severe viral respiratory infections in infants with cystic fibrosis

Severe Viral Respiratory Infections in Infants WithCystic Fibrosis

David Armstrong, FRACP,1* Keith Grimwood, MD,2 John B. Carlin, PhD,3Rosemary Carzino, BSc(Hons ),2 Jeremy Hull, MD,1 Anthony Olinsky, FRACP,1 and

Peter D. Phelan, MD1,2

Summary. Limited data in children with cystic fibrosis (CF) suggest that respiratory viral infec-tions during infancy result in substantial morbidity. Eighty of 101 (79%) infants with CF diag-nosed by neonatal screening during 1991–1996 were recruited into a prospective, multiple-birthcohort study. We aimed to perform an initial, then annual bronchoalveolar lavage (BAL) forbacterial and viral culture, cytology, IL-8, and elastolytic activity over the following 2 years. Whenpossible, BAL was also performed during any hospitalization for a pulmonary exacerbation, andadditional specimens for viral culture were collected by nasopharyngeal aspiration. Thirteeninfants undergoing bronchoscopy for congenital stridor served as disease controls.

During infancy, 31 children (39%) were hospitalized for respiratory disease and 20 (65%)cases had an etiologic agent identified. Respiratory viruses were detected in 16/31 (52%) cases,including four with simultaneous bacterial infection. Another four were infected with Staphylococcusaureus. Respiratory syncytial virus predominated and was found in seven infants. In the absence ofbacteria, those with viral infections had acute onset of respiratory distress, were not treated withantibiotics, and had an uncomplicated hospital course. Compared to noninfected CF subjects andcontrols, infected infants had elevated BAL inflammatory indices (P < 0.01). Eleven of 31 (35%)hospitalized infants followed for 12–60 months acquired Pseudomonas aeruginosa, compared withonly three of 49 (6%) subjects not hospitalized for respiratory symptoms during infancy (risk ratio 5.8,CI 1.9, 24). We conclude that respiratory viruses are important causes of hospitalization in CFinfants. While viral infections were self-limited, they were accompanied by airway inflammatorychanges, and admission to hospital was associated with early acquisition of Pseudomonas aerugi-nosa and persistent respiratory symptoms. Pediatr Pulmonol. 1998; 26:371–379.© 1998 Wiley-Liss, Inc.

Key words: cystic fibrosis; viral infections; bacterial infections; infant;bronchoalveolar lavage; treatment.

INTRODUCTION

In children and adults with cystic fibrosis (CF), respi-ratory virus infections are associated with pulmonary ex-acerbations,1–5 acquisition of Pseudomonas aerugi-nosa,1,5 and deterioration in lung function.2–6 However,viral infections are most common during early childhood.Previously, only one study examined the role of virusesin infant CF lung disease.7 It found that respiratory vi-ruses were an important cause of hospitalization duringthe first 12 months of life. Infection by respiratory syn-cytial virus (RSV) was followed by severe illness, pro-longed hospitalization, and chronic respiratory symp-toms. Clarification of the role of respiratory viruses in CFinfant lung disease might have implications for patientmanagement.8 If viral infections are self-limited, it maybe possible to reduce hospitalization and unnecessaryantibiotic therapy. Alternatively, if viral respiratory in-fections result in severe illness, bacterial superinfection,or chronic respiratory disease, interventions such as im-munization, antiviral agents, and antibiotics may becomenecessary.

1Department of Thoracic Medicine, Royal Children’s Hospital, Park-ville, Victoria, Australia.

2Department of Paediatrics, University of Melbourne, Parkville, Vic-toria, Australia.

3Clinical Epidemiology and Biostatistics Unit, Royal Children’s Hos-pital, Parkville, Victoria, Australia.

Presented in part at the American Thoracic Society/American LungAssociation International Conference, Seattle, 1995.

Current address of Keith Grimwood, M.D.: Department of Paediatricsand Child Health, Wellington School of Medicine, P.O. Box 7343,Wellington, New Zealand.

Current address of Jeremy Hull, M.D.: Department of Paediatric Mo-lecular Genetics, John Radcliffe Hospital, Oxford OX3 9DS, UK.

Grant sponsor: Royal Children’s Hospital Research Foundation; Grantsponsor: Smorgon Clinical Research Fellowship.

*Correspondence to: Dr. David Armstrong, Department of ThoracicMedicine, Royal Children’s Hospital, Parkville, Victoria 3052, Australia.

Received 19 January 1998; accepted 9 June 1998.

Pediatric Pulmonology 26:371–379 (1998)

© 1998 Wiley-Liss, Inc.

To help determine the effect of severe viral respiratoryinfections upon CF infant lung disease, we examined acohort of infants identified by a statewide newbornscreening program. All were diagnosed within the first 3months of life. At recruitment each infant underwentbroncholaveolar lavage (BAL) for the collection of lowerrespiratory tract specimens. BAL was then repeated at 12monthly intervals and if possible, at the time of admis-sion for severe or persistent respiratory symptoms. Thestudy aims were to observe the incidence of hospitaliza-tion for severe viral respiratory infection during the firstyear of life, the nature of the accompanying lower airwayinflammatory response, and the clinical course during,and for at least 12 months after, the respiratory illness.

