Occurrence and Anti-microbial Susceptibility Pattern of Extended Spectrum Beta ‐ Lactamase Producing Enterobacteriaceae in Waste Water Released from Governmental Hospital of Addis Ababa, Ethiopia

Background Worldwide, come out and dissemination of Extended-spectrum beta-lactamases (ESBLs) producing Enterobacteriaceae has been warning the ecacy of antibiotics to treat an infection. Hospital wastewaters were a reservoir of such kind of resistant bacteria. Currently, the predominant antibiotics used for the treatment of hospitalized patients infected by Gram negative bacteria are the 𝛽 -lactam antibiotics. So it is an important source to investigate the magnitude of ESBLs producing bacteria and their antimicrobial susceptibility pattern. The aim of this study was to determine the occurrence of ESBLs producing Enterobacteriaceae (ESBLs-pE) and their antibiotic susceptibility pattern in wastewater released from ve governmental hospitals in Addis Ababa, Ethiopia. A cross-sectional study was carried out from April 1 to May 31, 2020. A total of 100 wastewaters were collected from ve governmental hospitals in Addis Ababa using a grap-sampling technique. All Enterobacteriaceae were screened for ESBLs production using cefotaxime and ceftazidime as per 29th CLSI guideline. Each screen positive for ESBLs production was conrmed by the combination disk method (CDT) and their antibiotic susceptibility pattern were done using the Kirby-Bauer disk diffusion method on Muller Hinton agar (MHA). Data were entered and summarized using SPSS version 20 software. nutrient agar/TSA was sub cultured and identied based on the following biochemical tests: oxidase, indole, urea, motility, Lysine decarboxylase, citrate utilization and triple sugar iron tests as per the standards of microbiology procedure including E. coli ATCC 25922, as a control culture (26). Following purication and species identied, two –three puried colonies were preserved in Skimmy milk at -80 o C for further characterization, after each isolate was assigned a unique identication number. (ciprooxacine), (cefoxitin) and aminoglycoside (tobramycin) drug Categories. In addition, was tested combinantion drug (amoxaciline-clavunalate and pepracillin-tazobactam), phenicol (chloramphenicol), nitrofuran (nitrofurantoin) and folate pathway antagonist (trimthoprime-sulfamethoxazole) drug families. strains were mostly observed in the tested Enterobacteriaceae isolates by 64%. Almost similar MDR isolate results with ours were recorded in study carried out, in Northwest Ethiopia (Gondar) 81.5% (from hospital environment) (35), in Mekelle: 61.5% (from untreated hospital wastewater) (36), in Biratnagar, Nepal; 69.6% (37), in China: 85.5% (33). However, our report contradicted by the previous study conducted in South Eastern, Nigeria (from three hospital euents) where all the Enterobacteriaceae isolates recovered ( E. coli and Salmonella spp.) were MDR although their patterns of resistance varied (32). In the same talked in this study, a total of 83.3% Citrobacter spp., 64.3% E. cloacae, 62.7% E. coli, and 53.3% K. pneumoniae isolates were identied as the predominant MDR. Our nding, was concordant with other previous studies where the common MDR isolates were, in Addis Ababa; Citrobacter (100%), E. cloacae (66.7%) and E. coli (28.6%) (31), in Biratnagar, Nepal; Enterobacter spp. (100%), Citrobacter spp. (80%), E. coli (62.5%), Klebsiella spp. (33.3%) (37). However, our nding dissimilar with a study carried out in China (E. coli (91.4%) and K. pneumoniae (94.1%)) (33), Ibadan, Nigeria (E. coli (94.8%)) (39), and in Biratnagar, Nepal (Enterobacter spp. (100%)) (37), where the highest MDR proportion for E. coli, K. pneumonia and Enterobacter spp. were indicated.


