Abstract
This retrospective study evaluates the incidence and risk factors of community-acquired urinary tract infections (CA-UTIs) linked to extended-spectrum beta-lactamase–producing Enterobacterales (ESBLPE). The study was conducted in a tertiary hospital in northern Israel and included children younger than 18 years with CA-UTIs due to Enterobacterales who were admitted to the emergency department, during the years 2017 to 2019. Among the 570 children, 9.8% had ESBLPE-associated CA-UTIs. This prevalence remained steady over the study period. ESBLPE exhibited substantial resistance to amoxicillin/clavulanic acid (62.5% vs 20.4%, P < .001, odds ratio [OR] = 6.5), trimethoprim/sulfamethoxazole (58.9% vs 18%, P < .001, OR = 6.6), ciprofloxacin (33.9% vs 3.1%, P < .001, OR = 15.9), piperacillin/tazobactam (26.8% vs 7%, P < .001, OR = 4.9), and gentamicin (21.4% vs 4.3%, P < .001, OR = 6.1), compared with non-ESBLPE. Risk factors for ESBLPE-associated UTIs included recent antibiotic treatment within the past 3 months (P = .003, OR = 3.5) and colonization with ESBLPE (P < .001, OR = 12.8). Given the variable incidence of ESBLPE, relying on local epidemiology for antibiotic selection pending culture results is crucial. The study finding of a low ESBLPE incidence, coupled with global concerns regarding carbapenem resistance, supports cautious use of broad-spectrum antibiotics in nonsevere cases.
Introduction
Community-acquired urinary tract infections (CA-UTIs) caused by extended-spectrum beta-lactamase–producing Enterobacterales (ESBLPE) pathogens pose a significant threat to the pediatric population. The incidence of ESBLPE in children with CA-UTIs reaches 46% and varies in different regions of the world.1-5 The incidence in Israel ranged from 1% to 6% in studies conducted through 20136-8 and increased to 13.5% and 28% in 2015 to 2017.9,10
Resistance to beta-lactam agents is associated with concomitant resistance to other antibiotics, complicating the selection of appropriate empiric antibiotics and oral therapies at discharge9,11,12 and requiring preliminary knowledge of local antibiotic susceptibility patterns and severity of infection to make the appropriate choice of treatment.2,7,13,14 In the absence of such data, identifying factors associated with antibiotic resistance can be helpful. In recent years, several factors have been associated with an increased prevalence of ESBLPE, including underlying disease, vesicoureteral reflux, age less than 1 year, recent hospitalization, recent antibiotic use, recurrent UTIs, Klebsiella spp. infections, and antimicrobial prophylaxis with cephalosporin.7,15-17
Given the clear trend toward an increase in the incidence of ESBLPE in some studies2,7,9 and the data gap regarding the epidemiology of ESBLPE beyond 2017 in children in Israel, we conducted this study. The objective was to explore the current incidence and risk factors for ESBLPE in children admitted from the community with UTI caused by Enterobacterales and to assess the cross-resistance to other antibiotic regimens to facilitate optimal selection of empiric antibiotic treatment.
Methods
A retrospective cohort study was conducted at Ruth Rappaport Children’s Hospital and Rambam Health Care Campus, a 1000-bed tertiary referral hospital in northern Israel serving a population of over 2 million. The study included patients younger than 18 years with UTIs due to Enterobacterales, diagnosed up to 72 hours from hospital admission, between January 1, 2017, and December 31, 2019.
A positive urine culture was delineated by the EAU/ESPU clinical practice guidelines. This includes any bacterial growth during suprapubic aspiration, the presence of at least 10 000 colony-forming units (CFUs) per milliliter from a catheterized urine, or a growth of ≥100 000 CFU/mL from a clean catch urine of a single Enterobacterales species, coupled with clinical symptoms and signs indicative of UTIs. 18 The Enterobacterales included Escherichia coli, Klebsiella, Enterobacter, Citrobacter, Salmonella, Shigella, Leclercia, Raoultella, Proteus, Morganella, Providencia, Serratia, and other related species. For patients with recurrent episodes, only the first episode of CA-UTIs caused by Enterobacterales during the study period was considered for inclusion.
