Higher prevalence of ESBL producing uropathogenic Escherichia coli among Diabetic patients from a Tertiary Care Hospital of Kathmandu

Background: The increased rate of urinary tract infection (UTI) in immunocompromised patients especially diabetic patients is a major public health problem in adults. Moreover, the infection with multidrug resistant strains producing extended spectrum beta-lactamase (ESBL) is a key obstacle in disease management among such vulnerable population. An immediate proper treatment depends on rapid diagnosis of UTI and screening of antimicrobial resistant pattern with highly sensitive methods which also reduces the possible urinary complications among the diabetic patients. Hence, this study was aimed to determine the occurrence of antibiotic resistant genes for β-lactamases; blTEM and blaCTX-M in uropathogenic Escherichia coli isolates from UTI suspected diabetic and non-diabetic patients. attending Methods: A hospital-based cross-sectional study was conducted in Kathmandu Model Hospital in association with Central Department of Microbiology, TU from June to December 2018. A total of 1267 non-duplicate mid-stream urine specimens from diabetic and non-diabetic patients were obtained and processed immediately for isolation of uropathogens. The isolates were subjected for antibiotic susceptibility testing and ESBL conrmation. Finally, blaTEM and blaCTX-M ESBL genes were screened by using specic primers. Results: The overall prevalence of the urinary tract infection UTI was found to be 17.20%(218/1267) , out of which diabetic patients were signicantly more infected with UTI accounting for 32.29%(31/96) as compared to non-diabetic persons, 15.97%(187/1171). A total of 221 bacterial were from 218 culture positive specimens in which E . coli was a most predominate one; 67.9%(150/221.). Forty-four percent (66/150) of the total E. coli was MDR and 37.33%(56/150) were ESBL producers. Among 56 isolates, 92.3%(12/13) from diabetic patents and 83.0% (44/53) were from non-diabetic patients. Furthermore, 84.85% of the screened ESBL producers were conrmed to


Background
Diabetes mellitus is an emerging chronic non-infectious disease characterized by hyperglycemia occurs due to insu cient production of insulin from pancreas. Out of top ten diseases with higher mortality, diabetes is the one causing around four million deaths worldwide in 2017 in adults [1]. The patients with diabetes are highly susceptible to infectious diseases most frequently by urinary tract infection (UTI). The major predisposing factors for UTI in diabetic patients are impairment in immune system, poor metabolic control and incomplete bladder emptying due to autonomic neuropathy [2][3][4]. Older age is another crucial in uencing factor to enhance the risk of UTI in diabetic patients [5].
UTI involves the invasion of microbes, their multiplication, and colonization in genito-urinary organs [6]. The degree of UTI in the diabetic patients ranges from asymptomatic infection to various severe lower urinary tract infections including cystitis, pyelonephritis and urosepsis. UTI also remains widespread nosocomial infection, is commonly diagnosed in outpatients and inpatients [7,8]. Escherichia coli is the most dominant bacteria for causing UTI followed by other members of Enterobacteriaceae; Klebsiella pneumoniae, Citrobacter spp., Proteus spp., Enterobacter spp. etc [7,9,10]. Recently, the rapid widespread of Multi-Drug Resistance (MDR) bacteria has been the major public health problem especially caused by β-lactamases producing MDR strains. The production of Extended-Spectrum β-lactamases (ESBLs) has helped bacteria to expand their activity even against the newly developed β-lactam antibiotics [11]. ESBL producing microorganisms pose tremendous therapeutic consequences and signi cant clinical challenges if they remain undetected. They confer resistance or decreased susceptibility to narrow and expanded-spectrum cephalosporins and monobactams but do not affect cephamycin and carbapenem compounds [12]. They are usually resistant to uoroquinolones, aminoglycosides, and co-trimoxazole [13]. There are more than 300 different ESBL variants, TEM and SHV being the major type with increasing occurrence of CTX-M enzymes [14,15]. Almost all Enterobacteriaceae harbor ESBL with higher prevalence among E. coli isolates in community-acquired infections.
The increased risk of UTI among diabetic patients and enhanced urinary complication related with MDR strains may pose high morbidity and mortality. Therefore, screening for UTI and its causative agents in diabetic patients is very crucial to enable infections to be properly treated. Moreover, identifying exact drug resistant pattern is equally important to manage the disease and further prevent the development of urinary related complications in diabetes patients [16]. Hence, the study was carried out to evaluate microbiological agents and their resistant pattern by both phenotypic and genotypic methods of UTI in diabetics and non-diabetics. The aim of this study was to investigate if there are any differences between the resistance genotyping of E. coli isolated from Diabetic and non-diabetic patients with reference to bla TEM and bla CTX−M genes.

