Risk factors for resistance to antimicrobial drugs among extended-spectrum beta-lactamase producing Escherichia coli from urinary tract infection inpatients in tertiary health care facilities in Zanzibar: a retrospective health facility-based study

¹Department of Life Sciences, Open University of Tanzania, Dar es Salaam, Tanzania, ²Zanzibar Health Research Institute, Zanzibar, Tanzani, ³National Institute for Medical Research, Muhimbili Research Centre, Dar es Salaam, Tanzania, ⁴Department of Chemistry, University of Dodoma, Dodoma, Tanzania, ⁵Department of Medicine and Cosmetics Zanzibar Food and Drugs Agency, Zanzibar, Tanzania

^&Corresponding author
Muhiddin Hamada Omar, Department of Life Sciences, Open University of Tanzania, Dar es Salaam, Tanzania

Introduction    

Antibiotic resistance due to extended-spectrum beta-lactamases (ESBL) found in Enterobacteriaceae has rapidly become a public health threat [1]. In return, overuse of antibiotics accelerates the magnitude of the problem worldwide [2]. The predictive statistical models suggest an estimated 4.95 million bacterial-associated deaths, out of which the bacterial AMR claimed 1.27 million lives [3]. Africa is on the verge of facing the greatest AMR threat, with most countries currently dealing with infections caused by resistant organisms. The continent is threatened by a higher morbidity and mortality, expensive antimicrobial related illness resulting in extended hospital stays, and place a greater burden on health systems due to infections caused by microbial resistant organisms [4]. Enterobacteriaceae bacteria, including Escherichia coli (E. coli) have been pathogens of death associated with AMR, accounting for 3.57 million lives lost due to AMR in 2019 [5]. The overuse and misuse of extended-spectrum antibiotics contribute to the emergence of ESBL-producing E. coli. As a result, urinary tract infection (UTI) due to extended-spectrum β-lactamase (ESBL)-producing E. coli has become a significant public health problem worldwide [6]. The emergence of antibiotic resistance among ESBL-producing E. coli causing UTIs against commonly used antibiotics has been documented in different countries [7]. However, a few studies have been conducted in developing countries to determine the risk factors associated with resistance to ESBL-producing E. coli isolated from urine samples. Risk factors for antibiotic resistance differ due to prescription practices and the use of antibiotics in the community [8]. In Zanzibar, there has been an increasing level of antimicrobial resistance of ESBL-producing E. coli from urinary tract infection to third-generation cephalosporin antibiotics each year. Identifying risk factors will help improve the treatment through the prescription of the right and effective antibiotics and preventive measures to limit the transmission of resistant infectious pathogens in hospital and community settings to alleviate the magnitude of the problem.

Methods     

Study design: the retrospective analysis was carried out, covering three years from 2018 to 2020. All patients' medical information with UTIs admitted in tertiary hospitals was retrospectively recorded.

Setting: the study utilised data collected from five purposively selected clinical, microbiological laboratories that provide bacterial culture services and drug sensitivity testing and are also attached to the public and private tertiary hospitals in Zanzibar.

Participants: all documented UTI patients aged 2 to 72 years and who meet the inclusion criteria for the WHO definition of community and hospital-acquired UTI. The average age of patients is 33.33±11.62 years. In addition, 5220 (64.4%) participants were female, and 2812 males (35.1%) were male.

Variables: social demographics and clinical characteristic including, prior hospitalisation, recurrent UTI, prior antibiotic use, dose incompleteness, duration of hospital stay and prior use of a catheter were considered as dependent variables. Antimicrobial resistance of ESBL producing E. coli was considered as independent variable. A prior hospitalisation was defined as a case that had been hospitalised for more than one day before the diagnosis of a urinary tract infection [9]. Foley catheterisation history was defined as cases with catheterisation for more than three days before UTI diagnosis [10].

