Herd-level relationship between antimicrobial use and presence or absence of antimicrobial resistance in gram-negative bovine mastitis pathogens on Canadian dairy farms

Concurrent data on antimicrobial use (AMU) and resistance are needed to contain antimicrobial resistance (AMR) in bacteria. The present study examined a herd-level association between AMU and AMR in Escherichia coli (n = 394) and Klebsiella species (n = 139) isolated from bovine intramammary infections and mastitis cases on 89 dairy farms in 4 regions of Canada [Alberta, Ontario, Québec, and Maritime Provinces (Prince Edward Island, Nova Scotia, and New Brunswick)]. Antimicrobial use data were collected using inventory of empty antimicrobial containers and antimicrobial drug use rate was calculated to quantify herd-level AMU. Minimum inhibitory concentrations (MIC) were determined using Sensititre National Antimicrobial Resistance Monitoring System (NARMS) gram-negative MIC plate (Trek Diagnostic Systems Inc., Cleveland, OH). Isolates were classified as susceptible, intermediate, or resistant. Intermediate and resistant category isolates were combined to form an AMR category, and multivariable logistic regression models were built to determine herd-level odds of AMR to tetracycline, ampicillin, cefoxitin, chloramphenicol, trimethoprim-sulfamethoxazole combination, sulfisoxazole, streptomycin and kanamycin in E. coli isolates. In the case of Klebsiella species isolates, logistic regression models were built for tetracycline and sulfisoxazole; however, no associations between AMU and AMR in Klebsiella species were observed. Ampicillin-intermediate or -resistant E. coli isolates were associated with herds that used intramammarily administered cloxacillin, penicillin-novobiocin combination, and cephapirin used for dry cow therapy [odds ratios (OR) = 26, 32, and 189, respectively], and intramammary ceftiofur administered for lactating cow therapy and systemically administered penicillin (OR = 162 and 2.7, respectively). Use of systemically administered penicillin on a dairy farm was associated with tetracycline and streptomycin-intermediate or -resistant E. coli isolates (OR = 5.6 and 2.8, respectively). Use of cephapirin and cloxacillin administered intramammarily for dry cow therapy was associated with increasing odds of having at least 1 kanamycin-intermediate or -resistant E. coli isolate at a farm (OR = 8.7 and 9.3, respectively). Use of systemically administered tetracycline and ceftiofur was associated with cefoxitin-intermediate or -resistant E. coli (OR = 0.13 and 0.16, respectively); however, the odds of a dairy herd having at least 1 cefoxitin-intermediate or -resistant E. coli isolate due to systemically administered ceftiofur increased with increasing average herd parity (OR = 3.1). Association between herd-level AMU and AMR in bovine mastitis coliforms was observed for certain antimicrobials. Differences in AMR between different barn types and geographical regions were not observed.