Evaluation of bulk tank milk PCR and bulk tank milk modified ELISA tests for the detection of paratuberculosis at the herd level in goat and sheep dairies in Ontario, Canada

Early identification of dairy goat herds and dairy sheep flocks infected with Mycobacterium avium ssp. paratuberculosis is important for controlling this infection and minimizing economic losses. The objective of this study was to evaluate 2 bulk tank milk (BTM) paratuberculosis tests (PCR and modified ELISA) as potential herd-level tests. These tests were compared with the results obtained from testing 20 randomly selected lactating animals per farm (>2 yr) with an individual animal test (fecal culture, fecal PCR, serum ELISA, and milk ELISA). The study was conducted using 29 dairy goat herds and 21 dairy sheep flocks in Ontario, Canada, visited between October 2010 and August 2011. The sensitivity of the BTM PCR was poor in both the dairy goat herds (0.0%) and dairy sheep flocks (25.0%), but exhibited 100% specificity in both species. In comparison, the BTM modified ELISA demonstrated higher sensitivity. In goats, sensitivity ranged from 33.3 to 34.8% when fecal culture and PCR were the reference tests, respectively (specificities were both 100%), and 71.4 to 87.5% when the milk and serum ELISA, respectively, were the reference tests (specificities were 86.4 and 95.2%). The BTM modified ELISA in dairy sheep demonstrated comparable sensitivities, but lower specificities. When fecal culture and PCR were the reference test, sensitivities were 50.0 and 46.7%, respectively (specificities were 77.8 and 83.3%). The sensitivities when the milk and serum ELISA were the reference tests were 87.5 and 72.7%, respectively (specificities were 92.3 and 100%). Fecal PCR was the only individual animal test to identify significantly more farms as positive than the BTM PCR and modified ELISA test in both species. Therefore, whereas the BTM modified ELISA may provide an organization or control program with a high level of confidence that a BTM-positive farm is actually positive (high positive predictive value), if a producer wishes to increase the odds that a positive farm will test positive, so as not to miss an infection, then sampling and testing 20 animals with fecal PCR will better meet that objective.