Data source/measurements: the study used patient´s data collected from medical records at tertiary health centres. Specific patient information, such as socio-demographic and clinical information, was gathered from admitted health care facilities. Positive urine cultures and sensitivity records were obtained from five purposively selected laboratories that provide bacterial culture services and drug susceptibility testing and are also attached to tertiary health care facilities. Medical records of all patients documented with UTIs and found cultured positive for ESBL producing E. coli from corresponding laboratories were used to obtain information on ESBL-producing E. coli isolates. The patient's information was retrieved from the sample request form upon arrival in the laboratory.

Samples collection: in all the participated laboratories, standardized urine collection, transportation procedures were implemented. Urine samples were obtained by clean midstream catch method or balder catheterization. The samples were transported in cold boxes and kept at 2-8°C until ready for processing as described in the standard operating procedures.

Laboratory sample analysis: in all laboratories, urine specimens were processed following standard operating procedures and Clinical and Laboratory Standards Institute (CLSI) guideline [11]. Conventional microbiological methods were used to identify the ESBL-producing E. coli isolate. Moreover, screening of ESBL-producing E. coli was carried out through a zone of inhibition of ≤ 25 mm for ceftriaxone and/or ≤ 22 mm for ceftazidime and or ≤ 17 mm cefpodoxime and or 27 mm for cefotaxime considered as positive ESBL producer [12]. ESBL-producing E. coli was confirmed phenotypically using the combined disc method. Discs containing ceftazidime (30 μg) and a combination of discs containing clavulanic acid (30 μg + 10 μg) were placed 25 mm apart. The isolate was considered an ESBL producer when an increase in zone of the diameter of ≥5 mm in the zone of inhibitor for ceftazidime + clavulanic acid compared to ceftazidime alone was confirmed as ESBL-producing E. coli according to CLSI guideline [12]. For quality assurance, the reference strain of E. coli (ATCC 26122) and Klebsiella (ATCC 75674) were used in quality control for culture and susceptibility tests as well as for the detection of ESBL detection. Antibiotic's susceptibility test was carried out using Kirby-Bauer's disc diffusion method, and interpretation of the results, whether resistant, susceptible or intermediate, were based on recommendations by Clinical and Laboratory Standard Institute guidelines [12].

Bias: there was any bias carried out in this study. To ensure the validity and reliability of data, laboratories quality control measures were considered during this study. All laboratory procedures were done in accordance with standard operating procedures and CLSI guidelines. The reference strain of E. coli (ATCC 26122) and Klebsiella (ATCC 75674) were used in quality control for culture and susceptibility tests as well as for the detection of ESBL detection.

Study size: the study used retrospective data, therefore there was no statistic formula used to estimate sample size. A total of 8,104 urine samples of UTIs patients were collected for culture and sensitivity from five purposively laboratories, out of the 491294 suspicious cases of UTIs patients reported in five hospitals.

Data analysis: the data were analyzed using Statistical Package for Social Science version 21 (Chicago, Inc). The frequency, mean, media and standard deviation were used. The annually differences in antibiotics resistance were compared using the Chi-square test and Fischer's exact test. Univariates by binary logistic regression was performed to determine statistical significance with association of risk factors and resistance to ESBL producing E. coli. This was based on those variables with P-value 0.05 or less in the univariate logistic regression and was considered for the multivariable logistic analysis. Odds ratio (OR) and 95% confidence interval (95%Cl) were used to show the association between risk factors and antibiotics resistance of ESBL E. coli.

Ethics considerations: permission to use the data was obtained from the respective hospital administrations. Ethical approval was obtained from the Zanzibar Health Research Ethical Review Committee of the Zanzibar Health Research Institute, with approval number ZAHREC/O4/ST//32.

