Introduction
Antimicrobial resistance (AMR) occurs when microbes grow in the presence of a drug whose purpose is to eliminate these microbes. Reference Huemer, Mairpady Shambat, Brugger and Zinkernagel1,Reference Peri, Doi, Potoski, Harris, Paterson and Righi2 The rise of AMR has been catalyzed by the overuse of antimicrobial treatments in humans and animals Reference Manyi-Loh, Mamphweli, Meyer and Okoh3,Reference Walsh, Gales, Laxminarayan and Dodd4 and unsuccessful efforts to develop new alternative antimicrobials. Reference Lewis5,Reference Salam, Al-Amin and Salam6 AMR exacerbates negative health outcomes in patients, resulting in higher mortality rates and extended hospital stays. Reference Laxminarayan, Duse and Wattal7 As a result, labor supply and productivity diminish. Reference Dadgostar8,Reference Poudel, Zhu and Cooper9 Currently, AMR is one of the leading causes of infection-related deaths worldwide, exceeding annual deaths due to tuberculosis, malaria, and HIV/AIDs combined. Reference Laxminarayan, Duse and Wattal7,10 It is estimated that by the year 2050, AMR will be responsible for 10 million deaths per year worldwide, exerting a total global cost of US$100 trillion. Reference Neill11,Reference Varma, Oppong-Otoo and Ondoa12 Although AMR affects all nations, 10,13,Reference Founou, Founou and Essack14 developing nations are disproportionately affected. Reference Laxminarayan, Duse and Wattal7,Reference Okeke, Laxminarayan and Bhutta15 Of all Global Burden of Disease Study regions, sub-Saharan Africa (SSA) and South Asia had the highest mortalities (23.5 deaths per 100,000 and 21.5 deaths per 100,000, respectively) attributable to AMR. Reference Kariuki, Kering, Wairimu, Onsare and Mbae16 Generally, weaker infrastructure in these regions coincides with a high burden of infectious diseases, inadequate prevention measures, and other social determinants that contribute to high AMR rates. Reference Gwimile, Shekalaghe, Kapanda and Kisanga17,Reference Senn, Rarau and Salib18 In SSA, there is little regulation of the prescription and sale of antimicrobials, Reference Fuller, Aboderin and Yahaya19 poor adherence to treatment guidelines by professionals, and inadequate access to culture and sensitivity tests to identify drug-resistant bacteria. Reference Gulumbe, Haruna, Almazan, Ibrahim, Faggo and Bazata20 Although AMR surveillance systems are an integral part of public health, chronic underfunding and weak infrastructure are the most common problems in African countries. Reference Moyo, Moyo and Mangoya21 Less than half of these countries are enrolled in the WHO Global Antimicrobial Resistance and Use Surveillance System, Reference Fuller, Hamzat and Aboderin22 and even fewer have implemented AMR guidelines and reported surveillance data. Reference Gulumbe, Haruna, Almazan, Ibrahim, Faggo and Bazata20
Public health intervention efforts to address AMR focus on a “One Health” approach, which emphasizes the role of social and ecological determinants in creating systemic barriers to AMR. Reference Léger, Lambraki and Graells23–Reference Wernli, Jørgensen and Parmley25 This approach includes efforts to reduce the burden of disease; improve food consumption, water, sanitation, and hygiene access; and provide knowledge about antimicrobials. Reference Collignon, Beggs, Walsh, Gandra and Laxminarayan26 Studies in high-income countries have demonstrated how the examination of social and ecological factors reveals macro-level trends that can be monitored and addressed through intervention. Reference Bordier, Uea-Anuwong, Binot, Hendrikx and Goutard27,Reference Johnson, Hansen and Bi28 In contrast, studies in SSA focus on the magnitude of AMR but provide limited information to address social and ecological factors that contribute to it. Previous systematic reviews on AMR in Africa summarize the magnitude of AMR and its ecological distribution, Reference Bernabé, Langendorf, Ford, Ronat and Murphy29,Reference Escher, Muhummed, Hattendorf, Vonaesch and Zinsstag30 and a few reviews have synthesized social determinants trends of AMR in Africa. Reference Moyo, Moyo and Mangoya21,Reference Woldegeorgis, Kerbo, Obsa and Mokonnon31 To the best of our knowledge, no systematic review in Africa has examined both the social and ecological determinants of AMR. Researchers and decision-making bodies must be aware of the social and ecological influences on AMR and account for how they are intertwined to identify meaningful targets and implement effective interventions.
