Antimicrobial resistance (AMR) is regarded as one of the top ten threats to global health (WHO, 2020). This follows as the emergence and spread of drug-resistant pathogens that have acquired new resistance mechanisms continue to threaten the effectiveness of clinically important antibiotics to treat common infections (Koutsoumanis et al., 2021; Rahman et al., 2022). Globally, major organizations including the World Health Organization (WHO), Food and Agriculture Organization of the United Nations (FAO), World Organization of Animal Health (OIE) and the European Commission (EC) have recognized the need for further investigation and a multidisciplinary approach to combatting AMR (Koutsoumanis et al., 2021). However, in low and middle income countries (LMICs), such as those in Africa, challenges to implementing effective and sustainable AMR surveillance programmes remain (Elton et al., 2020). This follows as LMICs often lack the necessary infrastructural and institutional capacities and effective reporting systems to roll out sustainable surveillance programmes (Elton et al., 2020).

Many different sources and routes for human acquisition of antimicrobial resistant bacteria are recognized, including human-to-human transmission, direct contact with food-producing animals and pets, foodborne transmission as well as the environment (Koutsoumanis et al., 2021). In LMICs, the main driver of AMR is reported to be transmission and not antimicrobial use (Koutsoumanis et al., 2021). Non-human sources of pathogens such as extended-spectrum β-lactamase (ESBL)-producing Escherichia coli and plasmid mediated AmpC (pAmpC)-producing E. coli has been reported, with the need for longitudinal studies and continuous monitoring (Mughini-gras et al., 2019). Furthermore, the Centers for Disease Control and Prevention (CDC) recently reported that urgent AMR threats in the United States (US) include carbapenem-resistant Enterobacterales, while ESBL-producing Enterobacterales, drug-resistant nontyphoidal Salmonella, and drug-resistant Salmonella serotype Typhi are regarded as serious threats, among others (CDC, 2019). The Enterobacterales form part of the normal epiphytic microflora of fruit and vegetables, and include members ubiquitous in terrestrial and aquatic environments, as well as human foodborne pathogens including pathogenic E. coli and Salmonella spp. (Rajwar et al., 2015). Moreover, Enterobacterales are adapted to sharing genetic material and often clinically significant resistance genes through carriage on mobile genetic elements (Partridge, 2015). Recently, a comprehensive assessment of the global burden of AMR stated that the six leading pathogens for deaths associated with resistance included E. coli, followed by Staphylococcus aureus, Klebsiella pneumoniae, Streptococcus pneumoniae, Acinetobacter baumannii, and Pseudomonas aeruginosa (Murray et al., 2022).

In the last decade, an increased emphasis on the role of the environment in dissemination of AMR has been reported (WHO, 2015; Koutsoumanis et al., 2021). Furthermore, authors have reported on the importance of an integrated One Health approach for developing and implementing mitigation strategies in combatting AMR (White and Hughes, 2019; Ikhimiukor et al., 2022). The One Health concept in AMR mitigation strategies recognizes that humans, animals (including wildlife), environments, and ecosystems are key priorities (White and Hughes, 2019). To date, most AMR surveillance studies, especially in LMICs in Africa, have focussed on humans and animals. As an example, from 901 studies in LMIC-based studies in 2000-2018, the rapid increasing trends of AMR in the food-animal sector for common indicator pathogens such as E. coli, Campylobacter spp., Salmonella spp., and S. aureus have been reported (Ikhimiukor et al., 2022). The main objective of the current study was to elucidate the distribution and prevalence of AMR in clinically significant members of the Enterobacterales family isolated from the water-plant-food nexus in Africa.

The review and meta-analysis of occurrence included published articles from January 2010 – December 2022 and was compiled according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Page et al., 2021).

The role of the environment in dissemination of AMR is increasingly being reported (Koutsoumanis et al., 2021). This meta-analysis summarized the published data available on the occurrence of clinically significant Enterobacterales harboring antibiotic resistance genes, including genes expressing broad-spectrum β-lactamases, ESBLs and/or AmpC β-lactamases, isolated from the water-plant-food nexus in Africa. Environmental antibiotic resistant bacteria are typically classified as (i) carriers, that have a role in dissemination of resistance genes but cannot colonize the human body, and (ii) vectors, comprising of ARB that can colonize and invade the human body (Manaia, 2017). Globally, AMR surveillance programmes recognize ESBL- and carbapenemase-producing Enterobacterales as vectors of great importance in AMR gene dissemination (WHO, 2014, 2020). Based on eligibility criteria within the systematic approach used, 51 studies were identified for the current meta-analysis.

