JSTOR

Methods

The observation period was January 2001 through July 2006. The study group consisted of pediatric patients at Innsbruck Medical University (Innsbruck, Austria): 27 children with oncological disorders (9 with leukemia and 18 with solid organ tumors); 47 children with neurological disorders requiring long‐term care (10 with neurogenic bladder dysfunction, 14 with seizures, 10 with psychomental retardation, and 13 with tetraplegia); 33 children in intensive care because of severe, life threatening disorders; and 16 children who received solid organ transplants (kidney [7], liver [16], and intestines [2]). Diagnosis of UTI was confirmed if a urine sample obtained under sterile conditions yielded more than 10,000 colony‐forming units (cfu)/mL on culture. Additional inclusion criteria were leukocyturia (greater than 10 leukocytes/μL in urine obtained under sterile conditions), a rectal body temperature of 38°C or greater, C‐reactive protein levels of 0.70 mg/dL or greater, and clinical signs of UTI, such as suprapubic pain and dysuria. Only children who met all of these criteria were included in the study. One urine sample per patient was statistically evaluated. Patients with UTI caused by multiple pathogens were excluded. Patients with urogenital malformations or vesicoureteric reflux were also excluded from evaluation, as were neonates. Data regarding the patients and their episodes of UTI were prospectively collected from the patients' medical records and from attending physicians. The history of antimicrobial treatment prior to development of UTI was evaluated for each patient (Table 1). For individuals who were treated repeatedly with the same antimicrobial during the observation period, the antimicrobial was included only once in calculations of the percentage of individuals treated with that agent. All oncological patients received trimethoprim‐sulfonamide prophylaxis. The other patients did not receive any antimicrobial prophylaxis. All transplant recipients were receiving immunosuppressive therapy at the time of evaluation (14 patients were receiving tacrolimus, 4 were receiving sirolimus, 6 were receiving steroids, and 5 were receiving mycophenolate). The clinical characteristics of the patient groups are summarized in Table 1.

Table 1.  Clinical Characteristics of Pediatric Patients With Urinary Tract Infection, by Patient Group

Open New Window

Groups of patients with nosocomial UTI (those who had been treated in the oncology, neurology, and intensive care units, and patients who received transplants) were compared with 487 patients who had community‐acquired UTI but did not have underlying disease, who were randomly selected from the records of the outpatient clinic of the Department of Pediatrics at Innsbruck Medical University (Table 1). The inclusion criteria for patients with community‐acquired UTI were identical to those for patients with nosocomial UTI, except for hospitalization; the presence of underlying diseases, such as oncological or neurological disorders; and the necessity of intensive care or transplantation. None of the patients with community‐acquired UTI had neurogenic bladder dysfunction, used invasive devices, or had received antibiotic prophylaxis for vesicoureteric reflux or urogenital malformations. The study was performed according to the principles of the Declaration of Helsinki 2000 and was approved by the local ethics committee of Innsbruck Medical University.

Characterization of the pathogen and susceptibility testing for different antibiotics were performed according to standardized procedures of the laboratory of the Department of Hygiene, Microbiology, and Social Medicine at Innsbruck Medical University, by use of the disk diffusion technique. The antibiotics tested were penicillin (amoxicillin‐clavulanate), an aminoglycoside (gentamicin), cephalosporins (cephalexin, cefuroxime, and ceftazidime), a monobactam (aztreonam), a fluoroquinolone (ciprofloxacin), a carbapenem (imipenem), and trimethoprim‐sulfonamide. Empirical antimicrobial therapy was defined as inappropriate if the pathogen isolated was resistant in vitro to the antimicrobial agent(s) used. Patients were divided into 2 groups: those with UTI due to Escherichia coli and those with UTI due to an organism other than E. coli.

Binomial variables were analyzed with the χ² test or the Fisher exact test (SPSS for Windows, version 12.0 [SPSS]). A P value of less than .05 was considered to indicate statistical significance.