METHODS

Subjects and Controls

The State of Victoria, Australia (66,000 births peryear) has a two-tiered newborn screening program forCF9 based upon an estimation of the immunoreactivetrypsin from a blood spot Guthrie card, followed by ge-netic analysis for theDF508 deletion and three exon 11mutations: G542X, G551D, and R553X. A sweat chlo-ride concentrationù60 mEq/L confirms the diagnosisfor heterozygotes. Neonates with meconium ileus alsoundergo mutation analysis and sweat testing. Victorianchildren with CF are managed by a single clinical servicebased at the Royal Children’s Hospital, Melbourne.

For the purposes of this study, all CF infants bornbetween July 1, 1991–June 30, 1996 were recruited intoa previously described prospective study of the earlynatural history of CF lung disease.10–12 We aimed toperform an initial, then annual BAL for 2 additionalyears to collect lower respiratory tract secretions. If pos-sible, BAL was also performed at the time of hospital-ization for pulmonary exacerbations. Additional speci-mens for viral culture were collected by nasopharyngealaspiration (NPA). Subjects were seen every 3 months inthe CF Clinic. At each contact, subjects underwent aclinical evaluation, noting in particular respiratory symp-toms and antibiotic use. Physical signs evaluated in-cluded tachypnea, cyanosis, finger clubbing, chest wall

retractions, and signs of wheezes or crackles on auscul-tation.

The decision to hospitalize subjects for treatment ofacute respiratory infections was made by physicians car-ing for the individual patient, none of whom were di-rectly involved in the study. Criteria for hospitalizationincluded one or more of the following: severity of respi-ratory signs (audible wheeze with chest wall retractionsor oxygen saturation less than 92% in room air), failureto improve with oral antibiotics, feeding difficulties,weight loss, or deterioration in general health.13 Deci-sions regarding therapy with parenteral antibiotics, chestphysiotherapy, oxygen, bronchodilators, and the need foradditional investigations, including BAL and chest radi-ography, were determined on an individual basis by thephysicians responsible for the patient’s overall care.

Otherwise-healthy infants undergoing bronchoscopyfor evaluation of congenital stridor formed a disease con-trol group for the measurement of BAL inflammatoryindices. Subjects were excluded if there were symptomsor signs of respiratory infection or a history of antibioticuse within the previous 14 days.

Specimen Collection

As described previously,11 flexible bronchoscopy wasperformed under halothane anesthesia following topicaladministration of 4 mg/kg of 2% lignocaine hydrochlo-ride to the vocal cords. Briefly, the bronchoscope (Olym-pus model BF 3C20, Olympus Optical Co., Ltd., Tokyo,Japan; external diameter 3.6 mm, suction channel 1.2mm) was introduced into the lower airways through alaryngeal mask, avoiding use of the suction channel untilthe bronchoscope tip was below the carina. The tip waswedged in the right middle lobe bronchus, and 1 mL/kg(maximum 20 mL) of sterile nonbacteriostatic normalsaline was instilled. Using negative suction pressures of100–150 mm Hg, the saline was immediately aspiratedinto a sterile specimen suction set. The bronchoscopewas next wedged into the bronchus of the lingula, and afurther 1 mL/kg lavage was performed. The BAL fluidwas pooled and transported on ice to the laboratory forprocessing and storage at −70°C.

NPA was collected by a soft sterile plastic catheterpassed into the nasopharynx and, using negative suctionpressures of approximately 150 mm Hg, secretions con-taining epithelial cells were aspirated into a sterile speci-men suction set and sent immediately to the laboratoryfor analysis.

Laboratory Procedures

The NPA and 500mL of BAL fluid were tested byimmunofluorescence for RSV (RSV Direct IF, bioMer-ieux, Marcy L’Etoile, France), influenza virus (A and B),parainfluenza viruses 1–3, and adenovirus (Bartels Viral

Abbreviations

BAL Bronchoalveolar lavageCF Cystic fibrosisCFU Colony-forming unitsCI Confidence intervalIL-8 Interleukin-8mEq/L Milliequivalents per literNPA Nasopharyngeal aspirateRSV Respiratory syncytial virusSD Standard deviation

372 Armstrong et al.

Respiratory Screening and Identification Kit, Baxter Di-agnostics, Inc., Deerfield, IL) and cultured for respiratoryviruses.14 The cells inoculated for viral culture includedprimary monkey kidney (LLC-MK2), human epithelial(A549, HeLa), and fibroblast cell lines.

The remaining BAL fluid was gently vortexed withsterile glass beads for 5 sec, and 500mL were seriallydiluted in sterile phosphate buffered saline (pH 7.2) be-fore plating onto selective media for quantitative bacte-rial colony counts.11 Standard microbiologic techniquesidentified Staphylococcus aureus, Hemophilus influen-zae, Pseudomonas aeruginosa, Burkholderia cepacia,and other respiratory bacterial pathogens.15 Lower respi-ratory tract bacterial infection was diagnosed when BALfluid grew ù105 colony-forming units (CFU) per mL ofbacterial pathogens.11

One hundred microliters of BAL fluid were stainedwith toluidine blue, and the total cell count was estimatedusing a Neubauer hemocytometer (Weber, Teddington,UK). The differential cell count was estimated after cen-trifugation of 300mL BAL fluid at 900 rpm for 5 min ina cytocentrifuge (Shandon Southern Instruments,Sewickly, PA), Wright-Giemsa staining, and counting of300 cells.16 After centrifugation of BAL fluid at 500g for5 min at 4°C, the supernatant was stored at −70°C.