It is well known that the environment contains a natural antimicrobial resistance gene pool as well as resistance genes resulting from anthropogenic activities (Manaia, 2017). Previous research have further demonstrated that water is a unifying transmission pathway for dissemination of AMR throughout different environments (Liguori et al., 2022). Moreover, irrigation water is regarded as one of the main routes of transmission of human pathogenic bacteria onto fresh produce. At least 13 of the studies in the current analysis, from six different countries, all demonstrated that multidrug resistant potential pathogens such as E. coli, K. pneumoniae, Salmonella spp., Enterobacter spp., Serratia fonticola and Citrobacter freundii were present in fresh produce and associated irrigation water (Supplementary Table 1), which corresponds to global studies linking irrigation water to fresh produce contamination (Blaak et al., 2014; Ye et al., 2017; Vital et al., 2018; Banach and Van Der Fels-Klerx, 2020).

The current review also elucidated that water was mainly investigated in environmental AMR studies in Africa between 2010 and 2022. These studies had heterogeneous study locations and methods used, with E. coli predominantly isolated and characterized. The WHO recently published an integrated global surveillance protocol on ESBL-producing E. coli as an indicator (WHO, 2021). The procedures described were specifically designed to be conducted in a harmonized manner to provide the opportunity to increase capacities and to build national integrated surveillance systems for AMR within a One Health approach (WHO, 2021). The environmental aspect of this protocol proposes to detect and quantify ESBL-producing E. coli in contamination hotspot sources including surface water such as rivers that receive wastewater (WHO, 2021). However, the current metadata analysis showed that the methodology, including the frequency and number of samples analyzed, isolation, identification and characterization methods used (even within countries) differed, which contrasts the WHO proposed protocol to conduct research in a harmonized manner.

The results from the current review showed that for studies in South Africa and Algeria, MALDI-ToF was predominantly used for identification of the potential pathogens, while countries including Nigeria, Tunisia, Ghana, Egypt, Morocco and Kenya among others, predominantly used biochemical tests such as API strips for isolate identification. Recently, the reproducibility and accuracy of MALDI-ToF mass spectrometry was evaluated through comprehensive comparison studies in the clinical field (Hou et al., 2019). The authors concluded that the high-throughput, cost effective method with high accuracy resulted in MALDI-ToF mass spectrometry superseding previous conventional molecular or biochemical identification systems (Hou et al., 2019). Moreover, globally, MALDI-ToF is increasingly being used in food safety following approval by the U.S. Food and Drug Administration (Cheng et al., 2016). This follows as fast and consistent detection of foodborne pathogens in a cost-effective manner is vital in food safety analyses where time sensitivity represents an important factor (Cheng et al., 2016; Elbehiry et al., 2017). In many LMICs however, researchers still rely on biochemical tests for isolate identification due to infrastructural or funding constraints. This was evident in the current review where selected studies in Nigeria, South Africa and Tunisia as well as the studies from Benin, Ghana, Egypt, Morocco, Kenya and Tanzania all utilized biochemical tests, predominantly the API20E panels, which are reported to give efficient economical identification (Popov et al., 2022). From the current study, ESBL-producing Enterobacterales predominantly included E. coli and K. pneumoniae isolates. Of note was that 39.22% of the studies focussed on isolation and characterization of E. coli only, therefore, firm conclusions on the prevalence of ESBL-producers in the water-plant-food environment from these African countries cannot be drawn from this data alone. The WHO has reported that E. coli and K. pneumoniae, amongst other third-generation cephalosporin-resistant Enterobacterales, are categorized as a critical priority for global antimicrobial resistance research and development (WHO, 2017a).

