6–7, 9inflation 306infusate contamination 190

332 INDEX

infusion pumps, computerized small-volume173

infusion therapy team 193initial management 291initial therapy 64, 290

broad-spectrum �-lactams 57factors predicting outcome 172modifications 99recommendations 294

Institute of Medicine 175intangible costs 302integumental barrier disruption 3, 127intention-to-treat analysis 53interest costs 305interleukin levels 63International Antimicrobial Therapy

Cooperative Group 98International Bone Marrow Transplant Registry

133International Conference on Nosocomial and

Healthcare-Associated Infections 173–4intestinal tract 217, 218intolerance 177intraluminal colonization see luminal

colonizationinvasive fungal infection disease 226–7invasive hyalohyphomycosis 71, 72invasive molds 130, 269ionic silver 195isoniazid 102Italian Association of Pediatric Hematology and

Oncology 295–6itraconazole 132, 218, 228, 229

Japanese Infectious Disease Society 295joint costs 304–5

Kaplan–Meier survival curves 308Kaplan–Meier technique 48ketoconazole 228kidney 32Klebsiella spp. 128, 318

K. oxatoca 70K. pneumoniae 5, 16, 18, 68

laboratory evaluation 152of HSCT patients 134

patients with solid tumors 98laminar airflow units 113, 129–30late recovery period 126Legionella spp. 157

L. pneumophila 112Leptotricia spp. 62

Leptotrichia buccalis 218leukapheresis 266–7leukemias 111, 201

see also specific diseasesleukocyte compatibility 134leukocyte transfusions see granulocyte

transfusionsLinezolid 100, 137liposomal amphotericin B 32, 139

see also amphotericin BListeria monocytogenes 113, 118liver drug concentrations 32liver dysfuntion 38local microbiology 180local susceptibility/resistance 115, 180logistic regression analysis 48, 172logrank test 48long-dwelling intravascular devices 62long-term venous access devices 137–9low-risk patients

oral antibiotic therapy 176–7outpatient therapy 173–6

luminal colonization 190, 194lungs 4, 32lymphoid malignancies 251lymphomas 111, 253–6

macrolides 103macrophage colony-stimulating factors (M-

CSF) 247magnetic resonance imaging (MRI) 155

in children 206malabsorption 218Malassezia spp. 73

Malassezia furfur 190, 191management guidelines 134marginal costs 304masks 132MASCC see Multinational Association for

Supportive Care in Cancermaximum sterile barrier 193

INDEX 333

MD Anderson Cancer Center 75, 103, 139and catheter-related infections 193ciprofloxacin, oral 172documentation of infections 180empiric therapy trials 169outpatient therapy trials 178

medical devices 93medical effectiveness research 309medical errors, preventable 175medical instability 158, 159Medline 289meningitis 118

neoplastic 104mental status changes with neutropenic

pneumonia 157meropenem

for children 207empiric therapy 115

monotherapy 13, 204methicillin-resistant S. aureus 115methodological problems 53methods of care 180miconazole 228microbial evaluation 120microbial flora 5–6

endogenous, infection from 112local 98, 113

microbiologically documented infection 151Micrococcus 62, 64mid recovery period 125–6minimum bactericidal concentration 28minimum inhibitory concentration 28minocycline/rifampicin coating 195misclassification rates 163, 176

of MASCC Risk Index 163of Talcott model 160

moderate-risk patients 182molds

exogenous sources 219in HSCT patients 140invasive 130, 269

monitoringof clinical trials 47–8in outpatient therapy 180, 181–2

monotherapy 13–14cost-effectiveness 312–13current recommendations 23, 290, 294, 295

for empiric regimens 114in empiric therapy 115, 204outpatient 178patients with solid tumors 99

morbidity in children 202, 203mortality

assessment 51in children 202, 203, 207cytomegalovirus-related 231infection-related 50

mortality rates 45aspergillosis 141–2children with bacteremia 64early initiation of therapy 9, 10Gram-negative infections 7invasive fungal infections 71outpatient 178respiratory viruses 143

moxalactam 17Mucor 219Mucorales 140mucosal barrier 4mucosal barrier disruption 127mucosal damage 3, 60, 93mucosal infections, duration of 137mucositis 155

agents causing 93and bacteremia 62and fluoroquinalones 224–5following cancer therapy 7, 112gastrointestinal 156oral 217–18oropharyngeal 69and pharmacokinetics 38

multicenter studies 47, 48multi-colony stimulating factor (multi-CSF)