Methods
Literature search and selection
Predefined inclusion and exclusion criteria were used to identify relevant epidemiological studies using the PubMed database (Supplementary Table 1). To be considered, studies had to meet the following conditions: (i) publication in English; (ii) published within the past 10 years (after 2013); (iii) a cohort, case-control, cross-sectional study, or randomized controlled trial; (iv) identified social or ecological determinant exposures through interviews, hospital records, Area Deprivation Index, survey data, health-related social needs screening tools, livestock samples, water analysis, air quality assessments, soil examinations, and other relevant social or ecological factors; and (v) AMR determination using methods such as antimicrobial susceptibility testing, whole genome sequencing or metagenomics, bioinformatics, microarrays, commercial antibiogram techniques, immunochromatography, polymerase chain reaction (PCR) testing, MALDI-TOF mass spectrometry, AMR surveillance, or other applicable methods. Studies that appeared as letters, editorials, reviews, correspondences, and case reports were excluded from the analysis. Study eligibility was independently assessed by two researchers (CB, WR) through a meticulous 3-step process, beginning with a title examination, followed by an abstract review, and concluding with a full-text evaluation.
Study quality assessment
The quality of studies was assessed using an adapted version of the Newcastle–Ottawa assessment scale (NOS) for observational studies. Reference Wells, Shea and O’Connell32 The evaluation was based on 3 criteria: the selection of study groups, the comparability of groups, and the confirmation of exposure and outcomes. Cross-sectional and cohort studies were considered to be high quality if they had a score of 7 or greater out of 10.
Narrative analysis
Table 1 synthesizes the studies that were categorized as either social or ecological as determined by the exposure measurement. Thematic analysis resulted in the subcategories of social determinants of AMR (n = 7): demographic information, education history and awareness of AMR, study settings (urban vs rural), and the economic low-resource setting status of participants. The ecological determinants (n = 14) were categorized into themes such as water sanitation and hygiene (WASH), pollution, farming practices, food safety, water analysis, and wastewater treatment. Studies that discussed both social and ecological determinants (n = 3) had a combination of at least 1 social and 1 ecological determinant.
Results
Characteristics of included studies
A total of 24 studies were used in the systematic review, all of which were retrieved from PubMed (Figure 1). With the exception of one study, all studies used a cross-sectional study design (Table 1). Of the studies examined, 45.83% (n = 11) were conducted in East Africa, 37.5% (n = 9) in West Africa, and 16.67% (n = 4) in other regions in Africa (Supplementary Figure 1). No studies were located in North Africa. Overall, 58.33% (n = 14) of the studies focused on ecological determinants of AMR, 29.17% (n = 7) on social determinants of AMR, and 3 (12.5%) discussed both social and ecological determinants of AMR.
Quality appraisal results
Of the 24 studies assessed, 4 were deemed of moderate quality, and 23 were of high quality (Table 1). No studies received a poor-quality assessment.
Social determinants
Household and population size
There is some evidence that overcrowded living environments at the household and population level increase the prevalence of AMR and the abundance of antimicrobial resistance genes (ARGs). A study from Kenya showed that human carriage of ARGs from Escherichia coli in households that keep livestock was positively associated with increased household size. Reference Muloi, Hassell and Wee33 A metagenomic study on raw sewage in Cameroon found that the abundance of clinically relevant ARGs was higher in a city of 3.5 million inhabitants than in a city of 1.0 million. Reference Bougnom, McNally, Etoa and Piddock34 Specifically, 71 ARGs were common between the 2 cities, but an additional 61 ARGs were unique to the larger city, with only 4 ARGs unique to the smaller city. Reference Bougnom, McNally, Etoa and Piddock34
Pastoralist communities
Pastoralist communities may be at increased risk of AMR, given their close proximity to livestock and their use of prophylactic antimicrobials. A study in Nigeria showed that most (60%) pastoralists purchased antimicrobials from veterinary drug shops (as opposed to “animal drug hawkers”), and two-thirds of these pastoralists practiced arbitrary applications, rather than following the instructions given. Reference Alhaji and Isola35 In addition, most pastoralists did not know of AMR (67%), and few (30%) thought that antimicrobial misuse in livestock could result in the emergence of resistant pathogens. Reference Alhaji and Isola35 Low education, lack of funds for veterinary care, seasonal movement, and nonenforcement of antimicrobial distribution laws were all strong predictors of antimicrobial misuse. Reference Alhaji and Isola35
