Introduction
Acute urinary tract infection (UTI) is one of the most frequent bacterial infections in childhood and a leading cause of antibiotic prescription in pediatric practice [1-3]. If not diagnosed and treated appropriately, UTIs may progress to acute pyelonephritis and result in long-term complications such as renal scarring, hypertension, and chronic kidney disease. Early diagnosis of UTI in infants and young children is often challenging because of nonspecific clinical manifestations, which may lead to inappropriate or excessive antibiotic use. Consequently, antimicrobial resistance among uropathogens has increased worldwide, with substantial regional variation [3-5]. The emergence of extended-spectrum β-lactamase (ESBL)-producing organisms has further complicated empirical antibiotic selection for pediatric UTIs [4-6]. Several previous studies have reported changes in the epidemiology and antibiotic resistance patterns of pediatric UTIs; however, long-term time-trend data from a single center using consistent diagnostic criteria remain limited, especially in Korea [5]. At our institution, antibiotic susceptibility patterns and imaging findings of childhood UTIs were previously reported for the period 2003–2008 using comparable diagnostic and microbiological protocols, providing a unique historical reference for longitudinal comparison [7]. Therefore, the present study aimed to evaluate changes in uropathogen distribution, antibiotic susceptibility patterns, and imaging findings of childhood acute UTIs over an approximately 10-year interval at a single tertiary center in Korea, using a direct comparison with previously published institutional data.
Methods
Participants
We retrospectively reviewed the medical records of patients who visited the emergency room or outpatient clinic of Chungbuk National University Hospital between March 2015 and February 2020 and were clinically suspected of having a UTI. The 2015–2020 period was selected to represent the most recent 5-year cohort available at the time of study planning and to enable comparison with our previously published 2003–2008 institutional cohort across an approximately 10-year interval.
UTI was suspected in patients presenting with pyuria accompanied by clinical symptoms suggestive of UTI, including fever, dysuria, urinary frequency, or flank pain. Pyuria was defined as the presence of ≥5 white blood cells per high-power field on urinalysis. Fever was defined as a tympanic temperature of ≥37.5 °C.
A total of 1,015 patients underwent urine culture testing during the study period. Among them, 571 patients with a positive urine culture result, defined as the growth of ≥1.0×10⁵ colony-forming units/mL of a single urinary tract pathogen, were included in the analysis. Patients incapable of intact self-voiding (e.g., neurogenic bladder or indwelling Foley catheter), immunocompromised patients (including those receiving immunosuppressive therapy, with immunosuppressive diseases, or with hematologic or oncologic disorders), and patients with previously diagnosed vesicoureteral reflux (VUR) were excluded. The causative pathogens and their antibiotic susceptibility profiles were retrospectively identified and analyzed in the included patients.
Sample collection, pathogen identification, and antibiotic susceptibility testing
In boys incapable of self-voiding, the foreskin was gently retracted and the urethral meatus was cleansed using a sponge soaked in 2% boric acid, after which a sterile urine collection bag was applied. In girls incapable of self-voiding, the genital area and urethral opening were cleansed in the same manner before applying a urine collection bag. In children capable of self-voiding, the genital area was cleansed with 2% boric acid, and midstream urine was collected in a sterile, capped plastic container.
Urine specimens were inoculated with 0.001 mL of urine onto blood agar and MacConkey agar plates and incubated at 35–36 °C for 24 hours. Bacterial identification was performed using the VITEK 2 system (BioMérieux) during earlier periods and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS; Bruker) during the study period. Antibiotic susceptibility testing was conducted using the appropriate VITEK 2 susceptibility cards in accordance with the manufacturer’s instructions. For Gram-negative pathogens, susceptibility was assessed for ampicillin, amoxicillin/clavulanate, piperacillin/tazobactam, aztreonam, amikacin, gentamicin, cefazolin, cefoxitin, cefotaxime, ceftazidime, cefepime, ciprofloxacin, ertapenem, imipenem, trimethoprim/sulfamethoxazole, and tigecycline. For Gram-positive pathogens, susceptibility testing included penicillin, erythromycin, clindamycin, tetracycline, ceftriaxone, ciprofloxacin, imipenem, and vancomycin. ESBL production among Enterobacteriaceae was determined using the VITEK 2 system.
