International Journal of Antimicrobial Agents 23 (2004) 401–404

Frequency of Gram-negative bacterial pathogens in bloodstreaminfections and their resistance to antibiotics in the Czech Republic

PavelCermáka, Milan Kolár b,∗, Tomáš Látalc,on behalf of the Research Group1

a Institute of Clinical Microbiology, Faculty Hospital, Hradec Králové, Czech Republicb Institute of Microbiology, Faculty of Medicine, Palacký University, Street Hnevot´ınská 3, 775 15, Olomouc, Czech Republic

c Trios, spol. s.r.o., Zakouˇrilova 142, Prague, Czech Republic

Received 7 April 2003; accepted 29 July 2003

Abstract

A study performed at 12 hospitals in the Czech Republic in 2001 evaluated the Gram-negative bacterial pathogens most frequently associatedwith bloodstream infections and their susceptibility to a selection of antimicrobial agents. Of 831 Gram-negative strains, the most frequentlyisolated organisms wereEscherichia coli(32%),Klebsiella pneumoniae(24%) andPseudomonas aeruginosa(10%).E. coli isolates wererelatively susceptible to the antibiotics tested, whereasK. pneumoniaewere relatively resistant to all agents except meropenem, andP.aeruginosato all agents except gentamicin and amikacin. Other agents showed variable rates of resistance to penicillins, third-generationcephalosporins, aminoglycosides and ciprofloxacin.© 2004 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

Keywords:Bloodstream infection; Gram-negative pathogens; Frequency; Antibiotic resistance

1. Introduction

Community- and hospital-acquired bloodstream infec-tions (BSIs) due to bacterial pathogens are an importantsource of morbidity and mortality world-wide. These infec-tions are most frequently hospital-acquired and associatedwith patients in intensive care units. BSI accounts for

∗ Corresponding author. Tel.:+420-58-563-2407; fax:+420-58-5632966.

E-mail address:kolar@fnol.cz (M. Kolar).1 The members of the Research Group are: D. Burgetova (Department

of Microbiology, Teaching Hospital, Brno Bohunice); P.Cermak (Insti-tute of Clinical Microbiology, Teaching Hospital, Hradec Kralove); E.Chmelarova (Department of Microbiology, KHS Ostrava); M. Dovalova(Department of Microbiology, Hospital Opava); B. Heinigeova (Depart-ment of Microbiology, Hospital Jindrichuv Hradec); B. Horova (Depart-ment of Microbiology, Teaching Hospital Bulovka, Praha); H. Kremeckova(Department of Microbiology, Hospital Kyjov); M. Kolar (Institute of Mi-crobiology, Faculty of Medicine, Palacky University, Olomouc); T. Latal(Trios, spol. s.r.o., Praha); O. Nyc (Department of Microbiology, TeachingHospital Motol, Praha); V. Petkov (Department of Microbiology, IKEM,Praha), B. Prokupkova (Department of Microbiology, Hospital Kladno);and M. Štolbova (Department of Microbiology, HospitalUstı nad Labem).

approximately 10% of nosocomial infections and has anattributable mortality rate of approximately 15%[1].

The seriousness of BSIs necessitates effective antibacte-rial therapy. Since it is vital that antibacterial therapy beinitiated promptly upon admission or diagnosis, therapy isusually empirical and must be based on detailed knowledgeof the most likely bacterial pathogens and their suscepti-bilities to the antibiotic agents at the physician’s disposal.Gram-negative organisms pose a higher risk of sepsis andmortality than Gram-positive organisms[2–4] and requireclose surveillance.

The focus of this study was to evaluate Gram-negativepathogens most frequently associated with BSIs in the CzechRepublic and to characterize their susceptibility to a selec-tion of antimicrobial agents.

2. Materials and methods

2.1. Collection of isolates

The study was conducted at 12 hospitals coveringthe whole of the Czech Republic during the period

0924-8579/$ – see front matter © 2004 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.doi:10.1016/j.ijantimicag.2003.07.021

402 P. Cermak et al. / International Journal of Antimicrobial Agents 23 (2004) 401–404

June–October 2001. Bloodstream infection was char-acterized by bacteraemia and the following clinicalsigns: tremor, temperature >38◦C (hyperthermia) or<35.6◦C (hypothermia), pulse >90 min−1, tachypnoea (>20breaths min−1), left-shifted leukocytosis (>12,000 mm−3,>10% immature forms) and/or leukopenia (<4000 mm−3)and inflammatory markers (e.g. C-reactive protein,interleukins).