Results     

Prevalence of urinary tract infections based on demographical characteristics: the average age of the study participants was 33.33±11.62 years, with a minimum age of 2 years and maximum age of 72 years. In terms of high age, prevalence of UTI was reported at the age between 16-30 years at 550 (30.4) as compared to the lowest of 320 (17.7%) in the age group <15(p=0.01). With respect to sex, the prevalence was 1120 (61.8%) in females and 691 (38.1%) in males. A total of 8,104 urine samples were cultured. A total of 8,104 urine samples were cultured, while ESBL-producing E. coli was isolated in 421 (23.2%) samples.

Distribution of urinary tract infection within three years period: a total of, 491294 suspicious cases of UTIs were reported; the least number of cases reported in 2018 at 29.0%, followed by 33.7% in 2019, while in 2020, 37.2% (p = 0.02). The overall prevalence of UTI caused by ESBL producing E. coli from UTI in patients was 23.3% during 3 years´ time period.

Trend of antimicrobial resistance of ESBL producing E. coli isolated in urinary tract infection: from Table 1, ESBL producing E. coli showed the highest level of resistance to antibiotics reported to ampicillin 406 (95.9%), 398 (94.6%) tetracycline, 393 (93.2%) ciprofloxacin, 384 (90.8%) norfloxacin, 379 (89.5%) nalidixic acid respectively. Low level of resistance was reported to chloramphenicol 155 (51.6%) and least resistance to amikacin 116 (27.5%), and zero resistance to meropenem. Significant difference observed to ampicillin (p = 0.04) and ceftriaxone (p = 0.002) and nalidixic acid (p = 0.03).

Risk factors associated with antimicrobial resistance in ESBL producing E. coli: univariate logistic regression analysis of social-demographical and clinical characteristics with antimicrobial resistance in ESBL-producing E. coli shown in Table 2. The age, sex, recurrent urinary tract infection, previous use of the catheter, dose incompletion, prolonged hospital staying and history of taking antibiotics were subjected to regression model against antibiotics. All variables that had a statistical significance defined by a p-value<0.05 were further entered in a multivariate logistic regression model. The only risk factors that impose the effect of antimicrobial resistance among E. coli isolate from urinary tract infections commonly used antibiotics were the previous hospitalisation to ciprofloxacin (OR 1.40) and ceftriaxone (OR 2.43). Recurrent urinary tract infection is significantly associated with the resistance to ceftriaxone (OR 2.31), ampicillin (OR 1.86), ciprofloxacin (OR 2.51), cefpodoxime (OR 3.42). Prior to taking antibiotic to ampicillin (OR 3.2), ciprofloxacin (OR 2.12) and ampicillin- cloxacillin (OR 2.78). Extended hospital stay to ampicillin (OR 1.62), ceftriaxone (OR 3.35) and ciprofloxacin (OR 2.33) and ampicillin-cloxacillin (OR 2.74). Our study did not find a significant association between age, sex, dose non-completion, catheter use to antibiotic resistance (p>0.05) (Table 3).

Discussion     

In this study, a high prevalence of UTI was reported in the age group between 15-30 years compared to others; this is in line with a study from Nepal [13,14]. However, this contradicts studies which were conducted in Palestine [6]. This is the most reproductive age; hence are at high risk of UTIs [10,15]. Female patients reported the highest prevalence of UTIs compared to males. This finding is similar to studies reported in Korea [16]. However, other studies found statistical significance in male than female patients [15]. Having short urethra of the urinary tract in females makes it easy for uropathogens to reach the internal surface of the reproductive tract [17] and is easy for uropathogens to get into the internal surface of the female reproductive tract [18].

We found a greater significant association between antimicrobial resistance in urinary tract infections and previous hospitalisation, recurrent urinary tract infection, prolonged hospital stay, and previous antibiotic use. The risk factors have a strong association with resistance to ampicillin, ciprofloxacin, ceftriaxone, cefpodoxime, ampicillin-cloxacillin) respectively. Similar findings were reported in Tanzania Mainland [19] and China [20]. This could be explained as the indicator for hospital-acquired resistance associated with inadequate infection and prevention control measures in hospital settings. In addition, with received antibiotics in community pharmacies without a medical prescription [3]. Prolonged hospitalisation increases the duration of exposure to broad-spectrum antibiotics. This increased overuse of antibiotics causes mutations in gene sequences encoding for β-lactam antibiotics [21]. The association of recurrent UTIs with antibiotics resistance agrees with the study [22]. This suggests reducing the efficacy of most antibiotics for treating UTIs, including cephalosporin third generation.