Across all countries from the current meta-analysis, the Kirby-Bauer disk diffusion method was mainly used for phenotypic AMR analysis, with the CLSI criteria predominantly followed. However, the antibiotics included in analysis differed across studies and countries, therefore, no conclusion regarding phenotypic resistance patterns of potential pathogens isolated from the different matrices could be reported. Of note is that multidrug resistance (MDR), which is defined as resistance to more than one antibiotic class (Magiorakos et al., 2012), was reported in the majority of the studies. In 2014, the WHO released the first report on AMR surveillance and highlighted the need for an improved and coordinated global effort, while the FAO reported in 2018 on the need for increased environmental antimicrobial resistance surveillance (WHO, 2014; FAO, 2018). Phenotypic resistance to antimicrobials considered as critically important in clinical treatment were detected in the isolates from the water-plant-food nexus in the current meta-analysis. As an example, E. coli isolated from fresh produce and water across all countries showed resistance against β-lactams including penicillins like amoxicillin/clavulanic acid, cephalosporins, carbapenems and aminoglycosides including gentamycin, among others. Furthermore, both Abakpa et al. (2015) and Richter et al. (2020) reported on the occurrence of MDR Salmonella spp. isolated from irrigation water and fresh produce. Antibiotics from classes including β-lactams (amoxicillin/clavulanic acid, cefoxitin, imipenem), cephalosporins (cefepime), tetracylines, and aminoglycosides (gentamycin) amongst other, formed part of the antibiogram profiles in Nigeria and South Africa, respectively. Previous studies have reported that penicillins (specifically amoxicillin/clavulanic acid) are the most commonly prescribed antibiotics in hospital settings in Nigeria as well as South Africa (Okoro et al., 2019; Alabi and Essack, 2022). Tetracycline resistance was also evident in a high number of environmental bacterial isolates from the current review. The use of antibiotics in food animals and overuse in clinical settings have been linked to transmission and rise in environmental multidrug resistant bacteria (Jones-Dias et al., 2016). Along with spread of multidrug-resistance bacteria across all One Health domains, outbreaks of ESBL- and carbapenemase-producing bacteria present a serious challenge to clinicians, with the increasing occurrence in environments posing a public health concern.

The bla_{CTX − M} ESBL resistance gene was identified in all the studies that included PCR analysis in the current review. Similarly, Muthupandian et al. (2018) reported that class A and class D ESBLs are common in bacteria from clinical settings in Africa, with the CTX-M-15 gene being the most prevalent. Screening of resistance to carbapenem antibiotics were included in 25 (59.5%) of the studies from seven different countries in the current review. Imipenem, conferring resistance to Ambler Class B carbapenem-resistant metallo-β-lactamases (Sawa et al., 2020) was the dominant antibiotic included in phenotypic screening, followed by meropenem, which confers resistance to Ambler Class A carbapenemases (Sawa et al., 2020). Carbapenem antimicrobials are considered the last-resort antibiotics for treatment of infections caused by third- and fourth generation cephalosporin resistant and multidrug-resistant bacteria (Sheu et al., 2019). Similar to the results from the current review, Brunn et al. (2022) reported on the presence of carbapenem resistance in environmental reservoirs globally and suggested that wastewater treatment plant effluents were a primary environmental contaminating source. In the subset of studies that included genotypic analysis in the current review, only selected studies (n = 9) screened for presence of carbapenem resistance genes. The KPC genetic determinant was identified in isolates from water samples, while the GES carbapenem genetic determinant was detected in isolates from water and fresh produce samples.

Class A carbapenemases, which include KPC and GES, are plasmid-encoded and frequently detected in clinically significant Klebsiella spp. and Pseudomonas aeruginosa (Sawa et al., 2020). In the current review, these genes were present in E. coli and K. pneumoniae isolates in selected studies conducted in South Africa. Perovic et al. (2016) reported that the introduction of carbapenemase-producing Enterobacterales in South Africa was molecularly confirmed at the end 2011. Furthermore, following environmental introduction, that rapid gene dissemination occurs (Perovic et al., 2016). Recently, Ragheb et al. (2022) reported on the genetic environments of carbapenmases from a One Health perspective in Africa. The authors concluded that the most commonly found carbapenemases were variations of NDM, OXA-48 and VIM being reported from clinical, animal and environmental samples.

Class B carbapenemases are typically encoded on a plasmid, transposon, integron, or chromosome and include the IMP, VIM and NDM genetic determinants (Sawa et al., 2020), supporting rapid gene dissemination across the different one health domains. Interestingly, studies in South Africa from the current review only reported on presence of the VIM and IMP genetic determinants in E. coli isolates, while the same genetic determinants were reported in K. pneumoniae only, in similar studies in Tunisia. Additionally, the NDM carbapenem resistance gene was detected in soil K. pneumoniae isolates from Nigeria, water E. coli isolates from South Africa and fresh produce K. pneumoniae isolates from Egypt. The results from the current review reiterates that clinically significant antibiotic resistant bacteria are no longer restricted to hospital settings, supporting the WHO (2017c) findings that global research and development strategies should include antibiotics active against more common community bacteria.

