Risk Factors for and Outcomes of Bloodstream Infection Caused by Extended-Spectrum β-Lactamase–Producing Escherichia coli and Klebsiella Species in Children
Objective. The increasing prevalence of infections caused by extended-spectrum β-lactamase–producing Escherichia coli and Klebsiella species (ESBL-EK) has become a growing concern in the hospitalized patient population. Previous studies on risk factors for infection with ESBL-EK have mainly focused on adult populations, and these findings may not be relevant among the pediatric population that experiences a unique set of health care exposures and underlying conditions. The objective of this study was to define the risk factors and outcomes associated with ESBL-EK bloodstream infections in children.
Methods. We conducted a nested case-control study using data from the Children's Hospital of Philadelphia from May 1, 1999, to September 30, 2003. Eligible patients were identified from the hospital database of microbiology laboratory records. All patients with ESBL-EK bloodstream infections were compared to a random sample of patients with non–ESBL-EK bloodstream infections. Risk factors analyzed included prior antimicrobial use, comorbid conditions, and demographic characteristics. Pulsed-field gel electrophoresis was performed to determine genetic relatedness of the ESBL-EK isolates.
Results. Thirty-five cases and 105 control subjects were included in the study. The median age among the cases was 2 years (interquartile range: 0–11), compared with 1 year (interquartile range: 0–8) among control subjects. Patients with ESBL-EK infections were 5.8 times (95% confidence interval: 1.9–17.7) more likely to have had exposure to an extended-spectrum cephalosporin in the 30 days before infection than those with non–ESBL-EK infections. Other independent predictors of ESBL-EK infection were being female, infection with a Klebsiella species, and steroid use in the 30 days before infection. All ESBL-EK isolates were susceptible to carbapenem antibiotics. Pulsed-field gel electrophoresis analysis revealed that the ESBL-EK isolates were polyclonal. Although a substantially higher proportion of children with ESBL-EK died (in-hospital mortality: 36% vs 13%), this difference was not statistically significant.
Conclusions. Receipt of extended-spectrum cephalosporins in the 30 days before infection by an Escherichia coli or Klebsiella species is significantly associated with having an ESBL-EK infection in hospitalized children. Curtailed use of cephalosporins among high-risk groups may reduce the occurrence of ESBL-EK infections. Future studies on identifying high-risk children and investigating the impact of curtailed third-generation cephalosporin use to limit additional emergence of ESBL-EK infections should be undertaken.
Extended-spectrum β-lactamase (ESBL)-producing enteric Gram-negative bacilli with resistance to broad-spectrum oxyimino β-lactams were first detected in Western Europe in the mid-1980s1 and subsequently in the United States by 1989.2–4 ESBLs are plasmid-mediated enzymes that confer resistance to all penicillins and cephalosporins, including the sulbactam and clavulanic acid combinations and monobactams such as aztreonam.5 ESBL production is seen predominantly in Escherichia coli and Klebsiella species (ESBL-EK) but may also be seen in other Enterobacteriaceae.5 An increasing prevalence of ESBL-producing isolates of K pneumoniae was noted during the past decade in the Centers for Disease Control and Prevention National Nosocomial Infection Surveillance study, particularly in ICUs.6 The prevalence of ESBL-producing isolates among K pneumoniae ranges from 12% in ICUs in the United States to 44% among global bacteremia isolates.7
Infections caused by ESBL-EK species are of great concern for many reasons. First, they are often difficult to treat because they are multidrug resistant.5 Second, patients with ESBL-EK infections may experience a delay in the initiation of appropriate therapy.5, 8, 9 Third, patients with ESBL-EK infections have been shown to have significantly longer hospital stays and higher costs than patients without these infections.8, 9 Fourth, current methods of identification can result in the underestimation of the prevalence of these organisms.10 Finally, previous studies have shown that both adult and pediatric patients with ESBL-EK infections have an increased risk for clinical failure and death compared to patients with non-ESBL infections.11, 12
Most studies that have examined risk factors for ESBL-EK infections have focused on adult populations, and these identified risk factors may not be relevant in children who experience different underlying comorbidities. Although outbreaks of ESBL-EK infection in children have been reported,9, 13–20 information on risk factors and outcomes for ESBL-EK infection in the pediatric population is limited.