The neutrophil chemokine, interleukin 8 (IL-8), wasmeasured by an enzyme-linked immunosorbent assay(Medgenix Diagnostics, Fleurus, Belgium) which mea-sured free and receptor bound IL-8. The detection limitfor the assay was 0.7 pg/mL. After hydrolysis of thespecific chromogenic substrate, MeO-Suc-Ala-Ala-Pro-Val-pNA (Sigma Chemical Co., St. Louis, MO), freeneutrophil elastase activity was measured in BAL super-natant by a spectrophotometer at 405 nm.17 The absor-bance was compared with a standard curve derived frompurified human neutrophil elastase (Elastin Products Co.,Inc., Owensville, MO), and the concentration was calcu-lated from a log-logit plot. A specific inhibitor of humanneutrophil elastase, MeO-Suc-Ala-Ala-Pro-Val-CH2Cl(Sigma Chemical Co.), confirmed the specificity of theenzyme reaction. The sensitivity of the assay was 5.0mg/mL.

As recent studies have shown quantitation of solublecomponents within returned BAL fluid is difficult,18 theresults of quantitative culture, cytologic analysis, IL-8concentrations, and free neutrophil elastase activity wereexpressed per milliliter of BAL fluid.

Data Analysis

The cross-sectional analyses of the study compared CFinfants hospitalized because of viral respiratory tract in-fection and CF infants: 1) admitted to hospital with eitherdocumented bacterial or mixed viral and bacterial lowerrespiratory tract infections, 2) those hospitalized because

of severe respiratory symptoms, but in whom no respi-ratory pathogen was identified, 3) those with an originalnegative BAL culture who were not hospitalized for re-spiratory symptoms before age 12 months, and 4) con-trols. The longitudinal component compared outcomesfor CF infants hospitalized during their first year of lifebecause of a respiratory illness with those infants whowere not admitted to hospital.

Cytologic data and IL-8 concentrations were logarith-mically transformed and summarized as geometricmeans with 95% confidence intervals (CI). Group com-parisons for continuous variables were conducted usingt-tests or analysis of variance, and nonparametric testswere used to compare medians of severely skewed dis-tributions of length of treatment. Proportions were com-pared byx2 or Fisher’s exact tests. All analyses wereperformed using the Stata statistical software package(Stata Corporation, College Station, TX).19

This study was approved by the Human Ethics Com-mittee of the Royal Children’s Hospital, and written in-formed consent was obtained from the parents of eachchild before bronchoscopy.

RESULTS

Between July 1991–June 1996, 101 infants were diag-nosed with CF in Victoria, and 80 (79%) were recruitedinto the study before age 12 months. None received an-tiinflammatory drugs or the influenza vaccine. Duringthe first year of life, nine infants (11%) were admitted forgastrointestinal symptoms, whereas 31 (39%) infantswere hospitalized on 45 occasions because of severe orpersistent respiratory symptoms. Five in the latter groupwere admitted twice, and one had 10 admissions. How-ever, only data from the first hospitalization in these 31infants are presented, to avoid the potential confoundinginfluence of previous illness on the subsequent clinicaland bacteriologic course.

An etiologic agent was detected in 20/31 (65%) hos-pitalized infants. Respiratory viruses were identified in16 (52%), including four with simultaneous lower respi-ratory bacterial infections. An additional four infantswere shown to have lower airway infection withS. au-reus.

The clinical characteristics of these infants at the timeof hospitalization are shown in Table 1. There were nosignificant differences in age, gender, CF genotype,weight z-scores parental smoking, or prior antibiotic ex-posure between the four groups. Infants with either aviral lower respiratory infection or those in whom a caus-ative pathogen could not be identified had a shorter his-tory of illness prior to admission and were more likely tohave wheeze or respiratory distress than those with abacterial lower respiratory tract infection in whom per-sistent cough was the predominant symptom. The distri-bution of gender, CF genotype, and parental smoking in

Viral Respiratory Infections in CF Infants 373

hospitalized infants was comparable to that observed inthe 49 CF infants not admitted to hospital for respiratoryillness during the first year of life.

NPA was obtained during 26/31 (84%) admissions,and at least one respiratory virus was identified in 14(54%) of these specimens. Respiratory syncytial virus(RSV) was detected on six occasions and was the mostcommonly identified viral pathogen associated with hos-pital admission. Parainfluenza viruses were isolated infour infants, while rhinovirus accounted for another twocases. Influenza A and B was detected from one infanteach, respectively. A further patient hospitalized for 3weeks because of cough and wheeze had an enterovirusidentified from an NPA specimen, and one infant hadboth parainfluenza type 3 and RSV detected.

During hospitalization, BAL fluid was collected in 15(42%) infants. Bacterial lower respiratory infection wasdiagnosed in eight cases, four of whom also had a virusdetected in their NPA specimens.S. aureuswas the mostfrequently isolated lower airway bacterial pathogen.Colony counts ofS. aureusù105 CFU per milliliter wereobserved in seven BAL (47%) specimens. The mixedviral and bacterial infections found in these infants in-cluded two cases each with rhinovirus andS. aureus,another with influenza B andH. influenzae,and one withparainfluenza type 1 and bothS. aureusandH. influen-zae.Of the 12 infants with negative NPA immunofluo-rescence and culture, seven had simultaneously collectedBAL samples from which RSV and parainfluenza type 3were isolated in one child each, respectively.

Of the 11 CF infants admitted because of respiratorysymptoms, but in whom an etiological agent was notfound, seven had an NPA but only two underwent BAL.However, a further three of these infants had a BALwithin 2 weeks of discharge from hospital, when afterinitial improvement with parenteral antibiotic treatment

their symptoms of chronic cough returned. Two of theseinfants hadS. aureus,and the third had aP. aeruginosainfection diagnosed following culture of their BAL fluid.

