The use of loglinear model to evaluate factors associated with sero-prevalence of Campylobacter fetus in dairy cattle.
Campylobacter fetus is one of the leading infections causing bovine infertility and abortion. In a cross-sectional study, the associations of parity, herd, serological status against Leptospira hardjo and Haemophilus somnus with the sero-prevalence of Campylobacter fetus in dairy cattle were evaluated. Four hundred dairy cows from three herds were included in the study. Blood samples from the cows were collected in December 1986. An enzyme-linked immunosorbant assay (ELISA) was used to determine positive or negative status of cows against Campylobacter fetus , Haemophilus somnus and Leptospira hardjo antigens. Loglinear analysis was used to describe the relationship of four categorical variables with the sero-prevalence of Campylobacter fetus . The final loglinear model was interpreted as a logit model. The final loglinear model included a three-way interaction among Campylobacter fetus , Leptospira hardjo and herd and a two-way interaction between Campylobacter fetus and parity. The adjusted odds of being a positive Campylobacter fetus reading were highest for cows with parity >6 (e.g. 27.8, 95% confidence interval (CI) = 12.9, 60.1); it was next highest for cows in parity <4 (e.g. 18.0, 95% CI = 13.5, 24.2); and it was lowest for cows in parity 4 to 6 (e.g. 13.3, 95% CI 7.3, 24.4). The herd management practice that was associated with the carrier state in early and late parities is identified in the study, and the biological implication of a three-way interaction among Campylobacter fetus , Leptospira hardjo and herd are discussed.