247multidrug resistance see under resistancemultimodal therapy 318Multinational Association for Supportive Care

in Cancer (MASCC)on clinical trials methodology 298Risk Index 160–3, 176, 182

multiple entries 49multiple episodes 51–2multiple infections 319multiple logistic regression model 158, 160, 161

334 INDEX

multiple-lumen vascular access catheter 93musculoskeletal toxicity 137mycobacterial infections 102–3, 138myelodysplastic syndrome 60myeloid growth factor receptors 246, 247, 249myeloid growth factors 133–4, 208–9, 245,

246–65in acute leukemia 259–62benefit, evaluation of 250clinical events, effect on 251comparative activity of G-CSF and GM-CSF

288dose and schedule 264–5dose-intensive regimens 256–7economic evaluations 311, 312effectiveness, lack of 257, 259guidelines

for acute lymphoblastic leukemia 286–7for acute myeloid leukemia 285–6and chemotherapy dose increase 284with chemotherapy/irradiation 287dosing and administration route 288initiation and duration 288for leukemia in relapse 287for myelodysplastic syndromes 286and progenitor cell transplantation

284–5prophylaxis 282, 283therapy 283–4

management of chemotherapy patients251–3

mobilization 134as primary prophylaxis 253–6recombinant, biology of 247–50as secondary prophylaxis 257and stem cell transplantation 251–2, 262–4timing of treatment 264toxicity 265treatment of neutropenia 257–9see also granulocyte–macrophage colony-

stimulating factors; granulocyte colony-stimulating factors; hematopoietic colony-stimulating factors

myeloid progenitor cells 245myelopoiesis 246

nalidixic acid 207

National Cancer Comprehensive NetworkClinical Guidelines for Fever and

Neutropenia 128management guidelines 134

National Cancer Institute 7, 98, 203on neutropenia duration 168

National Comprehensive Cancer Network 178Practice Guidelines for Fever and

Neutropenia 196, 293–4necrotizing infections 69, 94, 101Neisseria spp. 216nephrotoxicity 13, 20

aminoglycoside 114amphotericin B 142

neutropenia 1–3, 134, 167–8in acute leukemia 58after stem cell transplant 262depth of 245duration of 59–60, 168–9, 170

with growth factors 112, 184, 245, 253, 257with PBSC 133and prophylaxis 60

febrile 11–20, 290–1, 293guidelines 128, 196, 293–4in hematologic patients 111in HSCT 127infection in 5patients at risk 318persistent 19–20predisposing factors 3with solid tumors 91, 92–3treatment of 257–9and vascular access device infections 192

neutropenia-associated infections 128neutropenic enterocolitis 69–70, 156, 218neutropenic patients 61, 318neutrophils 246, 267

counts 167–8, 245and bacteremia response rate 10and incidence of infection 1–2, 45

and G-CSF 247transfusions 23

Nocardia 118non-compliance 181, 208, 303non-Hodgkin’s lymphoma 59, 201, 253non-tunnelled central venous catheters 69non-tunnelled subclavian catheters 189

INDEX 335

nosocomial infections 173nosocomial outbreaks 74nosocomial pathogens 94–5

and coated catheters 195and prophylaxis 223

nystatin 227, 228nystatin, iposomal 142

obstruction 4, 93and infection 3

odynophagia 117oesophagitis 75ofloxacin, oral 172–3, 178Ommaya-reservoir-related infections 104opportunity costs 305, 306, 309oral antibiotic therapy 171–3, 291, 295

in children 206, 208contraindications 38and early discharge 176hospital-based 176–7outpatient 177–8

oral cavity 216damage to 217–18

oral lesions 69, 155oral therapy 118

versus IV therapy 21, 23oropharynx 3, 5oseltamivir 132–3otitis media 202, 203ototoxicity 13, 114outcomes (economic) 303–4outcomes (medical)

