Drug resistance and plasmid profile of uropathogenic Escherichia coli among urinary tract infection patients in Addis Abeba

Introduction: Urinary tract infection is a major cause of morbidity and mortality worldwide. Uropathogenic Escherichia coli bacteria are the most common cause of urinary tract infections. Drug resistant Escherichia coli is results in high levels of treatment failure and can be a significant threat to survival of patients. Methodology: Escherichia coli bacteria were isolated using culture and conventional biochemical tests. Antimicrobial susceptibility testing and plasmid profile were performed using the Kirby Bauer disc diffusion method and plasmid analysis. Data was processed with SPSS version 16.0 and Epi-info version 3.4.1 software. Results: The highest proportion of Escherichia coli isolates was resistant to (86.5%) to ampicillin, followed by ceftazidime (84%), ceftriaxone (80.5%), tetracycline (80%), trimethoprim-sulfamethoxazole (68.5%) and cefotaxime (66%). Escherichia coli isolates were most susceptible to meropenem (100%), imipenem (100%), amikacin (97.5%), nitrofurantoin (95%), ciprofloxacin (85.5%), norfloxacin (85%), chloramphenicol (83.5%), gentamycin (80%) and nalidixic acid (79%). Multidrug resistance (MDR) was observed in most (96.5%) E. coli isolates. Plasmid analysis revealed the presence of plasmid(s) in 165 (82.5%) of the E. coli isolates many of which had a plasmid size of 23 kb. Conclusions: The overall incidence of antibiotic resistance (including MDR) among E. coli in this study was high to commonly used antibiotics, but no drug resistance to meropenem and imipenem was observed. Periodic monitoring of the drug resistance pattern is essential for better management of urinary tract infections, which has direct impact on the outcome of the patient.


Introduction
Urinary tract infections (UTIs) are one of the most common human infections and a major cause of morbidity and mortality worldwide [1][2].UTIs are also a major cause of sepsis, which has a mortality rate of 25% and results in more than 36,000 deaths per year in the USA [3].It has been estimated that globally, symptomatic UTIs result in as many as seven million visits to outpatient clinics, one million visits to emergency departments, and 100,000 hospitalizations annually [4].In Africa, urinary tract infections are the most common causes of morbidity and mortality [5][6][7].Drug resistance in urinary tract infection, is a major concern in Africa including in Ethiopia [5][6][7].Escherichia coli is the main agent causing urinary tract infections, accounting for up to 80% of cases.Serotypes of Escherichia coli consistently associated with UTI are designated as uropathogenic E. coli (UPEC) [8].Uropathogenic E. coli are implicated in 70-90% of community acquired UTIs and 50% of nosocomial UTIs [9].Antibiotic resistance, including multidrug resistance, is an increasingly serious problem in UPEC [10].The high antimicrobial resistance of UPEC significantly reduces the therapeutic options and increases the treatment costs and mortality rates [11].Drug resistance ofUPEC to commonly used antibiotics such as ampicillin, amoxicillin-clavulanic acid, norfloxacin, cefuroxime, ceftriaxone and cotrimoxazole [6,7,11,12] has been reported.So, determination UPEC drug resistance patternsis important for appropriate treatment of urinary tract infections.The aim of this study was to assessdrug resistance, including plasmid profiles, of uropathogenic Escherichia coli among urinary tract infection patients in selected health facilities of Addis Ababa, Ethiopia.

Subjects and uropathogenic E. coli isolates
A total of 780 subjects in three hospitals (Tikur Anbessa Specialized Hospital, Yekatit 12 Hospital and Zewditu Hospital) in Addis Abeba, Ethiopia participated in this study.All subjects had been diagnosed with urinary tract infections.Sociodemographic and clinical data were collected by a questionnaire completed by interview.Mid-stream urine samples (10-20 mL) were collected using sterile containers.A sample was considered positive for UTI if a single organism was cultured at a concentration of ≥ 10 5 CFU (colony forming unit) per milliliter of urine.Escherichia coli isolates were presumptively identified by colonial morphology on MacConkey agar (Oxoid, Hampshire, UK), and further identified and confirmed by conventional biochemical tests.E. coli bacteria were recognized by the following: negative for urease, citrate utilization and hydrogen sulfide generation; and positive for motility, lysine decarboxylase, lactose fermentation, glucose fermentation and indole test [13].

