Drug resistance and extended-spectrum β-lactamase (ESBLs) - producing Enterobacteriaceae, Acinetobacter and Pseudomonas species from the views of one-health approach in Ethiopia: a systematic review and meta-analysis

Background Although antimicrobial resistance (AMR) bacteria present a significant and ongoing public health challenge, its magnitude remains poorly understood, especially in many parts of the developing countries. Hence, this review was conducted to describe the current pooled prevalence of drug resistance, multidrug- resistance (MDR), and Extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae, Acinetobacter, and Pseudomonas species in humans, the environment, and animals or food of animal origin in Ethiopia. Methods PubMed, Google Scholar, and other sources were searched for relevant articles as per the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines. A critical appraisal for screening, eligibility, and inclusion in the meta-analysis was made based on the Joanna Briggs Institute’s (JBI) essential appraisal tools. The meta-analysis was done on Statistical Software Package (STATA) version 17.0. Results A total of 33 research articles were included in this systematic review and meta-analysis. Escherichia coli, Klebsiella species, Acinetobacter, and Pseudomonas species were the most frequently reported bacteria from two or more sources. More than 50% of Klebsiella species and 25% to 89% of Escherichia coli from two or more sources were resistant to all analysed antibiotics, except carbapenems. Fifty-five percent (55%) to 84% of Acinetobacter species and 33% to 79% of Pseudomonas species from human and environmental sources were resistant to all analyzed antibiotics. Carbapenem resistance was common in Acinetobacter and Pseudomonas species (38% to 64%) but uncommon in Enterobacteriaceae (19% to 44%). Acinetobacter species (92%), Klebsiella species (86%), and Pseudomonas species (79%) from human sources, and Proteus species (92%), and Acinetobacter species (83%), from environmental sources, were the common multidrug-resistant isolates. About 45% to 67% of E. coli, Klebsiella, Acinetobacter, and Pseudomonas species from human and environmental sources were ESBL producers. Conclusion Our review report concluded that there was a significant pooled prevalence of drug resistance, MDR, and ESBL-producing Enterobacteriaceae, Acinetobacter, and Pseudomonas species from two or more sources. Hence, our finding underlines the need for the implementation of integrated intervention approaches to address the gaps in reducing the emergence and spread of antibiotic- resistant bacteria.


Background
Antimicrobial resistance (AMR) remains a significant One-Health problem, affecting humans, animals, and the environment [1].The infections caused by AMR bacteria are becoming more prevalent and can be difficult, and sometimes impossible to treat because the available drugs used to treat microbial infections have become less effective or ineffective.The AMR threat adds to the existing higher burden of bacterial infections, particularly in low-and middle-income settings in which there has been low access to adequate diagnostics, specifically at peripheral levels of the healthcare system.In addition to increased morbidity and mortality, resistant infections also add considerable costs to the healthcare system [1][2][3].
AMR gram-negative bacteria are the most frequently encountered bacterial isolates recovered from different clinical and non-clinical specimens [3].The emergence of ESBL-producing and carbapenem-resistant gram-negative bacteria, particularly Klebsiella pneumoniae, Escherichia coli, Acinetobacter baumannii, and Pseudomonas aeruginosa, are a matter of national and international concern as they are an emerging cause of healthcareassociated infection (HAI) that pose a significant threat to human and animal health [4,5].The infections caused by these bacteria may not be treated with the available antibiotics due to high levels of resistance and are associated with poor treatment outcomes.Importantly, although there are existing knowledge gaps in understanding the transmission pathway of AMR bacteria, there are various routes for widespread transmission of resistance bacteria and genes between humans, animals and the surrounding environment [1,6].Resistant bacteria can spread across humans and animal communities, the food supply, healthcare facilities, and the environment, which increases the burden of resistance and antibiotic-resistant infections [6,7].
Anyone of any age, in every country, can potentially be affected by the consequences of AMR.For instance, an estimated 4•95 million deaths were associated with bacterial AMR in 2019, and if not properly addressed, the numbers may be increase to 10 million per year by 2050 [8,9].The main factors exacerbating the issue of AMR in low-resource countries include limited access to quality antimicrobial drugs; antibiotics sold over the counter without prescriptions, or antibiotics used in feeding animals as prophylaxis or growth promoters.The issue of a lack of regulation and quality control of drugs, coupled with poor infection prevention and water, sanitation, and hygiene interventions, can accelerate the emergence and spread of drug-resistant microorganisms [10][11][12][13].
The ongoing public health threat of AMR bacteria was highlighted on the WHO list of critical-priorities for the need of new researches, discovery, and development of new antibiotics [14].Ethiopia has also implemented the One Health approach to respond to the existing and emerging health security threats, including AMR [15].However, poor integration among sectors, the institutionalization of One-Health as a good approach, limited research funds, and activities on One-Health are among the many challenges that need to be addressed.So far, no study has reported the current situation of AMR and ESBL-producing combinations in our country.Therefore, this systematic review and meta-analysis aimed to determine: I) the pooled prevalence of resistance to commonly prescribed broad-spectrum antibiotics; II) the pooled prevalence of MDR; and III) the pooled prevalence of ESBL-producing Enterobacteriaceae, Acinetobacter, and Pseudomonas species from humans, the environment, and animals, or food sources.
The searching process was filtered by year of publication, from January 2014 to October 2022, and full-text research articles.Additionally, relevant studies were manually searched from the bibliographies of eligible studies and from other meta-analysis studies.