Results

For all patients with nosocomial UTI, age, length of hospital stay, urine collection method, and previous antimicrobial treatment are summarized in Table 1. Urosepsis with blood culture positive for the same pathogen as that found in the urine culture at evaluation was observed in 25.9% of oncological patients, in 10.6% of neurological patients, in 6.1% of intensive care patients, in none of the transplant recipients, and in 1.2% of patients with community‐acquired UTI. Intermittent daily catheterization was performed for 17.0% of neurological patients. Among intensive care patients, 12.1% had indwelling devices.

For all groups, E. coli was the most common pathogen causing nosocomial UTI (in 37.1% of oncological patients, 46.8% of neurological patients, 42.4% of intensive care patients, and 43.8% of transplant recipients), followed by Enterococcus and Pseudomonas species in oncological patients (29.6% had UTI due to Enterococcus species and 14.8% had UTI due to Pseudomonas species), Enterococcus and Klebsiella species in neurological patients (19.1% had UTI due to Enterococcus species and 10.6% had UTI due to Klebsiella species), and Pseudomonas and Klebsiella species in intensive care patients (18.1% had UTI due to Pseudomonas species and 15.2% had UTI due to Klebsiella species) and patients who received transplants (25.0% had UTI due to Pseudomonas species and 12.5% had UTI due to Klebsiella species). Overall, the rate UTI due to E. coli was significantly lower in patients with nosocomial UTI, compared with patients with community‐acquired UTI (44.3% vs 77.7%; ). UTI due to an organism other than E. coli was more common in male than in female patients with community‐acquired UTI (35.0% vs 14.6%; ) and more common in male than in female patients with nosocomial UTI (66.0% vs 48.6%; ). Susceptibility testing results for E. coli strains and for pathogens other than E. coli are summarized in Table 2. Compared with E. coli, the other pathogens were significantly more resistant to most of the antimicrobial agents prescribed for community‐acquired UTI and for nosocomial UTI. Compared with the pathogens causing community‐acquired UTI, all pathogens causing nosocomial UTI showed a higher resistance to commonly used antibiotics.

Table 2.  Antimicrobial Resistance of Pathogens Causing Urinary Tract infection (UTI) in Pediatric Patients

Open New Window

Discussion

This survey of the spectrum and antimicrobial susceptibilities of pathogens causing nosocomial UTI in different groups of hospitalized pediatric patients revealed significant differences depending on the patients' underlying disease(s) and previous antimicrobial treatment. In this study, nosocomial UTI was predominantly caused by E. coli, in accord with results reported in the literature,^2,3,5‐7 followed by Enterococcus species in oncological and neurological patients, and by Pseudomonas species in intensive care patients and patients who received transplants. Epidemic outbreaks of infection due to Pseudomonas species have been described in intensive care patients,³ but were not observed among our patient population. In this study, nosocomial UTIs occurred as single, independent events that did not spread to other patients. Our results support observations that Enterococcus species have an increasing role in infections among hospitalized patients; this fact should not be underestimated, as it contributes to the overall increase in antimicrobial resistance among pathogens causing UTI.⁸ In oncological and neurological patients, Enterococcus species may be predominantly responsible for the high resistance rate to cephalosporins among the organisms other than E. coli, because Enterococcus species are intrinsically resistant to cephalosporins.

However, the differences in the frequency with which various pathogens are recovered from the groups of inpatients may also reflect differences in intrinsic risk factors, including the kind and severity of underlying disease. Device use and length of hospital stay have been described as the main risk factors for acquiring nosocomial UTI.⁹ In this study, the use of indwelling urinary catheters among a small number of intensive care patients played only a small role, but it did not cause differences in the frequency with which pathogens were recovered or in resistance rates. In these cases, contamination was unlikely, as cultures positive for multiple pathogens were excluded from evaluation. In addition, further inclusion criteria, such as collection of urine specimens under sterile conditions and the presence of clinical symptoms, together with positive laboratory infection parameters, helped to exclude contamination. The frequency with which E. coli was identified as the causative pathogen for community‐acquired UTI and nosocomial UTI was significantly increased among females in our study population, compared with males. Sex‐related anatomic factors are the main cause of an increased occurrence of UTI due to E. coli in females. The predominance of UTI due to pathogens other than E. coli in boys suggests that the source of infectious organisms is in the preputial sac or urethra, rather than in the bowel, as in girls.¹⁰