Study design, duration and site
A hospital-based cross-sectional study was carried out from 20th June to December 2018, among the clinically suspected Urinary Tract Infected (UTI) patients visiting the hospital. The clinical sample processing followed by identi cation of uropathogens were carried out in Microbiology Laboratory of Kathmandu Model Hospital, Kathmandu, while the nucleic acid extraction and PCR ampli cation were done in Department of Microbiology, GoldenGate International College, Kathmandu.

Inclusion and exclusion criteria
This study included the patients of both sexes and all age group attending the hospital with suspicion of UTI infection. The samples which were adequately collected and properly labeled were included in the study. Those samples which were not collected with standard collection procedure, inadequately collected, improperly labeled and samples with visible contamination were excluded in the study. A repeated sample was requested in such cases. The urine samples from the patients whose health status was not mentioned as the diabetic and non-diabetics and those who were taking antibiotic less than 24 hours before visiting hospital were also excluded.

Study variables
The study variables included were age, sex, out-patients, inpatients department, health status including diabetic or non-diabetic and history of antibiotic taking within 24 hours of visiting hospital using standard data form assigned by the hospital.

Sample size
Consent for participation was collected from all participants during enrolment and before data and sample collection. UTI patients or suspected patients were instructed to collect mid-stream urine sample in sterile, clean and leakproof vials. A total of 1267 non-duplicate mid-stream Urine (MSU) specimens were obtained and processed immediately in the Microbiology laboratory of Kathmandu Model Hospital. The adequately collected samples from both sexes and all age groups were included while the improperly collected and poorly labelled samples with visible contamination were excluded for the study. They were requested to repeat the sample collection. Besides demographic data, medical status of patients (diabetic or non-diabetic and inpatients or outpatients) had also been recorded.

Sample processing
The urine samples were cultured onto the Cysteine Lactose Electrolyte De cient (CLED) agar by the semiquantitative culture technique using a standard calibrated loop [17]. The agar plates were incubated at 37 0 C for overnight. The bacteria developed in the pure culture with a load greater than 10 5 CFU/ml were considered as signi cant growth and included in this study. The bacteria were identi ed by standard microbiological procedures including microscopy, colony microbiology and biochemical tests as described by American Society of Microbiology (ASM). Among different bacterial isolates, only E. coli isolates were further processed for molecular studies [18]. isolates [19]. Subsequently, the rate MDR E. coli was determined [20].

Phenotypic characterization of the ESBL producers
The E. coli isolates were screened for possible ESBL production using Ceftazidime (30 µg) and Cefotaxime (30 µg). The suspected ESBL producing E. coli were subjected to Combined Disk (CD) assay using Ceftazidime (30 µg) with Ceftazidime plus Clavulanic acid (30/10 µg) and Cefotaxime (30 µg) with Cefotaxime plus Clavulanic acid (30/10 µg) for phenotypic con rmation [19]. Molecular characterization of bla TEM and bla CTX−M genes: The plasmid DNA was extracted from phenotypically con rmed ESBL producing E. coli by the alkaline lysis method followed by the phenol: chloroform puri cation method [21,22]. A conventional linear PCR was used to amplify the bla TEM and bla CTX−M genes in the extracted plasmid DNA. The bla TEM gene was ampli ed by using a primer with forward nucleotide sequence 5 ' -GAGACAATAAGGGTGGTAAAT-3 ' and reverse nucleotide sequence 5 ' -AGAAGTAAGTTGGCAGCAGTG-3 ' [23].

Bacterial infection among different age and sex groups
Out of the 218 signi cant positive, 22.02% (48/218) were from the male while 77.98% (170/218) were from female. The prevalence of UTI was the highest in the patients of age group 21-40 and least in age group 1-20.
The culture positivity was signi cantly (p = 0.001) associated between gender and the age group (Table 1).

Frequency of uropathogenic bacteria
A total of 221 bacteria were isolated from 218 culture-positive specimens. Out of which, a single bacterium was isolated from each of 215 (98.62%) specimens whereas 3 (1.38%) specimens showed poly-microbial infection. Gram-negative and Gram-positive bacterial isolates were 89.14% (197/221) and 10.86% (24/221) respectively. E. coli was the most predominant Gram-negative bacteria accounting for 67.9% of total isolates followed by K. pneumoniae (14.94%). The coagulase-negative staphylococci were the pre-dominant Gram-positive bacteria. Gram-positive bacteria included 3.17% Enterococcus faecalis, 7.24% Staphylococcus saprophyticus and 0.45% S. epidermidis. There was no statistically signi cant association between the distribution of the bacterial isolates among the diabetic and non-diabetic UTI patients. E. coli was found to be the most predominant bacteria among both diabetic 67.2% and non-diabetic 71.86% patients followed by K. pneumoniae with 14.81% in non-diabetic and 15.63% in diabetic patients respectively.