The purpose of this study was to identify risk factors for bloodstream infection caused by ESBL-EK in children and to determine whether clinical outcomes differed between those with ESBL-EK versus non–ESBL-EK infections. Our secondary aims were to determine the antibiotic susceptibilities and patterns of transmission.
This case-control study was conducted at The Children's Hospital of Philadelphia (CHOP), which is an academic tertiary care center with 381 beds and ∼22000 admissions per year. Forty percent of the beds at CHOP are allocated to intensive care, and 42 beds are allocated to oncology. All case patients and control patients were identified retrospectively through records of the Clinical Microbiology Laboratory at CHOP, where screening for ESBL-EK is performed routinely on all blood culture isolates of E coli and Klebsiella species. All patients from whom blood culture results were positive for E colior Klebsiella species from May 1, 1999, to September 30, 2003, were eligible for inclusion in the study. Cases were defined as patients with a positive blood culture for ESBL-EK, and the control subjects were chosen through a 1:3 (case:control) random sample of all patients with a positive blood culture for non–ESBL-producing E coli and Klebsiella species bloodstream infection. All patients who were eligible to be control subjects were assigned a unique number on the basis of the medical record number and listed in numerical order. Subsequently, control subjects were selected through the use of a computer-generated random-number table. If multiple episodes of infection occurred during the study period, the patient was included as a study participant using only the first episode of infection.
Determination of ESBL Production
Minimum inhibitory concentrations were determined by antimicrobial susceptibility testing using Microscan Gram-negative microtiter panels (Dade Behring, Inc, Deerfield, IL). A potential ESBL-EK was defined by a zone of inhibition around a 10-μg cefpodoxime disk of ≤17 mm. Isolates that were identified as potential ESBL-EK underwent a confirmatory test using either the E-test ESBL screen (AB Biodisc, Solna, Sweden) or the double-disk diffusion test for synergy between clavulanic acid and both ceftazidime and cefotaxime. An increase of ≥5 mm in zone diameter for either antimicrobial agent tested in combination with clavulanic acid compared with its zone when tested alone was confirmation of the ESBL phenotype. Susceptibilities of the infecting organism to all antimicrobial agents were determined according to criteria of the National Committee for Clinical Laboratory Standards.18
Pulsed-field gel electrophoresis (PFGE) was used to determine the genetic similarities between all of the identified ESBL-EK isolates in the study and performed according to Gautom using Xbal digestion of genomic DNA (New England BioLabs, Beverly, MA) with slight modifications as follows.21 One hundred microliters of cell suspension (OD600 = 1.0–2.0) was transferred directly to 1.5-mL Eppendorf tubes and mixed with 100 μL of 1.2% InCert agarose. The bacterium-agarose mixture was added immediately to the plug molds. Plug slices were electrophoresed in 1% pulsed-field certified agarose (Bio-Rad, Hercules, CA). A CHEF Mapper apparatus (Bio-Rad) was used for electrophoresis with an initial switch time of 3 seconds and final switch time of 40 seconds, with a total run time of 21 hours. Each gel contained standard lanes using Lambda ladder PFGE marker (New England Biolabs). Band patterns were compared using Fingerprinting II Software, version 3.0 (Bio-Rad). Dendrograms were generated using the unweighted pair group method using arithmetic averages. Position tolerances were set at 1.0%; Dice coefficient was used for band similarity measurements.
Data on age, gender, race, hospital location, antimicrobial therapy in the 30 days preceding infection, length of stay in an ICU, comorbid conditions (including hepatic and renal dysfunction), previous surgery, previous trauma, mechanical ventilation, the presence of a central venous catheter, and the presence of a urinary catheter at the time of initial infection were abstracted from the medical record using a standardized data collection form. Microbiologic test results, including organism susceptibilities to antibiotics, any concurrent bacterial infections, and the presence of neutropenia at the time of initial infection, were also recorded.