Plasmid DNA extraction and analysis
Plasmid DNA was extracted from E. coli isolates by the alkaline lysis method using Wizard Plus SV Minipreps DNA Purification Systems kit (Promega Corporation, Madison, USA).The manufacturer's instructions (protocol from Promega, USA for Wizard Plus SV Minipreps DNA Purification Systems) were strictly followed.The alkaline lysis method makes plasmid DNA linearized before it is placed in gel electrophoresis.
Twenty microliters (20µl) of the extracted plasmid DNA was mixed with 5µl of 6x loading dye on parafilm and loaded on slots of 0.7% agarose gel electrophoresis stained with 10μL 10,000x GelRed.After electrophoresis for 4 hours at 100 Volt on TAE buffer system, the gel was imaged under UV light (E-gel Imager; life technologies, Waltham, USA).Plasmid number/s wereobtained by counting the number of bands observed on the agarose gel.Standard DNA molecular weight markers (1 kb DNA ladder (Invitrogen, Waltham, USA and Lambda (λ) DNA/HindIII marker (Promega Corporation, Madison, USA) were used to estimate the plasmid size [23].
Data analysis SPSS version 16.0 and Epi-info version 3.4.1 were used for data analysis.A p value <0.05 was considered as significant.

Ethical approval and informed consent
The proposal of this study was approved by the Institutional Review Board of Addis Ababa University, College of Health Sciences.Permission was also obtained from the medical directors of Tikur Anbessa Specialized Hospital, Yekatit 12 Hospital and Zewditu Hospital.Written informed consent was obtained from each patient who participated in the study.

Results
The mean age of study participants was 33.95 years ± 14 SD; two thirds were female (Table 1).Most of the patients came from the Outpatient Department (OPD) but others were inpatients on the wards (13.5%); there was no difference in the E. coli isolation rate between these two groups (χ 2 0.067 p = 0.796).No significant association was found between gender, age, educational status or marital status, and Escherichia coli isolation rate).There was no significant association between site of data collection and E. coli isolation rates

Clinical data of study participants
The study participants had at least two of the following urologic symptoms; the most common complaint was dysuria, followed by urine urgency, urine incontinence, and flank pain at 50-65%, then suprapubic pain and a smaller proportion with fever and chills (Figure 1).

Bacterial isolation
Urine samples were cultured from all patients, and the 200 (25.6%)Escherichia coli isolates were identified by biochemical tests.

Antimicrobial susceptibility patterns
The antibiotic resistance and susceptibility patterns of the 200 Escherichia coli isolates revealed highest resistance to ampicillin followed by ceftazidime, ceftriaxone and tetracycline at similar levels, and to a lesser extent, to trimethoprim-sulfamethoxazole and  cefotaxime.The E. coli isolates were most susceptible to meropenem, imipenem, amikacin, nitrofurantoin, and somewhat less to ciprofloxacin, norfloxacin, chloramphenicol, gentamycin and nalidixic acid (Table 3).Antibiotic resistance profiles showed that almost all of the local uropathogenic E. coli were resistant to two or more antibiotics, i.e., were multidrug resistant.

Multidrug resistance (MDR) in uropathogenic E. coli
The most common multidrug resistance combinations found were to TMP, CRO, AMP, TE, CAZ and CTX, followed by CRO, AMP, TE and CAZ, and CIP, NOR, TMP, CRO, AMP, TE, CAZ and CTX combinations (Table 4).

Plasmid profile
Plasmid analysis showed presence of plasmid/s in more than 80% of the E. coli isolates but they were absent in nearly one-fifth.
One kilo base plus (1 kb plus) DNA ladder and Lambda/HindIII markers were used to determine the size of plasmids (Figure 2).The majority of the isolates (44.5%) carried 1 to 2 plasmids, while the maximum number of plasmids found was 10 (Figure 3).The plasmid size carried by E. coli isolates varies from 0.5 kb to >23 kb.The plasmid size carried by E. coli isolates varied from 0.5 kb to > 23 kb.The majority, about half, carried plasmid sizes of 1.6 kb to 5 kb, while a third had a plasmid size of 23 kb.A few isolates had plasmids of a large size (> 23 kb) (Figure 4).
We found no significant association between drug resistance and presence of plasmids (Table 5).