The maximum number of consecutive Gram-negative iso-lates complying with the study specification that were iso-lated from non-polymicrobial samples collected at each hos-pital was 100. Patients were routinely required to provideblood samples. Bloodstream infection was confirmed usingautomated systems (Bactec, Becton Dickinson; BacT/Alert,BioMerieux).

2.2. Identification of isolates andsusceptibility testing

Gram-negative bacterial strains were identified usingstandard microbiological procedures as well as supplemen-tary tests[5]. Strains of Gram-negative rods that couldbe connected to bloodstream infection by the attendingphysician were eligible for antibiotic susceptibility anal-ysis. Resistance to antimicrobial agents was assessed us-ing standard methods of microdilution[6]. The followingbreakpoint values were used: ciprofloxacin, 1 mg/l; ce-fotaxime, ceftazidime, cefepime, meropenem, gentamicinand netilmicin, 4 mg/l; ampicillin–sulbactam, amikacin,cefoperazone and cefoperazone–sulbactam, 8 mg/l (in thecases of ampicillin–sulbactam and cefoperazone–sulbactam,the concentration refers to the ampicillin or cefopera-zone component); piperacillin and piperacillin–tazobactam,16 mg/l (the concentration refers to piperacillin). ForPseu-domonas aeruginosa, a value of 64 mg/l for piperacillin andpiperacillin–tazobactam was used as the breakpoint.E. coliATCC 25922,E. coli ATCC 35218 andP. aeruginosaATCC27853 were used as reference strains for quality-controlprotocols.

Table 1Resistance of Gram-negative bacterial isolates to selected antimicrobial agents

Organism Resistance (% isolates)

AMS PIP PPT CPR CTX CTZ CPS CPM MER CIP GEN NET AMI

Escherichia coli 21.2 27.7 3.0 7.2 0.4 0.8 2.3 1.5 0.0 6.8 6.1 4.5 2.3Klebsiella pneumoniae 53.3 56.3 33.5 39.6 18.8 29.4 24.4 18.3 1.0 31.5 47.2 41.6 22.3Pseudomonas aeruginosa 100.0 25.3 25.3 31.3 97.6 38.6 31.3 50.6 37.3 47.0 2.5 21.7 4.8Enterobacter cloacae 98.1 44.4 40.7 38.9 42.6 38.9 38.9 5.6 1.9 3.7 1.9 1.9 0.0Klebsiella oxytoca 34.1 31.7 22.0 31.7 7.3 4.9 12.2 0.0 0.0 7.3 14.6 4.9 4.9Acinetobacterspp. 22.6 45.2 25.8 87.1 58.1 38.7 9.7 38.7 0.0 35.5 32.3 16.1 12.9Salmonella enteritidis 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0Citrobacter freundii 96.8 65.6 62.5 53.1 56.3 53.1 50.0 16.7 0.0 20.8 25.0 25.0 8.3Proteus mirabilis 10.5 31.6 0.0 10.5 0.0 0.0 0.0 0.0 0.0 21.1 47.4 36.8 0.0Serratia marcescens 94.7 73.7 52.6 36.8 15.8 26.3 26.3 5.3 0.0 26.3 68.4 68.4 10.5

AMS, ampicillin–sulbactam; PIP, piperacillin; PPT, piperacillin–tazobactam; CPR, cefoperazone; CTX, cefotaxime; CTZ, ceftazidime; CPS,cefoperazone–sulbactam; CPM, cefepime; MER, meropenem; CIP, ciprofloxacin; GEN, gentamicin; NET, netilmicin; AMI, amikacin.

3. Results

A total of 831 strains were isolated from 831 patients(432 men, 399 women) investigated within the study period.Polymicrobial infections were not included. Most patients(75%) were aged≥50 years. The period of time betweenthe date of initial hospitalization and the collection of bloodsample was≥3 days in 773 (93%) patients.

The most frequently identified Gram-negative bacterialrods causing bloodstream infections wereE. coli (31.8%of isolates),Klebsiella pneumoniae(23.7%) andP. aerug-inosa (10.0%). Other species with an incidence of≥2.0%includedEnterobacter cloacae(6.5%), Klebsiella oxytoca(4.9%), Acinetobacterspp. (4.6%),Citrobacter freundii(3.9%), Salmonella enteritidis(3.5%), Proteus mirabilis(2.3%) andSerratia marcescens(2.3%).