Exposure to antibiotics was assessed using the following drug classifications: third-generation cephalosporins (cefotaxime, ceftazidime, and ceftriaxone), extended-spectrum penicillins, carbapenems, aminoglycosides, anti-anaerobic agents (defined as ticarcillin/clavulanate, piperacillin, ampicillin/sulbactam, imipenem, meropenem, metronidazole, clindamycin, and ceftriaxone), trimethoprim-sulfamethoxazole, and fluoroquinolones. Univariate analysis on potential risk factors for ESBL-EK infection was performed. Continuous variables were compared using the Wilcoxon rank sum test, and categorical variables were compared using the Fisher’s exact test. Multivariate analysis was performed using logistic regression to adjust for the presence of confounding. All variables with a P < .20 on univariate analysis were considered for inclusion in the multivariate model. The presence of confounding was defined by a difference of ≥15% in the odds ratio of the primary exposure, extended-spectrum cephalosporins, when adjusting for a given covariate. A 2-tailed P < .05 on multivariate analysis was considered statistically significant. All statistical calculations were performed using the statistical package STATA version 8.0 (Stata Corp, College Station, TX).
Human Subjects Oversight
The conduct of this study was approved by the Committees for the Protection of Human Subjects at CHOP and the University of Pennsylvania.
We identified a total of 296 E coli and Klebsiella bloodstream infections during the study period. Thirty-five (12%) were ESBL-EK. The prevalence of ESBL-producing isolates was 7% among all identified E coli isolates and 18% among all identified Klebsiella isolates. Of the 35 ESBL-EK cases, 8 were caused by E coli, 17 were caused by K pneumoniae, and 10 were caused by K oxytoca. A total of 105 control subjects with non–ESBL-EK bloodstream infection were randomly selected from the 261 non–ESBL-EK bloodstream infections originally identified. Patients who had significant portions of their medical chart missing or those who failed to meet inclusion criteria were excluded from the analysis and systematically replaced by another randomly selected patient. The medical chart retrieval rate was 92%.
Descriptive Statistics and Univariate Analysis
Patient demographics are listed in Table 1. The characteristics that were significantly associated with ESBL-EK infection on univariate analysis were infection with a Klebsiella species, mechanical ventilation at the time of infection, receipt of total parenteral nutrition at the time of infection, having a blood-related cancer (leukemia or lymphoma), previous organ transplantation, steroid and other immunosuppressant use in the 2 weeks before infection, and longer length of hospital stay before infection.
The presence of a central venous catheter or urinary catheter at the time of infection was not a significant predictor for an ESBL-EK infection, although a greater proportion of cases than control subjects had these catheters in place at time of infection. Similarly, co-infection with other bacterial organisms was also not statistically significant, although more cases did have a polymicrobial infection at the time of E coli or Klebsiella infection.
Previous antibiotic use was also assessed (Table 2), which revealed that exposure to third-generation cephalosporins, extended-spectrum penicillins, aminoglycosides, and trimethoprim-sulfamethoxazole in the previous 30 days all were significantly associated with infection by an ESBL-EK species on univariate analysis. Although a greater proportion of patients with ESBL infection were exposed to anti-anaerobic antibiotics and fluoroquinolones than patients with non-ESBL infections, these differences were not statistically significant.
Multivariate regression analysis revealed that cases were almost 6 times more likely to be exposed to third-generation cephalosporins in the 30 days before infection than control subjects (odds ratio: 5.82; 95% confidence interval: 1.92–17.68). After also controlling for type of malignancy and total parenteral nutrition before infection, other independent predictors of ESBL-EK bloodstream infection were female gender, infection by a Klebsiella species, and steroid use in the 30 days before infection (Table 3).
Antibiotic Susceptibility Profile of ESBL Organisms
The antibiograms of the 35 ESBL-EK isolates are shown in Fig 1. Imipenem was the only antibiotic agent to which all isolates were susceptible. The majority (80%) of isolates were resistant to gentamicin, with an additional 14% of isolates demonstrating intermediate resistance; 20% of ESBL-EK isolates were also resistant to fluoroquinolones.
The PFGE results revealed that for both E coli and Klebsiella species, there was little genetic similarity among the isolates (Fig 2). The mean values of genotypic similarity all were <50% within each individual species, suggesting a low likelihood of patient-to-patient transmission of ESBL-EK within the hospital.
Despite a higher proportion of cases than control subjects (36% vs 13%) who died during the same admission as that in which infection occurred, there was no statistically significant difference in all-cause mortality between patients with ESBL-EK and non–ESBL-EK bloodstream infection (P = .180). Although cases also experienced a longer mean length of hospital stay after first documentation of infection than control subjects (8.2 days), this difference was not statistically significant on univariate analysis (P = .182).
