Prevalence and risk factors for extended-spectrum β-lactamase- and AmpC-producing Escherichia coli in dairy farms.

A cross-sectional study was conducted to evaluate the prevalence of extended-spectrum β-lactamase (ESBL)- and plasmid-mediated AmpC-producing Escherichia coli and associated risk factors in dairy herds. One hundred dairy herds were randomly selected and sampled to study the presence of ESBL- and AmpC-producing E. coli in slurry samples. The sensitivity of testing slurry samples for ESBL/AmpC herd status is less than 100%, especially for detecting herds with a low ESBL/AmpC prevalence. Therefore, whereas herds that tested positive for ESBL/AmpC-producing E. coli in slurry were defined as positive herds, herds with negative slurry samples were defined as unsuspected. Isolates of ESBL- and AmpC-producing E. coli were further characterized by detection and typing of their ESBL/AmpC gene. At the initial sampling, a comprehensive questionnaire was conducted at the participating farms. The farmers were asked questions about management practices potentially associated with the ESBL/AMPC herd status. Also, data on antimicrobial purchases during 2011 were acquired to evaluate whether the animal-defined daily dose of antimicrobials per year at farm level was associated with the ESBL/AmpC herd status. Multivariable logistic regression models were used to determine the association between management practices and the ESBL/AmpC herd status. Six months after the initial slurry sampling, 10 positive herds and 10 herds that had an unsuspected ESBL/AmpC herd status during the first visit were resampled. At each farm, slurry samples and feces from 24 individual cows were collected to evaluate within herd dynamics. During the first sampling, ESBL/AmpC-producing E. coli were isolated from the slurry samples collected at 41% of the herds. In total, 37 isolates were further characterized, revealing 7 different ESBL genes (blaCTX-M-1, -2, -14, -15, -32, -55 and blaTEM-52), 1 plasmid-encoded AmpC gene (blaCMY-2), and 1 chromosomally encoded ampC gene (ampC type 3). The total animal-defined daily dose of antimicrobials per year at farm level was not significantly different between ESBL/AmpC-positive and unsuspected dairy herds. The use of third- and fourth-generation cephalosporins, however, was found to be associated with ESBL/AmpC status, with higher use of these antimicrobials resulting in a significant higher odds to be ESBL/AmpC-positive. Management factors that were associated with a higher odds of being ESBL/AmpC-positive were treatment of all cases of clinical mastitis with antimicrobials, a higher proportion of calves treated with antimicrobials, not applying teat sealants in all cows at dry off, and the use of a floor scraper. This last association, however, was considered a methodological effect rather than a true risk factor. On 5 of the 10 initially positive farms, no ESBL/AmpC-producing E. coli were cultured from the slurry or any of the individual cow samples collected during the second sampling. In 4 of the initially unsuspected farms, slurry or individual cow samples tested positive during the second sampling. In conclusion, ESBL/AmpC could frequently be cultured from slurry samples collected from Dutch dairy farms and the ESBL/AmpC genes carried by the isolates were consistent with those reported earlier. The use of third- and fourth-generation cephalosporins appeared to be associated the ESBL/AmpC herd status.