in clinical trials 46assessment 48measurement and reporting 50, 53predictors 50

of febrile neutropenia 57and lesion characteristics 156–7and neutrophil count 167–8

outcomes research 304, 309–11outpatient management 162outpatient therapy 173–6

dosing route 173, 177–8and growth factors 253guidelines 294, 295in hematologic patients 118for HSCT patients 143

and nosocomial infections 100pediatric 178

outpatient therapy trials 50, 179oxacillin resistance 65

paired quantitative blood cultures 192parasitic infections 96, 102, 135partial decontamination 222pathogens 5

nosocomial 94–5shift in predominance 20–1, 62sources 127

patient evaluation 143–4patient selection 178, 180, 181patient/home factors 180patients at risk 317–18patients’ preferences 309, 311

pediatrics 312patterns of disease see epidemiologypeak concentrations 29penicillin resistance 64penicillins 12Penicillium spp. 219perianal area 4perianal lesions 9perianal/perirectal cellulitis 70periodontal infections 4, 155peripheral blood stem cells (PBSC) 133peripheral blood stem cell transplants (PBSCTs)

262, 264peripherally inserted catheters 189pH as barrier 216phaeohyphomycosis 72pharmacoeconomics 301, 313

of growth factors 264pharmacogenetic assessment 321pharmacokinetic:pharmacodynamic (PK:PD)

ratio 30–2antibiograms 30

pharmacokinetics 27, 29, 32hepatic dysfuntion 38renal dysfunction 32, 38

pharmocodynamics 27, 28pharyngitis 9phase I trials 45phase II trials 45, 47phase III trials 45, 47

336 INDEX

phase IV trials 45piperacillin–tazobactam 207pipericillin 207pipericillin–tobramicin 312–13Pizzo criteria for response 297PK:PD ratios 39–41platelet support 135pneumococcal bacteremia 64pneumococci 68Pneumocystis carinii 96, 118

prophylaxis against 222, 228in respiratory infections 157

pneumonia 4, 93epidemiology 68in hematologic patients 117neutropenic 157with Ps. aeruginosa 65

pocket infections 196polymerase chain reaction 230polymicrobial bacteremia 63polymicrobial infections 96, 103, 105

in hematologic patients 112, 117of perirectum/groin 156

ports 193–4posaconazole 142positive time preference 305post-antibiotic effect 28post-antibiotic leukocyte enhancement 28potency, antimicrobial 27, 28Practice Guidelines Committee 290predictors of outcome 50pre-engraftment period 125, 126, 127

risk factors 140, 141presumption of infection 151pretransplant period 125, 126prevention of infection 222–4

bacterial 130–1fungal infections 131–2from vascular access devices 192–5viral 132–3see also prophylaxis

procalcitronin serum levels 63prognosis

of bacteremic infections 62factors 49, 172

progression of infection 4prophylaxis 215–32

antibacterial 131, 222–6antifungal 60, 63, 131–2, 226–30

problems with 215recommendations 294

antimicrobial 60, 293antiviral 230–2, 294for BSCT and autoPBSCT 60in children 208effect of 112, 113efficacy 226–30with growth factors 209and growth factors 251primary 253secondary 257side-effects 224–5tubercular 102see also prevention of infection

prosthetic devices, surgically implanted 93protein binding 39protocols 46Pseudoallescheria 140pseudomembranous colitis 117Pseudomonas bacteremias 9, 10Pseudomonas infections 8Pseudomonas spp. 219

Ps. aeruginosa 112, 318bacteremia 65–6and bacterial translocation 5colonization 216empiric therapy 5in HSCT 128, 138infections 95, 157invasive potential 10management 101in pneumonia 68and prophylaxis 112, 222synergy versus single agents 16, 17, 18, 19translocation 218and typhilitis 156in vascular access device colonization

191, 196Ps. fluorescens 66Ps. paucimobilis 66

pulmonary infections 94in hematologic patients 117–18, 120post-obstructive 103–4secondary 122

INDEX 337

quality of life 50and early discharge 176and growth factors 251, 252, 256and risk-adjusted therapy 181

quality-adjusted life-years 308, 309quality-of-life studies 309quality-of-life variables 308quantitative culture techniques 192quinolones 7, 103

in children 207, 208in empiric therapy 115monotherapy 178oral 177

quinpristin/dalfopristin 100

radiation therapy 217–19radiographs see X-raysrandomized trial design 46, 47, 48, 49rating system of recommendations 290,

