Prevalence, antibiotic sensitivity pattern and genetic analysis of extended spectrum beta lactamase producing Escherichia coli and Klebsiella spp among patients with community acquired urinary tract infection in Galle

Introduction Community acquired urinary tract infections (CA-UTI) are commonly caused by Escherichia coli and Klebsiella spp which are known extended spectrum beta lactamase (ESBL) producers. Objectives To determine, the prevalence and characteristics of ESBL producing of E. coli and Klebsiella spp in the community, and the association of risk factors with ESBL CA-UTI. Methods Descriptive cross-sectional study with urine cultures performed from clinically suspected CA-UTI patients by CLSI standards. Conventional multiplex PCR was performed for gene analysis. Results Cultures were positive in 178 (38%) patients from 465. Majority were females (103, 58%). Most frequently isolated was E. coli (149, 84%) with 68 (46%) ESBL producers followed by 16 (9%) Klebsiella pneumoniae with 4 (25%) ESBL producers. Majority of patients with ESBL CA-UTI were >50 yrs (35/72, 49%) and 13 (18%) children <10 years were present. ABST of ESBL producers revealed high resistance rates for quinolones (41%) and >80% sensitivity for nitrofurantoin, fosfomycin, mecillinam, aminoglycosides and carba-penems. Presence of ESBL genes were 83% CTX-M, 71% OXA, 24% TEM and 9% SHV with one organism often producing more than one gene in 29 isolates (71%). Haematuria and structural abnormalities of urinary tract were significantly associated with increased ESBL CA-UTI (p<0.01). gene was Haematuria and structural abnormalities of urinary tract were significantly associated with ESBL CA-UTI.


Introduction
Extended beta lactamase (ESBL) producing organisms were identified in 1960 due to plasmid mediated antibiotic resistance (CTX-1 gene). Since then several other genes causing ESBL resistance were identified such as CTX-M, SHV, TEM, VEB, and PER. ESBLs render resistance to penicillin, cephalosporins and monobactams. Unlike most of such genes, CTX-M type is known to be linked with multi-drug resistance which includes aminoglycosides, quinolones and cotrimoxazole [1] as well.
ESBLs are basically produced by the enterobacteriaceae family; in particular Klebsiella pneumoniae and Escherichia coli [2]. Other enterobacteriaceae and nonfermenting bacteria like Acinetobacter and Pseudomonas also produce ESBL [3] but prevalence is low. Most of the urinary-tract infections (UTI) are due to Escherichia coli and Klebsiella spp and studying community acquired (CA) UTI is more practical with non-invasive urine cultures comparatively. And these results often reflect antibiotic sensitivity pattern in the community setting.
There had been similar studies in the past. However, many previous studies on risk factors for ESBL were based on the samples received in the routine culture laboratory. Information were often taken from the request forms or bed head tickets are highly unsatisfactory when consi-Vol. 64 Further, in Sri Lanka a lot of patients with CA UTI are treated with empirically at the out-patient department and at general practitioners (GP) and therefore most of these patients were not included in the previous studies as they were not admitted. However, this study was designed to include such patients as well with active sampling and direct interviews with the patient (in both OPD and GP centers) to gather data to see the prevalence of community acquired UTI by ESBL producing coliforms and risk factors associated with the food habits, lifestyle and others.

Objectives
In this study, we investigated the prevalence of ESBL producing Escherichia coli and Klebsiella species in CA UTI. Antibiotic sensitivity pattern and genetic analysis of ESBL producing bacteria, and the association of risk factors and the clinical presentation with CA ESBL UTI were also described.

Materials and methods
A descriptive cross-sectional study was carried out for 4 months from December 2016 to March 2017. Study sites were Teaching Hospital Karapitiya (THK) and 10 selected General practitioners' (GP) centers in the Galle district. THK is the 3 rd largest tertiary care hospital in Sri Lanka which usually sees about 1500 OPD patients and nearly 1500 newly admitted patients daily.

Study population
Galle district population in the area surrounding Teaching Hospital Karapitiya, who had CA-UTI during this time period considered as the study population.

Inclusion criteria
Patients who came to OPD, clinics and GP centers with clinically suspected CA UTI according to the history were included in the study.

Exclusion criteria
Patients who had hospital admission within the last 3 months, or admitted to the hospital for more than 48 hours or patients on urinary catheters were excluded because we wanted to exclude recurrences due to partially treated urinary tract infections and health care associated infections. Patients who were on antibiotics other than prophylaxis were also excluded.

Sample size
Sample size was calculated as 173 according to the study by Dissanayake et al, (2010) [4] in which the ESBL prevalence among CA-UTI patients was 13%. For the calculation the standard formula n=Z 2 pq/d 2 was used with margin of error (d) of 5%.

Data extraction sheet
By an interviewer based structured questionnaire, information on basic demography, clinical features, about past history of UTI, prophylactic antibiotics, structural anomalies, and about possible risk factors (diabetes, malignancy, chronic renal failure, steroid drugs, visit to India, fresh recreational swimming, consumption of fish / chicken, family history of UTI, health care worker in the immediate family) were gathered. We assumed that an average person in Galle district would consume about 50g of chicken per week and about 50g of fish 3 times per week according to the household expenditure survey report 2012/13 [5].

Ethical approval
Ethical approval was obtained from ethics review committees of the Faculty of Medicine, Galle and the Medical Research Institute, Colombo 08.

