Multivariate logistic analysis of repeated cross-sectional surveys of Campylobacter fetus in dairy cattle

Saeed Akhtar, Hans P. Riemann, Mark Thurmond, Thomas B Farver, Charles E. Franti

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


Five repeated cross-sectional serological surveys BLEED of 790 dairy cattle on four dairy herds between December 1985 and February 1987 provided an opportunity to study the changes in the seroprevalence of Campylobacter fetus (CF) across the five surveys and with respect to some demographic and disease variables. The demographic variables included were AGE (heifers or cows), FARM representing two groups of herds (two herds in each group, located in the Central and Northern Valleys of California). Disease variables CFETUS, HSOMNUS and LEPTO were the serological evidence of natural exposures to CF and Haemophilus sommus and presumed vaccinal titers against Leptospira hardjo, respectively. Enzyme-linked immunosorbent assay (ELISA) was used to define levels of CFETUS, HSOMNUS and LEPTO. Logistic regression analysis was used to compute maximum-likelihood estimates of adjusted odds ratios (OR) and their respective 95% confidence intervals (CI). The adjusted odds of being CF seropositive for cattle did not vary significantly at samplings 2-5 relative to sampling 1. The relative odds for prevalent CF seropositivity were stratified on BLEED based on the interaction terms in the final multivariate model. Cows and heifers had about the same adjusted odds of being CFETUS positive at BLEED 1-4 but at BLEED 5 the relative adjusted odds of being CFETUS positive forr cows compared with heifers were 3.55 (95% CI = 1.07, 11.73). Cattle from the two herds in Central Valley had almost the same likelihood of CF seropositivity as cattle from two herds in Northern Valley at BLEED 1, 2 and 3. However, at BLEED 4 and 5 cattle from herds in the Central Valley were protected compared with those from herds in the Northern Valley (adjusted OR = 0.02, 95% CI = 0.01, 0.08; adjusted OR = 0.19, 95% CI = 0.07, 0.55 respectively). HSOMNUS-positive cattle were about two times more likely to be CFETUS positive at sampling 1 (adjusted OR = 1.92, 95% CI = 1.20, 3.05). However, this association was not significant at samplings 2-5. Being LEPTO-positive had no impact on CF seroposivivity of cattle at BLEED 1, 2, 3 and 5 but at BLEED 4 the relative adjustedodds for LEPTO-positive cattle were 7.22 (95% CI = 1.97, 26.42). The possible biological mechanisms of the observed associations are discussed.

Original languageEnglish (US)
Pages (from-to)15-24
Number of pages10
JournalPreventive Veterinary Medicine
Issue number1-2
StatePublished - 1990

ASJC Scopus subject areas

  • Animal Science and Zoology
  • veterinary(all)


Dive into the research topics of 'Multivariate logistic analysis of repeated cross-sectional surveys of Campylobacter fetus in dairy cattle'. Together they form a unique fingerprint.

Cite this