In contrast to hospitalized infants, respiratory viruseswere rarely detected in stable CF subjects undergoingelective BAL during the first few months of life. Onlyone respiratory virus, RSV, was detected in a 3-month-old stable patient; however, 18 infants (8 females; meanage, 5.1 (SD 3.9) months) had lower airway bacterialinfections diagnosed as outpatients by BAL.S. aureuswas identified in 10 infants, three hadH. influenzae,twohad mixedS. aureusandH. influenzae infection,a furthertwo M. catarrhalis,and one hadP. aeruginosadetected.As these infants were either asymptomatic or had onlymild to moderate cough, they were treated as outpatients.

Analysis of the inflammatory indices in the 13 hospi-talized infants who underwent BAL provided five casesadmitted with viral lower airway infection and eight withbacterial infection, the latter including four with bothrespiratory viruses and bacteria. Of the five infants withpure viral respiratory infections, three had RSV and twohad parainfluenza type 3. These data were compared withthose from 22 CF infants categorized as noninfected ac-cording to BAL culture criteria11 who were not hospital-ized during infancy for respiratory illness, and 13 diseasecontrols. Their data are shown in Figure 1 and summa-rized in Table 2. Compared to noninfected CF infants anddisease controls, CF infants admitted with either viral orbacterial respiratory infections had a marked lower air-way inflammatory response, characterized by an influxof neutrophils, high IL-8 concentrations, and increasedelastolytic activity. In contrast, BAL inflammatory indi-ces in CF infants who had not been hospitalized weresimilar to those of controls.

The hospital course for the 31 infants, including 20 inwhom an etiologic agent was identified, is shown in

TABLE 1—Clinical Characteristics of CF Infants Hospitalized With Severe or Persistent Respiratory Symptoms 1

Virus(n 4 12)

Virus + bacteria(n 4 4)

Bacteria(n 4 4)

No pathogens(n 4 11)

Mean age in months (SD) 5.5 (4.0) 6.3 (4.0) 2.0 (0.7) 5.1 (2.9)Females (%) 6 (50) 1 (25) 2 (50) 5 (45)Mean weight z-score (SD) −0.44 (1.30) 0.44 (0.66) −1.0 (0.71) −0.65 (1.40)CF genotype

HomozygousDF508 9 1 2 5HeterozygousDF508 3 3 1 6No copiesDF508 0 0 1 0

Paternal smoking 2 1 1 0Median symptom days before admission (range) 4.0 (2–23) 24.5 (4–56) 47.5 (14–90)* 4.0 (1–35)Number (%) receiving antibiotics before admission 6 (50) 2 (50) 0 8 (73)Wheeze and chest wall retractions (%) 11 (92) 3 (75) 0* 7 (64)

1Virus, respiratory viruses identified in NPA or BAL specimens by direct immunofluorescence or culture.14 Virus + bacteria, respiratory virusesidentified in NPA or BAL specimens by direct immunofluorescence or culture andù105 CFU/mL bacterial respiratory pathogens in BALfluid.11,14 Bacteria,ù105 CFU/mL bacterial respiratory pathogens in BAL fluid.11 Not pathogens, no respiratory viruses identified in NPAspecimens or BAL cultures <105 CFU/mL, four of whom, including one subject with croup, did not undergo either NPA or BAL.*Significant difference (P < 0.01, Kruskal-Wallis test) between bacterial infection and virus infection groups.


Table 3. While there were no significant differences inoxygen requirement or length of hospitalization, infantswith viral respiratory infections were less likely to re-ceive parenteral antibiotics. None required mechanicalventilation, and no infants with RSV infection receivedribavirin.

Although their initial BAL cultures had not grownP.

aeruginosa,four of 16 (25%) hospitalized infants withviral respiratory infection later acquired lower airwayinfection with this organism. After an admission for RSVinfection at age 5 months, one child hadP. aeruginosaina BAL culture at age 14 months. Following intravenousand inhaled antibiotic therapy, the respiratory symptomsresolved and a BAL at 24 months failed to grow the

Fig. 1. BAL fluid data for CF infants and controls. Groups as defined in Table 2.

TABLE 2—Clinical Characteristics and Bronchoalveolar Lavage (BAL) Inflammatory Indices in Hospitalized CF InfantsWith Severe Viral or Bacterial Infection, Compared to Those Without Infection, and Controls

Virus(n 4 5)

Bacteria(n 4 8)1

Not infected(n 4 22)

Controls(n 4 13) P value

Mean age in months (SD) 6.5 (5.3) 3.6 (3.3) 2.6 (1.5) 11.7 (10.2)CF Genotype

HomozygousDF508 2 4 17 NAHeterozygousDF508 3 3 4 NANo copiesDF508 0 1 1 NA

BAL parameters2

Total cell count (×103/mL)3 645 (201, 2,072) 537 (167, 1,728) 133 (91, 195) 143 (80, 253) <0.001Neutrophils (×103/mL)3 169 (16, 1,809) 260 (62, 1,084) 16 (8, 34) 7 (2, 22) <0.001Interleukin-8 (pg/mL)3 1,372 (561, 3,358) 548 (64, 4,721) 47 (21, 108) 32 (12, 84) <0.001

Percent with neutrophil elastase activity 60 (15, 95) 75 (35, 97) 23 (8, 45) 23 (5, 54) <0.03