Introduction
Wastewater refers to water used by humans and loses its quality and usefulness that embrace waste liquid of domestic, agricultural, commercial sources, industries, and hospital sources. Every these anthropogenic activity generates wastewater as a result of physical, chemical and biological quality of water deterioration (1). The e uent from hospital contains a lot of drug-resistant pathogens, a larger form of chemicals, solvents, disinfectants and a lot of risky materials like pharmaceuticals and radionuclides than domestic sewerage (2). As a result of this hospital e uent contains antibiotic residues that are enough to kill susceptible bacteria and at the same time increases the number of resistant bacteria (3). The presence of antibiotic-resistant microorganism in e uent and sewage system line is a growing public health concern (4). Since one of the main ways of diffusion of pathogenic and/or antibiotic resistant microorganism is thru water, soil and air. This phenomenon results multi-drug resistant (MDR) microorganisms that have been revealed in different water sources including rivers, lakes, groundwater and drinking water (5)(6)(7)(8). The discharge of resistant bacteria to the receiving aquatic environment will create public health impact through, carrying transmissible gene, by acting as a vector or reservoir of resistant gene (9,10).
The members of the Enterobacteriaceae are gram-negative bacilli, which are usually resident in the gastrointestinal tract. Instance of such organisms consisting of E. coli, K. pneumoniae, Enterobacter cloacae, Citrobacter freundii and Proteus mirabilis. In patients hospitalized in intensive care units (ICUs), the Enterobacteriaceae holds for about one third of all cases of ICU-acquired pneumonia, one third of all cases of ICU-acquired urinary tract infection, and 10% to 15% of ICU-acquired bloodstream infections (11).
to the dissemination of ESBL-producing bacteria. The permanent presence of detectable antimicrobial levels within the hospital wastewater treatment may powerfully in uence the environment, creating a selective pressure that would be ultimately responsible for the dominance of resistant microorganisms among those present in the habitat.
Our rst objective was to generate information on the occurrence of ESBLs producing Enterobacteriaceae (ESBLs-pE) and their antibiotic susceptibility pattern in Wastewater Released from Governmental Hospital of Addis Ababa, Ethiopia. Since currently almost all hospitals in Addis Ababa neither have wastewater treatment plants (WWTPs) nor functional waste stabilization bond. Our second objective was to assess the frequency of the common Enterobacteriaceae and their antimicrobial susceptibility pattern.

Study setting
A cross sectional study design was employed from March 1 to April 30, 2021 in Addis Ababa which is the Capital City of the Federal Democratic Republic of Ethiopia, the diplomatic capital of Africa and the seat of different international and regional organizations. The city administration is divided in to ten sub city and 116 woreda. It hosts an estimated 3.2-38 million people, which is a 17% share of Ethiopia's total urban population. Presently, Addis Ababa is experiencing an annual growth rate of 3.8% and is speculated to reach 4.7 million residents by 2030. The city covers a Landmass of 540 square kilo meters. The city is located at the heart of the country, at an altitude ranging from 2,100 meters at Akaki in the south to 3,000 meters at Entoto Hill in the North (21).
Within the city there are 13 government hospitals ( ve federal, six under Addis Ababa health bureau, one owned by police force and one armed force hospital) distributed throughout 10 sub cities. For this study, it was necessary to pick out major hospitals with a lot of instrumentation, a variety of medical services, higher bed and large run of patients. Hence, hospitals having 200 or more beds and those which provide 10 or more type of medical services are considered as major. Seven governmental hospitals in Addis Ababa meet these criteria. Out of these, 5 hospitals are selected randomly to making a sample size of 38.5%.
Ten percent or more samples is considered as a good sample size for small populations (22). Laboratory analysis was conducted at Ethiopian public health institute (EPHI) in clinical bacteriology and mycology national referral laboratory in collaboration with food microbiology laboratory. The laboratory has been accredited by Ethiopian National Accreditation O ce.