However, according to our results E. coli remains the predominant pathogen causing nosocomial UTI and community‐acquired UTI, but awareness of differences in the frequency with which various pathogens are recovered from different groups of hospitalized patients and differences in pathogen susceptibility are important. The higher antimicrobial susceptibility rates observed among the pathogens causing UTI in neurological and intensive care patients, compared with oncological patients, may be due to the relatively short hospital stay and more limited use of antimicrobial agents among the former group. Despite the risks that immunosuppression creates for acquiring infectious complications, the pathogens causing UTI in transplant recipients demonstrated higher antimicrobial susceptibility rates than did the pathogens causing UTI in oncological patients—probably as a result of fewer previous antibiotic treatments among the former group. In oncological patients at our institution, neutropenic fever is usually treated with a combination of imipenem and gentamicin, administered intravenously. Cephalosporins and gentamicin are usually administered intravenously as empirical therapy for severe systemic infection in transplant recipients. Cephalosporins are also the first choice of oral therapy for oncological patients and patients who have received transplants; amoxicillin‐clavulanate is the first choice of oral therapy for neurological patients and outpatients. Critically ill children with community‐acquired UTI receive amoxicillin‐clavulanate or cephalosporins intravenously, and in combination with gentamicin when infection due to Pseudomonas species is suspected.

However, it appears that the development of antibiotic resistance in pathogens causing UTI in hospitalized children is mainly driven by the frequency of infections and antibiotic treatment habits. This observation is supported by the fact that the pathogens recovered from oncological patients who received antimicrobial prophylaxis with trimethoprim‐sulfonamide showed significantly increased resistance to trimethoprim‐sulfonamide, compared with pathogens recovered from other patient groups. Although evidence is limited in some cases by the relatively small number of patients in each group (as a result of strict inclusion criteria to exclude contamination), our study allows discussion of empirical therapy in different groups of hospitalized children and comparison with community‐acquired UTI. Generally, resistance to trimethoprim‐sulfonamide, amoxicillin‐clavulanate, cefuroxime, and gentamicin—the last an intravenous antibiotic that is predominantly restricted to hospital use—is increasing.⁵ In nosocomial UTI and community‐acquired UTI, antimicrobial therapy with trimethoprim‐sulfonamide is insufficient treatment because of high resistance rates. Our results suggest imipenem should be used as the initial therapy for nosocomial UTI in oncological patients, because it is effective against E. coli and most of the UTI‐causing pathogens other than E. coli. Empirical oral therapy with amoxicillin‐clavulanate or oral cephalosporins for localized UTI, or systemic therapy with imipenem, cephalosporins, and gentamicin for systemic UTI would be appropriate for inpatients in services other than oncology. Although ciprofloxacin is not recommended for children, it is occasionally used as a short‐term, third‐line antibiotic because of the complexity of many of the patients' underlying conditions and the prevalence of other pathogens in the urinary tract that have low rates of resistance to ciprofloxacin. But, in view of emergence of multidrug‐resistant E. coli strains that produce extended‐spectrum β‐lactamases against third‐generation cephalosporins in high‐risk patients with severe underlying diseases, the use of antibiotics that may cause coselection, such as ciprofloxacin, trimethoprim, and cephalexin, should be restricted.¹¹

To summarize, hospitalized pediatric patients were revealed to be a heterogeneous group with regard to the spectrum and antimicrobial susceptibility of pathogens causing UTI. Therefore, these patients' underlying diseases, changes in the rates at which pathogens are recovered, and changes in antibiotic resistance rates under the influence of previous antimicrobial treatment have to be considered to avoid the failure of empirical antimicrobial therapy. Future research should be directed at determining the sources from which patients acquire these organisms, and the mechanisms underlying the acquisition of resistance by these pathogens.