Selection and eligibility criteria
The systematic and comprehensive literature review methods were used to identify, select, and critically appraise relevant research and to collect and analyze data from the studies that are included in the review.Those research articles conducted in Ethiopia and published in English as research articles in the years 2014 to 2022, and those articles focusing on the reports of antimicrobialresistant Enterobacteriaceae, Acinetobacter, and Pseudomonas species in humans, animals, or food of animal origin, and those that provided details on the number of studied isolates, are used as criteria for eligibility for the review.On the other side, those articles that did not provide full information on the outcomes of interest, provided data on gram positives only, conducted molecular investigations of AMR molecular markers only, were not freely accessible as a full text, and those reviewed articles on AMR were excluded.In order to guarantee the quality of studies, two independent reviewers were assigned to select the articles throughout each stage of the review (i.e., screening, eligibility, and inclusion in meta-analysis).

Article quality assessment
The article selection process was done based on the preferred reporting items for systematic reviews and metaanalysis (PRISMA) guidelines [16] (Fig. 1).The quality assessment and enrollment of each article were made by two independently critical appraisers based on the Joanna Briggs Institute (JBI) critical appraisal tools [17] and the Cochrane Handbook for Systematic Reviews [18].The criteria for quality assessment include: whether the research question is clear and adequate to the study; whether the study design used is appropriate to the set research question; whether descriptions of the setting, including periods of recruitment, and the sampling method are appropriate for the set research question and design; and whether the collected data was properly managed and analyzed.In addition, a comprehensive search strategy was made in order to reduce the impact of publication bias on the results of the review.

Data extraction
An Excel database was designed for the purpose of extracting data from the included studies.The first author, publication year, study region or area, study or data collection period, study design, study subjects, type of sample, type and numbers of selected gram-negative bacteria, the number of isolates tested for antimicrobial resistance, the number of isolates reported as MDR, and if reported, the numbers of ESBL producers were extracted.Additionally, the investigation method (phenotypic or genotypic) was extracted.The data extraction process was done independently and in duplicate using piloting forms to ensure double-checking.

Data analysis
The total number of each bacterium species and the number of isolates tested for antimicrobial resistance from each source were extracted, and meta-analysis was done on STATA version 17.The pooled prevalence of AMR, MDR, and ESBL production for each bacterium was analyzed using the random-effects model.Cochran Q tests and the I 2 statistic were used to analyze the heterogeneity of the studies, and significant variation was considered at p-values < 0.05 and I 2 > 50% [19].For the studies on the environment and food-producing animals, the meta-analysis was done if the outcome of interests was reported in at least three studies, whereas at least four studies were considered in the case of human sources.The pooled percentage for each reported resistant gram-negative species was then deduced from the total number of tested isolates.A categorical meta-analysis for each antibiotic resistance isolate was made based on their sources.Begg's and Mazumdar rank correlation test was performed to assess the publication biases across the studies, and statistical significance was considered at a p-value < 0.05.Testing for publication bias and heterogeneity was carried out to check the extent of the variation in study outcomes between the included studies and whether the results of the studies were valid for systematic reviews and meta-analyses.Finally, the results were narrated in words and presented in figures and tables that were best suited for readers.

