Bacterial associated urinary tract infection, risk factors, and drug susceptibility profile among adult people living with HIV at Haswassa University Comprehensive Specialized Hospital, Hawassa, Southern Esthiopia

People living with human immunodeficiency virus (HIV) are more likely to develop urinary tract infections (UTI) due to the suppression of their immunity. The aim of this study was to determine the prevalence, risk factors of UTI, and drug susceptibility pattern of bacteria isolated among peoples infected with HIV. A hospital-based cross-sectional study was conducted among 224 HIV positive individuals attending Hawassa University Comprehensive Specialized Hospital (HUCSH) from September 17 to November 16, 2018. Midstream urine was collected from all study participants and inoculated on to Blood and MacConkey agar. Bacterial isolates were characterized by Gram stain and standard biochemical tests. Kirby-Bauer method was used for antimicrobial susceptibility testing. Sociodemographic and clinical data were collected by a semi-structured questionnaire. Data were analyzed using SPSS version 20. A bivariate and a multivariable regression model were employed to determine the association between dependent and independent variables. From the total 224 study participants, 23 (10.3%) (95% CI 6.7–14.7) had culture-confirmed UTIs. The distributions of the bacteria were as follows: Escherichia coli 16 (69.6%), Staphylococcus aureus 2 (8.7%), Klebsiella pneumoniae 2 (8.7%), Enterobacter aerogenes 2 (8.7%) and Pseudomonas species 1 (4.3%). UTI prevalence was also high among study participants with a previous history of UTI and CD4+ count < 200/mm3. Female study participants were about five times more likely to have UTI (AOR 5.3, 95% CI 1.5–19.2). Ninety-three percent of bacteria isolated were susceptible to nitrofurantoin, ceftriaxone, and gentamycin; 87.5% were susceptible to meropenem and norfloxacin; whereas 93.8%, 68.8%, and 62.5% of isolates were resistant to ampicillin, tetracycline, and cotrimoxazole respectively. Multidrug resistance (MDR) was seen in 18 (78.3%) of bacterial isolates.

www.nature.com/scientificreports/ and margin of error (d) 5%, 95 confidence interval, and 10% non-response rate. Based on the above assumption the total sample size was 224.
Operational definition. Urinary tract infection: is the presence of pathogenic microorganisms within the urinary tract in a significant quantity (≥ 10 5 cfu/ml) 23 .
Multi-drug resistance bacteria: are bacteria resistant for greater than two different classes of drug categories 24 .
Sampling technique. To recruit study participants we used a systematic random sampling method. We followed patient flow at ART clinic for one week. The average patient flow per day was 22 and the data collection period was 2 months. By dividing the sample size (N = 224) for the data collection period (48 days) we arrived at the sample size that could be collected per day, which is equal to 5. K value was calculated by dividing the average number of participants per day (n = 22) to participants recruited by day (n = 5), K was 4. By using a lottery method one participant was selected from the 1st 4 attendants then systematically every 4 participants were selected until the required sample size was obtained.
Data collection. Sociodemographic, associated factors and clinical data were collected by attending nurses using semi-structured questionnaire. The participants' current CD4 + cell value was taken from their medical records (It was performed by using BD FACSPresto).

Sample collection.
A urine sample was collected after adequate explanation/information was provided by attending laboratory professionals. Participants were instructed to collect about 30 ml of midstream urine (MSU) for microbiological examination by giving a sterile, dry, wide-necked, leak-proof container. Urine samples were processed immediately at HUCSH microbiology laboratory. If there were a delay the samples were stored in the refrigerator at 2-8 °C.
Urine culture and biochemical test. Using

Data quality assurance.
To ensure the quality of sociodemographic and clinical data, the semi-structured questionnaire was pretested and data collectors were trained. Sterility of culture media was checked by incubating 5% of culture media overnight at 35-37 °C without specimen inoculation. The performance of culture media was checked by suing control strains. Any physical changes like cracks, excess moisture, color, hemolysis, dehydration, and contamination was assessed and expiration date was also checked. Standard strains of E. coli (ATCC 25922) and S. aureus (ATCC 25923) were used as quality control throughout the study for culture and antimicrobial susceptibility tests.
Data processing and analysis. The data was analyzed using SPSS version 20. The bivariate regression model was employed to examine the associations between dependent and independent variables. Based on the bivariate analysis a variables with P value ≤ 0.25 were selected for further analysis using a multivariable regression model. A P value < 0.05 was considered as statistically significant and results were presented by using odds ratio and 95% level of confidence.