Antibiotic susceptibility pattern of E. coli
The highest number of the E. coli isolates was resistant to amoxicillin followed by the nor oxacin and third generation cephalosporins whereas the highest number of E. coli was sensitive to nitrofurantoin. The amikacin was the most effective drug against the MDR isolates. Among 127 E. coli from the non-diabetic patients, the highest number of the bacteria (65.35%) was resistant towards amoxicillin. The least number of the bacteria (4.72%) were resistant towards nitrofurantoin. Similarly, out of 23 E. coli from diabetic patients, the highest number of bacteria (65.22%) and the least number of bacteria (8.7%) were observed to resist towards amoxicillin and nitrofurantoin respectively. Among 44 MDR E. coli from non-diabetic patients, the highest number (86.36%) of bacteria was resistant to amoxycillin clavulanic acid and the least number of the bacteria (13.64%) to amikacin and carbapenems. Similarly, out of 12 MDR E. coli from diabetic patients, the highest number (66.67%) of bacteria was resistant to cefepime and the least number of the bacteria (8.33%) to piperacillin/tazobactam and carbapenems. The amikacin was found to be 100% effective. E. coli isolates resisting second line drugs; 7 E. coli from non-diabetic patients and 3 E. coli from diabetic patients showed 100% susceptibility towards colistin, polymyxin B and tigecycline (  (Table 7).
Antimicrobial resistance pattern among the non-ESBL and ESBL E. coli Among the non-diabetic patients, amoxicillin was the most resisted antibiotic whereas 2.4% non-ESBL producing E. coli and 9.1% ESBL producing E. coli showed the least resistance towards nitrofurantoin. Similarly, among the diabetic patients, the non-ESBL E. coli were 100% resistance to the gentamicin and nitrofurantoin while the cephalosporins were the most sensitive antibiotics. The ESBL producing E. coli showed 100% resistance to amoxicillin and the nitrofurantoin was observed to be the most effective antibiotic. The non-ESBL MDR strains among the non-diabetics showed the highest resistance to the amoxicillin-clavulanic acid and cefperazone/sulbactam whereas the ESBL strains showed the highest resistance to the amoxicillin-clavulanic acid ( Table 6).

Detection of ESBL genes
Out of the 56 phenotypically con rmed ESBL producing isolates of E. coli, bla TEM gene was detected in 53.57% isolates during the ampli cation process by PCR. Similarly, the bla CTX−M gene was detected in 87.5% isolates.
Both bla TEM and bla CTX−M genes were detected in 50%. Two (3.57%) strains harboured only bla TEM gene while 37.5% harboured only bla CTX−M . Both the genes were absent in 8.93% isolates. Although, the occurrence of ESBL producers were higher among diabetic patients, screening of ESBL genes showed comparatively less detection than that of non-diabetic patients ( Table 7, Fig. 3).