Sampling Frequency and Sampling Technique
Hospitals that have higher bed and serving many patients were selected in ten sub-city of Addis Ababa. Based on these ve governmental hospitals were included for this study. In each hospital, six sampling sites are employed to collect hospital e uent. These are a manhole used to collect a wastewater originate from adult ward, pediatric ward, labor ward, laboratory unit, laundry unit and a manhole of septic tank (which hold e uent from all source and we termed Mixed). In addition, one sampling unit viz. MDR TB ward was incorporated to collect hospital e uent only in alert hospital.
A total of 100 discrete (that is 50 at the morning and 50 at the afternoon) wastewater samples were collected from each ward/unit wastewater collecting manholes at the sampling site with four hour interval in the study period. This encompassed discrete e uent of 24 from each Y12HMC and MIIRH, 20 from each SPHMMC and ALERT, and 12 discrete e uents from TASH.
A "Grab sampling technique" was applied to collect the most representative samples according to guidelines of wastewater sampling techniques stated on Environment Protection Authority (EPA) (24) and American Public Health Association (APHA) (25). Discrete samples were collected in two rounds in each hospital for two months. The 1 st and 2 nd round samples were collected within 15 day interval. In each round, the discrete samples were collected two times a day with four hour interval from each sampling sites in each hospital in 150 ml cleaned and sterile microbiological glass bottles. Here 150 ml sterile glass container was used to collect 125 ml wastewater samples.
Hospitals' e uent Samples were collected during their maximum activity period (usually 10:00 am-2:30 pm) according to the method used by Nuñez and Moretton (10). In addition, samples were collected near the center of ow channel, at approximately 10-15 centimeter of the water depth, where the turbulence was at maximum and the possibility settling was minimized. Grazing (skimming) the water surface or slogging the bottle was avoided. The rst sample was collected in the morning 10 -10:30 AM whereas the second sample was collected at 2-230 PM from each sampling site. After taking the sample, the neck of the bottle was wiped with 95% alcohol then the sample bottle labelled with date, code number, and time and sampling site. All samples were collected manually and transported instantly to Ethiopia Public Health Institute (EPHI) food microbiology and clinical laboratories with cold chin (4˚C) for bacteriological analysis within six hours of collection. A pair of new, non-powdered, disposable glove, a suitable gown and eye google were used each time while, collecting samples to avoid personal contamination. In the same token, heavy duty glove was used to clean and pick up the cover of manholes at time of sample collection.

Sampling Site
In present study, there were different wards and units in the selected hospitals. Each ward/unit generated wastewater having different characteristics. So in order to locate in which sampling site the isolate found, the hospital e uents were collected at different manhole of each units/wards of hospital. The hospital e uents were collected at manhole of the adult ward, pediatric ward, delivery ward, laboratory unit, laundry service unit and at sampling site named MIXED (for this study purpose only). Here mixed sample indicates a hospital e uent in which its origin holding from different ward/unit and ow together hence it was di cult to identify at which speci c unit/ward it came. In addition, wastewater was collected from MDR TB ward in case of Alert hospital only. The sampling site manholes located just at the outlet of wastewater of each ward/unit before discharged in to receiving water/ collecting septic tank and at the side of each ward/unit building. The geographical position of the sampling site/ unit was obtained and documented.

Data Collection Procedure
The important information's were recorded using a pre-developed data collection form by asking the authorized body (about sampling unit/site, wastewater disinfection and disposal procedures), from record book/ le (e.g. number of patients served during study period) and using google map application (for geographical location of the sample). After the sampling unit/site at each hospitals identi ed and its location as well as the source of wastewater, aseptically with care, the cover of each manholes are lifted up to collect the wastewater then covered immediately. The name of the hospital and sampling unit, time and round of collection, as well as its geographical location are recorded on pre-developed data collection format in addition to, labeling collecting bottle at time of sample collection. All of this information was collected by the principal investigator.

