Risk factors associated with Neospora caninum seropositivity in randomly sampled Canadian dairy cows and herds

Our objective was to determine cow- and herd-level risk factors associated with seropositivity for Neospora caninum in a large number of randomly selected Canadian dairy herds, controlling for important confounding variables and co-infections with bovine leukemia virus (BLV), bovine viral diarrhea virus (BVDV) and Mycobacterium avium subspecies paratuberculosis (MAP). Serum samples were obtained from 30 randomly selected cows, where available, in 240 herds using monthly milk testing, within 6 of 10 provinces, and these samples were tested for antibodies against BLV, MAP and N. caninum using commercially available ELISA test kits. Five unvaccinated cattle >6 months old from each herd were tested for antibodies to BVDV using virus neutralization. Most herd-level predictors were obtained through personal interviews with questionnaires administrated to each farm manager. A mixed logistic-regression model was built using N. caninum serostatus at the cow-level as the outcome variable, with herd as a random effect and province as a fixed effect. A BLV seropositive cow was 1.50 times more likely to be seropositive for N. caninum than a BLV-seronegative cow, and this was the only cow-level variable to remain in the final model. Regarding herd-level variables, with "no on-farm dogs" as the baseline, "presence of dogs but not known to eat placentas and/or fetuses" increased the odds of seropositivity for N. caninum by a factor of 1.66. For "presence of dogs known to eat placentas and/or fetuses", the odds ratio (OR) was 2.75, demonstrating a dose-response relationship. "Using embryo transfer" (OR=0.69), "asking for a BVDV-negative test before introducing an animal" (OR=0.30), "using monensin in dry cows" (OR=0.71), and "heifers having nose-to-nose contact with calves" (OR=0.73) were all dichotomous variables negatively associated with seropositivity for N. caninum. "Number of milk cows on the farm" (OR=0.99), and "area (acres) used for forage production" (OR=0.99) were continuous variables negatively associated with N. caninum seropositivity.