Discussion
E. coli has been reported to be the most common cause of urinary tract infections [15].We found a higher proportion of UTI in females than in males, as previously reported [16,17,25], and, possibly because of the shorter urethra in females, or injury during sexual intercourse and proximity to the anus [16,25].The highest incidence of UTI was observed in the age groups 26-45, also reported from other studies [18,25].
The participants in our study had at least two of the following urologic symptoms; the most frequent complaint was dysuria followed by urine urgency, urine incontinence, flank pain, suprapubic pain, and a few with fever and chills.These findings are in agreement with results from studies conducted in South Korea [21] and Nigeria [24].
The overall incidence of antibiotic resistance in the E. coli in this study was high and almost all of the strains were resistant to two or more antibiotics.This result is comparable to the high rates of resistance found in India (92.5%) [19], (100%) [20], China (78.1%) [16], (82.6%) [25], but quite a lot higher than observed in South Korea (21.9%) [21], Mexico (30.2%) [22] and another study in India (50%) [23].The high prevalence of multidrug resistance E. coli strains in our study may be explained by the fact that he majority of the isolates has been exposed to several.
The most frequent resistance found was against ampicillin, at a similar high rate of over 80% as was found in other countries: Mexico [22], Nigeria [26], India [18], and Ethiopia [7].Resistance to trimethoprim-sulfamethoxazole, also frequently used in the treatment of UTI, was also common, found in more than two-thirds of strains; comparable frequencies were reported from Nigeria [24] and India [18].Resistance to Ceftriaxone was high, over 80%, as observed in China [16] and Nigeria [26], all of which were much higher than results from other studies in Mexico at 10.2% [22] and Nigeria at 23.3% [24].The high frequency of resistance to Ceftazidime at over 80% was similar to India [27] and China [16], but much higher than found in Mexico (8.5%) [22] and Nigeria (15.8%) [24].These differences in antibiotic resistance patterns could be due to variations in antibiotic prescribing habits among different countries; for example, in some areas people can purchaseand use antibiotics without a prescription and may use it incorrectly, leading to resistance (see below).
Indiscriminate use of antimicrobials by healthcare providers or because of self-prescribing and over-thecounter availability are major risk factors for the development of high levels of antimicrobial resistance, which is common in both developed and developing countries [5,6,10].Other factors contributing to resistance include incorrect diagnosis, unnecessary prescriptions, improper use of antibiotics by patients, and the use of antibiotics as livestock food additives for growth promotion [5,6,24,30].Therefore, proper use of antibiotics could be helpful to tackle antibiotic resistance.That means both implementing good prescribing and dispensing practices, and correct use by patients, who should take the antibiotics for specified period and within a specified time interval as prescribed by the physician.
Clinical isolates of E. coli are known to harbor plasmids of different molecular size ranging from 2-3 kb to 6.5 kb, with a maximum 26 kb [23].The majority of our E. coli isolates carried plasmids with the size between1.6kb and 5 kb, as was also found in India [23].;Similarly, studies in Nigeria revealed that clinical isolates of E. coli which showed multiple drug resistance harbored plasmids with molecular sizes ranging from 2 kb to 6.5 -23 kb to a maximum of 26 kb [24,30].In Iran, reported plasmid sizes ranged from 1 kb to 33 kb [28], similar to our results and to findings in Nepal where the 23 kb plasmid size was common [31].We found no significant association between the presence of plasmids and drug resistance, which is in agreement with a study conducted in Nepal [31].

Conclusion
In this study E. coli isolates from urinary tract infections werehighly resistant to one or more of: Trimethoprim-sulfamethoxazole, Ampicillin, Ceftriaxone, Ceftazidime and Cefotaxime.They still had susceptibility to Amikacin, Nitrofurantoin, Ciprofloxacin, Norfloxacin, Chloramphenicol and Gentamycin.There was no resistance to Meropenem and Imipenem.It is important to periodically monitor the antibiotic resistance patterns to support the choice of treatments for better management of urinary tract infections.The judicious use of antibiotics and the correct implementation of an antibiotic policy in hospitals will help in limiting the emergence and spread of drug resistance.

Figure 1 .
Figure 1.Clinical data of study participants.

Figure 3 .
Figure 3. Number of Plasmid/s carried by uropathogenic Escherichia coli isolates.

Table 1 .
Study participants' socio-demographic status and site of data collection, and association with E.coli isolation rate.

Table 2 .
Association between Escherichia coli isolation rate and clinical data, 2017.
Antimicrobial agents Number of resistant (%) Number of intermediate (%) Number of susceptible (%)

Table 5 .
Association between presence of plasmids and drug resistance of uropathogenic E. coli, 2017.