Isolation and Characterization of Pure Cultures
For isolation the bacteria, a loopful of a well-mixed sample suspension was inoculated using sterile inoculating loop on to MacConkey agar pate (Oxoid LTD, Basingstoke, Hampshire, England) and incubated aerobically at 37 o C for 18-24 hours.
After incubation for 24 hr. at 37 o C, bacterial colonies with distinct coloration and morphology were randomly picked up and sub cultured on to another MacConkey Agar plate for further puri cation. Then puri ed colonies with distinct presumptive colonies of each suspected bacterial species and fermentation on MacConkey agar are sub cultured on tryptic soya agar (TSA)/nutrient agar (Oxoid LTD, Basingstoke, Hampshire, England) depending on the availability of media for biochemical test.
For identi cation pure colony from non-selective nutrient agar/TSA was sub cultured and identi ed based on the following biochemical tests: oxidase, indole, urea, motility, Lysine decarboxylase, citrate utilization and triple sugar iron tests as per the standards of microbiology procedure including E. coli ATCC 25922, as a control culture (26). Following puri cation and species identi ed, two -three puri ed colonies were preserved in Skimmy milk at -80 o C for further characterization, after each isolate was assigned a unique identi cation number.

Screening isolates for ESBLs Producing
Those Enterobacteriaceae that was resistant or reduced susceptibility to the screening indicator cephalosporin (cefotaxime and/or ceftazidime) was considered as suspicious of ESBLs production. In other word, isolates that showed an inhibition zone size of ≤27 mm for cefotaxime (30μg) and/or ≤ 22 mm for ceftazidime (30μg) were considered as suspicious ESBL producers and selected for con rmation for ESBLs production.

Con rmation of ESBLs Producing Enterobacteriaceae
Con rmation of suspicious ESBL-producing isolate was veri ed by the combination disk method (CDT) as delineated by the 29 th edition CLSI guide line (27).
The test was performed using two cephalosporin antibiotics: ceftazidime (30μg), and cefotaxime (30μg) alone and in combination with beta-lactam inhibitor ((ceftazidime-clavulanic acid (30/10μg), and cefotaxime-clavulanic acid (30/10μg)) by dispensing on 0.5 McFarland turbidity bacterial suspension inoculated Muller Hinton agar (MHA) plate (Oxoid LTD, Basingstoke, Hampshire, England) and then incubated overnight (18-24 hours) at 37 o C as per 29 th edition CLSI guideline. ESBL production was considered positive when ≥ 5mm increase in the zone diameter for the ceftazidime or cefotaxime tested in combination with clavulanic acid versus its zone when tested alone (27). E. coli ATCC 25922 was used as a negative control throughout the tests as a non-ESBL culture.

Antimicrobial Susceptibility Testing
Once the bacteria were isolated and identi ed from each sample collected, all Enterobacteriaceae isolates were assessed for non-susceptible pattern for 12 antibiotic agents by using the Kirby-Bauer disk diffusion method on MHA in line with 29 th edition CLSI guideline (27). Bacterial inoculum was prepared by suspending the freshly grown bacteria in 4-5 ml sterile normal saline and the turbidity was adjusted to that of a 0.5 McFarland standard. Then a prepared bacterial inoculum suspension (0.5 McFarland standards) was streaked on MHA using sterile swap applicator stick and antimicrobial discs were placed. The antibiotic discs used for this study were: trimethoprim⁄ sulphametoxazole (SXT, 1.25/3.75ug), cipro oxacin (CPR, 5ug), tazobactam+piperacillin (TZP, 30ug), cefoxitin (CXT, 30μg), chloramphenicol (CHL, 30μg), nitrofurantoin (F, 300 ug), amoxicillin/clavulanic acid (AMC, 20/10 μg), tobramycin (TOP, 10ug), meropenem (MER, 10μg), cefotaxime (CTX, 30ug), cefepime (CFP,30 μg), and ceftazidime (CAZ, 30 μg ). The antibiotic discs used were the product of Abtek Biologicals Ltd, Liverpool, United Kingdom. Inhibition zones were measured using ruler and isolates were categorized as: resistance, intermediate and susceptible for each antimicrobial agent using the break point as set in line with 29 th edition CLSI guidelines (27). The isolates were going to be considered as MDR when they were non-susceptible for three or more classes of antibiotics (28). E. coli ATCC 25922 and P. aeruginosa ATCC 27853were used for quality control throughout the antimicrobial susceptibility tests as recommended by 29 th edition CLSI.