All eligible isolates were evaluated for susceptibility toantimicrobial agents. Susceptibility data for the 10 most fre-quently isolated organisms are shown inTable 1. The min-imal inhibitory concentrations for 50 and 90% of isolatesof E. coli, K. pneumoniae, andP. aeruginosaare given inTable 2.

The results showed high variability in susceptibility ratesamong pathogens and among antimicrobial agents. Withregard to pathogens,E. coli appeared generally resistant topenicillins but susceptible to cephalosporins, meropenem,ciprofloxacin and aminoglycosides. The addition of tazobac-tam to piperacillin considerably restored the activity ofthe latter.K. pneumoniae, C. freundii, and Acinetobacterspecies proved moderately resistant to all agents exceptmeropenem;Acinetobacterremained relatively susceptibleto cefoperazone–sulbactam.

P. aeruginosa, the third most frequently isolated Gram-negative rod, demonstrated considerable resistance to allagents tested except amikacin, whereasE. cloacaewasresistant to penicillins and most of the cephalosporins butwas susceptible to cefepime, meropenem, ciprofloxacin andthe aminoglycosides.K. oxytocawas resistant to penicillinsbut susceptible to the newer cephalosporins (cefotaxime,

P. Cermak et al. / International Journal of Antimicrobial Agents 23 (2004) 401–404 403

Table 2Minimal inhibitory concentrations (MICs) for 50% (MIC50) and 90% (MIC90) of isolates ofEscherichia coli, Klebsiella pneumoniaeand Pseudomonasaeruginosa

Antimicrobial agent Escherichia coli Klebsiella pneumoniae Pseudomonas aeruginosa

MIC50 MIC90 MIC50 MIC90 MIC50 MIC90

AMS 4 16 16 32 32 64PIP 2 256 128 256 8 256PPT 2 8 4 256 8 256CPR 0.5 8 4 64 8 32CTX 0.125 0.5 0.125 8 16 32CTZ 0.125 1 0.5 16 4 16CPS 2 4 2 64 8 64CPM 0.25 0.5 0.25 16 8 32MER 0.25 0.5 0.5 0.5 4 32CIP 0.125 0.5 0.125 16 1 16GEN 0.5 2 0.5 32 0.5 32NET 1 2 1 32 2 32AMI 2 8 2 32 1 8

AMS, ampicillin–sulbactam; PIP, piperacillin; PPT, piperacillin–tazobactam; CPR, cefoperazone; CTX, cefotaxime; CTZ, ceftazidime; CPS, cefoperazone–sulbactam; CPM, cefepime; MER, meropenem; CIP, ciprofloxacin; GEN, gentamicin; NET, netilmicin; AMI, amikacin.

ceftazidime, cefoperazone–sulbactam) and remaining ag-ents, while P. mirabilis was resistant to ciprofloxacin,gentamicin and netilmicin but susceptible to amikacin.S.marcescenswas resistant to all agents except meropenemand amikacin. OnlyS. enteritidiswas susceptible to allagents tested.

From the perspective of antimicrobial agents, mostpathogens were resistant to the penicillins studied. Theactivity of cephalosporins against Gram-negative organ-isms was mixed: cefepime was often more active than theother cephalosporins, but had high resistance rates for 4of the 10 most frequently isolated organisms. High levelsof ciprofloxacin resistance were recorded for many of thepathogens. Resistance rates for ciprofloxacin tended to cor-respond with rates of resistance to the aminoglycosides gen-tamicin and netilmicin. Among the aminoglycosides, onlyamikacin was active against most organisms. Meropenemproved the most active agent for the 10 most frequentlyisolated organisms; onlyP. aeruginosademonstrated a highrate of resistance to this agent.

4. Discussion

Nosocomial BSIs are currently associated with an increas-ing frequency of bacterial pathogens characterized by a highlevel of resistance to antibiotics[7–10]. Therefore, it is clearthat the choice of antibiotics for the treatment of BSIs mustbe carefully considered and well informed.

The choice of antibiotic regimen should be based on,among other factors, microbiological investigation of the pa-tient, the frequency of bacterial pathogens in his/her blood,and the antibiotic susceptibility profiles of suspected blood-stream pathogens, based on local and up-to-date surveil-lance data. The multicentre study described above evaluatedthe antibiotic susceptibility profile of Gram-negative BSI

pathogens from Czech patients. No single centre was per-mitted to provide a large proportion of isolates for analysis,so as to ensure that the data were representative of the wholeof the Czech Republic.