Asymptomatic bacteriuria, urine cultures obtained >72 hours after admission, urine cultures obtained from patients with permanent urinary catheters or intermittent catheterization, and urine cultures with isolation of bacteria other than Enterobacterales were excluded.
Cases were identified using a computerized microbiology laboratory database. Clinical data extracted from electronic medical records included age, sex, ethnicity, clinical symptoms and signs of infection, length of hospital stay, empiric antibiotic therapy, underlying diseases, urogenital abnormalities, antibiotic prophylaxis, and previous UTI episodes. Microbiologic data included the method of urine collection, organisms isolated, resistance to antibiotics and known ESBLPE colonization by rectal screening or isolation from a clinical specimen in the 6 months prior to admission without evidence of eradication. Missing data were subtracted from the total case number presented in the tables.
The Vitek 2 instrument (bioMérieux, Inc Hazelwood, Missouri) was used to identify and determine the antibiotic susceptibility of Enterobacterales isolated from urine cultures. Minimum inhibitory concentrations values were reported according to the clinical and laboratory standards institute guidelines for interpretation.
Susceptibility to cephalosporins was evaluated by measuring the inhibition zone diameter using the Kirby-Bauer disk diffusion test on Mueller-Hinton agar plates (Hylabs, Rehovot, Israel).
The zone of inhibition of cefotaxime 30 g and ceftazidime 30 g alone and in combination with clavulanic acid (Oxoid, Basingstoke, UK) was measured after incubation of the plates at 37°C for 18 to 24 hours. When the presence of clavulanic acid increases the diameter of the zone of inhibition by 5 mm or more, the bacterium produces ESBL.
All data were analyzed using SPSS version 26. Categorical variables were presented as counts and percentages and continuous variables as means and standard deviations or median and interquartile range (IQR) according to their distribution. Univariate analysis using the χ2 and multivariate analysis using binary regression analysis were performed to identify risk factors for ESBLPE-associated CA-UTIs. Parameters were included in a log-likelihood ratio test model, with P < .1 from the univariate analysis or clinical relevance. The magnitude of association between these variables was approximated by calculating odds ratios (OR) and 95% confidence intervals. All statistical tests are two-tailed; P < .05 is considered statistically significant (95% confidence interval). The study was approved by the Rambam Health Care Campus institutional review board.
Results
Demographic and Clinical Characteristics
A total of 570 patients with Enterobacterales-associated CA-UTI were detected during the study. The median age of the included patients was 1.3 years (IQR: 0.69-5.9, range: 5 days to 17.9 years). Females represented 87.9% of the patients, and urogenital anomalies were reported in 62 patients (10.9%). Of all the collected samples, 358 patients (62.8%) were obtained through catheterization, while the remaining samples were collected via midstream voiding. Forty-four patients (7.7%) had recent antibiotic treatment. Of 570 patients admitted to the emergency department, 339 patients (59.5%) were hospitalized, whereas the rest were discharged as indicated in Table 1.
Demographic and Clinical Characteristics of Patients With Community-Acquired Urinary Tract Infection Caused by Enterobacterales.
Urinary tract infection in the past 3 months.
In the past 3 months.
Hospitalization lasting a minimum of 72 hours in the last month.
Rectal screening or isolation from a clinical specimen in the 6 months prior to admission without evidence of eradication.
Frequency of Enterobacteriaceae Isolates and Their Antibiotic Resistance Profile
The most prevalent bacterial isolates identified in the study were E coli, accounting for 82.1% of cases, followed by Klebsiella spp. at 8.8% and Proteus spp. at 5.8% (Table 2). Among all Enterobacterales, ESBLPE were detected in 56 patients (9.8%). The incidence of ESBLPE-associated CA-UTIs varied between 8.4% and 11.2% over the study period.