Laboratory Data Quality Assurance
Sample collection, handling, transportation and microbiological analysis and interpretation of results were carried out using standard operating procedures.
Before the tangible procedure; reagents, antimicrobial disks, and media were checked for damage, storage problems and expiry date. Laboratory equipment's are appropriately cleaned and sterilized before use. Media's was prepared according to the respective manufacturer's instruction. Five percent of prepared media per batch was incubated overnight for sterility checkup. Quality control for new batch was performed using ATCC 25922 E. coli standard control to cross check the quality of antibiotics disks and culture media. For ESBLs con rmatory test E. coli ATCC 25922 (ESBLs negative) and K. pneumoniae ATCC 700603 (ESBLs positive) standard control strains are served at time of the procedure (27).

Data analysis and interpretation method
Data was entered and summarized using SPSS version 20 software (IBM Corporation, Armonk, NY, USA). Frequency and percentages of isolates, antibiotic susceptibility pattern of Enterobacteriaceae and ESBLs-pE were calculated. Tables and gures have been employed for data demonstration.

Statistical Data Quality Assurance
Before data entry, data from the data collection form was cross checked for its completeness and accuracy. Culture isolates and antibiotics susceptibility test results had been documented consciously ahead of entry to SPSS. Furthermore, data cleaning and double-data entry was implemented to assure quality of the data.

Results
Distribution of gram-negative bacteria isolates against sampling unit and hospitals A total of 100 hospital e uent samples were collected and analyzed for the presence of Enterobacteriaceae family. Of these samples 87% were tested positive and contained one or more than one type of isolates. Meanwhile, 183 non-duplicate Gram-negative bacteria were picked from MacConkey agar, 80.3% (147) belonging to Enterobacterial species. The remaining isolates included Pseudomonas spp. (2.2%), Acinetobacter spp. (4.4%), and other unidenti ed Gramnegative bacteria (13%).
Of 147 Enterobacteriaceae family isolates recovered, the highest distributions were from; laboratory unit 32 (21.8%) and mixed source wastewater 30(20.4%). Whereas the least isolates identi ed from MDR TB ward e uent (4.8%), which was only collected from ALERT. In this study, the highest number of isolated bacteria, irrespective of total sample collected, were recovered from Y12HMC (34), MIIRH (32) and SPHMMC (32); while, the least isolates were obtained from TASH (21) ( Table 1).

Distribution of MDR and ESBLs producing Enterobacteriaceae against the variables
From all hospital wastewater collected for this study purpose, the highest ratio of ESBLs-pE within sampling unit was observed in adult ward e uent (66.7%) followed by laundry unit (58.8%) and labor ward (47.6%) respectively; whereas, the least proportion was recovered in pediatric ward (36.9%) and laboratory unit e uent (40.6%) ( Table 5). Similarly with less difference; the elevated MDR isolates within sampling unit were identi ed in adult ward e uent (71.4%) pursued by laundry unit (70.6%) and laboratory unit (62.5%) correspondingly; while the lowest ratio was found, in pediatric ward (52.6%) ( Table 6). In MDR TB ward wastewater which was collected only in ALERT hospital, the proportions of ESBLs-producing and MDR Enterobacteriaceae within sampling unit were 57.1% and 85.7% respectively (Table 5 and Table 6). Generally, of all MDR Enterobacteriaceae, 73.4% (69/94) were ESBLs producer, whereas only 26.6% (25/94) of them were non-ESBLs producer Enterobacteriaceae.
The magnitude of ESBLs producing and MDR Enterobacteriaceae in the wastewater were different in the ve hospitals. The highest occurrence of ESBLs-pE within hospital according to CDT identi cation method was found in wastewater from the ALERT (67.9%), followed by TASH (52.4%) and Y12HMC (47.1%) respectively. Whereas the least ESBLs-pE within hospital occurred in MIIRH (31.2%) ( Table 5). In contrary, the elevated MDR isolates within hospital were observed in wastewater of TASH (76.2%) and ALERT (71.4%); while, the lowest ratio was found in the same way as ESBL producer isolates in MIIRH (53.1%) ( Table 6).
The magnitude of MDR and ESBLs-pE obtained from all wastewater samples were higher at the afternoon than the morning wastewater collected. At morning e uent the occurrence of MDR isolates and ESBL producers within time of wastewater collection were 59.2% and 47.9%, whereas at the afternoon were 64.5% and 52.1% respectively. On the contrary, they were higher in the rst round of e uent collection than the second. In the rst round MDR isolates and ESBL producers were 77.6% and 48.8%, while in the second round e uent collection they were 62.7% and 40% correspondingly (Table 5 and Table 6). Antibiotics Susceptibility pattern of ESBLs producing Enterobacteriaceae to potentially active drugs The antibiotic resistant pro le for ESBLs producing and non ESBLs producer Enterobacteriaceae are displayed in Figure 3 below. The antibiotics susceptibility pattern of ESBLs-pE were also performed in relation to potential active antibiotics like carbapenems (meropenem), quinolone/ uroquinolone (cipro oxacine), cephamycine (cefoxitin) and aminoglycoside (tobramycin) drug Categories. In addition, it was tested against for combinantion drug (amoxaciline-clavunalate and pepracillin-tazobactam), phenicol (chloramphenicol), nitrofuran (nitrofurantoin) and folate pathway antagonist (trimthoprime-sulfamethoxazole) drug families.

