Bloodstream infections caused by extended spectrum β-lactamases (ESBL)-producing Enterobacteriaceae often have unfavorable outcome in terms of survival, complications, length of hospital stay, and expenses compared with infections caused by non-ESBL isolates, in both adults and children.1–8 However, the outcome may differ depending on the site of infection. Hyle et al6 showed that inadequate initial antimicrobial therapy was an independent risk factor for mortality for nonurinary infections caused by ESBL-producing bacteria but not for urinary tract infections (UTIs) caused by such isolates in adults. There are few data on the outcome of UTIs caused by ESBL-producing bacteria in infants and children. The purpose of the present study was therefore to compare the clinical, microbiologic, and imaging outcomes of UTIs caused by ESBL-producing bacteria with that of UTIs caused by non-ESBL-producing bacteria in pediatric patients.
A matched case-control study was conducted in a tertiary-care level general hospital. The medical records of all children admitted with UTI between January 2001 and December 2008 were retrospectively reviewed.
Inclusion criteria were age ≤14 years and urine culture positive for bacterial species. Patients with a positive urine culture for ESBL-producing isolates were defined as cases. Patients of the same gender and similar age group, who were admitted before and after each case patient and had a positive urine culture for non-ESBL-producing isolates, were defined as controls. We intended that age difference between cases and controls should be <20% of cases' age. Every case patient was matched with 2 control patients with priority to gender, then admission date, and then age.
The exclusion criterion was lack of more than one group of patient's data. These were recorded and organized in the following 3 groups: clinical information (age, gender, presenting symptoms, antimicrobial agent administration, day of symptoms' resolution, length of hospital stay), microbiologic information (species identification and susceptibility, results of follow-up urine cultures), and imaging data (voiding cystourethrography, 6-month 99Tc DMSA renal scanning). According to the purpose of this study, there was no need to exclude patients with previous UTIs.
Urine samples were obtained with suprapubic aspiration or bladder catheterization for children <3 years old and by the clean-catch method for older children; positivity was interpreted according to standard criteria. Identification and susceptibility testing of microorganisms was performed with Vitek 2 automated system (Biomerieux, France). Antimicrobial susceptibilities and ESBL production were determined according to Clinical and Laboratory Standards Institute guidelines.9
Antimicrobial therapy was classified as empiric (administered before results of urine cultures) or definite (following antimicrobial susceptibility results) and as appropriate or inappropriate, depending on whether the causative organism was susceptible or resistant, respectively, to any of the prescribed antimicrobial agents.
Outcome was evaluated with regard to clinical response (days from therapy initiation to resolution of fever or other symptoms), microbiologic response (days to first negative urine culture after treatment initiation), and results of 99Tc DMSA scanning. According to standard guidelines and the clinicians' judgment, children with UTI underwent voiding cystourethrography and the febrile ones underwent renal scanning with 99Tc DMSA 6 months after the UTI episode.
Continuous data were analyzed with Student t test (parametric) or Mann-Whitney U test (nonparametric) and categorical variables were compared using the Mantel-Haenszel test. P values were 2-tailed (statistically significant: P < 0.05). Statistical analyses were performed with Epi Info (version 3.5.1).
A total of 28 cases and 56 controls were included in this study; demographic characteristics are presented in Table 1. Our initial goal was for age difference between cases and controls not to exceed 20% of cases' age. However, as priority was given to gender and admission date matching and the number of included patients was relatively small, the initial goal was not always achieved. Provided that gender and admission date matching was satisfied, age matching was conducted with caution that controls had the closest possible age with cases. As a result, every patient was matched to 2 controls of the same age group, that is, infants were matched to infants and toddlers to toddlers, with caution that ages differed the least possible.
Escherichia coli and Klebsiella pneumoniae were the isolated uropathogens in all cases and most of the controls. There were significant differences in the prevalence of these species between the 2 patient groups (Table 1). Additional species isolated from controls included Klebsiella oxytoca, Proteus mirabilis, Enterococcus faecalis, and Pseudomonas aeruginosa (one each).
Fever was the most common presenting symptom and was more frequent in controls than cases. There was no significant difference among groups in the overall rate of other symptoms (vomiting, hematuria, decreased food intake, diarrhea, and urine incontinence) (Table 1). Although the overall percentage of asymptomatic children was similar among cases (24%) and controls (23.2%), there was significant difference regarding absence of symptoms between matched pairs (P = 0.01 by Mantel-Haenszel test).
Most case (26/28) and control (55/56) patients received empiric antimicrobial therapy intravenously, including second-generation cephalosporin (12 [42.8%] cases and 26 [46.4%] controls), combination of ampicillin and amikacin (13 [46.2%] cases and 28 [50%] controls), third-generation cephalosporin (2 [7.1%] cases and 1 [1.7%] control), and amoxicillin/clavulanate (1 case and 1 control). Of 13 cases, 10 had MIC for amikacin indicative of resistance. Fourteen percent of cases (4/28) received appropriate empiric therapy compared with all (100%) controls, based on in vitro susceptibility results subsequently available. Antimicrobial agents were intravenously administered to the majority (22/24) of case patients who received inappropriate empiric therapy. Of the 4 cases who received appropriate empiric therapy, 3 received a combination of ampicillin and amikacin, and 1 received amoxicillin/clavulanate.