1Includes four infants with mixed viral and bacterial infections; see Table 1. Not infected, infants without evidence for infection in their NPAand BAL specimens and who were not hospitalized during infancy for respiratory illness. Controls were subjects with congenital stridor.2Results expressed as geometric mean and 95% confidence intervals, except for percent with neutrophil elastase, where arithmetic mean and 95%confidence intervals are reported.3Comparison between the virus-infected group and the noninfected group and disease controls by ANOVA for TCC, neutrophils, and IL-8, andFisher’s exact test for percent neutrophil elastase.


organism or show signs of airway inflammation. Thesecond infant was originally admitted at age 2 monthswith parainfluenza type 3 infection, and again at age 8months, when rhinovirus was isolated from a NPA speci-men. However, mucoidP. aeruginosaand S. aureuswere also present in BAL culture. Despite aggressiveantimicrobial therapy and mechanical ventilation, he hadprogressive respiratory failure and died 2 months later.The third child was admitted at age 4 months with rhi-novirus on NPA andS. aureusin BAL fluid. Sputumculture at 31⁄2 years grewP. aeruginosa,and he has fin-ger clubbing, poor growth, and a chronic cough. Another3-year-old child with several weeks of cough, and para-influenza type 3 infection at age 2 months, has recentlyhadP. aeruginosaidentified in BAL culture.

None of the remaining 12 hospitalized infants who hadrespiratory viruses detected in their NPA or BAL speci-mens have developedP. aeruginosainfection or chronicrespiratory symptoms and signs. This includes six of theseven infants admitted with RSV infection. Antibiotictherapy for these children is restricted to times of inter-current respiratory infection, and none have required fur-ther hospital admission.

In contrast, three of the four infants admitted withS.aureus lower respiratory infection during the first 3months of life have a chronic cough and persistent in-fection with mucoid strains ofP. aeruginosa.The firstdeveloped a chronic, productive cough withP. aerugi-nosarepeatedly isolated from her sputum by age 3 years.Another infant was clinically stable until age 5 years,when she presented with a 2-month history of cough andincreasing dyspnea;P. aeruginosawas grown from hersputum cultures. The third child was hospitalized at age4 years following a 6-month history of weight loss and1-month history of productive cough. Sputum culturegrewP. aeruginosaandB. cepacia.Four of 11 hospital-ized infants who did not have any causative pathogensidentified, subsequently hadP. aeruginosainfection di-agnosed within the following 12 months. Only one ofthese infants had undergone BAL at the time of admis-sion to hospital and two, whose primary symptoms werecough, hadP. aeruginosaidentified shortly afterwards.

Overall, the 31 CF infants hospitalized for the man-

agement of persistent or severe respiratory symptomshave so far been followed for 12–60 (mean 43.9 (SD14.3)) months. At a mean age of 50.8 (SD 15.9) months,11 had acquiredP. aeruginosa(n 4 10) andB. cepacia(n 4 2) in their lower airways. Infection byP. aerugi-nosawas diagnosed on average 23 (SD 19) months fol-lowing admission to hospital. In contrast, of the 49 CFinfants who had not been admitted to hospital for respi-ratory disease during the first year of life, only three hadacquiredP. aeruginosaby a mean age of 44.9 (SD 17.9)months (risk ratio, 5.8; CI, 1.9, 24). The first of thesethree children had a nonmucoidP. aeruginosastrain iso-lated at age 17 months, but following treatment the or-ganism was not detected in BAL cultures taken 12months later. However, mucoid strains ofP. aeruginosapersist in the other two children who acquired the organ-ism at 37 and 40 months, respectively. Both have chroniccough and finger clubbing.

DISCUSSION

This longitudinal, observational study of an unselectedcohort of CF infants found that almost 40% were hospi-talized for respiratory illness during the first year of life.More than 50% of respiratory admissions were associ-ated with viral infections; these infants usually presentedwith a short history of cough, wheeze, and respiratorydistress and did not receive parenteral antibiotic therapy.In contrast, those with bacterial infections had a persis-tent cough and all had antibiotic treatment. Infants withviral lower respiratory tract infections had airway inflam-matory responses similar to those with bacterial infec-tions. Moreover, infants hospitalized with severe or per-sistent respiratory symptoms were at increased risk ofacquiringP. aeruginosaduring the first 5 years of life.

The distribution of respiratory viruses found in theseCF infants, including the predominance of RSV, wassimilar to that observed in non-CF children.20 The onlyother prospective study of severe viral lower respiratoryinfections in CF infants was conducted in 48 subjectsdiagnosed by a neonatal screening program in Colorado.7

RSV was identified in seven cases, accounting for onethird of hospitalizations during infancy. However, in

TABLE 3—Hospital Course of CF Infants With Respiratory Illnesses

Virus(n 4 12)1

Virus + bacteria(n 4 4)

Bacteria only(n 4 4)

No pathogens(n 4 11)

Median inpatient days (range) 8 (2–15) 12.5 (9–17) 10.5 (7–102) 11.0 (3–19)Oxygen therapy n (%) 3 (25) 1 (25) 1 (25) 4 (36)Median duration in days (range) 7.0 (3–8) 7.02 1.02 7.5 (4–14)Intravenous antibiotics n (%) 1 (8)* 4 (100) 4 (100) 10 (91)Median duration in days (range) 1.0 11 (5–13) 7 (7–14) 10 (5–16)

1See Table 1 for group definitions.2Range not given, as single observation only.*Significant difference (P < 0.001, Kruskal Wallis test) between virus and bacterial infection groups.