Discussion
The occurrence of antibiotic resistant and ESBLs-pE from hospital wastewater could be exceptionally problematic because of the ability of nosocomial pathogens to transfer antibiotic resistance genes among different hosts and environments (29). The dissemination of MDR bacteria via hospital wastewater is a usable cause for concern (30), because it is reasonable that MDR bacteria are selected mostly in hospitals and take away by wastewater.  (37). These variations might be due to sample type (inanimate object and swage of hospital), study period, sample size and type of pathogen infecting patients at time of sample collection.

Antibiotics resistant pattern of Enterobacteriaceae isolates
In the present study, the overall prevalence of antimicrobial resistant pattern for Enterobacteriaceae isolated ranged from 8.2 to 77.6% in wastewater isolates, with most of the strains susceptible to meropenem (MER) and nitrofruntonine (F). This nding was in line with study conducted in Rio de Janeiro, Brazil, with 0 to 83% resistant range for Gram negative isolates and most strains susceptible to meropenem(38).

Magnitude of Extended Spectrum B-lactamase producing Enterobacteriaceae
In the present study, of all Enterobacteriaceae 55.1% were suspected as potential ESBLs producing and 87.7% of them were con rmed ESBLs producing isolates. A little comparable result was reported in Dubai, UAE by Khan MA. et al 2020: among all isolates from municipality wastewater 57.4% suspicious and 25.7% con rmed ESBLs producer Enterobacteriaceae were reported (40), in Northern Italy: 45.4% beta-lactamases producing Enterobacteriaceae were recovered from WWTPs (41). The difference with our result might be due to the type of sample used (hospital Vs municipality wastewater), method used to con rm potential ESBLs producer (CDT Vs DDST) and sample size.
According to the present study, the overall magnitude of ESBLs-pE were 48.3% which is almost in line with a study conducted in Rio de Janeiro, Brazil (38), and Nepal (37) with ESBLs producer isolates of 39%, and 30.4% respectively. In contrast to the current study, other studies conducted in Ethiopia and other countries reported lower prevalence of ESBLs-pE, in Addis Ababa: 25% from hospital wastewater (31), in Northwest, Ethiopia: 14.8% from hospital environment (42), and in Austria: 27.4% from activated sludge (43), were recovered (43). The difference in the prevalence of ESBLs producer in different studies from wastewater isolates might be due to difference in geographic areas, source of sample, period of study (ESBL rapidly changing over time), sample size, method of ESBL detection and an infection control system.

