Pediatric Febrile Urinary Tract Infection Caused by ESBL Producing Enterobacteriaceae Species

Background Over the past decade, drug resistance pattern has worsened for many of the uropathogens due to overuse of antibiotics for empiric treatment. The burden of extended spectrum beta-lactamase (ESBL) producing Enterobacteriaceae associated urinary tract infections (UTI) has become increasingly more common, limiting treatment options among children presenting with febrile UTI. We investigated the burden and correlates of ESBL producing Enterobacteriaceae associated UTI among children and antibacterial resistance pattern. Methods 284 midstream urine specimens were collected using standard aseptic techniques from 284 children who were diagnosed with suspected UTI. Urine culture and bacteria isolation were performed following standard bacteriological techniques. The Kirby-Bauer disk diffusion technique and the double-disc synergy test were used to investigate antibiotic susceptibility and presence of ESBL production. Results UTI was confirmed using a positive urine culture for a relevant pathogen in 96/284 (33.8%) of the cases. Enterobacteriaceae accounted for 75% (72/96) of etiologies of UTI in children. The most frequent Enterobacteriaceae spp. were E. coli, 44.4% (32/72) and K. pneumonia, 27.8% (20/72). The overall multidrug resistance rate was 86.1% (62/72). ESBL-producers accounted for 41.7% (30/72) of the isolated Enterobacteriaceae. ESBL producing K. pneumonia and E. coli isolates accounted for 70% (14/20) and 37.5% (12/32), respectively. History of UTI in the past 1 year (adjusted odds ratio (AoR) = 0.08, 95%CI (0.01 − 0.57)) and medium family wealth index (AoR = 0.03, 95%CI (0.00 − 0.27)) protected from infection with ESBL-producing Enterobacteriaceae. Conclusion. ESBL production was more common in K. pneumonia and appeared to be a major factor contributing drug resistance UTI in children. The findings call for the need to incorporate ESBL testing in the routine clinical practice. The resistance level to commonly prescribed first-line antibiotics observed within Enterobacteriaceae was alarming calling for strengthened antimicrobial stewardship.


Introduction
Urinary tract infection (UTI), which is commonly caused by Enterobacteriaceae like Escherichia coli (E. coli) and Klebsiella species and other Gram-negative bacteria, is one of the commonest causes of febrile illnesses in children [1,2]. Extended spectrum β-lactamase (ESBL) producing Enterobacteriaceae causing UTI have been associated with prescription of broad-spectrum antibiotics, which further exacerbate the challenge of antimicrobial resistance (AMR) [3].
In several settings, the occurrence of ESBL-producing bacterial infections like E. coli and Klebsiella species were described as a pandemic [1,4,5]. ESBLs are plasmid-mediated enzymes, which confer resistance to Gram-negative bacteria [6], through three mechanisms-first, β-lactam cannot reach penicillinbinding proteins (PBP); second, decreased affinity to PBPs; and third, β-lactamases destroy the drug [7]. β-lactamase enzymes destroy inactivate β-lactam containing antibiotics, and ESBLs contain serine. These enzymes can hydrolyze penicillin, cephalosporin, aztreonam, and monobactam antibiotics making them ineffective. Cephamycin groups including cefoxitin and cefotetan and carbapenems including imipenem, meropenem, and ertapenem are not hydrolyzed by ESBLs and are hence considered as the drugs of choice for treating ESBLproducing Enterobacteriaceae [8]. However, these drugs are not readily available; and, in settings where diagnostics for ESBL producing Enterobacteriaceae are limited, targeting treatment becomes a challenge.
Circulation of ESBL-producing Enterobacteriaceae in the community and healthcare settings is a significant global challenge as this could be associated with increasing trends of AMR, which is even more significant in the sub-Saharan African region [9]. Furthermore, high levels of AMR related to ESBL-producing Enterobacteriaceae complicate individual patient care increasing the mortality and morbidity associated with common infectious diseases in children like UTI [10]. Of note, the alarming AMR trend is partly due to the irrational use of antibiotics in resource-limited settings which further leads to more selective pressure drug resistance [7,11].
There is limited data on the extent of infections among children caused by ESBL-producing Enterobacteriaceae. We investigated the prevalence and correlates of pediatric UTI caused by ESBL-producing Enterobacteriaceae at a tertiary hospital. The findings will have a critical importance to guide empiric use of antibiotics in the pediatric population in the setting. It will also shed light on the importance of routine testing for ESBL-production in pediatric urine samples.

Methods and Materials
2.1. Study Setting and Participants. A cross-sectional study was conducted from February 1, 2018, to July 30, 2018, among children presenting to the outpatient department of Hawassa Comprehensive Specialized Hospital, which is a tertiary referral facility in southern Ethiopia.
Included were children below the age of 15 years with clinically suspected UTI defined as those with at least one of the signs and symptoms of urinary tract infection including frequency, urgency, dysuria, abdominal pain, back pain, and fever (>38.0°C). Participants who received antibiotics within two weeks before presentation to the hospital were excluded.

