Given the difficulty of measuring pathogen transmission in wildlife, epidemiological studies frequently rely on cross-sectional seroprevalence. However, seropositivity indicates only exposure to a pathogen at an unknown time. By allowing to obtain repeated test results from individuals sampled multiple times over an extended period, longitudinal data help reduce this uncertainty. We used capture-mark-recapture data on bank vole (Myodes glareolus) individuals collected at four sites over ten years in northeastern France to investigate the impact of environmental variables on seroprevalence and incidence of Puumala orthohantavirus (PUUV). PUUV causes a chronic infection without apparent symptoms, that may however impair survival of its rodent host in the wild. Viral transmission between rodents may occur through direct contact or via the environment. Principal component analysis was used to deal with multicollinearity among environmental variables. Incidence and seroprevalence were investigated with either generalized estimating equations or Poisson regression models depending on the number of observations for each season. In spring, only the factor site was found to be significant for seroprevalence, while a principal component including meteorological conditions of the previous winter and the normalized difference vegetation index (NDVI) of both the previous winter and spring had a significant effect on incidence. In autumn, only the factor site was significant for incidence, while two principal components, including either the meteorological conditions of the autumn and previous spring or NDVI of the autumn significantly affected seroprevalence. We discuss these results in light of the particular demography of small mammals. We encourage other researchers to investigate the relationships between demographic parameters of wild host populations and the environment, by using both incidence and seroprevalence.
Read full abstract