Abstract
Five repeated cross-sectional serological surveys of 790 dairy cattle in 4 dairy herds between December 1985 and February 1987 provided an opportunity to study the changes in the seroprevalence of Haemophilus somnus across the 5 surveys and with respect to some demographic and disease variables. The demographic variables included were age (heifers or cows) and farm, representing two groups of herds (two herds in each group, located in the Central and Northern Valleys of California). The serological status of cattle as either negative or positive against H. somnus, Campylobacter fetus and Leptospira hardjo were determined with enzyme linked-immunosorbent assays. Logistic regression analysis was used to compute maximum likelihood estimates of adjusted odds ratios and their respective 95% confidence intervals. The baseline risk of being H. somnus seropositive in the cattle observed at first sampling did not vary significantly during the study period after adjustment for the effects of covariates. Only at first sampling were cows about twice as likely to be H. somnus seropositive than heifers. At samplings 1 and 3, but not at 2 and 5, being in the herds of the Central Valley appeared protective. In contrast, at sampling 4 the cattle in herds in the Central Valley were about 7 times more likely to be H. somnus seropositive. C. fetus-positive cattle were about 3 times more likely to be H. somnus seropositive at sampling 1 only. The relationship between H. somnus status and L. hardjo was not significant during the study period.