Sample
Size. Sample size was calculated using singlepopulation proportion formulae considering the prevalence of extended spectrum β-lactamase (ESBL) producing Enterobacteriaceae of 78.6% [12]. Using survey methods, 10% nonresponse rate and 95% confidence interval, a final sample size of 284 subjects with suspected UTI was estimated.

Data Collection
2.3.1. Sociodemographic and Clinical Data. Sociodemographics including age (years), sex, residence, parental education and occupation, living standard, and clinical history of participants such as hospital admission and history of UTI within the past 12 months were recorded.

Laboratory Data Collection.
A total of 284 early morning midstream (MSU) and catheterized urine samples were collected using properly labelled sterile, clean, transparent, screw-capped, wide-mouth plastic cups. Samples were transported to the microbiology laboratory within two hours of collection. Standard wire loop of 1 μl diameter was used to inoculate approximately 0.001 mL urine on 5% sheep Blood agar (OXOID Ltd. England) and MacConkey agar (OXOID Ltd., Basingstoke, United Kingdom) plate which were incubated aerobically at 37°C for 24 hrs.
Significant bacterial growth was determined on Blood agar plate when a single midstream or catheterized urine culture yields >10 5 CFU/mL [13]. Macroscopic examination for hemolysis, changes in physical appearance on differential media, and colony characteristics were recorded to help in identification. Furthermore, Gram reaction, morphology, and colony arrangement were recorded. Biochemical tests including Indole production, sugar fermentation, H 2 S and gas production, citrate utilization, motility test, mannitol test, and urease and oxidase test were also done to further identify Enterobacteriaceae isolates. If not processed immediately, we kept the isolated bacteria at 2-8°C in a nutrient broth for not more than 24 hrs until the antimicrobial sensitivity test was done.

Detection and Confirmation of ESBL Producing
Enterobacteriaceae. Standard disc diffusion method was used to assess ESBL production among isolated Enterobacteriaceae which was then confirmed by the double-disc synergy test (DDST) using third-generation cephalosporins and modified double-disc synergy test using cefepime along with the third-generation cephalosporins, following standard microbiological procedures and the Clinical and Laboratory Standards Institute (CLSI) guidelines [14].

Procedures for Double-Disc Synergy Test (DDST).
Antibiotic discs of 20/10 μg of amoxicillin/clavulanic acid centrally and 30 μg each of cefotaxime and ceftazidime were placed center to center at 15 mm apart, followed by incubation at 37°C for 18-24 hours. ESBL production was considered when inhibition zones of cefotaxime and ceftazidime extending towards clavulanic acid disc. Modified double-disc synergy test using Cefepime antibiotic disc was used to further detect AMPC lactamase coproducing false-negative isolates.
2.6. Procedures for Modified Double-Disc Synergy Test (MDDST). Similar concentrations of amoxicillin-clavulanate antibiotic disc with cephalosporins namely ceftriaxone, cefotaxime, cefepime, and ceftazidime are used in MDDST. Any distortion or increase in the zone of inhibition of the cephalosporin discs towards the amoxicillin-clavulanate disc was considered as confirmatory for ESBLs production.
2.7. Antimicrobial Susceptibility Testing. Kirby-Bauer disk diffusion method was used to assess antimicrobial susceptibility as recommended by the CLSI guidelines [13,14]. Ampicillin 10 μg). Zone of inhibition was measured after overnight incubation at 37°C. Nonsusceptibility to three or more classes of antibiotics defined multidrug resistance.

Data
Analysis. Descriptive data were presented as frequency (percentage), mean (standard deviation), median (range), and using tables and figures. Predictors of infection with ESBL producing Enterobacteriaceae were assessed using bivariate logistic regression. Multivariate logistic regression models were run using all variables with a P value < 0.25 in univariate analysis. Covariates with a P value of <0.05 in the multivariate models were considered as independent predictors of infection with ESBL producing bacteria. Anthropometric Z-scores were generated using WHO Anthro plus, while family wealth status was assessed using principal component analysis (PCA).

Ethical
Considerations. The study received ethical approval from the institutional review board (IRB) of Hawassa University (IRB Number: IRB/156/10). Written informed assent was obtained from study participants or caretakers of children. Patient privacy was protected by deidentification of records. Names of patients were coded. All data obtained during the study were kept confidential and were used solely for the purpose of the study. Positive laboratory result from the study participant was communicated to their physicians for appropriate treatment or management.    Table 2).
We then performed bivariate analysis using paternal occupation, maternal education, paternal education, age, place of residence, patient type, history of UTI within the past 12 months, and family wealth index against ESBL infection. Next, we pooled the variables with a P value < 0.025 to identify independent predictors of infection with ESBL producing Enterobacteriaceae. Children with a history of UTI within the past 12 months were less likely to be infected with ESBL producing Enterobacteriaceae (AOR = 0.076 with 95% CI (0.010-0.569). Family wealth index of medium was associated with lower risk of infection by ESBL-producing Enterobacteriaceae as compared to those with poor family wealth index (AOR = 0.029 with 95% CI (0.003-0.265) ( Table 3).