A study from Tanzania highlights the complexity of understanding patterns of antimicrobial use in livestock against the backdrop of variations in cultural and livelihood practices. Reference Caudell, Quinlan and Subbiah36 Ethnic groups varied in how often they sought veterinary consultations (Chaaga = 96%, Arusha = 45%, Maasai = 36%), which was inversely related to the use of self-administered antimicrobials (Chaaga = 1%, Arusha = 21%, Maasai = 74%). Additionally, all pastoralists self-administering antimicrobials did not observe the recommended withdrawal period from consumption of milk and meat. Maasai, who purchased over-the-counter medicines from veterinary drug shops, were less likely to consult with a veterinarian for livestock care. Arusha, owning transportation or a cell phone or with higher levels of education, were more likely to self-administer antimicrobials. Maasai, who owned a radio or who had electricity, were more likely to observe posttreatment withdrawal from the consumption of milk and meat. Reference Caudell, Quinlan and Subbiah36 In a follow-up study, Reference Caudell, Mair and Subbiah37 AMR from human stool samples was highest for Maasai and Arusha households and lowest for Chagga households, with >40% of isolates resistant to ampicillin, tetracycline, trimethoprim, sulfamethoxazole, and streptomycin in Maasai and Arusha households, compared to 20%–30% resistance in Chagga households. Reference Caudell, Mair and Subbiah37 However, antimicrobial use in humans and livestock was not associated with resistance to any antimicrobial. Reference Caudell, Mair and Subbiah37
Rural versus urban
Four studies explicitly assessed the impact of rural versus urban living on antibiotic use, AMR and ARG prevalence, and knowledge of AMR. The Tanzanian study noted above Reference Caudell, Quinlan and Subbiah36 found a negative association between distance to urban centers and the likelihood of consulting a veterinarian for livestock care and purchase of antimicrobials in Arusha populations but not Chagga or Maasai populations. Reference Caudell, Quinlan and Subbiah36 A Kenyan study of communities and hospitals in an urban (Nairobi) and rural (Siaya) setting showed that extended-spectrum cephalosporin-resistant Enterobacterales were more prevalent in the urban setting (52%) compared to rural setting (45%) and in urban hospitals (70%) compared to rural (63%) hospitals. Reference Ita, Luvsansharav and Smith38 A large global burden of antibiotic resistance study including 1589 metagenomes from 26 countries showed that increased access to improved water and sanitation was associated with a lower abundance of ARGs; this association was stronger in urban compared to rural settings. Reference Fuhrmeister, Harvey and Nadimpalli39 A Tanzanian study on antibiotic residues in foods reported that 42% of egg and Chinese cabbage consumers were unaware of the likelihood of antibiotic residues in their food, Reference Mongi, Meshi and Ntwenya40 even though 73% knew that animal waste containing antibiotics was used for cultivating fruits and vegetables and 78% knew that antibiotic residues could be harmful to humans. Educational status and urban living were linked to awareness of antibiotic residues in foods. Reference Mongi, Meshi and Ntwenya40
Clinical-based settings
A study of patients in urban and rural hospitals in Kenya found a high prevalence of extended-spectrum cephalosporin-resistant Enterobacterales in nasal and fecal specimens and a very low prevalence of carbapenem-resistant Enterobacterales and methicillin-resistant Staphylococcus aureus, with a higher prevalence of resistant Enterobacterales in urban compared to rural hospitals. Reference Ita, Luvsansharav and Smith38 A study of a teaching hospital in Kigali, Rwanda, revealed that gram-negative bacteria are highly resistant to antibiotics that are usually prescribed. Additionally, patients who had undergone surgery, taken antimicrobials in the past 30 days, or were transferred from another hospital had a greater chance of developing AMR. Reference Sutherland, Mpirimbanyi and Nziyomaze41
In Ghana, a study exploring antimicrobial prescribing behavior and the knowledge of teams treating dental patients found that antimicrobials were viewed as a fundamental aspect of dental care and were being prescribed both therapeutically and prophylactically, with amoxicillin, metronidazole, and amoxicillin/clavulanic acid being the most commonly prescribed. Reference Sneddon, Thompson and Kpobi42 Some respondents shared concern about the cleanliness of facilities and the sterility of instruments used as a rationale for the extensive use of antimicrobials to prevent postoperative infections. They also wanted prescribers to follow the prescription advice of senior colleagues rather than advice from the national standard treatment guidelines. Reference Sneddon, Thompson and Kpobi42 Most respondents had a good general knowledge of AMR, but this knowledge did not affect their decision-making around antimicrobial prescribing, which was influenced more by the perception of the patient’s ability to pay. Reference Sneddon, Thompson and Kpobi42
Ecological determinants
Water, sanitation, and hygiene (WASH)