Similar to Ragheb et al. (2022), the results from the current review showed that detection of ESBL- and carbapenemase-producing Enterobacterales in Africa were relatively low as only nine countries from the entire continent have done environmental AMR surveillance in the current review. Moreover, only studies from six countries have included PCR analysis of β-lactamase genes. However, it must be reiterated that the low prevalence of these multidrug-resistant critical priority pathogens does not represent low environmental occurrence, but rather limited research on the role of the environment in dissemination of AMR within Africa. In selected studies from the current review, multidrug resistant potential pathogens were detected in different samples along the fresh produce farm-to-fork continuum. The mobility of antimicrobial resistance genes in association with mobile genetic elements and the subsequent potential of these genes to move across species, highlights the vital importance of standardized methods, achievable for LMICs as well as, to establish internationally comparable baseline environmental occurrence data (Ikhimiukor et al., 2022; Liguori et al., 2022). From a food safety perspective, knowledge of AMR dissemination along the food supply chain, especially in food crops commonly consumed raw, is critically important for risk communication, intervention and public health (WHO, 2017b).

The inclusion of WGS has been reported as a promising tool for estimation of ARB in a one health context (Aslam et al., 2021). In the current review, only four studies included WGS analysis (Adelowo et al., 2020; Le Terrier et al., 2020; Zekar et al., 2020; Altayb et al., 2022), conducted in Nigeria, Algeria and Sudan, respectively. Additionally, a follow-up WGS analysis study for further characterization of isolates previously published was found in South Africa (Richter et al., 2021), however, this study was not included in the current meta-analysis, to avoid duplication of isolate occurrence information. Whole genome sequencing has successfully been incorporated in integrated food safety surveillance programs in high-income countries (Brown et al., 2019). Across borders, platforms such as GenomeTrakr and PulseNet International are used to share outbreak investigation data, however, there is still a long way to go in the ability to attribute sporadic illness to specific food categories (Brown et al., 2019). As sequencing costs reduce and methods for designing and interpreting metagenomic studies improve, the use of WGS in surveillance studies, especially in infrastructure and capacity-limited LMICs, are increasingly becoming a possibility (Ikhimiukor et al., 2022). The globalization of food supply chains necessitates an integrated surveillance system and platform to share molecular data for food safety purposes. The addition of WGS analysis to phenotypic AMR surveillance studies can aid in understanding the genetic basis of AMR mechanisms and distinguish isolates with phenotypically identical antibiograms, in addition to virulence genes and single-nucleotide polymorphism (SNP) analysis typically used in food safety surveillance and foodborne outbreak investigations (Brown et al., 2019; WHO, 2020). Subsequently, valuable information on the pathways of AMR dissemination across all sectors can be obtained.

The available data on occurrence of multidrug-resistant Enterobacterales in environmental settings in Africa emphasizes the need for improved surveillance and documentation of resistance gene dissemination across the farm-to-fork continuum globally. This follows as clinically significant bacterial isolates were found in various water sources and fresh produce. Furthermore, these human pathogenic microbes harbored resistance traits that corresponded to antibiotics often used in clinical settings as well as animal husbandry. The information obtained from the current review could however not be used to determine the extent of the human health risk in consumption of fresh produce where ESBL-producing potential pathogens were present. In addition to a need for harmonized methodology, the cost-effectiveness of One Health AMR surveillance systems, especially in LMICs, should also be considered and further investigated to inform the development and effective and efficient systems in Africa. The results further posed a challenge in comparative studies, as standardized methods were not utilized across the board. Although limited environmental AMR surveillance studies were found in comparison to published data on AMR surveillance in human and animal health, this review showed the vital importance of including information from the water-plant-food nexus in food safety surveillance programs related to AMR in a One Health context. It was further highlighted that comparable indicators to monitor AMR in food crop value chains is necessary. Establishing WGS as a surveillance tool in addition to phenotypic data in AMR surveillance studies will provide comprehensive information to inform comparable national and international actions plans against AMR.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

LR, ED, SD, and LK contributed to the conceptualization of the study. LR and ED performed the data extraction. LR summarized the data, prepared all figures, and performed the statistical analysis. All authors contributed to the manuscript revision, read, and approved the submitted version.