Abstract
Little data exist on the levels of antimicrobial resistance from bacteria isolated from British sheep and beef cattle. The aim of this study was to investigate antimicrobial resistance patterns on sheep and beef farms in England and Wales using multiple interpretation methods. Fecal samples (n = 350) from sheep and beef cattle were collected from 35 farms. Disk diffusion antimicrobial susceptibility testing against ten antimicrobials was carried out for 1115 (699 sheep, 416 beef) β-glucuronidase-positive Escherichia coli isolates. Susceptibility was interpreted using clinical breakpoints, which determine clinically resistant bacteria, and epidemiological and livestock-specific cut-off values, which determine microbiological-resistant bacteria (non-wild type). Using livestock-specific cut-off values, a high frequency of wild type for all ten antimicrobials was observed in isolates from sheep (90%) and beef cattle (85%). Cluster analysis was performed to identify patterns in antimicrobial resistance. Interpretation of susceptibility using livestock-specific cut-off values showed a cluster of isolates that were non-wild type to cefotaxime and amoxicillin/clavulanic acid, whereas clinical breakpoints did not. A multilevel logistic regression model determined that tetracycline use on the farm and soil copper concentration were significantly associated with tetracycline non-wild type isolates. The results suggest that using human clinical breakpoints could lead to both the under-reporting and over-reporting of antimicrobial resistance in sheep and beef cattle.