We found that the use of third-generation cephalosporins in the 30 days before infection was associated with infection by ESBL-EK species in hospitalized children. Our findings of an association between antibiotics and ESBL-EK infection are supported by the results of molecular typing studies. The ESBL-EK isolates were distinct strains, suggesting that minimal cross-transmission occurred between patients and that antibiotic selective pressure was the most important factor in the emergence of these isolates.
The antibiograms of the ESBL-EK isolates revealed a limited group of effective antibiotics for the treatment of infections caused by these organisms. Carbapenem-class antibiotics are the most reliably effective empirical therapies for these infections, whereas extended-spectrum penicillins are not likely to be successful in treating the infection. Gentamicin susceptibility was low among these isolates and should be used with caution when treating empirically. In addition, 20% of patients with ESBL-EK infection were found to be resistant to ciprofloxacin. This is a surprising finding because, although the most commonly identified risk for fluoroquinolone resistance is recent use of a fluoroquinolone, ciprofloxacin is rarely used in the routine clinical practices of our pediatric hospital. Fluoroquinolone resistance in children with ESBL-EK may provide a unique opportunity to elucidate the emergence of fluoroquinolone resistance.
Our findings are similar to those reported by investigators in Taiwan, who found that previous use of extended-spectrum cephalosporins was associated with the development of ESBL-EK bloodstream infection in children.11 In addition, a case-control study that investigated an outbreak of ESBL-producing Klebsiella infections in a pediatric ICU transplant population did not find an association between extended-spectrum cephalosporins and ESBL infection. However, exposure to extended-spectrum cephalosporins was rare in that study.18
Past attempts in adult populations to identify risk factors for infection caused by ESBL-producing organisms have come to very different conclusions. This may be partly explained by previous studies failing to distinguish colonization from such pathogens with infection.4, 22–24 Recent reports have suggested that risk factors for ESBL-EK colonization may differ substantially from risk factors for ESBL-EK infection.25 In support of this, 2 studies that were performed by our group in an adult patient population found discrepant results when investigating colonization versus infection.8, 26 In a matched case-control study, Lautenbach et al8 found that total antibiotic exposure was the only independent predictor of ESBL-EK infection. However, the same investigators found no association between prior antibiotic use and ESBL colonization in the same institution.26 To identify children with an increased likelihood of developing a clinical infection caused by ESBL-EK, we have emphasized the elucidation of those factors that predict true infection.
Previous studies in adults and children indicate that ESBL-EK invasive infections are associated with increased morbidity and mortality, largely as a result of the absence of or delayed appropriate antibiotic therapy.11, 12, 27, 28 In our analysis, children with ESBL-EK infections experienced higher rates of mortality and longer lengths of stay after infection than children with non–ESBL-EK infections. However, these differences were not statistically significant. A small sample size may have hampered our ability to detect significant differences in mortality between cases and control subjects.
Previous molecular typing studies have documented that isolates from outbreaks have represented both clonal and polyclonal strains.2, 29–31 Clonal strain transmission emphasizes that infection control measures are important and such measures have interrupted the spread of ESBL-producing organisms.25 In addition, antibiotic control measures, specifically decreases in ceftazidime use, have been effective in controlling ESBL emergence in several instances.22, 32 The finding of female gender as a significant risk factor for ESBL-EK infection is puzzling but may indicate that gender is a surrogate for another important risk factor that was unobserved in this study. In addition, the finding that ESBL-producing isolates are more common among Klebsiella species may be a result of a higher predisposition toward the development of antibiotic resistance on a molecular level.
There were several potential limitations to our study. Although the possibility of selection bias is normally of concern in a case-control study, all cases and control subjects were identified through the same microbiology laboratory records. All patients with ESBL-EK infection and a random sample of patients who were not infected with ESBL-EK species were included. Thus, the potential for selection bias is expected to be small. Harris et al33 demonstrated that the choice of control group in case-control studies affects the identification of use of antimicrobial agents as risk factors and the magnitude of the effect. Furthermore, Harris et al suggested that control groups that consist of patients with an antibiotic-susceptible form of the organism may not be appropriate. However, we propose that control group selection depends on the question being asked. In our study, we wanted to identify patients who had clinical infection with E coli or Klebsiella and had an increased likelihood of having an ESBL-EK infection. Therefore, we chose the control group that would best allow us to answer this specific question.