General characteristics of the included studies
In this systematic review and meta-analysis, a total of 33 studies were included; of these, 14 were human studies, 11 were on environmental studies, and 8 were related to animals or foods of animal origin (Fig. 1).The included studies were published from 2014 to 2022, and 30 studies were done with a cross-sectional study design; two studies were retrospective and one was a cohort study.Based on the study area, half of human studies (50.0%) were from the Amhara region, 4 (36.4%) of the environmental studies were from southern Ethiopia, and 3 (37.5%)studies on animals or foods of animal origin were from Addis Ababa (Table 1).Out of the 14 included studies on humans, 10 studies involved various clinical samples for the detection of drug-resistant bacteria from patients with multiple infections.Bloodstream infections (BSIs), urinary tract infections, nosocomial infections, and other conditions are commonly considered medical conditions from which drug-resistant bacteria isolates were reported.In studies involving animals or foods of animal origin, raw milk, meat or carcass swabs, animal feeds, and chicken droppings and caecum were the most frequently considered specimens in the detection of drug-resistant isolates.Swabs from hospital contact surfaces and mobile phones, indoor air, and waste/river water are the sources of samples for environmental studies.The detailed characteristics of the studies are presented below in Table 1.
In this review, Begg's and Mazumdar rank correlation test showed that no significant effect of publication bias was observed among the included studies (p-value > 0.05).However, the variation in the study methodology, setups, study periods, and study populations could have an effect on the heterogeneity among the included studies.

The frequency of selected bacterial isolates
In this review, a total of 12 species of gram-negative bacteria were extracted; however, the meta-analysis  was computed for eight gram-negative bacteria from studies in humans, the environment, and animals, or food of animal origin.Escherichia coli (n = 716), Klebsiella species (n = 543), Pseudomonas species (n = 401), and Acinetobacter species (n = 366) were the most frequently reported species from two or more sources (Fig. 2).

The pooled prevalence of ESBL-production
In this review, the rate of ESBL production was also computed from the forest plots for each bacterium.Among human isolates, the highest proportion of ESBL production was recorded by Pseudomonas species (0.67; 95% CI: 0.  4).