contrast to our experience, RSV infection was associatedwith prolonged hospitalization. Three infants requiredmechanical ventilation, and at discharge five (71%)needed continuous home oxygen therapy for as long as 5months. Furthermore, 2 years later all seven infants withRSV infection had chronic respiratory symptoms and ra-diological signs of lung injury. It is not apparent whyRSV infection in CF infants from Denver and Melbourneshould vary so markedly in severity and outcome. Dif-ferences in either severity among RSV subtypes21 ormanagement22 are unlikely explanations. Reliance uponoropharyngeal, rather than BAL, cultures meant that theinfants with lower airwayP. aeruginosainfection maynot have been identified.1,11 Alternatively, the randomeffects of studying small case numbers or selection ofsicker infants studied in a tertiary referral hospital mayexplain some of the variance in disease severity betweenthe two centers. The findings of the present study areconsistent with a recent Canadian study which did notidentify CF infants as being especially vulnerable to se-vere-life threatening RSV infection.23

Flexible bronchoscopy and BAL have become impor-tant research tools in the investigation of infant CF lungdisease.10–12,24 However, their role in patient manage-ment is undefined. This requires early resolution, as arapid diagnosis ofP. aeruginosainfection allows initia-tion of aggressive antimicrobial therapy and possibleeradication before the development of chronic infectionand irreversible lung injury.25 Unfortunately, coughswabs and oropharyngeal cultures do not reliably predictlower airway bacterial pathogens,11 and frequent or in-appropriate use of antipseudomonal antibiotics is associ-ated with emergence of resistant organisms.26 Instead, ithas been suggested that early BAL in young childrenwith CF may help identify the cause of persistent respi-ratory symptoms and ensure that therapy is directed at theinfecting pathogen.27,28 In the present study, BAL al-lowed for the early diagnosis of lower airway bacterialinfection in CF infants hospitalized with protracted re-spiratory symptoms. Furthermore, consistent with earlierobservations, two infants with negative NPA tests hadpositive immunofluorescence for respiratory viral anti-gens in their BAL specimens.29 Previously, bacterial in-fection was assumed to be present in all these hospital-ized CF patients, mandating at least a 7–10-day course ofinpatient parenteral antibiotic therapy. A BAL culture<105 CFU/ml in CF infants with a short history of coryzaand wheeze, and inspiratory crackles on auscultation, butwith a respiratory virus detected on NPA or BAL allowedparenteral antibiotics to be discontinued, and avoided re-peated intravenous line placements. Although not achiev-ing statistical significance in this study, earlier dischargefrom hospital may also be possible. Despite RSV beingrapidly identified, none of the CF infants received theantiviral agent, ribavirin. This is consistent with concerns

over its efficacy and safety in those with severe RSVdisease,28 and with a recently modified recommendationof its use in CF infants by the Committee on InfectiousDiseases of the American Academy of Pediatrics.30,31

Thirteen of the 31 subjects underwent BAL duringhospitalization. These infants had either protracted coughor mild respiratory distress. Infants hypoxic in room airor demonstrating moderate chest wall retraction were in-eligible. Bronchoscopy was well-tolerated. It was con-ducted in the operating room under general anesthesiagiven by an experienced pediatric anesthetist. Pulse rate,oxygen saturation, and in most cases expired carbon di-oxide were measured continuously, and 100% oxygenwas administered via a laryngeal mask. Minor adverseevents encountered during the study were occasional epi-sodes of transient laryngospasm during introduction ofthe bronchoscope, and decreased oxygen saturation at thetime of lavage. Laryngospasm responded rapidly to with-drawal of the bronchoscope and administration of higherconcentrations of halothane. Hypoxemia was brief andtransient, and oxygen saturation did not fall below 85%.No subject developed bradycardia or required endotra-cheal intubation or mechanical ventilation. All recoveredrapidly after BAL and were able to return to the respi-ratory ward from the postoperative recovery room.

Analysis of BAL inflammatory parameters demon-strated that viral lower respiratory infections stimulate anintense host reaction, characterized by an increase in totalcell count, composed predominantly of neutrophils. Theinflux of neutrophils is associated with markedly el-evated levels of the proinflammatory cytokine, IL-8, andincreased free neutrophil elastase activity. This pattern oflower airway inflammation is similar to that observedwith lower airway bacterial infection in CF infants.10–12

The marked neutrophil predominance in BAL fluid fromCF infants with RSV infection is also present in otherinfants with severe RSV bronchiolitis.32 Viral infectionof respiratory epithelial cells results in the release ofseveral cytokines and adhesion molecules, including IL-8, which recruit and retain neutrophils within the air-ways.33 We previously showed that following eradica-tion of respiratory pathogens, airway inflammation in CFinfants improves with time,12 but whether it is com-pletely reversible in this young patient population re-mains unknown. As diffusion of urea from the serum intoBAL fluid during bronchoscopy results in overestimationof epithelial lining fluid volume,18 we chose not to nor-malize our BAL inflammatory indices to urea. Our re-sults therefore cannot be directly compared to other pub-lished studies which have used this method.

