Toxoplasma gondii sero-positivity

Stensgaard, A. S., Sengupta, M. E., Chriel, M., Nielsen, S. T. & Petersen, H. H., 2022, Sero-prevalence and risk factors of Toxoplasma gondii infection in wild cervids in Denmark, International Journal for Parasitology: Parasites and Wildlife 17, pp. 288-294 : 291

3.2. Risk factors associated with Toxoplasma gondii sero-positivity

In the initial exploratory univariate mixed effects logistic regression models, implemented for each species of cervids individually, a significant increase in sero-positivity with age for red deer only (OR: 13.22; CI: 5.96–33.7) was observed ( Table 1). Risk factors associated with T. gondii sero-positivity in the initial exploratory, univariate analysis (all cervid species) were mainly individual level risk factors (age, sex and species), as well as the estate level factors “average annual minimum temperatures”, proportion of the land cover types (“artificial area” and “forest”) and fencing practices (yes/no) (see Table 2 for details).

After the backward elimination procedure (p <0.05 level) in the final multivariate, two-level hierarchical logistic regression model the following predictor variables were found to be significantly associated with T. gondii sero-positivity (yes, no): age-group (3 categories); species and fencing (yes/no) ( Table 2). The random effect of estate ID was small, accounting only for 5% (rho = 0.049) of the variance in the model, but significant (p = 0.051 in LR test of rho = 0), and hence retained in the final model. The odds of sero-positivity was 4.5 times higher in roe deer than fallow deer, and 3.0 times higher in red deer than in fallow deer. The prevalence was significantly different between the age groups (p <0.0001). The odds of sero-positivity in adults (3–5 years) was 10 times higher than in calves (<1 year), and 4 times higher than in yearlings (1–2 years). The only significant other risk factor associated with T. gondii sero-positivity in wild cervids in the multivariate analysis was fencing, with the highest exposure associated with deer from non-fenced hunting areas (OR = 2.21(1.05–4.99), p = 0.037). There was no interaction between the variables species and fencing (p = 0.749).