Although misclassification bias is likewise of concern in case-control studies, cases and control subjects were drawn from the same hospitalized patient population and were identified solely on whether ESBL-EK was isolated from culture. Because these tests were conducted without previous knowledge of the patient's status regarding possible exposures of interest, there was unlikely to be any differential misclassification bias.
Despite a high medical chart retrieval rate, the study may also have been limited by the restricted ability to capture all outpatient data on antibiotic exposures, particularly for cases and control subjects who developed infection within 1 week of admission. However, the majority of children who are seen at our hospital are chronically ill and have multiple previous inpatient visits, for which home medications are documented in the discharge record of the previous inpatient stay. In addition, we would not expect any differential ascertainment of exposure information between cases and control subjects that would bias our results away from the null hypothesis. Finally, our study was conducted at a large, academic, tertiary care children's hospital, and our results may not reflect those at other, dissimilar institutions.
We found that previous exposure to third-generation cephalosporins in the 30 days before infection, female gender, infection by a Klebsiella species, and previous steroid use all were significant risk factors for the development of ESBL-EK bacteremia in hospitalized children. Furthermore, molecular epidemiologic analysis revealed that the ESBL-EK isolates were not closely related. These results suggest that the incidence of ESBL-EK infections in children may be reduced with curtailed use of third-generation cephalosporins among high-risk groups. Future studies on identifying children who are at high risk and investigating the impact of curtailed third-generation cephalosporin use to limit additional emergence of ESBL-EKs should be undertaken.
This study was supported by the Centers for Education and Research on Therapeutics grant from the Agency for Healthcare Research and Quality (U18-HS10399), Merck and Co. Inc, and an educational grant from Elan Pharmaceuticals, Inc.
We thank Huong Ung for performing PFGE and Xiangquin Mao for antimicrobial susceptibility testing.
- Accepted August 30, 2004.
- Reprint requests to (T.E.Z.) Division of Infectious Diseases, Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, PA 19104. E-mail:
No conflict of interest declared.
- ↵Rice LB, Willey SH, Papanicolaou GA, et al. Outbreak of ceftazidime resistance caused by extended-spectrum beta-lactamases at a Massachusetts chronic-care facility. Antimicrob Agents Chemother.1990;34 :2193– 2199
- Quinn JP, Miyashiro D, Sahm D, Flamm R, Bush K. Novel plasmid-mediated beta-lactamase (TEM-10) conferring selective resistance to ceftazidime and aztreonam in clinical isolates of Klebsiella pneumoniae. Antimicrob Agents Chemother.1989;33 :1451– 1456
- ↵Diekema DJ, Pfaller MA, Jones RN, et al. Survey of bloodstream infections due to gram-negative bacilli: frequency of occurrence and antimicrobial susceptibility of isolates collected in the United States, Canada, and Latin America for the SENTRY Antimicrobial Surveillance Program, 1997. Clin Infect Dis.1999;29 :595– 607
- ↵Lautenbach E, Patel JB, Bilker WB, Edelstein PH, Fishman NO. Extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae: risk factors for infection and impact of resistance on outcomes. Clin Infect Dis.2001;32 :1162– 1171
- ↵Tenover FC, Mohammed MJ, Gorton TS, Dembek ZF. Detection and reporting of organisms producing extended-spectrum beta-lactamases: survey of laboratories in Connecticut. J Clin Microbiol.1999;37 :4065– 4070
- ↵Kim YK, Pai H, Lee HJ, et al. Bloodstream infections by extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae in children: epidemiology and clinical outcome. Antimicrob Agents Chemother.2002;46 :1481– 1491