Identified Enterobacterales Species Causing Urinary Tract Infection.
Number of patients with a given isolate and their proportion of all Enterobacterales isolates.
Number of ESBL-positive Enterobacterales and their proportion of the total number of the respective isolate.
Higher rates of cross-resistance were observed in cultures with Enterobacterales positive to ESBL compared with cultures of Enterobacterales negative to ESBL (Table 3). Resistance to commonly used oral antibiotics such as amoxicillin/clavulanic acid was documented in 62.5% versus 20.4% (P < .001, OR = 6.5), trimethoprim/sulfamethoxazole in 58.9% versus 18% (P < .001, OR = 6.6), ciprofloxacin in 33.9% versus 3.1% (P < .001, OR = 15.9), and the intravenous antibiotic piperacillin/tazobactam in 26.8% versus 7% (P < .001, OR = 4.9). Amikacin resistance was observed in 1% of patients with ESBL-negative but not in those with ESBL-positive Enterobacterales. However, significantly higher resistance to gentamicin was found in patients with ESBL-positive isolates, 21.4% versus 4.3% (P < .001, OR = 6.1). Neither group exhibited resistance to carbapenems (Table 3).
Comparing Antibiotic Resistance Rates in ESBL-Positive and ESBL-Negative Enterobacterales.
Number of resistant isolates divided by the total number of isolates tested for the antibiotic (%).
Risk Factors for ESBLPE
In the univariate analysis, patients aged 3 years or younger demonstrated a lower rate of ESBLPE (7.8%) compared with those in the 3- to 10-year age group (16%, P = .013). However, this rate was similar to that in children between 10 and 18 years (11.3%, P = .262). Patients with ESBLPE associated CA-UTIs exhibited a higher prevalence of urogenital abnormalities compared with patients with ESBL-negative Enterobacterales (25% vs 9.3%, P < .001), higher likelihood of having received antibiotic treatment for any indication within the past 3 months (23.2% vs 6%, P < .001), and antibiotic prophylaxis to prevent UTI (7.1% vs 0.8%, P = .021). In addition, known colonization with ESBLPE during the preceding 6 months was more prevalent in patients with ESBLPE-associated CA-UTIs (17.9% vs 1.2%, P < .001) (Table 4). The specific antibiotic treatment has been documented for 22 of 44 patients (50%), in which 10 cases were treated with first-generation cephalosporin (all were ESBLPE negative), 4 cases were treated with amoxicillin/clavulanic acid (3 of 4 were ESBLPE positive), 4 cases were treated with amoxicillin (1 of 4 was ESBLPE positive), and 4 cases were treated with other antibiotic agents (2 of 4 were ESBLPE positive).
Factors Associated With Increased Risk for Acquisition of ESBLPE associated UTI, Univariate and Multivariate Analysis.
In the preceding 3 months.
Hospitalization lasting a minimum of 72 hours in the last month.
Urinary tract infection.
Rectal screening or isolation from a clinical specimen in the 6 months before admission without evidence of eradication.
A multivariate analysis that included age, sex, urogenital abnormalities, antibiotic treatment in the last 3 months, and known ESBLPE colonization showed that ESBLPE colonization was the strongest risk factor (P < .001, OR = 12.8) of ESBLPE-associated UTIs and that antibiotic treatment in the past 3 months was also associated with an increased likelihood of ESBLPE-associated UTIs (P < .002, OR = 3.7). Due to a high correlation with previous antibiotic treatment in the past 3 months, recent UTI within the same time frame was not included in the multivariate analysis.
