Use of survival analysis to compare cull rates between bovine leukemia virus seropositive and seronegative dairy cows

Bovine leukemia virus (BLV) infection and culling of cows in a commercial dairy herd were evaluated to determine whether a relation existed between the 2 factors. Cattle from the study population, a Holstein dairy herd consisting of approximately 400 milking cows, were tested for antibodies to BLV, using the agar gel immunodiffusion test, semiannually for 2 years, annually for 2 years, and when cattle were culled. Complete records of BLV test results were available for 849 (79%) of the 1,078 cattle that had at least 1 test during the study period. Using the Cox hazard model, the cull hazard rates (culls/cow-months) were greater for BLV seropositive cows than for seronegative cows > 36 months old. Hence, among older dairy cows, BLV-infected cows were culled prematurely, compared with uninfected cows.