Discussion
This systematic review and meta-analysis was conducted to estimate drug-and multidrug-resistant bacteria from one-health perspective in Ethiopia.It also determined the prevalence of ESBL-producing gram-negative bacteria in human and environmental isolates.From human sources, more than 60% resistance was reported to commonly prescribed β-lactam antibiotics, ciprofloxacillin, gentamycin, and co-trimoxazole.In addition, the highest rates of MDR were found in Acinetobacter spp.(92%), followed by Klebsiella species (86%), and Pseudomonas species (79%).With some exceptions, almost consistent findings were reported in a review of findings in Ethiopia [53,54], and in Cameroon [55], and East Africa [56].Hence, this review suggests that, as infections caused by antibioticresistant bacteria are becoming more prevalent, serious concerns should be given to the use and choice of antibiotics for effective management of infections in Ethiopia.
Gram-negative bacteria use several mechanisms to develop resistance to antimicrobials.Mutations and recombination of genomic materials allow these bacteria to disseminate genes encoding for antimicrobial resistance within and across species [57].Actions in the human and animal healthcare sectors are all considered to be contributing to the development of pathogen resistance to current available antimicrobials [57][58][59][60].Frequent use of antibiotics may create favorable conditions for selective pressure, which leads to the further development of resistance.For instance, the production of β-lactamase that hydrolyzes the β-lactam ring is the most common resistance mechanism for these bacteria against β-lactam antibiotics.Gram-negative bacteria that produce ESBLs carry plasmid-encoded enzymes that can hydrolyze and confer resistance to a variety of β-lactam antibiotics, as well as fluoroquinolones, aminoglycosides, and trimethoprim-sulfamethoxazole [57,61,62].
In this review, above 50% of E. coli, Klebsiella, Pseudomonas, and Acinetobacter species from human sources were ESBL producers.The presence of bacteria in human and animal bodies as carriers may result in frequent exposure to antimicrobials used for treatment and prophylactic purposes [57,59,60,62,63].There is no question that the widespread use, overuse, and misuse of antimicrobials have been associated with the explosion of antimicrobial resistance.A study confirmed that those who had exposure to third-generation cephalosporins, carbapenems, and fluoroquinolones had three-to-four times greater risk for infections with extended-spectrum β-lactamase-producing bacteria [64].Therefore, updated and effective measures, such as antimicrobial stewardship which promotes the careful and responsible use of antimicrobials and prevents antimicrobial overuse and misuse in hospital and community settings, and infection prevention, are the most effective ways to reduce the spread and development of antimicrobial resistance and to protect patients from harms caused by unnecessary antibiotic use.
Antimicrobial susceptibility testing appeared to be inconsistent and low in animal, food, and environmental sources of isolates compared with humans.From environmental sources, E. coli, Klebsiella spp., and Acinetobacter spp.were recorded with more than 60% rates of AMR to ampicillin, amoxicillin-clavulanic acid, ceftriaxone, and co-trimoxazole.The rate of MDR was above 50% for five bacterial species.Mutation of bacterial genomes by different mechanisms, such as frequent antibiotic use or misuse in long-care facilities, may provide a selective advantage to the emergence of resistant variants [57,59,61].For instance, in this review, 10 to 66% of the ESBLproduction rate was found in environmental isolates, with the highest rates found in Acinetobacter (66%) and Klebsiella spp.(62%).Most of the included environmental studies were from hospital settings, specifically hospital surfaces, indoor air, and wastewater, suggesting a need for control of resistant gram-negative infections through a comprehensive approach, including detection and identification of resistant organisms and implementation of effective infection-control and prevention strategies in healthcare settings.
In isolates from animals or food of animal origin, the analysis for drug resistance was done only for E. coli and Salmonella species.Accordingly, greater than 50% of E. coli and Salmonella species were resistant to ampicillin, AMC, and co-trimoxazole, and the rate of MDR was 36% and 29%, respectively.A higher pooled estimate of antibiotic resistance (86%) and multidrug resistance (73%) was also reported in a review study in Africa [65].Surface contamination with fecal matter, animal excreta, and water or soil sources may allow the transmission of drugresistant bacterial populations to raw meat and carcasses, which could be transmitted to humans through consumption of animal products [66][67][68][69].Additionally, the frequent contact between humans, dairy cattle, and poultry may also be a good opportunity for the bidirectional transmission of AMR bacteria such as E. coli [60,69,70].Hence, the frequent contact with dairy cattle and poultry products as well as the habitual consumption of raw meat and milk may be contributing factors in the acquisition of resistance bacteria.
In general, in this review study, the prevalence of AMR, MDR, and ESBL-producing bacteria was higher in isolates from human samples as compared to other environmental and animal samples.However, some isolates from hospital environments showed comparable rates of AMR, MDR, and ESBL production.This may be indicated by the frequent exposures of humans to most antibiotics and the healthcare sectors, which can be contributing factors to the development of resistance and the possible transmission of antimicrobial-resistant bacteria from humans to the hospital environment and vice versa.Therefore, implementation of the integrated approaches, such as best regulation of the use of antibiotics, effective infection prevention, improving food safety, and preventing zoonotic disease infections, are important measures for the prevention and control of these complex AMR development and transmission cycles.

Conclusion
This review report consists of the most recent situation of AMR with commonly prescribed antibiotics from a one-health perspective in Ethiopia.The review indicated that the high pooled prevalence of antibiotic resistance, MDR, and ESBL-production was in Enterobacteriaceae, Acinetobacter, and Pseudomonas species isolated from humans, the environment, and animals or food of animal origin.Therefore, to address the gaps related to measures taken to reduce the emergence and spread of AMR bacteria in humans, animals, and the environment, it is time to implement a harmonized and multidisciplinary onehealth approach.

Fig. 1
Fig. 1 Flow diagram depicting the selection process of included articles

Fig. 2
Fig. 2 Type and frequency of bacteria isolated from humans, environment and from animals or food of animal origin

Table 2
Estimated AMR gram-negative bacteria isolated from humans, animals/food, and the environment

isolates Types of antibiotics and estimated resistance (95% CI) AMP AMC CRO CTX CTZ CEF CIP CN SXT MRP
= 19)because, the total number of isolates tested for antimicrobial resistance from a two or more sources was < 50."ND" was used to indicate susceptibility testing was not performed to calculate pooled prevalence of AMR NBAMP Ampicillin, AMC Amoxicillin-clavunilic acid, CRO Ceftriaxone, CAZ Ceftazidime, CTX Cefotaxime, CEF Cefepime, CIP Ciprofloxacin, CN Gentamycin, SXT Trimethoprim-sulphamethoxazole, MRP Meropenem, and Pooled prevalence of AMR was not calculated for Shigella spp.(n = 24), Serratia spp.(n = 33), Providentia spp (n = 37) and Morganella spp (n

Table 3
Estimated rate of MDR in gram-negative bacteria from humans, animals/food, and the environment

Table 4
Estimated ESBL-producers among gram-negative bacteria isolated from humans and the environment