Results
Sociodemographic characteristics. In  Factors associated with the prevalence of UTI among HIV positive individuals. Based on bivariate regression analysis variables with P value ≤ 0.25 (sex, age and marital status) were selected for further multivariable regression analysis and female study participants showed at least five times more likely to have significant bacteriuria (AOR 5.3; 95% CI 1.5, 19.2) when compared to male study participants (P = 0.012). However, there was no statistically significant association with age and marital status (P > 0.05) ( Table 2). Based on bivariate logistic analysis variables with P value ≤ 0.25 such as fever, diabetics, previous history of UTI, previous history of catheterization and CD4 count were further selected for multivariable logistic analysis. Previous history of UTI (AOR 4.4; 95% CI: 1.6, 11.7) and CD4 count less than 200 (AOR 4.9; 95% CI 1.2, 18.5) were significantly associated with UTI. However, there was no statistically significant association between UTI and fever, diabetes and previous history of catheterization (P > 0.05) ( Table 3).
Multi-drug resistance pattern of isolated bacteria. None of the isolates was susceptible or resistant to all the drugs in the testing panel. Among the total isolates (n = 23), 18 (78.3%) were MDR. From 21 g negative bacterial isolates 16 (76.2%) showed MDR of these, 12 (57.1%) E. coli were MDR (Table 5). www.nature.com/scientificreports/