Distribution of MDR and ESBLS Producing Enterobacteriaceae against the independent variables
In the present study, of all MDR Enterobacteriaceae, 73.4% were ESBLs producer, whereas only 26.6% of them were non-ESBLs producer Enterobacteriaceae. The magnitude of ESBLs producing and MDR Enterobacteriaceae in the wastewater were different in the ve hospitals. The highest occurrence of ESBLs-pE within hospital according to CDT identi cation method were found in wastewater from the ALERT (67.9%), followed by TASH (52.4%) and Y12HMC (47.1%) respectively, whereas; the least ESBLs-pE were detected in MIIRH (31.2%). In contrary; the elevated MDR isolates within hospital were observed in wastewater of TASH (76.2%) and ALERT (71.4%), while; the lowest ratio was found in the same way as ESBL producer isolates in MIIRH (53.1%). It is di cult to compare directly the occurrence of MDR and ESBLs-pE in hospital e uent from one country to other because of the presence of difference in geographical zone, the epidemiology of disease (the severity and disease type), the number of patients served, the service provided in the hospital and wastewater disposal police from country to country. However, like our country there were different ESBLs producer occurrence within country hospital e uent, in Ibadan, Nigeria: more ESBLs producer was found ina privately-owned hospital (33.3%) than a State Government-owned hospital (29.1%) (39), in Europe: the elevated ESBLs-pE was found in e uents from the Slovenian general hospital, followed by the Austrian private rehabilitation clinic and the Austrian private surgery clinic (44).
From all hospital wastewater collected for this study purpose, the highest ratio of ESBLs-pE within sampling unit was observed in adult ward e uent (66.7%) followed by laundry unit (58.8%) and labor ward e uents (47.6%) respectively; whereas, the least proportion was recovered in pediatric ward (36.9%) and laboratory unit e uent (40.6%). Similarly with less difference; the elevated MDR isolates within sampling unit were identi ed in adult ward e uent (71.4%) pursued by laundry unit (70.6%) and laboratory unit e uents (62.5%) correspondingly, while; the lowest ratio was found in pediatric ward e uent (52.6%). In MDR TB ward wastewater which was collected only in ALERT hospital, the proportions of ESBLs-producing and MDR Enterobacteriaceae within sampling unit were 57.1% and 85.7% respectively. Almost all the preceding publication on antibiotic resistant pro le of pathogenic microbes has been focused towards crude hospital wastewater rather than at each re ned source of it. As a result, it was di cult to compare our result directly with other studies conducted from hospital wastewater, anyway a study conducted in hospital environment in Gondar, reported from inanimate object of medical ward, surgical ward and Gynobs ward ESBLs-pE of 52.6%, 10.5% and 5.3% respectively (42). The variation in ESBLs-producing and MDR Enterobacteriaceae proportion within sampling unit in the present study might probably be attributed to the difference in type of patients served, length of patient stay, type of medical service provided, infection prevention and control procedure in each department/unit.
In this study, the magnitude of MDR and ESBLs-pE obtained from all wastewater samples were higher at the afternoon than the morning wastewater collected.
At the afternoon e uent, the occurrence of MDR and ESBL producer isolates within time of wastewater collection were 64.5% and 52.1%, whereas; at the morning e uent, were 59.2% and 47.9% respectively. This difference most probable happen because of the majority medical activity performed around the afternoon and outpatients number increase at the afternoon due to transportation and other reasons. On the other hand, MDR and ESBLs-pE were higher in the rst round of e uent collection than the second. In the rst round, they were 77.6% and 48.8%, while in the second round e uent collection, they were 62.7% and 40% correspondingly.

Antibiotics susceptibility pattern of ESBLs producing Enterobacteriaceae
In current study, the predominant ESBLs-pE were found to be more than 85% resistant to the antibiotic like cefotaxime (95.8%), AMC (93%), ce peme (90.1%), and ceftazidime (87.3%). These were in close agreement with other study done in, Northwest Ethiopia; amoxicillin/clavulanic acid (100%) and ceftazidime (100%), Dubai, UAE; from municipality wastewater, cefotaxime (86%) and ceftazidime (77%) (40), Austria; amoxicillin/clavulanic acid (53.1%) (43)  respectively. We observed in the present study ESBLs producer isolates were more resistant to the tested antibiotics than non-ESBLs producer Enterobacteriaceae. This difference might be from the resistant gene on ESBLs producer also contributed the isolate to develop resistance to other antibiotic too.
Strength of the study Our study tried to collect the wastewater, at their utmost source, which enables to take preventive measure and at large sample size (in relative to previous study) to be representative. This study conducted at different governmental higher hospitals to display the extent of distribution of MDR and ESBLs-pE in each hospital sampling units/sites.

