Metagenomic surveillance of AMR in the veterinary environment

Introduction There has been an increased recognition of the importance of the human- animal interface as a driver of AMR. Companion animals such as dogs and cats can be reservoirs of AMR microorganisms. Multi-drug-resistant organisms (MDROs) that commonly cause human-healthcare associated infections (HHAIs) are increasingly being isolated from animals and the environment. Most previous literature looking at MDROs and AMR transmission in veterinary environments has utilised phenotypic testing to identify species and AMR genes and there are currently no standardised tools or guidelines for genomic surveillance of One Health (OH) AMR. In the UK, whilst there is an AMR surveillance programme for both clinical data and farm animals, there is no national surveillance programme for AMR in companion animals, and only limited data exists. This project aimed to utilise a metagenomic approach to identify AMR genes and MDROs present in a small animal hospital environment and compare these with bacterial organisms identified from clinical samples based on phenotypic speciation and antibiotic susceptibility test (AST) results from infection sample sites from animal patients (dogs and cats) from the same veterinary hospital during the previous year. Methods Swabs and wastewater samples were collected from sites around a large, small animal veterinary hospital in North London. Metagenomic DNA was extracted, and a sequencing library prepared, before sequencing using Oxford Nanopore Technologies' MinION platform. The sequencing data were analysed for AMR genes, plasmids and clinically relevant pathogen species. These sequencing data were then compared to phenotypic speciation and ASTs of bacteria isolates from patients in the preceding 12 months. Results In the hospital environment, the most common resistance genes identified were aph (n=101 resistance genes isolated across all 48 metagenomic samples), sul (84 genes), blaCARB (63 genes), tet (58 genes) and blaTEM (46 genes). In clinical isolates, a high proportion of phenotypic resistance to the ß-lactams was identified, especially in Staphylococcal spp. Areas of the veterinary hospital with the greatest mean number of resistance genes identified per swab site were the medical preparation room, dog ward, surgical preparation room, veterinarian workstation room and the consultation room. When stratified by specific sample site, the kennel waste drain had the greatest mean number of resistance genes (n=64), followed by the consulting tables (n=29), bins (n=20), sink plug holes (n=16) and door handles (n=12). Twenty-four Gram-positive plasmids were identified across the 48 metagenomic samples, mostly associated with staphylococci. Four enterobacterial plasmids were identified, carrying carbapenem and quinolone resistance genes. Sequencing reads matched with 14/22 (64%) of the bacterial species identified phenotypically; Pseudomonas aeruginosa and Escherichia coli were common clinical isolates that were also regularly identified in the environment. Discussion This study identified AMR genes, plasmids and MDROs of relevance to human and animal medicine in metagenomic samples from the veterinary hospital environment. High animal-handling areas and areas where dogs were present, such as the dog ward, preparation rooms and consultation rooms were more likely to harbour AMR genes, plasmids and MDROs in the environment. When considering infection prevention and control (IPC) measures, adherence to, and frequency of, cleaning schedules, alongside potentially more comprehensive disinfection of these areas, especially surfaces that come into contact with animals or animal fluids such as consultation tables and bins, may reduce the number of potentially harmful bacteria present.