Studies in older children show no evidence for viralrespiratory infections being more common in those withCF,2 but rather such infections are accompanied by acutedeterioration in lung function and an increased risk ofacquiringP. aeruginosa.1,5 We observed an association


between hospitalization during the first year of life andsubsequentP. aeruginosainfection which was not influ-enced by CF genotype. Many, but not all of these infantshad been admitted because of a viral infection. Whileintercurrent viral respiratory infections also result in air-way inflammatory changes,34 our experience withP. ae-ruginosasuggests that those with the severest symptomshave the greatest degree of lower respiratory tract inflam-mation.35 Injury to respiratory epithelial cells from eitheran inciting viral or bacterial infection, followed by re-current or persistent airway inflammation, may predis-pose to colonization and infection byP. aeruginosamonths or years later.36,37

The present study was confined to hospitalized infants.Ambulatory subjects were excluded, as not everyonewith mild respiratory symptoms had an NPA or wereseen at the CF clinic. Based on the findings of the presentstudy, earlier community-based studies in CF childrenwhich identified respiratory viruses in only a minority ofacute respiratory episodes may have underestimated theprevalence of viral infections.2–6 The long-term conse-quences of hospitalization for respiratory symptoms inCF infants is not yet known. Recent advances in infantpulmonary function testing suggest that such measure-ments may now be sufficiently reproducible to be usefulin monitoring lung disease in CF infants. Meanwhile, thechildren in this cohort are undergoing spirometry testingwhen they reach age 6 years. Using plain chest radio-graphs to follow CF lung disease in young children is oflimited value, and although radiological scores have beenreported to be persistently abnormal following RSV in-fection,7 this has not been our experience.10–12 Instead,low-dose, high-resolution computed tomographic imag-ing of the lungs, a sensitive technique for detecting bron-chiectasis and mucus impaction, may provide more use-ful information.

In conclusion, respiratory viruses, especially RSV, areimportant causes of hospitalization in CF infants. How-ever, in the present study when respiratory bacterialpathogens were absent, viral infections were self-limitedand not treated with parenteral antibiotics. Independentof etiology, lower respiratory infection was accompaniedby inflammation and, with hospitalization, there was anincreased risk of earlyP. aeruginosaacquisition result-ing in persistent respiratory symptoms and, in somecases, premature death.38,39

ACKNOWLEDGMENTS

David Armstrong was the recipient of a SmorgonClinical Research Fellowship. We thank Suzanna Vid-mar for assistance with statistical computing.

REFERENCES1. Petersen NT, Hoiby N, Mordhorst CH, Lind K, Flensborg EW,

Brunn B. Respiratory infections in cystic fibrosis patients caused

by virus, chlamydia and mycoplasma-possible synergism withPseudomonas aeruginosa.Acta Paediatr Scand 1981;70:623–628.

2. Wang EL, Prober CG, Manson B, Corey M, Levison H. Associa-tion of respiratory viral infections with pulmonary deterioration inpatients with cystic fibrosis. N Engl J Med 1984;311:1653–1658.

3. Hordvik NL, Konig P, Hamory B, Cooperstock M, Kreutz C,Gayer D, Barbero G. Effects of acute viral respiratory tract infec-tions in patients with cystic fibrosis. Pediatr Pulmonol 1989;7:217–222.

4. Smyth AR, Smyth RL, Tong CYW, Hart CA, Heaf DP. Effect ofrespiratory virus infections including rhinovirus on clinical statusin cystic fibrosis. Arch Dis Child 1995;73:117–120.

5. Collinson J, Nicholson KG, Cancio E, Ashman J, Ireland DC,Hammersley V, Kent J, O’Callaghan C. Effects of upper respira-tory tract infections in patients with cystic fibrosis. Thorax 1996;51:1115–1122.

6. Ramsey BW, Gore EJ, Smith AL, Cooney MK, Redding GJ, FoyH. The effect of respiratory viral infections on patients with cysticfibrosis. Am J Dis Child 1989;143:662–668.

7. Abman SH, Ogle JW, Butler-Simon N, Rumack CM, Accurso FJ.Role of respiratory syncytial virus in early hospitalisations forrespiratory distress of young infants with cystic fibrosis. J Pediatr1988;113:826–830.

8. Prober CG. The impact of respiratory viral infections in patientswith cystic fibrosis. Clin Rev Allergy 1991;9:87–102.

9. Balnaves ME, Bonaquisto L, Francis I, Glazner J, Forrest S. Theimpact of newborn screening on cystic fibrosis testing in Victoria,Australia. J Med Genet 1995;32:537–542.

10. Armstrong DS, Grimwood K, Carzino R, Carlin JB, Olinsky A,Phelan PD. Lower respiratory infection and inflammation in in-fants with newly diagnosed cystic fibrosis. Br Med J [Clin Res]1995;310:1571–1572.

11. Armstrong DS, Grimwood K, Carlin JB, Carzino R, Olinsky A,Phelan PD. Bronchoalveolar lavage or oropharyngeal cultures toidentify lower respiratory pathogens in infants with cystic fibrosis.Pediatr Pulmonol 1996;21:267–275.

12. Armstrong DS, Grimwood K, Carlin JB, Carzino R, Gutierrez JP,Hull J, Olinsky A, Phelan EM, Robertson CF, Phelan PD. Lowerairway inflammation in infants and young children with cysticfibrosis. Am J Respir Crit Care Med 1997;156:1197–1203.

13. Phelan PD, Bowes G. Cystic fibrosis in Melbourne. Thorax 1991;46:383–384.

14. Uren E, Eslum R, Jack I. A comparative study of the diagnosis ofrespiratory virus infections by immunofluorescence and virus iso-lation in children. Aust Paediatr J 1977;13:282–286.

15. Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH,editors. Manual of Clinical Microbiology, 6th ed. Washington,DC: American Society for Microbiology; 1995. p 246–662.