Discussion
The prevalence of UTI was found to be 17.20% which was lower than the studies by Kattel et al., and Rijal et al. [25,26]. MSU specimens being collected from the patients under treatment, infections due to slow-growing bacteria or those which might not be able to grow in routine culture media used in the study might be the possible causes of a low rate of culture positivity [27].
The prevalence of UTI in female was 77.98% which was signi cantly higher than in male (22.08%). The results were in agreement with the ndings by many other investigators [28][29][30]. The close proximity of the vaginal and anal opening in the female and the absence of the prostatic uid which has the bactericidal property in the female may make them more susceptible to the UTIs [29,31]. The patients of age group (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40) [34]. Similarly, another study found the prevalence of UTI in diabetic and non-diabetic patients to be 32% and 13% respectively [35]. Mubarak et al., 2012 reported higher prevalence of UTI among diabetic patients [36]. It might be due to the diabetic nephropathy and incomplete bladder emptying in hyperglycaemic condition of diabetic patient [37]. Every 10 years of diabetes duration increases 1.9-fold prevalence of bacteriuria [38]. However, the culture positivity among diabetic and non-diabetic patients were not much different accounting for 43.8% and 42.9% respectively in a study conducted in Bangladesh [39]. The prevalence of UTI in female diabetic patients was 46% which was higher compared to prevalence in diabetic male patients 43% [40]. These variations in the results may be due to the differences in the sample size and clinical conditions among the study population as mentioned in many studies.
Among the heterogeneous bacterial etiological agents of UTI, member of Enterobacteriaceae family remains the predominant pathogens. The members of Enterobacteriaceae, being the normal ora of the human intestine, can easily invade and attach the uro-epithelium causing UTI infections. The total Gram-negative and Gram-positive bacteria involved in the UTI were found to be 89.14% and 10.86% respectively. Among the total Gram-negative bacteria, E. coli and Klebsiella pneumoniae were the most predominant ones. Similar results were present in other studies as well [40][41][42]. However, 19.64% of E. coli was involved in bacteriuria followed by 2.7% K. pneumoniae in the study conducted in International Friendship Children Hospital [43], out of which 33.33% were MDR isolates. Similarly, 7.04% E. coli and 2.3% K. pneumoniae were isolated from signi cant bacteriuria in the study conducted by Chander and Shrestha [44].
The bacteria involved in UTI were similar in both diabetic and non-diabetic patients. E. coli was the predominant pathogen involved with 71.86% in diabetic patients and 67.2% in non-diabetic patients followed by Klebsiella pneumoniae i.e. 15.63% in diabetic patients and 14.80% in non-diabetic patients which was in accordance to a study by Jankhwala et al. [35]. Bonadia et al., 2006 found that E. coli constituted 32.5% and 31.4% followed by Enterococcus spp. 9.4% and 14.5% respectively among diabetic and non-diabetic patients [45]. They have also reported Pseudomonas spp. as the third most frequently isolated bacteria in diabetic and non-diabetic male patients with occurrence of 8.5% and 17.2% respectively. Similar to the results in male, the female patients were found to be infected with 54.1% and 58.2% E. coli followed by Enterococcus spp. 8.3% and 6.5% and Pseudomonas spp. 3.9% and 4.7% respectively with and without diabetes mellitus. Akbar 2001 showed that though E. coli was the most common bacteria in both CA-UTI and hospital-acquired UTI in non-diabetics, Pseudomonas spp. was the most common bacterium among the hospital-acquired UTI in diabetic patients [46].
Regarding the antimicrobial susceptibility testing of E. coli isolates, most of them are resistant to amoxicillin and third-generation cephalosporins which could be due to the production of Beta lactamases [47]. The highest number of bacteria was sensitive to nitrofurantoin. Amikacin was another most effective drug against the MDR isolates. Nearly half of E. coli isolates was found to be MDR strains out of which 84.85% of the screened ESBL producers were con rmed to possess either single or both of the bla TEM and bla CTX−M genes. The incidence of ESBL producing E. coli among the outpatients (89.1%) was higher than inpatients (63.64%) which were in contrast to the ndings [43], which revealed that incidence of ESBL producing E. coli among the inpatients was higher than in outpatients. In this study, the ESBL producing E. was found comparatively higher in number than in the studies by Parajuli et al. and Seyedjavadi et al. [48,49]. Pokhrel et al., 2014, andChander andShrestha 2013 [44, 50] revealed even less ESBL producing E. coli. Diabetic patients showed an increase in number of infections as compared to non-diabetic patients. The number, however, was not statistically signi cant. The emergence of more and more ESBL producing E. coli isolates form UTI patients in the recent years in comparison to the previous study can be related with disease chronicity and long antibiotic therapy among the patients suffering from diabetic, heart diseases and other non-communicable complications. These resistant strains are rapidly spreading in different clinical setting. These ESBL producing strains are highly resistant to the oxy-imino-cephalosporins. However, certain β-lactams may not be totally resisted by these strains which create the problem in the treatment procedures.
The prevalence of bla CTX−M gene was found to be higher than that of the bla TEM gene which was in accordance with the study by Pokhrel et al., 2014 andParajuli et al., 2016 [48, 50]. Many other studies have also shown that the genes responsible for the production of CTX-M-type was more prevalent among the tested strains in comparison to the genes encoding other ESBL like SHV-type or TEM-type β-lactamases. Ojdana et al., Vidhya and Sudha 2013 and Valenza et al., too found the higher prevalence rate of the bla CTX−M genes [51][52][53].
However, in contrast to this study higher prevalence of the bla TEM genes was detected in E. coli followed bla CTX−M by Rezai et al. in the North of Iran [54]. The repression of the genes in the bacteria might cause a variation in their prevalence rate. However, the presence of genes is not associated with diabetic condition of patients.

Conclusions
Diabetes mellitus is one of the most rapidly increasing non-infectious diseases among the older age groups (> 40 years). Opportunistic infections including urinary tract infections is another burden for diabetic patients. Moreover, increasing antimicrobial resistance of bacterial uropathogens is causing a huge problem in disease management leading to high rate of mortality and morbidity of diabetic patients. Identifying multidrug resistant causative agents and proper diagnosis in UTI among diabetic patients are crucial for disease management.
However, few routinely used techniques such as disk diffusion are not su cient to con rm such MDR strains including ESBL. So, to avoid treatment failure due to misdiagnosis, new approaches of diagnosis including molecular tools have to be adopted.

Limitations
Due to limited molecular resources and time constraint, all the isolates could not be con rmed for production of ESBL enzymes by CD test and those ESBL producers were not screened for all possible ESBL genes due to lack of primers and limited PCR master mix.

Consent for publish
All authors are read and reviewed the nal manuscript and give consent for publication.

Availability of data materials
Raw data in the excel le was provided.

Competing interests
The authors declared no competing interests.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.