293ravuconazole 142readmission rate 176recombinant engineering 247remission induction therapy 2renal dysfunction 32, 38renal excretion 208renal toxicity 178, 206reporting of clinical trials 48resistance see antimicrobial resistanceresistant organisms 5, 6, 98

antibiotic use and 94Gram-positive 99in HSCT 127in vascular access devices 194

respiratory mucosa 3respiratory syncytial virus 75, 143, 203respiratory tract infections 94, 141respiratory viruses 142–3response to therapy 50reverse barrier isolation 219Rhizopus 219Rhodococcus spp. 99Rhodotorula 73, 191ribavirin, aerosolized 143risk

assessment 151, 157–63definition 158

factors 63, 169scoring system 160, 161stratification 184, 206, 318, 321

risk threshold estimates 312risk-adjusted clinical trials 171risk-adjusted management 179–80risk-adjusted therapy 180, 181–2risk-assessment 253risk-based stratification 209risk-based therapy 100route for antibiotic therapy 151, 312

safety monitoring 47Salmonella typhi 5, 156sample size 53saprophytic organisms 62sarcoma patients 184seasonal respiratory viruses 157secondary infections 122secondary malignancies 256selective decontamination 112, 222sensitivity

of MASCC Risk Index 163, 182of Talcott model 160

sensitivity versus specificity 161septic shock 63septic thrombosis 196sequential designs 51sequential therapy 177serial microbiological surveillance cultures

98serodiagnosis 120serum albumin changes 38serum bactericidal activity 16–18serum/plasma concentrations 31Shigella 156shock syndrome 118, 119, 156side-effects

anti-bacterial prophylaxis 224–5anti-fungal prophylaxis 228, 229anti-viral prophylaxis 232granulocyte transfusions 267–8myeloid growth factors 266

silver iontophoretic catheters 195Simon two-stage design 47simultaneous blood culture methods 192sinusitis 4

338 INDEX

sites of care 143–4, 151, 180cost-effectiveness 311–12

sites of infection 3–5, 9, 94, 96primary 152with Ps. aeruginosa 65

site-specific assessment 152–7esophagus 155hepatobiliary tract 155intestines/colon 156mouth/oropharynx 155perirectum/groin 156respiratory tract 157sinus/nasal passages 152, 154, 155skin/soft tissues 156–7urinary tract 157

site-specific cultures 135site-specific investigations 139site-specific therapy 294skin 216, 217

damaged 4skin barrier, breach of 112skin infections 117skin insertion site 190soft tissue infections 117solid tumors

and catheter-related infections 193chemotherapy for 59and growth factors 184, 253–6and infection 91–107

risk factors 91–4inflammatory responses with 96and low risk 176and neutropenia duration 168and S. aureus infection 65

specificityof MASCC Risk Index 163, 182of Talcott model 160

spectrum of infection 94–5spleen 32spores 70, 73–4sputum cultures, regular 120Staphylococcus spp.

in hospital-acquired infections 96St. aureus 70, 319

and CRBSIs 196in HSCT 128, 138infection in neutropenia 5

in pneumonia 68prophylaxis against 222in sinuses/nasal passages 155and soft tissue infection 157in surgical wound infections 95synergy versus single agents 16, 18in vascular access device colonization

190, 191St. epidermidis 64

and catheter-related infection 112commensal 216in HSCT 128, 129infection in neutropenia 5in vascular access device colonization

190, 191statistical design 46–7stem cell exhaustion/steal 247stem cells 246

sources 125stem cell transplantation 251–2, 262–4Stenotrophomonas spp. 112

St. maltophila 66, 115, 219in HSCT 128, 138and soft tissue infection 157in vascular access device colonization

191, 196stepdown therapy 177

cost-effectiveness 313steroids 59stomach 216stomatitis 38Stomatococcus spp. 62

St. mucilaginosus 218, 220storage of granulocytes 265, 270stratification 49–50

risk 170, 171, 174, 184, 264stratified analyses 49streptococcal bacteremia 59streptococcal infections 7, 128, 207Streptococcus spp.