The aforementioned metagenomic global burden of antibiotic resistance study Reference Fuhrmeister, Harvey and Nadimpalli39 found that the abundance of ARGs was highest in Africa, followed by Southeast Asia and South and Central America. Increased access to improved water and sanitation was negatively associated with the abundance of ARGs for tetracycline and trimethoprim but not with the abundance of ARGs conferring resistance to beta-lactams, fluoroquinolones, aminoglycosides, or streptogramins. A study in Nairobi, Kenya, showed that children eating soil and the presence of communal handwashing stations within a block of households were associated with an increased load of antimicrobial-resistant E. coli. Reference Omulo, Lofgren and Lockwood43 However, rainfall was the best predictor of a decreased load of antimicrobial-resistant E. coli, suggesting that transmission through unsanitary living conditions overwhelms the incremental effects of prior antibiotic use. Reference Omulo, Lofgren and Lockwood43 A study of water quality in Guinea-Bissau (West Africa) documented that samples from the wet season had 67% more fecal coliforms than samples from the dry season and had a higher proportion of samples resistant to 3 or more antibiotics (wet = 22%; dry = 9%) but a lower percentage of isolates resistant to at least 1 antibiotic (wet = 59%; dry = 72%). Reference Machado and Bordalo44
Several studies in different parts of Africa examined wastewater samples. In South Africa, raw sewage water had higher levels of ARGs than treated wastewater discharge. Reference Hamiwe, Kock, Magwira, Antiabong and Ehlers45 In Uganda, most (61%) Salmonella typhimurium and Salmonella enteridis isolates collected from wastewater treatment plants and slaughterhouses were susceptible to all antimicrobials tested. However, there were high levels of resistance to a variety of commonly used antibiotics, and shared Salmonella genotypes were found in human, livestock, and environmental sources, indicating that zoonotic and environmental transmission to humans is likely. Reference Afema, Byarugaba, Shah, Atukwase, Nambi and Sischo46 Another study comparing the wastewater from hospitals in Benin and Burkina Faso revealed a higher relative abundance of ARGs in wastewater from hospitals in these countries compared to 6 hospitals in Finland. Reference Markkanen, Haukka and Pärnänen47 In Zanzibar, plastic litter from 4 rural sites that had a history of cholera outbreaks, and poor sanitation facilities contained multiple multidrug-resistant enteric pathogens, including Klebsiella pneumonia, Enterobacter cloacae, Citrobacter freundii, and Vibrio cholerae. Reference Rasool, Saavedra and Pamba48 In a Kenyan study, the availability of WASH resources like toilet handwashing stations, proper waste separation, clean toilet facilities, appropriate distance between hospital beds, and drinking water storage might explain the differences in patterns of AMR across hospitals. Reference Maina, Tosas-Auguet and McKnight49
Farming
There is ample evidence of AMR associated with livestock. The Wakiso District in Uganda has the highest livestock production in the county, and when environmental, human, and livestock samples were screened for extended-spectrum-beta-lactamase-producing E.coli, 80% of households had at least 1 positive sample. Seventy percent had animals carrying the drug-resistant bacteria, 62% had humans carrying the drug-resistant bacteria, and 10% had environmental samples carrying the drug-resistant bacteria. Reference Muleme, Musoke and Balugaba50 The use of a protected water source for drinking, using containers with lids, and having a clean household were negatively associated with the presence of drug-resistant bacteria. Reference Muleme, Musoke and Balugaba50 A survey in Tanzania revealed that poultry farmers consistently use oxytetracycline via drinking water for disease prevention, treatment, or growth promotion, and a majority of cabbage farmers use manure from livestock treated with antibiotics to enrich their soil. Reference Mongi, Meshi and Ntwenya40 Similarly, a study conducted in a Zambian poultry farm revealed that 10% of the samples contained cefotaxime-resistant E. coli and that all of these samples were also resistant to at least 2 other antimicrobial classes. Reference Shawa, Furuta and Paudel51 Further analysis showed that 20% of the poultry isolates were closely related to 25% of human inpatient isolates, sharing 2 plasmids that contained 14 distinct AMR genes, pointing to transmission between poultry and humans. Reference Shawa, Furuta and Paudel51 Another study conducted on lettuce farms in Ghana found E. coli in all lettuce samples, regardless of the water source used for irrigation, with most (82%) of the samples resistant to more than 1 antimicrobial. Reference Quarcoo, Boamah Adomako and Abrahamyan52 About one-fifth of raw meat samples (beef, sheep, goat, and chicken) from slaughterhouses in Ethiopia contained E. coli that were resistant to commonly used antibiotics, with nearly one-half showing multiple drug resistance. Reference Messele, Abdi, Yalew, Tegegne, Emeru and Werid53
Food safety