The overall prevalence of UTI is 23.3% is relatively higher compared to the studies in Kenya at 15.7% [23], Uganda 19.5% [10]. The elevated levels of UTIs may be linked to the fact that the island, to a larger extent, relies on imported animal feeds and feed products from other places where hygiene is limited. A high level of resistance to the commonly used antibiotics, tetracycline, is similar to the recent reports in Nigeria [24]. Ultimately, most of these drug groups are relatively cheap and are more likely to be dispensed over the counter. As a result, the over-use of antibiotics increases selection pressure and accelerates antimicrobial resistance [25]. Moreover, the broad-spectrum nature of these antibiotics such as ciprofloxacin, and their widespread used in the treatment of opportunist bacterial infections in AIDS patients and typhoid fever, can accelerate the risk of resistance [26]. These findings suggest the critical need to implement a hospital stewardship program to facilitate monitoring antimicrobial consumption and use in community and hospital settings for better treatment outcomes.

Cephalosporin resistance is the broad-spectrum third-generation antibiotics, including ceftriaxone and cefotaxime, and ampicillin-cloxacillin is attributed to their β-lactam nature, targeted by β-lactamase enzymes, resulting in the inactivation of the ESBL enzymes and hence accelerated resistance [13,27]. Similar findings were previously reported in Spain [28]. Nonetheless, a low level of resistance was reported in a study reported in Korea [29]. The low resistance rate to amikacin and chloramphenicol is similar to the findings in Egypt [30]. This is attributed to its clinical use through an intravenous (IV) or intramuscular (I.M.) administration in injectable formulation for the treatment of complicated and recurrent UTIs, and thereby, they are less available over the counter without a prescription. In addition, chloramphenicol, is rarely prescribed due to its high risk of aplastic anaemia. Surprisingly, as with the study by Moyo et al. (2020), this study could not find any resistance to meropenem, suggesting its best performance in treating UTIs and a drug of choice for ESBL.

Conclusion     

The prevalence of urinary tract infections has risen steadily within three years. Antimicrobial resistance of ESBL-producing E. coli to commonly used antibiotics has been reported to be very high and alarming. Meropenem, cefotaxime, chloramphenicol, nalidixic acid, norfloxacin, and gentamicin are drugs of choice against ESBL-producing E. coli. Previous hospitalisation, extended hospital stay, recurrent urinary tract infection, and prior antibiotic use are significant risk factors for increasing antibiotic resistance in Zanzibar. A strategic plan and intervention are required, including establishing a national antimicrobial surveillance system and a hospital antimicrobial stewardship program. In addition, improvement of the existing laboratory capacity for detecting antimicrobial resistance pathogens with specific ESBL producing Enterobacteriaceae pathogens

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Muhiddin Omar, Andrew Kilale, Emmanuel Kigadye, Angaza Gimbi contributed to the study design and data review and wrote the manuscript; Muhiddin Omar collected data. Isaack Onoka and Burhan Simai analysed the data and prepared the tables. All the authors have read and agreed to the final manuscript.

Acknowledgments     

The author wishes to express their gratitude to all staff at all hospitals and laboratories contributing to the availability of data. This includes physicians, microbiology laboratory scientists and data scientists to facilitate data collection and analysis.

Tables     

Table 1: trend of antimicrobial resistance of ESBL producing E. coli isolated in urinary tract infection (2018-2020)

Table 2: univariate analysis of risk factors for ESBL E. coli antibiotics resistance

Table 3: multivariable analysis of independent risk factors for ESBL E. coli antibiotics resistance

References