Imaging studies
All patients with confirmed UTI who provided consent for imaging studies underwent evaluation to assess potential urinary tract abnormalities. Renal ultrasonography was initially performed, followed by 99mTc-dimercaptosuccinic acid (DMSA) renal scintigraphy during hospitalization in the acute phase of infection to assess renal cortical defects. Voiding cystourethrography (VCUG) was recommended for patients younger than 1 year or for those with abnormal findings on renal ultrasonography (e.g., hydronephrosis), renal cortical defects on DMSA scintigraphy, or recurrent UTI, and was performed when caregivers agreed after clinical counseling.
Outcome measures
The distribution of causative pathogens and their antibiotic susceptibility patterns was analyzed retrospectively. ESBL-producing strain prevalence and antibiotic susceptibility were evaluated according to pathogen type. Comparisons of pathogen distribution and antibiotic susceptibility patterns were made between the study population and a previously analyzed UTI subgroup. Imaging findings from renal ultrasonography, DMSA scans, and VCUG were also analyzed according to pathogen type. Additionally, antibiotic susceptibility patterns were compared between patients with and without VUR.
Results
Baseline characteristics of patients
A total of 571 children aged 0–18 years with culture-confirmed acute UTI were included in the study. The median age was 0.33 years (range, 0–18 years). Of these patients, 374 (65.5%) were male and 197 (34.5%) were female, with a male-to-female ratio of 1.9:1. Infants younger than 1 year accounted for the majority of cases (500 patients, 87.6%). The proportion of patients younger than 1 year was significantly higher during 2015–2020 compared with 2003–2008 (87.6% vs. 70.2%; P<0.001). There was no significant difference in sex distribution between the two periods (Table 1). Presenting symptoms are summarized in Supplementary Table 1 (available online). Fever was the predominant symptom, observed in 569 patients (99.6%).
Distribution of uropathogens
Among the 571 culture-positive urine samples, 561 (98.2%) yielded Gram-negative organisms. Escherichia coli was the most frequently isolated pathogen, accounting for 509 cases (89.1%). The proportion of E. coli significantly increased from 81.4% (349/429) in 2003–2008 to 89.1% (509/571) in 2015–2020 (P<0.001). Other Gram-negative organisms included Klebsiella spp. (21 cases, 3.7%), Enterobacter spp. (17 cases, 3.0%), and Proteus spp. (8 cases, 1.4%). Gram-positive organisms were identified in 10 cases (1.9%) (Table 2).
Changes in antibiotic susceptibility over time (2003–2008 vs. 2015–2020)
All bacteria
When comparing the two study periods, antibiotic susceptibility significantly decreased during 2015–2020 for aztreonam (96.2% to 87.1%), cefazolin (81.2% to 58.8%), cefoxitin (93.7% to 80.8%), gentamicin (93.0% to 61.5%), trimethoprim/sulfamethoxazole (71.8% to 55.2%), and third-generation cephalosporins (95.6% to 66.1%) (all P<0.001). In contrast, susceptibility to imipenem remained unchanged at 100% across both periods. No significant differences were observed for piperacillin/tazobactam, ciprofloxacin, cefepime, or ampicillin (Fig. 1).
Escherichia coli
Among E. coli isolates, susceptibility significantly decreased during 2015–2020 compared with 2003–2008 for aztreonam (97.9% to 87.4%), cefazolin (86.5% to 59.5%), cefoxitin (97.4% to 82.3%), gentamicin (79.0% to 58.2%), ciprofloxacin (91.9% to 83.9%), trimethoprim/sulfamethoxazole (68.7% to 51.7%), and third-generation cephalosporins (97.1% to 64.2%) (all P<0.001). Imipenem maintained 100% susceptibility in both periods, and no significant change was observed for piperacillin/tazobactam (Fig. 2).
Klebsiella pneumoniae
Antibiotic susceptibility trends in K. pneumoniae are presented in Supplementary Fig. 1 (available online).
ESBL-producing organisms
Among Gram-negative isolates, ESBL-producing organisms were identified in 116 cases (20.8%) during 2015–2020, representing a marked increase compared with 2003–2008 (4.2%; P<0.001). A similar increase was observed among E. coli isolates, with ESBL positivity rising from 4.6% to 22.0% (P<0.001). Although ESBL prevalence also increased in K. pneumoniae, the difference did not reach statistical significance (Table 4).