A survey of ready-to-eat foods in Accra, Ghana, showed that E. cloacae was present in 17% of tested food samples and showed resistance to commonly used antibiotics in the region. Reference Dela, Egyir and Behene54 Similarly, a study of ready-to-eat shrimp in Nigeria showed that 93% of ready-to-eat shrimp contained Vibrio species, with the most common being Vibrio parahaemolyticus (38%), Vibrio vulnificus (12%), and Vibrio fluvialis (10%). Reference Beshiru, Okareh, Okoh and Igbinosa55 All isolates were 100% sensitive to colistin and gentamycin but broadly resistant to other antibiotics commonly used in the region. Reference Beshiru, Okareh, Okoh and Igbinosa55 Likewise, a study from a different part of Nigeria showed that 13% of barbecued beef samples from local markets contained E. coli. Reference Datok, Ishaleku, Tsaku, Agya and Adoga56 All the E. coli strains were resistant to ampicillin, and 75% of the cultures that tested positive for E. coli showed resistance to multiple different antimicrobials. Reference Datok, Ishaleku, Tsaku, Agya and Adoga56
Discussion
This systematic review of studies assessing social or ecological determinants of AMR in Africa points to some common themes. There is some evidence that overcrowded living environments at the household and population level could increase the prevalence of antimicrobial-resistant bacteria and the abundance of antimicrobial-resistant genes circulating in the human population. Additionally, indicators of poor water sanitation and hygiene-related variables, framing practices, and consumption of farm products were among the common ecological factors linked with AMR in Africa. It seems clear that our ability to understand patterns of AMR, and ultimately the factors driving its emergence, is going to depend on a nuanced and integrative investigation of not only factors like antimicrobial use, population density, water quality, and sanitation but also critically factors like variation in cultural practices, how livestock are raised, farming practices, and the management of healthcare facilities.
Strengths and limitations
One of the strengths of this review was the comprehensive screening process. Eligible studies were chosen based on clearly defined inclusion criteria, and 2 reviewers (CB and BT) worked together to determine the final set of studies. This review also adds to the existing literature on the determinants of AMR by using a combination of social and ecological lenses rather than individual factors alone.
There are limitations associated with our review, however. First, PubMed was the only search engine used in our study, raising the possibility of selection bias. Another key limitation was the lack of research on social and ecological determinants of AMR in some parts of Africa, such as North Africa. Furthermore, the cross-sectional nature of the included studies in this review limits our ability to draw conclusions about the causal relationship between the exposure (social and ecological determinants) and the outcome (AMR). It is difficult to determine whether the reported AMR followed the social and ecological factors. Having this information would be crucial to establishing the temporal relationship between social and ecological factors and AMR. Although most of the included studies explored the impact of potential confounding variables on the observed association between social and ecological factors and AMR, the possibility of reverse causation is difficult to eliminate without baseline data fully. We recommend future longitudinal studies to rule out the possibility of reverse causation.
Given these limitations, our review provides an impetus to assess further complex sociodemographic and ecological factors contributing to African AMR. Furthermore, the review is particularly relevant to attempts to inform public health professionals about how social-demographic and environmental factors are interconnected and facilitate AMR in Africa.
Implications
The principal current strategy to combat AMR among clinicians and the pharmaceutical industry is to prescribe alternative antimicrobials and develop new drugs [63]. However, this strategy is difficult to sustain. It requires ongoing monitoring and detection of AMR and testing, which makes it challenging to provide most African healthcare facilities with the resources necessary to tackle AMR effectively. Additionally, the spread of AMR bacteria is usually nonlinear and involves many social-demographic and ecological determinants. Therefore, public health experts must simultaneously design and implement holistic AMR prevention measures addressing AMR’s social and ecological determinants.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/ash.2024.375
Author contribution
BT conceived and designed the study and wrote the final version of the manuscript. CB and BT searched databases and reviewed studies. CB and WR cleaned the data and prepared the initial draft of the manuscript. FF and RU participated in writing the final version of the manuscript and critically reviewing it. All authors read and approved the final manuscript.
Financial support
This systematic review received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Catherine Bennett received a travel grant from the Colgate University Office of Undergraduate Research to present this work at the 2024 National Conference on Undergraduate Research (NCUR), Long Beach, CA.
Competing interests
We declare that we have no conflicts of interest. The views expressed are those of the author(s) and not necessarily those of Colgate University.