In our study, ESBL producing Enterobacteriaceae were isolated in 16 of the 65 (24.6%) environmental samples. In published reports, the percentage of inanimate hospital environments contaminated with ESBL are significantly varied ranging from 3% to 33% of surfaces in patients’ rooms on regular hospitals wards, as reported from Gonder in Ethiopia (14.8%) , Tunisia (4%) , Israel (9%) , Tunisia (4%) , UK (3.1%)  and Pakistan (33%) . These discrepancies could possibly be due to differences in patient colonization load, hospital’s cleaning/disinfection protocols, study design including timing of sample collection and laboratory method used [5, 11, 20, 21].
We found a high proportion of ESBL producers among E. coli and K. pneumoniae isolates which has similarly been reported from studies in Gondar, Ethiopia , Zimbabwe , Gaza in Palestine  and Algeria . Given the fact that Klebsiella spp and Escherichia coli have capacity to survive up to 30 and 16 months, respectively, on inanimate dry surfaces . Moreover, isolation of such Enterobacteriaceae is highly indicative of fecal contamination and poor hand washing practices amongst health workers and patients .
The most prevalent ESBL producer detected in the current study was E. coli (50%) as compared to K. pneumoniae (18.8%) which has similarly been reported from studies in Gaza, Palestine (E. coli 40% vs K. pneumoniae 20%) . In contrast to our result, different findings from most reports in other parts of the world ESBL production was more predominantly found among Klebsiella spp isolates: Zimbabwe (Klebsiella spp 60% vs E. coli 40%) , Algeria (K. pneumoniae 48.5% vs E. coli 22.8%) , Gondar, Ethiopia (K. pneumoniae 42.1% vs E. coli 35.1%) , Sudan (K. pneumoniae 37% vs E. coli (0%)  and on clinical isolates from another study in Ethiopia (K. pneumoniae 72.7% vs E. coli 22.7%) .
The Pediatrics and Medical-Surgical ICU exhibited the highest ESBL producing Enterobacteriaceae, each with (37.5%, 6/16). This may lead to cross-infections between patients in the same ICU because patients present in these ICUs are likely to have reduced immune system due to illness, surgical and mechanical manipulation, and/or the use of immune-suppressors and other therapeutic drugs, all of which increasing patients susceptibility to infections [27, 28].
Bed linens, ventilators, beds and sink were inanimate surfaces observed to be associated with variable degree of ESBL producing Enterobacteriaceae positivity, with 31.3%, 12.5% and 6.5%, respectively. The identification of ESBL producing bacteria on linens, sink, bed and other environmental surface is consistent with reports from the literatures [5, 6, 12, 20]. It is generally assumed that ESBL producing Gram-negative bacteria require moist or damp sites for enhanced longevity [12, 24]. Moreover, bed linens are considered to be high patient-contact surfaces and therefore the detected pathogens might have been shed by the infected/colonized patients occupying the particular beds.
In regards to level of antimicrobial resistance among our ESBL producing isolates, significantly high resistance level was recorded to penicillin groups such as ampicillin (100%) and cephalosporin groups such as cefuroxime (96%), ceftazidime (92%), cefepime (86%), azetronome (80%) and cefotaxime (72%). These results were mostly comparable with results from other studies conducted in Ethiopia (cefpirome, cefpodoxime, ceftazidime, ceftriaxone and amoxicillin with clavulanic acid each with 100% resistance level)  and Algeria where ampicillin resistance level was reported at 98.1% . Probably lack of antibiotic resistance screening and confirmatory testing prior to using these drugs could be responsible for accumulation of such high level of resistance among ESBL producing bacteria in the hospital environments.
High resistance rate was also recorded form ESBL-producing isolates for non-beta-lactam antimicrobials such as, cotrimoxazole (81.3%), ciprofloxacin and chloramphenicol each with (50%). This was in close agreement with studies done in Ethiopia (ciprofloxacin 56%) , from another study in Addis Ababa from clinical isolates (cotrimoxazole 77%, ciprofloxacin 46.3%) ; Tikur Anbessa Specialized Hospital, from clinical isolates, cotrimoxazole (83.6%) and Chloramphenicol (61.8%) . The most active drugs for ESBL-producing isolates were amikacin (81.2%). Low resistance rate of amikacin in our study could be explained by the dearth in usage of amikacin in our study setting.
Reports from other studies showed that multidrug resistance (MDR) to the commonly prescribed antimicrobial agents are more common among ESBL producers found from inanimate hospital surfaces [29, 30]. Hence, unwise use of antibiotics can inadvertently favor emergence of multidrug resistant bacterial strains [11, 28]. The finding from the current study showed that (15/16, 93.8%) of the ESBL producing bacterial isolates were resistant to at least 3 antibiotics, showing how rampant MDR Enterobacteriaceae are in the hospital environments. This finding is higher than reports from Northwest Ethiopia (75%) , Zimbabwe (75%)  and Iran (79.4%) .