The prevalence of UTI in the present study was 16.7% (95% CI 12.7-20.4). It is in line with the study done in Addis Ababa 15.8% . However, it is lower than the findings of the studies in Gondar, 26.45%  and Hawassa, 27.5% . This difference might be a female predominance to Gondar University Hospital (60.94%) in contrast to our study participants (44.8%). The majority of male study participants (71.7%) in the Hawassa University Hospital were not circumcised. Therefore, a tight foreskin may interfere with the normal passage of urine and can prevent fully emptying of the bladder [25, 26].
Though UTI can be caused by both Gram-negative and Gram-positive bacteria, Gram-negative bacteria are the most common cause of the infection, because the agents are the normal constituent of the normal intestinal microbiota . Acquisition of UTI starts with periurethral contamination by a uropathogen inhabiting in the gut, followed by colonization of the urethra and successive migration of the pathogen to the bladder . Among these, E. coli is the predominant one [14, 29, 30]. Similarly, the present study reveals that 88% (44/50, 95% CI 80-96%) of the isolate are Gram-negative bacteria. Of these, E. coli accounts for 63.63% of Gram-negatives followed by Klebsiella species (15.9%), and Citrobacter species (13.63%) are the predominantly isolated uropathogens. This predominance might be due to their unique structures such as flagella and pili, which help for their attachment to the uroepithelium increases risk for infection . This finding is considerably similar to studies done in Nigeria  for Citrobacter species, 10.7%, E. coli 57.4% in Iran  and E. coli 56% and Klebsiella species 19% in Scotland . A study in Bangladesh reported the same prevalence of E. coli, 63.3% of the total Gram-negative bacterial isolates .
About 60% (30/50) of the isolate and 72% (20/28) of E. coli and Klebsiella species were isolated from females. This might be due to poor hygienic conditions, proximity of anal and urethral openings and relatively wide urethra .
All Gram-negative bacterial isolates were 100% resistant to ampicillin followed by augmentin (68.18%) and tetracycline (65.91%). Similar findings were seen in four hospitals of Scotland  93% and 100% in Pakistan  for ampicillin respectively. The reason may be continuous use of these drugs for many years, easily available without professional prescription from pharmacy, self-prescription and the tendency of patients using relatively cheaper antibiotics for all types of infection and misuse (overuse or underuse). Meropenem 97.3% were sensitive, it might be the unavailability of this drug in the area. As table 2 shows: ciprofloxacin, cefoxitin, and ceftazidime were showed the best performance for E. coli. This is almost similar results to the findings from Gondar Hospital and Addis Ababa, Yekatit 12 hospital . According to the findings in Gondar Hospital, the sensitivity of gentamicin and cotrimoxazole to E. coli was 80%, 73%  respectively. However, to the present study, the susceptibility to the above-listed drugs were 46.4% and 42.9% respectively. This might be due to the gradual increase of drug resistance/selective pressure of bacteria to the drug /mutation, the difference in antibiotic practices in the study area. Similar research from Iraq revealed that 75% sensitive for chloramphenicol .
All Gram-positive isolates showed 100% sensitive to nitrofurantoin. Similar results were obtained in Gondar hospital (77%)  and Addis Ababa,Yekatit 12 hospital (100%) . The reason for the effectiveness of this drug might be, due to the nature of having multiple mechanisms and site of actions of the drug with a non-specific attack of protein synthesis, would reduce the ability of bacteria to produce resistance. Limited access to the drug, narrow-spectrum nature of the drug. However, it is 71.4% resistant to the finding in Hawassa Hospital . Higher resistance to this antibiotic is perhaps due to their widespread and wrongly use as empirical therapy.
Furthermore, the overall MDR prevalence was 66% (95% CI 52-78) which comprises Gram-negative (70.5%) and Gram-positive bacteria (33.33%). This is comparable to the findings in the Gondar hospital pediatric patients 58.53% , the Addis Ababa, Yekatit 12 hospital 73.7% . Although antimicrobial resistance comes primarily as a result of selective pressure to place on susceptible microbes by the use of therapeutic agents, there are also further multiplying factors for the spread of resistance. These are antimicrobial prescription in many resources limited countries is almost entirely empirical and based on surveillance data obtained during the survey. This result is only shown at a time of the situation, not for another time. Broad-spectrum agents are frequently used due to lack of susceptibility data, easy availability of antimicrobials in non-controlled pharmacy gives the chance to buy easily as a commodity: these drugs might be sub-standard/poor quality, not finishing the full course of treatment when they feel better, over-prescription due to a poor diagnostic set-up or fear of loss of follow-up .
The emergence of resistant strains among uropathogens is alarmingly increased with different resistance patterns . Acquisition of resistance might be either mutational (i.e. changing the target site of a bacteria within its genetic material) or acquisition of new genetic material from other bacteria. This problem is also magnified by an irrational use and poor administration of drugs. Once a patient acquires resistant strain bacteria, then it transfers antibiotic resistance genes to other bacteria, by this means the problem becomes challenging to control .
The present study also demonstrated the prevalence of ESBL which was 19.4% (95% CI 6.5-32.3%). The prevalence is increasing; due to delayed initiation of appropriate treatment and the spread of the strain to the community or in hospital environment. Inappropriate use of carbapenems, increased use of third-generation cephalosporins and quinolones in the community, stool-mediated infections, and the existence of the ESBL gene in the plasmid are other possible factors . The ESBL producing strains are commonly resistant to other antimicrobial agents because of mobile genetic elements encoding other antimicrobial resistance determinants and/or chromosomal mutations. The co-resistance to other agents limits the antimicrobial treatment options available and may enable selection for ESBLs by non-beta-lactam antimicrobials such as the aminoglycosides and fluoroquinolones .
Furthermore, the history of previous UTI and male uncircumcision was the independent risk factors for the acquisition of UTI. There are two possible reasons for the previous history of UTI as the cause of recurrent UTI: these are; frequent repeat ascending infections and persistent infections in the bladder. Gram-negative bacteria are the predominant cause of UTI. Of these uropathogenic E. coli is the leading strain . In studies using pulsed-field gel electrophoresis, 52%–77% of recurrent UTIs were caused by uropathogenic E. coli strain which was identical to the primary infecting strain . This due to E. coli has an adhesion structure for the progression of UTI .
Another reason suggests that, E. coli could replicate intracellularly, form a loose collection of bacteria, and then escape into the bladder and could form a complex of intracellular bacterial communities (IBCs) within the superficial umbrella cells of the bladder, which could be formed after 4–16 h of bacterial infection and then develop a persistent quiescent intracellular reservoir after 2 weeks . These IBCs could be quiescent for extended periods, despite antibacterial therapy and then re-emerge to cause recurrent UTI .