Time to define One Health approaches to tackling antimicrobial resistance

A study of extended spectrum beta-lactamase (ESBL) producing E. coli by Public Health England (now UK Health Security Agency) compared isolates from human faeces, sewage, farm slurry, and retail foodstuffs to human bloodstream infection (BSI) isolates and found little overlap between non-human reservoirs and isolates from invasive human disease². Similarly, a genomic surveillance of E. coli in the UK found distinct lineages and mobile genetic elements of E. coli between human BSI and livestock samples³. A recent study of Klebsiella pneumoniae from a wide range of clinical, community, animal and environmental settings in Italy, described by the authors as “a hotspot for hospital-acquired carbapenem non-susceptible Klebsiella”, found no genotypic or phenotypic evidence for non-susceptibility to carbapenems outside the clinical environment⁴. In these studies, a lack of relatedness between bacterial isolates found in human infection samples and isolates from other ecological sources may not equate to a lack of transmission between ecological compartments. The sampling strategies employed including sampling animals from closed farms alongside unlinked humans from a large geographical area, may have influenced the ability to detect significant or potential transmission events⁵.

On the other hand, some studies have shown evidence for a strong association between poor environmental health infrastructure and AMR, particularly affecting the most vulnerable populations in the world⁶. The Drivers of Resistance in Uganda and Malawi (DRUM) study, collected demographic, geospatial, clinical, animal husbandry and environmental health (including WASH infrastructure and practice) data from households in urban, peri-urban and rural settings in Uganda and Malawi⁷. Longitudinal human, animal and environmental sampling at each household was used to isolate ESBL E. coli and ESBL K. pneumoniae. Multivariable models illustrated that human ESBL-producing E. coli colonisation was associated with the wet season and close animal interaction, and that without adequate efforts to improve environmental health, ESBL-producing Enterobacteriaceae transmission is likely to persist in these settings^8,9.

E. coli is a highly diverse species, and the resolution offered by whole genome sequencing (WGS) reveals that sub-groups of E. coli, even at the level of discrimination offered by multi-locus sequence typing are generally distinct in different ecological sources. Despite this, there is flow of antimicrobial resistant strains or AMR genes between one health compartments. This is particularly evident in sub-Saharan Africa (sSA), where analysis of E. coli genomes from the UrbanZoo study in Kenya demonstrated that although transmission of the general E. coli population is often within host species¹⁰, there is significant flux of accessory genes on mobile genetic elements (MGE), especially AMR genes, between humans and animals^11,12. Sharing of the accessory genomes across different sources was also observed among K. pneumoniae isolates from clinical, environmental, and animal sources in Ghana, an indication of transmission of K. pneumoniae between the different compartments¹³.

In Malawi and Uganda, comparison of E. coli genomes revealed high diversity of E. coli, MGE and ESBL genes that were distributed independent of ecological compartments¹⁴. Detailed lineage level SNP analysis of these genomes further indicated putative transmission events of ESBL-E. coli between humans, animals and the environment¹⁴. Outside sSA and Europe, evidence of phylogenetic intermixing and sharing of AMR genes between humans, animals and soil has been reported in Bangladesh¹⁵. Similarly, in Ecuador, significant clonal and AMR gene sharing of third generation cephalosporin resistant E. coli was observed between children and animals, suggesting interlinkage of AMR in animals and humans¹⁶. The degree of sharing of bacteria strains and AMR determinants between different ecological compartments reported in studies such as these is certainly underestimated. This is particularly the case because carriage studies involving isolate WGS do not fully account for within sample diversity. Nevertheless, even without capturing that full diversity, the available evidence suggests exchange of AMR bacterial strains and AMR determinants between the different ecological compartments is common, particularly in low- and middle-income countries (LMICs).

Detection of closely related ESBL-producing Enterobacteriaceae in humans, animals and the environment within LMICs not only illustrates the relevance of the One Health paradigm to key AMR human pathogens, but is also a manifestation of the poor environmental health standards and practices combined with the close interactions between humans, animals and the environment. Although not a surprise to the WASH community, the prominence of AMR as a global health issue can be used to shine new light on the urgency to address Sustainable Development Goal 6: ensuring the universal and equitable access to WASH for all. This could be effectively utilised to catalyse renewed enthusiasm and commitment for investment in WASH infrastructure, with aligned effective social and behaviour change programmes to reduce the interactions between these different ecological sources. On the other hand, the fact there is limited evidence for ESBL-E. coli and K. pneumoniae transmission between humans, animals and the environment in settings such as Europe where WASH infrastructure is strong, highlights the central role context appropriate approaches to improving environmental hygiene must play. The presence of these bacteria in healthcare settings informs us that IPC strategy in these settings is failing, placing patients at risk of adverse outcomes. Indeed, a high prevalence of healthcare associated infection, in part, illustrates a failure of environmental hygiene measures and practices in healthcare settings¹⁷. The response to AMR must not only be contextually appropriate, but also take a whole system approach, understanding that infection prevention is broader than the practice of IPC in healthcare facilities, and should embrace WASH infrastructure and practice in domestic, public and institutional settings. Indeed, to prevent severe bacterial infection from ESBL-E. coli, asymptomatic transmission of ESBL-enteric bacteria needs to be interrupted. AMR does not recognise the disciplinary silos in which we typically operate and the presence of widespread ESBL-enteric bacteria indicates that we are failing to create effective, equitable and sustainable access to basic WASH infrastructure and services for the most vulnerable populations in the world.

Evidence for a One Health Framework for AMR remains incomplete. While some studies have provided evidence of the need to take a One Health approach to AMR, such approaches will differ according to setting. Acknowledging however, that environments in LMIC are commonly contaminated by AMR bacteria such as ESBL–producing E. coli and K. pneumoniae and therefore unsafe, is a critical starting point. It demonstrates the need for a whole system approach to environmental hygiene, in which we will see catalytic and reinforcing advantages from transdisciplinary working to develop context appropriate and sustainable solutions that tackle global public health, economic development, dignity and wellbeing simultaneously by improving WASH infrastructure and practice. Bacterial populations must be segregated by introducing transmission bottlenecks between human, animal and environmental compartments to reduce overall AMR transmission rates. To achieve this, there is an urgent need for designing, validating and understanding how best to implement such approaches in the face of competing public health priorities in all societies, and especially in LMICs.