The frequencies of occurrence of pathogens were largelyin agreement with those reported from large internationalsurveillance studies[7–10]. Escherichia coli, K. pneu-moniaeand P. aeruginosawere the three most commonGram-negative pathogens, accounting for more than 60%of the Gram-negative isolates. Susceptibility data for 13antibiotic agents available in the Czech Republic for thetreatment of BSIs indicated wide variations in susceptibilityamong agents and among organisms. Strains ofE. coli wererelatively susceptible to most of the agents tested, beingless susceptible to ampicillin–sulbactam and piperacillin,whereasK. pneumoniaeisolates were resistant to all agentsexcept meropenem, andP. aeruginosaisolates to all agentsexcept gentamicin and amikacin. Other organisms showedvariable rates of resistance to penicillins, third-generationcephalosporins, aminoglycosides and ciprofloxacin.

While the relatively high susceptibility ofE. coli isolatesto all tested antibiotics in the present study is encouraging,the higher frequency ofK. pneumoniaeisolates resistantto third-generation cephalosporins, aminoglycosides andciprofloxacin is alarming. Given the simultaneous resistanceof these strains to both aminoglycosides and third-generationcephalosporins, it is reasonable to suppose that many ofthe strains are capable of producing the AmpA extended-spectrum�-lactamases (ESBL). High-level resistance tothird-generation cephalosporins was also documented forE. cloacaeandC. freundii. However, the resistance of thesestrains is likely to have been caused by the overproductionof broad-spectrum AmpC�-lactamases.

The need to consider local bacterial prevalence patternsand antibiotic susceptibility data in selecting initial em-pirical therapy for BSIs is clear. The results of this study

404 P. Cermak et al. / International Journal of Antimicrobial Agents 23 (2004) 401–404

demonstrate high levels of antibiotic resistance among fre-quently encountered Gram-negative BSI pathogens. Thisreinforces the need to consider the use of antimicrobialagents that are active against these organisms.

Acknowledgements

The study was designed and organized by Trios, spol.s.r.o., and was performed and supported by a grant (ATB PF2001) from Pfizer, spol. s.r.o.

References

[1] Wenzel RP, Edmond MB. The impact of hospital-acquired blood-stream infections. Emerg Infect Dis 2001;7:174–7.

[2] Sofianou DC, Constandinidis TC, Yannacou M, Anastasiou H, Sofi-anos E. Analysis of risk factors for ventilator-associated pneumoniain a multidisciplinary intensive care unit. Eur J Clin Microbiol InfectDis 2000;19:460–3.

[3] Gómez J, Simarro E, Baños V, Requena L, Ruiz J, Garcıa F, et al.Six-year prospective study of risk and prognostic factors in patientswith nosocomial sepsis caused byAcinetobacter baumannii. Eur JClin Microbiol Infect Dis 1999;18:358–61.

[4] Holzheimer RG. Antibiotic induced endotoxin release and clinicalsepsis: a review. J Chemother 2001;13:159–72.

[5] Holt JG, Krieg NR, Sneath PH, Staley JT, Williams ST. Bergey’sManual of Determinative Bacteriology. 9th ed. Baltimore, Maryland:Williams & Wilkins; 1994.

[6] National Committee for Clinical Laboratory Standards. Performancestandards for antimicrobial susceptibility testing. Sixth informa-tional supplement. Villanova, PA: Nat Comm Clin Lab Standards;1995.

[7] Richards MJ, Edwards JR, Culver DH, Gaynes RP. Nosocomial in-fections in medical intensive care units in the United States. Na-tional Nosocomial Infection Surveillance System. Crit Care Med1999;27:887–92.

[8] Fluit AC, Schmitz FJ, Verhoef J. Frequency of isolation of pathogensfrom bloodstream, nosocomial pneumonia, skin and soft tissue, andurinary tract infections occurring in European patients. Eur J ClinMicrobiol Infect Dis 2001;20:188–91.

[9] Diekema DJ, Pfaller MA, Jones RN, Doern GV, Kugler KC,Beach ML, et al. Trends in antimicrobial susceptibility of bacterialpathogens isolated from patients with bloodstream infections in theUSA, Canada and Latin America. SENTRY Participants Group. IntJ Antimicrob Agents 2000;13:257–71.

[10] Pfaller MA, Jones RN, Doern GV, Kugler K. Bacterial pathogensisolated from patients with bloodstream infection: frequenciesof occurrence and antimicrobial susceptibility patterns from theSENTRY antimicrobial surveillance program (United States andCanada, 1997). Antimicrob Agents Chemother 1998;42:1762–70.