Global trends in antimicrobial resistance in animals in low- and middle-income countries

Background: Since the 1950s, the global increase in demand for meat and dairy has driven the use of antimicrobial drugs in agriculture. This practice has led to the development of antimicrobial resistance in animals and food products, with potentially harmful consequence for agricultural productivity, and human health. In low- and middle-income countries (LMICs), trends in AMR are poorly documented, and surveillance network are at best nascent. Thus far, expert opinion has prevailed over an evidence-based assessment of the AMR status of LMICs. In this context, point prevalence surveys can be used as surrogates to systematic surveillance to provide a baseline of AMR levels, and guide interventions in LMICs. Methods and materials: We extracted twelve thousand resistance rates from point prevalence surveys conducted in LMICs on common foodborne pathogens. Data on AMR was identified for Escherichia coli, non-Typhoidal Salmonella, Campylobacter spp., and Staphylococcus aureus. Resistance rates were manually extracted and curated across all the studies in a public database named RESBANK. We accounted for potential differences in accuracy of antimicrobial susceptibility testing between regions using the WHO External Quality Control System, as well as for temporal revisions of the breakpoints used for susceptibility testing. For each study, we calculated the proportion of drugs tested with resistance levels higher than 50% (P50), and used ensemble geospatial modelling (stacked generalization) to produced global maps of P50, at 10Km resolution. Results: From 2000 to 2018, the proportion of antimicrobials with resistance higher than 50% increased twofold in chickens, and threefold in pigs. China, Northeast and South India represented the largest hotspots of resistance, while new hotspots are emerging in Central India, Brazil, and Kenya. Our maps suggest that worldwide a substantial proportion of chicken, cattle and pigs are raised in hotspots of AMR in 2013. Conclusion: We report a rapid but geographically heterogenous increase of AMR in animals in LMICs. These trends call for urgent actions to preserve the efficacy of existing drugs used in animal agriculture, limit the future economic burden of AMR on farmers. Our global maps of AMR provide a baseline to outline priorities for interventions in LMICs, and monitor their efficacy in the future.