Limitation of the study
In this study only wastewater was used, hence it was unable to differentiate the source of resistant bacteria either it was from clinical isolates or sewage system.
Our study was unable to select all antimicrobial agents commonly used for resistance evaluation due to the fact that some of them were not available during the period of study.
The major limitation of the study was that ESBL detection was only performed phenotypically using CDT method, it was better to include genotypic method of detection.
Some source of hospital wastewaters were not incorporated in the study. So, to generalize the distribution of MDR and ESBLs-pE in hospital wastewater, it was better assessing all source of wastewater in selected hospitals.
It would have a better gurative data if the study was also include private hospitals wastewater found in Addis Ababa and WWTP of the city.

Conclusion
The hospital wastewater released directly into urban sewerage systems without appropriate disinfection or treatment is the serious environmental attention of the day. In this study, there was high magnitude of MDR & ESBLs-pE (≈65% Vs 50%) from the wastewater of selected governmental hospitals in Addis Ababa, Ethiopia. Majority of them were in adult ward and laundry unit e uents. In addition, the most frequent ESBLs-pE were E .coli, K. pneumoniae and citrobacter spp. In addition, ESBLs-pE showed high rate of resistance to all tested antibiotics as compared to non-ESBLs-E. This is a warning threat to such infection via contamination of food and water from rivers and urbane drainage which will be polluted through untreated hospital wastewater. Hence, infection prevention and control implementation at each hospital is mandatory One of the persistence concerns due to the existence of ESBL-producing bacteria in water bodies are related with the transmission of conjugative plasmids, which additionally carry genes of resistant to sulfonamides and aminoglucosides, giving the bacteria multi-resistant patterns (45). This phenomenon was observed in our study, where the most common higher co-resistant rates among the ESBLs-pE identi ed for cipro oxacin (74.6%) and SXT (73.2%). However in this study, the most active drugs for ESBLs-pE were meropenem (94.4%), nitrofurantoin (88.7%) and piperacillin/tazobactam (87.3%) During sample collection in this study we noticed the selected hospitals had neither WWTP nor wastewater stabilization pond except in TASH nonfunctional waste stabilization pond was observed. Since the present study detected a high proportion of pathogenic, resistant and ESBLs-pE, indicating a higher probability potential risk of microbial pollution of water bodies in the community, hence accelerate spreading of resistant microorganisms into community.

Recommendation
The high occurrence of MDR and ESBLs-pE in the present study from hospital e uent may have a severe consequence in the public health. These bacteria can carry several genetic determinants which can be transmitted to another bacterium including pathogens. This indicates a need for a highly committed and consistently hygienic treatment of e uent (implementation of nal disinfection procedure to minimize the microbial burden) at each respective wards and units to inhibit the transmission of the antibiotic resistant to another enteric pathogenic bacterium. Since all selected hospitals for this study purpose were hadn't any either wastewater treatment plant or functional wastewater stabilization pond, therefore every concern bodies should take urgent measures to minimize the devastating outcome from the discharge of hospital wastewaters into the community drainages without getting appropriate treatment. We suggest additional studies to take place on molecular epidemiology of ESBLs producing Enterobacteriaceae and their effects on patient recovery and health care burden in Addis Ababa, Ethiopia.  The percentage of ESBLs positive, ESBLs negative and non-potential ESBLs Enterobacteriaceae at selected governmental hospitals wastewater, Addis Ababa, Ethiopia April 1 to May 31, 2020.

Figure 2
Frequency of total, potential and con rmed ESBLs producing Enterobacteriaceae species from different sampling units at selected governmental hospitals wastewater, in Addis Ababa, Ethiopia from April 1 to May 31, 2020.