- ↵Paterson DL, Ko WC, Von Gottberg A, et al. Outcome of cephalosporin treatment for serious infections due to apparently susceptible organisms producing extended-spectrum beta-lactamases: implications for the clinical microbiology laboratory. J Clin Microbiol.2001;39 :2206– 2212
- Royle J, Halasz S, Eagles G, et al. Outbreak of extended spectrum beta lactamase producing Klebsiella pneumoniae in a neonatal unit. Arch Dis Child Fetal Neonatal Ed.1999;80 :F64– F68
- Hibbert-Rogers LC, Heritage J, Gascoyne-Binzi DM, et al. Molecular epidemiology of ceftazidime resistant Enterobacteriaceae from patients on a paediatric oncology ward. J Antimicrob Chemother.1995;36 :65– 82
- Siu LK, Lu PL, Hsueh PR, et al. Bacteremia due to extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae in a pediatric oncology ward: clinical features and identification of different plasmids carrying both SHV-5 and TEM-1 genes. J Clin Microbiol.1999;37 :4020– 4027
- ↵Rebuck JA, Olsen KM, Fey PD, Langnas AN, Rupp ME. Characterization of an outbreak due to extended-spectrum beta-lactamase-producing Klebsiella pneumoniae in a pediatric intensive care unit transplant population. Clin Infect Dis.2000;31 :1368– 1372
- Szabo D, Filetoth Z, Szentandrassy J, et al. Molecular epidemiology of a cluster of cases due to Klebsiella pneumoniae producing SHV-5 extended-spectrum beta-lactamase in the premature intensive care unit of a Hungarian hospital. J Clin Microbiol.1999;37 :4167– 4169
- ↵Gautom RK. Rapid pulsed-field gel electrophoresis protocol for typing of Escherichia coli O157:H7 and other gram-negative organisms in 1 day. J Clin Microbiol.1997;35 :2977– 2980
- ↵Rice LB, Eckstein EC, DeVente J, Shlaes DM. Ceftazidime-resistant Klebsiella pneumoniae isolates recovered at the Cleveland Department of Veterans Affairs Medical Center. Clin Infect Dis.1996;23 :118– 124
- Piroth L, Aube H, Doise JM, Vincent-Martin M. Spread of extended-spectrum beta-lactamase-producing Klebsiella pneumoniae: are beta-lactamase inhibitors of therapeutic value? Clin Infect Dis.1998;27 :76– 80
- ↵Pena C, Pujol M, Ardanuy C, et al. Epidemiology and successful control of a large outbreak due to Klebsiella pneumoniae producing extended-spectrum beta-lactamases. Antimicrob Agents Chemother.1998;42 :53– 58
- ↵Lucet JC, Chevret S, Decre D, et al. Outbreak of multiply resistant Enterobacteriaceae in an intensive care unit: epidemiology and risk factors for acquisition. Clin Infect Dis.1996;22 :430– 436
- ↵Bisson G, Fishman NO, Patel JB, Edelstein PH, Lautenbach E. Extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella species: risk factors for colonization and impact of antimicrobial formulary interventions on colonization prevalence. Infect Control Hosp Epidemiol.2002;23 :254– 260
- ↵Schiappa DA, Hayden MK, Matushek MG, et al. Ceftazidime-resistant Klebsiella pneumoniae and Escherichia coli bloodstream infection: a case-control and molecular epidemiologic investigation. J Infect Dis.1996;174 :529– 536
- ↵Nouvellon M, Pons JL, Sirot D, Combe ML, Lemeland JF. Clonal outbreaks of extended-spectrum beta-lactamase-producing strains of Klebsiella pneumoniae demonstrated by antibiotic susceptibility testing, beta-lactamase typing, and multilocus enzyme electrophoresis. J Clin Microbiol.1994;32 :2625– 2627
- Gazouli M, Kaufmann ME, Tzelepi E, Dimopoulou H, Paniara O, Tzouvelekis LS. Study of an outbreak of cefoxitin-resistant Klebsiella pneumoniae in a general hospital. J Clin Microbiol.1997;35 :508– 510
- ↵Bingen EH, Desjardins P, Arlet G, et al. Molecular epidemiology of plasmid spread among extended broad-spectrum beta-lactamase-producing Klebsiella pneumoniae isolates in a pediatric hospital. J Clin Microbiol.1993;31 :179– 184
- ↵Harris AD, Samore MH, Lipsitch M, Kaye KS, Perencevich E, Carmeli Y. Control-group selection importance in studies of antimicrobial resistance: examples applied to Pseudomonas aeruginosa, Enterococci, and Escherichia coli. Clin Infect Dis.2002;34 :1558– 1563
- Copyright © 2005 by the American Academy of Pediatrics