16. Saltini C, Hance AJ, Ferrans VJ, Basset F, Bitterman PB, CrystalRG. Accurate quantification of cells recovered by bronchoalveolarlavage. Am Rev Respir Dis 1984;130:650–658.

17. Bieth J, Spiess B, Wermuth CG. The synthesis and analytical useof a highly sensitive and convenient substrate of elastase. Bio-chem Med 1974;11:350–357.

18. Walters EH, Gardiner PV. Bronchoalveolar lavage as a researchtool. Thorax. 1991; 46:613–618.

19. StataCorp. Stata statistical software: release 4.0. College Station,TX: Stata Corporation. 1995.

20. Phelan PD, Olinsky A, Roberston CF. The epidemiology of acuterespiratory infections. In: Phelan PD, Olinsky A, Roberston CF,editors. Respiratory illness in children. Oxford: Blackwell Scien-tific Publications; 1994. p 27–51.

21. McIntosh DG, De Silva LM, Oates RK. Clinical severity of re-spiratory syncytial virus group A and B infection in Sydney, Aus-tralia. Pediatr Infect Dis J 1993;12:815–819.


22. South M, Henning R, Walker C, Uwyyed K. Intensive care man-agement of infants with severe bronchiolitis. Anaesth IntensiveCare 1994;22:229.

23. Wang EEL, Law BJ, Stephens D, other members of PICNIC.Pediatric investigators collaborative network on infections inCanada (PICNIC) prospective study of risk factors and outcomesin patients hospitalized with respiratory syncytial viral lower re-spiratory tract infection. J Pediatr 1995;126:212–219.

24. Ramsey BW, Boat TF. Outcome measures for clinical trials incystic fibrosis. J Pediatr 1994;124:177–192.

25. Frederiksen B, Koch C, Hoiby N. Antibiotic treatment of initialcolonization withPseudomonas aeruginosapostpones chronic in-fection and prevents deterioration of pulmonary function in cysticfibrosis. Pediatr Pulmonol 1997;23:330–335.

26. Denton M, Todd NJ, Littlewood JM. Role of anti-pseudomonalantibiotics in emergence ofStenotrophomonas maltophiliain cys-tic fibrosis patients. Eur J Clin Microbiol Infect Dis 1996;15:402–405.

27. Brownlee KG, Crabbe DCG. Paediatric bronchoscopy. Arch DisChild 1997;77:272–275.

28. Connett GJ, Doull IJM, Keeping K, Warner JO. Flexible fibre-optic bronchoscopy in the management of lung complications incystic fibrosis. Acta Paediatr 1996;85:675–678.

29. Derish MT, Kulhanjian JA, Frankel LR, Smith DW. Value ofbronchoalveolar lavage in diagnosing severe respiratory syncytialvirus infections in infants. J Pediatr 1991;119:761–763.

30. Moler FW, Steinhart CM, Ohmit SE, Stidham GL, for the Pedi-atric Critical Care Study Group. Effectiveness of ribavirin in oth-erwise well infants with respiratory syncytial virus-associated re-spiratory failure. J Pediatr 1996;128:422–428.

31. American Academy of Pediatrics. Section 3. Summaries of infec-tious diseases. In: Peter G, editor. Red book: report of the Com-

mittee on Infectious Diseases, 24th ed. Elk Grove Village, IL:American Academy of Pediatrics; 1997. p 443–447.

32. Everard ML, Swarbrick A, Wrightham M, McIntyre J, Dunkley C,James PD, Sewell HF, Milner AD. Analysis of cells obtained bybronchial lavage of infants with respiratory syncytial virus infec-tion. Arch Dis Child 1994;71:428–432.

33. Bonfield TL, Panusaka JR, Konstan MW, Hilliard KA, HilliardJB, Ghnaim H, Berger M. Inflammatory cytokines in cystic fibro-sis lungs. Am J Respir Crit Care Med 1995;152:2111–2118.

34. Gern JE, Galagan DM, Jarjour NN, Dick EC, Busse WW. Detec-tion of rhinovirus RNA in lower airway cells during experimen-tally induced infection. Am J Respir Crit Care Med 1997;155:1159–1161.

35. Grimwood K, Carzino R, Armstrong DS, Carlin JB, Gutierrez JP,Olinsky A, Phelan PD. population transcript accumulation ofPseudomonas aeruginosatoxA, lasB, and algD in bronchoalveo-lar lavage samples from young children with cystic fibrosis. AustN Z J Med 1996;26:605.

36. Birrer P, McElvaney NG, Rudeberg A, Wirz Sommer C, Leichti-Gallati S, Kraemer R, Hubbard R, Crystal RG. Protease-antiprotease imbalance in the lungs of children with cystic fibro-sis. Am J Respir Crit Care Med 1994;150:207–213.

37. de Bentzmann S, Plotkowski C, Puchelle E. Receptors in thePseudomonas aeruginosaadherence to injured and repairing air-way epithelium. Am J Respir Crit Care Med 1996;154:155–162.

38. Demko CA, Byard PJ, Davis PB. Gender differences in cysticfibrosis: Pseudomonas aeruginosainfection. J Clin Epidemiol1995;48:1041–1049.

39. Grimwood K, Armstrong D, Carlin J, Carzino R, Kyd J, Moore R,Olinsky A. Acquisition ofPseudomonas aeruginosain young in-fants. Pediatr Pulmonol [Suppl] 1997;14:133–134.


severe viral respiratory infections in infants with cystic fibrosis

Documents