Discussion
The incidence of CA-UTIs caused by ESBLPE in this study ranged from 8.4% to 11.2% during the study years. This rate was comparable to a concurrent study conducted in Italy 19 but considerably lower than the rates reported in other recent studies worldwide, where rates ranging from 25% to 80% were observed.2,20,21 Notably, the rate identified in this study was lower compared with 19.4% and 28% documented in 2 prior studies conducted in Israel until 2017.9,10 However, it aligns with the 6% rate from the Dayan et al study until 2011 7 or the 11% rate observed in cultures of pediatric patients with neurogenic bladder and asymptomatic bacteriuria between 2010 and 2018. 22 The lower prevalence of ESBLPE-associated CA-UTIs detected in the present research underscores the diversity of ESBLPE rates within the same country. This is significant as ESBLPE has extended its reach to inhabit individuals without significant health care interactions, essentially becoming the primary source of ESBLPE in both community and hospital-related infections. 23 Understanding the distinct antibiogram patterns within the population of each institution is vital in making informed choices for empirical antibiotic treatment. 10
Both ESBLPE colonization and recent antibiotic treatment were identified as risk factors for ESBLPE-positive UTIs in this study. Several studies have also demonstrated the selective pressure of antibiotic treatment on the risk of ESBLPE-positive UTIs.7,10,24,25 A previous study has shown that the specific use of beta-lactam/beta-lactamase inhibitors before infection in children and urinary tract catheterization are risk factors for ESBLPE-positive UTIs in colonized patients. 26 However, this study could not establish a link between particular antibiotic treatment and ESBLPE-associated UTIs, potentially attributed to underreporting of the antibiotic treatment given in the community. The correlation between rectal colonization with ESBLPE and ESBLPE-associated UTIs was demonstrated in a study conducted on elderly patients. 27 Another study in adult kidney transplant recipients also showed a strong association between gut colonization with ESBL-positive Klebsiella pneumoniae and UTIs caused by this organism. 28 As surveillance of ESBLPE is not routinely performed, this association prompts inquiries about the necessity of conducting surveillance studies in regions with a high prevalence of ESBLPE.
The resistance observed in ESBLPE isolates toward oral antibiotics can be attributed to the presence of ESBL-encoding plasmids, which might also carry genes conferring resistance to other antibiotic groups. 29 In addition, the unexpectedly high resistance rate to piperacillin/tazobactam, in contrast to findings in other studies,30,31 could potentially complicate treatment selection. With the relatively low ESBLPE rate presented in this study and concerns regarding increased resistance against carbapenem, 32 we do not warrant a shift toward broader antibiotic coverage for patients with nonsevere infections.
Remarkably, the high percentage of patients diagnosed with CA-UTIs required hospitalization. This finding underscores the cautious strategy adopted for handling febrile UTIs. Typically, the initial therapeutic protocol for these situations entails the utilization of aminoglycosides as the primary treatment, followed by third generation of cephalosporins as an alternative option when kidney function is uncertain. These treatment approaches are conventionally administered within a hospital setting. However, the study’s findings raise questions about the appropriateness of these treatments, particularly in light of the substantial resistance to gentamicin observed in around one fifth of the patients.
The primary limitation of this study lies in its retrospective framework, drawing upon medical records, which could introduce information bias and result in data gaps. Moreover, this study is confined to a single center, implying that patients might have received treatment at different medical facilities. In addition, the study was conducted within a hospital setting rather than in community clinics, potentially introducing selection bias toward more intricate cases. Despite this limitation, the study revealed a lower rate of ESBLPE than the majority of prior investigations.
In conclusion, this study revealed a reduced incidence of ESBLPE in comparison with prior research, underscoring the varying incidence of ESBLPE in the same country. Established ESBLPE colonization and recent antibiotic usage stand out as risk factors for ESBLPE-induced UTIs, bolstering the notion of selection pressure as a precursor to clinical infection. These risk factors should promote implementation of antibiotic stewardship programs and surveillance mechanisms for tracking ESBLPE rates within the community in endemic regions. In addition, ESBLPE exhibits substantial resistance to numerous other antibiotic classes. Nonetheless, considering the modest ESBLPE rate and concerns regarding increased resistance to broad-spectrum antibiotics, there is no justification for altering empirical antibiotic treatment in nonseverely ill patients.
Author Contributions
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
This study was approved by the Rambam Health Care Campus institutional review board (0132-20-RMB).