Methodology
With the help of medical professionals all clinically suspected CA-UTI patients were identified in the OPD and the GP centers. GP centers were selected according to the consent of the particular GP, facilities available for onsite collection and the distance to the laboratory. After taking the informed written consent, the questionnaires were completed by the principal investigator. Only patients with the history suggestive of CA-UTI were included and sample collection was carried out throughout the intended 4 months to complete the target sample size. Patients were guided on the correct way of collecting urine for culture. One sample was collected from each consented patient into sterile bottles and were immediately transported to the laboratory during day time or kept in the 4°C refrigerator if collected after 5pm. Storage time did not exceed 12 hours in any case.
Sample processing was done according to the standard laboratory manual [6]. Urine samples were plated on HiCrome UTI Agar and the plates were read after overnight aerobic incubation at 37°C. Pure cultures of colony count >10 were taken as significant as the samples were taken only from the symptomatic patients. Bacterial Original article species were initially identified using Gram stain and basic laboratory tests and the Gram negatives were identified by using REMEL's RapID TM ONE system to the species level.
Also, cefpodoxime 10 µg, ceftazidime 30 µg, aztreonam 30 µg, and ceftriaxone 30 µg were added to the primary ABST. If their zone diameters were less or equal to the standards, ESBL confirmatory test was performed for all E.coli and Klebsiella spp.
These organisms were further tested with amikacin, imipenem, meropenem, and netilmicin. Only Escherichia coli isolates were checked with fosfomycin, and mecillinam. Identified organisms were stored in the nutrient agar broth at -70°C for conventional multiplex PCR later at the Genetech Institute (Colombo) according to the in-house protocols by the principal investigator.

Data analysis
The comparison of bacteriological profiles was analyzed by Statistical Package for Social Sciences (SPSS) software version 22. To analyze the association between categorical variables Pearson's chi-square test was used.

Results
During the study period, data and urine samples were collected from 465 of clinically suspected symptomatic CA UTI patients from OPD (240), GP centers (150) and from hospital clinics (75). We continued to collect samples throughout the intended study period despite the calculated sample size being 173. We had 178 (38%) culture positives and were recruited in the study.
From that collection, 72 (40.4%) ESBL forming Escherichia coli and Klebsiella pneumoniae, 93 non-ESBL Escherichia coli and Klebsiella pneumoniae and 13 other bacterial species were identified (Table 1).  Majority was females among the total positives (103/ 178, 58%) and among the total ESBL positives (40/72, 56%) but the association was not statistically significant (p=0.637). The mean age was 39.58 years with 90yrs being the maximum and most of ESBL producing organisms were isolated from age between 51-60 years (Figure 1)

ESBL gene analysis
Only 43 samples were tested by conventional PCR due to limited funds. Therefore, only 39 ESBL E.coli isolates (randomly selected) out of 68 totals, and all 4 ESBL Klebsiella isolates were analysed.

Clinical features
Statistical significance was analysed for features of UTI such as fever, dysuria, frequency, abdominal pain, renal angle tenderness, haematuria and anuria. There was significant association between haematuria and ESBL CA-UTI.

Risk factors for ESBL CA-UTI
Among the risk factors we analysed, only structural abnormalities of the urinary tract was significantly associated with ESBL UTI in the community.

Discussion
Community acquired urinary tract infection is one of the major reasons to attend to the medical practitioners. Due to antibiotics resistance of major causative organisms of UTI the antibiotic options in the community has been rendered limited.
Only 39.03% of clinically suspected CA-UTI urine samples yielded positive cultures with Escherichia coli (83%) as the main pathogen. Majority of culture positive patients were females (57.86%). These facts are compatible with most of previous CA-UTI studies [4].
In this community, ESBL producing Escherichia coli and Klebsiella pneumoniae prevalence was 40.44% while Mohomed et al in India [8] found almost same in 2007. However, according to the local study done by Dissanayake et al in 2010 [4], the ESBL prevalence found among the CA-UTI patients was 13%.
We noted that, many primary antibiotics such as coamoxiclav (59% resistance), norfloxacin (48.5%) and cotrimoxazole (43.6%) were shown resistant for E.coli and Klebsiella spp. According to the standard guidelines, to use an antibiotic as an empirical antibiotic it must have been tested >90% sensitive on the causative E.coli of that community [9]. According to that criterion, only mecillinam, nitrofurantoin and fosfomycin can be used as oral empirical antibiotics in this community. However, fosfomycin is currently unavailable in Sri Lanka.
When comparing the current study and the study by  Dissanayake et al [4], while imipenem and amikacin sensitivity has been preserved same, nitrofurantoin and ciprofloxacin sensitivity has decreased by 7% and 24% respectively during the 8 years gap. No previous studies had tested fosfomycin or mecillinam in Sri Lanka.
It had been identified that most of ESBL producing organisms are having CTX-M gene. We found CTX-M gene on 83% of samples while OXA (70.73%), TEM (24.4%) and SHV (9.8%) were also detected. High prevalence of CTX-M can explain the high rates of multiresistance to antibiotics [1].
According to the clinical features we analysed, statistically significant relationship was noted only between haematuria and ESBL CA-UTI. However, several studies have concluded that the clinical picture of UTI cannot exactly predict whether it is ESBL UTI or not [10].
Among risk factors, only structural abnormality of urinary tract was statistically significant in the association with ESBL UTI. In 2013 Søraas et al [11] found several independent risk factors increased the probability of ESBL UTI, namely, travel to Asia, Middle East or Africa either during the past six weeks to 2 years, recent use of fluoroquinolones or beta lactam antibiotics, diabetes mellitus, and freshwater swimming in the past year. They further concluded that increasing number of fish meals will reduce the risk of ESBL producing UTI. In the present study we could not find statistically significant relationship between those same risk factors and some others and ESBL UTI.