The association between previous 305-day milk yield and disease in New York State dairy cows

Logistic regression was used to examine the association between milk yield and disease. We studied 8070 cows of second or higher parity from 25 New York State Holstein herds, calving between June 1990 and November 1993. Previous 305-d mature equivalent milk yield was used because most of the disorders studied occurred too early in lactation for current milk yield to be considered. Seven disorders were studied: retained placenta, metritis, ovarian cyst, milk fever, ketosis, abomasal displacement, and mastitis. A separate model was used for each disorder. Each model included, in additional to milk yield, parity, calving season, and heard. Results showed that higher milk yield was not a risk factor for any disease except mastitis. However, the association between higher previous milk yield and mastitis does not necessarily imply causation.