Discussion
Our findings show that urine culture growth was observed in a third of cases presenting with clinically suspected cases of UTI. The majority of bacterial isolates identified in the culture were Gram-negative bacteria, a finding comparable to findings from another tertiary care hospital in Ethiopia [12]. The culture confirmation rate of suspected cases in the current study higher than studies elsewhere [10,15,16], reflecting the variability in clinical index of suspicion and  the healthcare settings. Important limitations of the current study include that certain clinical features such as ICU admission and circumcision (for boys) were not assessed as potential factors for infection with ESBL producing Enterobacteriaceae. Molecular epidemiological characterization of ESBL producing Enterobacteriaceae was not possible. The prevalence of ESBLs producers within the Enterobacteriaceae in the study was 41.7%, which is similar to findings in Ethiopia (38.0-51%) [17,18], Nigeria (47.1%) [19], and India (38.2%) [20]. However, the prevalence of ESBL production in the current study is much lower than that reported from a study at a tertiary facility in Ethiopia [12], which included Enterobacteriaceae from specimens other than urine and reported a much smaller number of isolates. The ESBL prevalence was also lower than those reported in Burkina Faso 58%, from clinical samples [21] and Nepal 62.31% from urinary isolates [22], both of which included adults. On the other hand, the prevalence of ESBL observed in our study was higher than another study in Ethiopia (25%) [23], and studies in Morocco (25.5%) [24], Iran (28.4%) [25], and Korea (16.7%) [26]. The higher prevalence seen in our study compared to developed countries like Greece (10.4%) [27] could be attributed to differences in infection control policies and practices, duration of hospitalization, and improved nursing barriers that decrease acquisition and spread of ESBL producing strains.
Similar to previous reports, E. coli was the most common cause of UTI followed by K. pneumonia [18,28,29]. Even  [17,30], while more ESBL production in E. coli as compared to K. pneumonia in another study in Ethiopia [12].
A worrying level of multidrug resistance was reported in our study which adds to previous reports in other facilities in Ethiopia, for example, 87.4% in Northwest Ethiopia [1] and 68.3% in Addis Ababa, Ethiopia [18]. Studies elsewhere also reported similarly high isolation rates of multidrug-resistant bacteria, for instance in Chicago, USA, 76% [31], Nepal 64.04% [32], and Iran 52.7% [29]. Unsurprisingly, ESBLproducing Enterobacteriaceae are associated with multidrug resistance in nearly all the cases (96.7%)-a finding very close to another survey in Addis Ababa, Ethiopia, with 96.3% multidrug resistance [18]. Nitrofurantoin demonstrated a lower level of resistance among E. coli and K. pneumoniae isolates, favoring the use of this antibiotic for empiric treatment of UTI in children. The lower resistance levels could point to the lower utilization of the antibiotic [33]. Our finding together with similar previous studies underscore the contribution of ESBL producing Enterobacteriaceae spp in multidrug resistance outbreaks in healthcare facilities. The findings call for an urgent assessment of the national burden of ESBL-producing Enterobacteriaceae which cause UTI and other clinical diseases including the antimicrobial susceptibility patterns. Such efforts would have paramount importance to guide local treatment and care guidelines.
Antibiotic resistance levels of Enterobacteriaceae are reported to be high in several resource-limited settings. For example, in Iran, resistance to trimethoprim-sulfamethoxazole (93.6%), ciprofloxacin (40.4%), and tetracycline (84.5%) [29] was reported. Similarly, in Bangladesh, 97.8% for trimethoprim-sulfamethoxazole   BioMed Research International was reported [34]. These figures are considerably lower in high-income countries, for example, in the United States, resistance to ampicillin (55%), amoxicillin/clavulanic acid (10%), trimethoprim-sulfamethoxazole (24%), and nitrofurantoin (14%) of Enterobacteriaceae was reported [31]. These findings present a concerning public health problem, which requires a coordinated action to generate more data in understanding the magnitude of the problem and designing interventions that help to mitigate the issue.

Conclusion
Our research finding evidenced that ESBL-producing bacteria are prevalent among children with UTI. From the total bacteria species isolated, Enterobacteriaceae contributed to the majority of the isolates. Most of the Enterobacteriaceae were isolated from patients who attended inpatient department. ESBL-producing Klebsiella species were the most frequent Enterobacteriaceae followed by E. coli. Majority of Enterobacteriaceae had resistance to commonly prescribed antibiotics. The drug with preserved efficacy for ESBL producers and non-ESBL producers is nitrofurantoin.

Data Availability
Data can be provided by the principal investigator upon reasonable request.

Conflicts of Interest
The authors declare that they have no competing interests.

Authors' Contributions
AA conceived the study, analyzed and interpreted the data, and prepared the manuscript draft. BTT guided the analysis, interpreted the data, and helped in drafting the manuscript. MW and DN conceived the study and helped during analysis, and BL helped during analysis.