Discussion
The overall prevalence of UTI among HIV positive individuals in the current study was 10.3%. The result of this finding is consistent with other studies carried out in Gondar, Ethiopia (10.7%) 12 and Jimma, Ethiopia (12%) 11 . While a high prevalence of UTI was recorded from India (77.5%) 27 , South Africa (48.7%) 19 , Warsaw (23.2%) 5 , and Nigeria (21.1%) 6 . On the other hand, a low prevalence of 5.8% was recorded from Jos metropolis, Nigeria 10 . This difference may be due to the difference in sample size, the degree of the immune status of the study participants, ART use and geographical variation. Even if CD4 cell count rise as a result of initiation of HIV treatment other factors such as old age, other chronic disease can increase the risk of UTI. In this study, HIV infected females had about 5 times the chance of developing UTIs compared to HIV infected males (P = 0.012). The finding of this study is in line with reports from other parts of Ethiopia. A study from Jimma, Ethiopia reported a high prevalence of UTI among females than males HIV positive individuals 11 . According to a study from Addis Ababa, Ethiopia HIV infected female study participants were three times more likely to have significant bacteriuria 14 . Additionally, a study from Gondar found a high prevalence of UTI among females 12 . Evidence from various epidemiological studies showed that UTIs were more common in females than in males 2,28 . High prevalence of UTI among female participants may be due to females have shorter and wider urethra, lack of prostatic fluid, and having moist urethra. Additionally, mechanical introduction of pathogens into the bladder and trauma increase the risk of UTI among females irrespective of their HIV serostatus 3,29 . But the finding of the current study is not comparable to the study conducted in Nigeria that reported a high prevalence of UTI in males than females 30 .
In the current study, participants with the previous history of UTI were about 4 times more likely to develop UTI (P = 0.004). This finding agrees with the studies conducted in Gondar, Ethiopia 12 and Addis Ababa, Ethiopia 14 . This might be due to the presence of resistant strains from those who had the previous history of UTI.
Urinary tract infections appear to be multifactorial in patients with HIV infections as CD4 + level declines 13 . In the current study, the distribution of UTI according to CD4 + count showed that study participants with CD4 + count < 200/mm 3 had a chance of 4.9 times to develop UTI (P = 0.017).This finding was supported by studies  30 . These results imply that as CD4 + value declines the risk of UTI increases. There was no significant association between age, residence, educational status, occupation, and marital status with UTI in this study (P > 0.05).
In the current study, 91.3% of UTI was caused by Gram negative bacteria. We have noted that non-typical bacteria among UTI in our study. The predominant bacterium isolated in the current study was E. coli (69.6%). A similar E. coli predominance was reported from Jimma, Ethiopia (54.3%) 11 and Gondar, Ethiopia (56.1%) 12 . E. coli predominance may be due to E. coli is the most common microorganism in the vaginal and rectal area 31 . In contrast, this study was inconsistent with the finding reported in Ebony State, Nigeria in which the predominant isolates were S. aureus (45.33%) 4 , Cape Coast, Ghana the predominant isolates were S. aureus (40%) and S. saprophyticus (21.8%) 2 . Most of the isolates reported from Tamil Nadu, India were P. aeruginosa (41.9%) 27 . Our finding of K. pneumoniae (8.7%), S. aureus (8.7%) and Pseudomonas species is greater than report from Warsaw 5 . The variation in the type of bacterial isolate may be due to sample collection technique and personal and environmental hygiene, and underlying conditions 13 .
In the present study, 80% of Gram negative bacteria were susceptible to ciprofloxacin, gentamycin, nitrofurantoin, and norfloxacin. The finding of this study is similar to findings reported from other areas 4,12,14 . Whereas 95.2% Gram negative bacteria in this study showed resistance to ampicillin and 57.1% of them were resistant to cotrimoxazole and 69.9% were resistant to tetracycline. Among Gram-negatives isolates, 93.8% of E. coli demonstrated resistance to ampicillin followed by tetracycline (68.8%) and co-trimoxazole (62.5%). Whereas, all isolates of E. coli were susceptible to ceftriaxone and nitrofurantoin followed by gentamicin (93.8%), ciprofloxacin and norfloxacin (87.5%) each and augumentin (68.8%). All K. pneumoniae were resistant to ampicillin Table 3. Bivariate and multivariate analysis of clinical characteristics and UTI among people living with HIV who were attending Hawassa University Comprehensive Specialized Hospital ART clinic, Hawassa, Ethiopia, from September 17 to November 16,2018 (n = 224). AOR adjusted odds ratio, COR crude odd ration, AOR adjusted odd ratio, CI confidence interval, n number. www.nature.com/scientificreports/ and augumentin and all of them were susceptible to co-trimoxazole, gentamycin, ceftazidime, nitrofurantoin, and norfloxacin. All S. aureus isolated in the current study were susceptible to ciprofloxacin, norfloxacin, gentamycine, penicillin, whereas, all of them were resistant to co-trimoxazole and tetracycline. Generally, ciprofloxacin, ceftriaxone, gentamicin, nitrofurantoin, and norfloxacin were most effective for bacteria isolated in the current study while ampicillin, cotrimoxazole and tetracycline were less effective. This agrees with the finding from Ethiopia 14 and South Africa 19 .
In this study, MDR was observed among 78.3% of the isolated bacteria. This was higher compared to the finding reported in Mysore, India (58.3%) 7 and Bangalre, India (48.55%) 16 . But it was lower than a report from Gondar, Ethiopia (95%) 12 and Portharcourt, Nigeria (92.8%) 18 . The high prevalence of MDR seen to commonly prescribed antibiotics in this study might be due to the easy availability of drugs in the community, and inappropriate use of antimicrobial agents.

conclusion
In the current study, the overall prevalence of UTI among people living with HIV was 10.3%. Factors such as sex, CD4 + count < 200/mm 3 , and previous history of UTI were significantly associated with the prevalence of UTI. The isolated bacteria were E. coli, K. pneumoniae, S. aureus, E. aurogenes, and Pseudomonas spp. E. coli was the  www.nature.com/scientificreports/ predominant bacteria. Most of the bacterial isolates were susceptible to ciprofloxacin, ceftriaxone, gentamicin, nitrofurantoin, and norfloxacin. Most of the isolates were resistant to ampicillin, tetracycline and co-trimoxazole. Multi-drug resistant bacteria were common in the current study. As susceptibility of bacteria to various antibiotics vary, management of UTI among HIV infected individuals should be guided by antimicrobial susceptibility testing.
Ethics approval and consent to participate. Ethical clearance was obtained from the institutional review board (IRB) of Hawassa University College of medicine and health science. Then support letter were obtained from the hospital administration. Written informed consents were obtained from each study participants. All methods were carried out in accordance with relevant guidelines and regulation as mentioned by Declaration of Helsinki.