infection in neutropenia 5in surgical wound infections 95

S. mitis 216infection 217respiratory distress syndrome 220

S. oralis 216

INDEX 339

Streptococcus spp. – cont.S. pneumoniaea

prophylaxis against 222in respiratory infections 96in sinuses/nasal passages 155

S. sanguis 216Strongyloides 156sub-MIC effect 28sunk costs 305superinfection 50supportive care 111surgical intervention 101surgical wound infections 94surrogate markers 68surveillance cultures 142survival 58susceptibility patterns 11, 13Synercid 100, 137synergistic activity 100synergy 17

�-lactam and aminoglycoside 12, 14carbenicillin and gentamicin 7–8and neutropenia 14, 16

Talcott clinical prediction rule 158–60, 176, 294versus MASCC Risk Index 160, 162–3

Talcott derivation and validation study 159,160

tap water 74targeted patient population 46T-cell defects 157T-cell recovery 126teicoplanin 115tetracyclines 207therapeutic index 38three-stage trial designs 47thrombin sheath formation 190thrombocytopenia 192ticarcillin 17, 18time to clinical response 182–4time to defervescence 62time to onset of fever 60time-dependent killing pharmacodynamics 28time-to-event distributions 48tissue concentrations 32tissue penetration 32tissue samples 120

total body irradiation 218, 219totally implanted catheters 69toxicity

bone marrow toxicity 137cartilage 207, 208in children 207ganciclovir 232and granulocyte transfusions 267gut 218musculoskeletal toxicity 137of myeloid growth factors 265nephrotoxicity 13, 20, 114, 142ototoxicity 13, 114as outcome 50renal 178, 206

Toxoplasma gondii 118translocation, bacterial 5, 60, 218treatment

efficacy 220plan 46response 296–8

Trichosporon 73trimethoprim–sulfoxamethazole 103, 118,

208guidelines 293tissue concentrations 32see also co-trimoxazole

Tsukamurella paurometabolum 220tuberculosis 102–3tumor necrosis 102tunnel infections 137, 138, 196, 223, 224tunneled catheters 69, 189, 192tunneling 193–4two-site culturing, review of 294two-stage trial designs 47typhilitis 69–70, 156, 218

ultrasound 155United States Food and Drug Administration

207, 296upper respiratory tract infection 202, 207ureteral obstruction 93urinary catheters 93urinary tract infections 94, 118urine cultures, regular 120utility assessment 309, 310utilization, unit of 305

340 INDEX

vaccination of healthcare workers 129valacyclovir prophylaxis 133validation

of MASCC Risk Index 160of Talcott model 160

validity 53vancomycin

and catheter-related infections 196in empiric therapy 20–1, 22, 136–7, 139

in children 204, 296guidelines for use 290prophylaxis 223tissue concentrations 32

vancomycin-dependant enterococci 68vancomycin-resistant enterococci 20, 67–8, 70,

100, 136in HSCT patients 128, 138

varicella zoster virus 75in children 203, 208reactivation 76and soft tissue infection 157

vascular access device infections 189–98epidemiology 189evaluation/diagnosis 191–2infusion therapy team 193management/treatment 195–7maximum sterile barrier 193microbiology 191, 196–7pathogenesis 189–91placement site selection 194prevention 192–5sources 189–90

vascular access devices 189–90, 191, 192–5, 196

venipuncture as damage to integument 3venous catheter 138–9, 189viral infections 61, 75–6, 96

diarrhea 135in HSCT 127patients with solid tumors 101–2

viral pneumonia 68viridans streptococci 20, 112, 218, 319

and bacteremia 64in HSCT 128prophylaxis 222treatment efficacy 220

visual analog score for MASCC Risk Index161

volume of distribution 32, 38vomiting, chemotherapy-induced 4voriconazole 139, 142

withdrawals from trials 49, 52–3within-antibiotic-class comparisons 45within-patient correlation 51–2wound infections 94

X-rays, chest 116, 152in children 203of HSCT patients 134

yeast -derived GM-CSF 265yeasts 140, 216

zanamivir 132Zygomycetes 155

INDEX 341

febril netropeni

Documents