Clinical and microbiologic responses occurred in 17 of 18 (94%) and 19 of 20 (95%), respectively, of cases and in all of the controls. Cases had an earlier clinical response than controls, while time to microbiologic response was comparable between the 2 groups (Table 1).
The only case patient clinically and microbiologically unresponsive to empiric therapy was a 5-year-old girl receiving oral cefuroxime due to febrile illness for 6 days before admission. Urine culture drawn on admission grew ESBL-producing E. coli while fever continued. She was treated with amoxicillin/clavulanate intravenously on day 9 of fever and responded clinically and microbiologically 2 days later. She had bilateral I-II grade vesicoureteral reflux and renal defects on early-phase 99Tc DMSA scan but was lost to follow-up.
On the basis of in vitro susceptibility test results, with an average delay of 4.5 days after admission, clinicians changed to appropriate therapy in 22 of 24 (92%) cases receiving inappropriate empiric therapy. The remaining 2 cases continued on intravenous cefuroxime and ceftriaxone, respectively; both of them had responded clinically and microbiologically on hospitalization day 1 and 3, respectively. Length of hospital stay did not differ significantly between cases and controls (Table 1).
High-grade (≥3) vesicoureteral reflux was detected more frequently in cases than matched controls (Table 1). In 13 (46.4%) of case patients (12 febrile and 1 with other symptoms) and 43 (76.7%) of controls (42 febrile and 1 with other symptoms), 6-month renal 99Tc DMSA scan was performed. Renal parenchymal defects were noted in none of the cases and one of control patients.
In this study, we compared the outcome of patients with UTI caused by ESBL-producing strains to matched controls suffering from UTI caused by non-ESBL-producing bacteria. Significantly, more case patients received inappropriate empiric therapy than controls. Nevertheless, clinical and microbiologic outcomes as well as formation of renal scars did not differ between the 2 groups. These findings agree with those of Hyle et al in adults.6
A possible explanation of the favorable outcome of UTIs caused by ESBL-producing bacteria, despite inappropriate empiric therapy, might be the achievement of pharmacodynamic/pharmacokinetic goals (ie, time above the MIC for β-lactam drugs) in the urinary tract of these patients. There is evidence of achievement of very high concentrations (200–2000 μg/mL) of cefuroxime in the urine of adults with normal renal function intravenously treated with cefuroxime.10 Although it is not known whether such high concentrations can be achieved with orally administered antimicrobial agents, most of our patients received empiric antibiotics intravenously. It is not clear whether the outcome would be different if they had received the same antibiotics orally. It is also unknown whether cases would still have the same long-term favorable outcome (manifested by normal 6-month 99Tc DMSA scan) if they were continued on inappropriate definite therapy for a full course.
Limitations of this study are that matching for hospital admission was conducted from a consecutive patient series, rather than a random selection, its retrospective character, and the restricted number of cases due to low prevalence of UTIs due to ESBL-producing strains.
In conclusion, UTIs caused by ESBL-producing bacteria in children do not appear to have worse outcome than UTIs caused by non-ESBL-producing bacteria, despite inappropriate empiric antimicrobial therapy most commonly administered intravenously. More data are needed to confirm these findings. In this context, the need to switch therapy, after notification of ESBL-producing uropathogens, as well as comparison of outcomes (ESBL vs. non-ESBL) in pediatric patients treated for UTI with initial empiric oral therapy, should also be investigated.
1.Tumbarello M, Spanu T, Sanguinetti M, et al. Bloodstream infections caused by extended-spectrum-beta-lactamase-producing Klebsiella pneumoniae
: risk factors, molecular epidemiology, and clinical outcome. Antimicrob Agents Chemother
2.Kang CI, Kim SH, Kim DM, et al. Risk factors for and clinical outcomes of bloodstream infections caused by extended-spectrum beta-lactamase-producing Klebsiella pneumoniae
. Infect Control Hosp Epidemiol
3.Lautenbach E, Patel JB, Bilker WB, et al. Extended-spectrum beta-lactamase-producing Escherichia coli
and Klebsiella pneumoniae
: risk factors for infection and impact of resistance on outcomes. Clin Infect Dis
4.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
5.Zaoutis TE, Goyal M, Chu JH, et al. Risk factors for and outcomes of bloodstream infection caused by extended-spectrum beta-lactamase-producing Escherichia coli
species in children. Pediatrics
6.Hyle EP, Lipworth AD, Zaoutis TE, et al. Impact of inadequate initial antimicrobial therapy on mortality in infections due to extended-spectrum beta-lactamase-producing Enterobacteriaceae: variability by site of infection. Arch Intern Med
7.Kang CI, Kim SH, Park WB, et al. Bloodstream infections due to extended-spectrum beta-lactamase-producing Escherichia coli
and Klebsiella pneumoniae
: risk factors for mortality and treatment outcome, with special emphasis on antimicrobial therapy. Antimicrob Agents Chemother
8.Paterson DL, Ko WC, Von Gottberg A, et al. Antibiotic therapy for Klebsiella pneumoniae
bacteremia: implications of production of extended-spectrum beta-lactamases. Clin Infect Dis
9.National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A5 and informational supplement M100-S10. Wayne, PA: National Committee for Clinical Laboratory Standards; 2000.
10.Bundtzen RW, Toothaker RD, Nielson OS, et al. Pharmacokinetics of cefuroxime in normal and impaired renal function: comparison of high-pressure liquid chromatography and microbiological assays. Antimicrob Agents Chemother