Imaging findings and their association with microbiological results
Renal ultrasonography was performed in 549 patients (96.1%), and VCUG was performed in 462 patients (80.9%). VUR was detected in 100 patients (21.6%), with 49% classified as grade IV–V. No significant differences in causative organisms were observed according to ultrasonography or VCUG findings (Supplementary Tables 2 and 3, available online). Antibiotic susceptibility patterns did not differ significantly according to the presence or severity of VUR, except for ciprofloxacin, which showed lower susceptibility in patients with VUR compared with those without VUR (72.0% vs. 87.5%; P<0.001) (Supplementary Table 4, available online). Among 526 patients who underwent DMSA renal scanning, renal cortical defects were identified in 234 patients (44.5%). No significant differences in causative organisms were observed according to DMSA scan findings (Supplementary Table 5, available online).
Discussion
In this single-center study, we analyzed long-term changes in microbiological characteristics, antibiotic susceptibility patterns, and imaging findings of childhood acute UTIs by directly comparing recent data with a previously published cohort from the same institution [7]. This study design minimizes inter-institutional variability and allows for a more accurate assessment of temporal trends in antibiotic resistance. The predominance of E. coli as the primary uropathogen observed in the present study is consistent with previous reports [3,5]. Importantly, despite similar pathogen distribution, antibiotic susceptibility patterns changed substantially compared with the earlier institutional cohort, indicating a true temporal shift rather than differences attributable to patient population or diagnostic practices [7].
Across the overall bacterial cohort, substantial declines in susceptibility were observed for several commonly used antibiotics over the 10-year interval, particularly third-generation cephalosporins, gentamicin, and trimethoprim/sulfamethoxazole. Similar reductions were also evident within E. coli isolates. These changes paralleled a marked increase in ESBL-producing strains. Similar upward trends in resistance among pediatric E. coli isolates have been reported in both Korean multicenter studies and international surveillance data, where increased antimicrobial pressure and widespread cephalosporin use have been suggested as contributing factors [8-11]. Our findings extend these observations by demonstrating a pronounced temporal decline within a single institutional setting using consistent diagnostic criteria over an approximately 10-year interval [7]. Although the proportion of E. coli increased in the later cohort, subgroup analysis demonstrated a significant decline in antibiotic susceptibility within E. coli isolates themselves, indicating that the observed resistance trends cannot be explained solely by changes in pathogen composition. The increasing prevalence of ESBL-producing organisms further complicates empirical antibiotic selection. Resistance to third-generation cephalosporins closely paralleled the rise in ESBL prevalence, supporting the clinical relevance of cephalosporin resistance as a surrogate marker for ESBL production [4-7].
Age was stratified at 1 year because infancy represents a clinically distinct stage in pediatric UTI, characterized by a higher prevalence of VUR and different diagnostic considerations compared with older children [1,2]. Notably, the proportion of infants younger than 1 year was significantly higher in the 2015–2020 cohort than in the earlier period. This shift may reflect evolving diagnostic practices, including increased evaluation of febrile infants for UTI. Because age distribution influences the clinical profile of the study population, this demographic change should be considered when interpreting temporal trends in antimicrobial resistance. Although imaging abnormalities such as VUR and renal cortical defects were frequently identified, no significant association was observed between imaging findings and causative organisms or antibiotic susceptibility patterns. These results are consistent with current guideline recommendations that imaging findings should not be used to guide empirical antibiotic selection [2,12].
A major strength of this study is the direct comparison with previously published data from the same institution using similar diagnostic criteria, which enhances the internal validity of the observed temporal trends [7]. However, several limitations should be considered. This study was retrospective and conducted at a single tertiary center, which may limit external generalizability. Recurrence data were not systematically collected, precluding analysis of a potential relationship between ESBL-producing organisms and recurrent UTI. In addition, microbiological identification methods changed during the study period (from the VITEK 2 system to MALDI-TOF MS), which may have introduced minor methodological differences. Moreover, although conducted within a consistent institutional setting, subtle shifts in clinical practice patterns over time, including imaging utilization or indications, may have influenced the findings.
In summary, this study provides institution-specific longitudinal evidence of evolving antimicrobial resistance in pediatric UTIs over an approximately 10-year interval. While increasing antimicrobial resistance has been widely reported, our direct comparison of two cohorts from the same center using consistent diagnostic criteria uniquely demonstrates a significant increase in antibiotic resistance within the predominant pathogen over time. These findings underscore the importance of continuous local surveillance to guide empirical antibiotic selection.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
