Epidemiology of Neospora caninum in Canadian dairy farms

Neospora caninum was first recognized in dogs in Norway in 1984 and in 1988, a new protozoan species, N. caninum, was proposed under a new genus, Neospora. It is an apicomplexan protozoan, with dogs and coyotes as the definitive host.

The overall objective of this thesis was to acquire a more complete appreciation of the epidemiology of N. caninum infection in Canadian dairy farms, by answering important questions about the epidemiology of this disease. The epidemiology of neosporosis was investigated through four different studies and one extensive literature review.

The review covers the life cycle of the agent, its mechanisms of transmission, clinical signs, and tests for diagnosing the infection. Data on the prevalence of the infection in Canadian dairy and beef cattle are reviewed and briefly compared with estimates from other parts of the world. Most importantly for Canadian bovine practitioners, the impacts of the infection, risk factors for its occurrence and methods of control are also discussed.

A total of 6,662 blood samples were collected from cows on 240 dairy farms selected randomly from 6 provinces in Canada. Enrollment in a monthly, individual milk-testing regimen through a dairy herd improvement (DHI) program was a prerequisite for participation. Serum samples were stored at -20°C, and then tested for antibodies to N. caninum using an indirect ELISA.

The overall Canadian cow level prevalence was 11.9%, while for the maritime provinces it was 19.5% and for the western provinces it was 10.4%. The herd level, prevalence was 81.9% at the national level, and 86.7% and 80.0% for maritimes and western provinces, respectively. N. caninum infection was spatially clustered in both western and maritime Canadian provinces, occurring more commonly in Alberta in the western provinces and New Brunswick in the Maritimes.

A study was carried out to investigate the impact of serostatus for Neospora caninum (NC), bovine viral diarrhea virus (BVDV), bovine leukosis virus (BLV) and Mycobacterium avium subspecies paratuberculosis (MAP) and their possible interactions on reproductive efficiency (specifically, the ability to conceive and assumed fetal loss) in dairy cows. A Cox proportional hazards model with shared (herd-level) frailty was used to analyze the calving to conception interval data while controlling for herd-clustering effects. In this model, only BLV serostatus was associated with calving to conception intervals (CCI). Peak milk production before 73 days, SCC (linear score) and peak milk production after 73 days also had significant effects on the interval between calving and conception.

Logistic regression models of CCI greater than 200, 250, and 300 days were built, controlling for herd as a random effect. N caninum-seropositive cows had a 1.27 times higher risk of exhibiting a CCI exceeding 200 days, a 1.37 times higher risk of a CCI exceeding 250 days, and a 1.54 times higher risk of a CCI exceeding 300 days. BLV status also interacted with lactation numbers in the CC1200 model only, with older seropositive cows being less likely to have a CCI over 200 than 1st lactation seropositive cows. Neither BVDV nor MAP showed any significant effect in the models.

A logistic regression model of first service conception (FSC) revealed a significant interaction between N. caninum and BVDV infections. Increased peak milk production also reduced FSC. BLV and MAP had no significant impact on FSC.

Another study had the objective of determining important cow and herd level risk factors for seroprevalence for N. caninum in a population of randomly selected Canadian dairy cattle. 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. BLV serostatus was the only cow-level variable that remained in the model. Dogs being present on the farm, but not eating placentas and/or fetuses increased the odds of a farm having N. caninum by a factor of 1.54. If the dogs also ate placentas and/or fetuses, the odds ratio was 2.22. Other variables that remained (and appeared to be protective) in the final model included: "farmer asks for BVDV-negative exam before introduction of animal"; "number of milk cows on farm"; "dry cows receive Rumensin (monensin) in diet"; "embryo transfer used on farm"; "area of farm used for forage production per cow"; and "heifers have nose-to-nose contact with calves". These findings indicate management factors that could be considered to help control N. caninum seroprevalence and its related impacts.

An economic analysis that estimated the production losses (and their ranges) due to N. caninum in the Canadian dairy industry and the possible range of losses for individual Canadian dairy herds was conducted. For the Canadian dairy industry, annual losses attributable to N. caninum were estimated to be $1,838.76 per 100 cows ($18 per cow) (95% CI = $7 to $32 per cow). For herds (assumed to be 61 cow herds), losses were estimated to be $1,494.38 but there was a wide range of losses (95% CI = $37 to $5,309), primarily due to the variation in the within-herd prevalence.