Incidence and persistence of classical swine fever in free-ranging wild boar (Sus scrofa)

Abstract

Although veterinary authorities aim to limit persistence of classical swine fever (CSF) in wild boar (Sus scrofa), to avoid potential transmission to pigs, factors influencing CSF transmission and persistence are not clearly understood. Here we analyse incidence and persistence in a CSF epidemic that occurred in the French Vosges Forest. Higher incidence was found in large forests compared to smaller isolated ones, being highest near the starting point of the epidemic, but poorly related to the local density. We hypothesize that the spatial and social structure of wild boar populations may be responsible for this variability of incidence over space. Persistence was highest near the starting point of the epidemic and where initial density was highest. We hypothesize that persistence was favoured by the abundance of young wild boar, itself encouraged by CSF. Our results allow us to propose management measures aimed at limiting CSF persistence.

Long-term monitoring of classical swine fever in wild boar (Sus scrofa sp.) using serological data

“…We observed the same pattern from virological data [33] and propose that, in the absence of a physical barrier within the Vosges forest, the apparent decrease of incidence over space may be due to differences in densities of wild boars, or to the low contact rate between wild boars of different social groups [33]. This observation is in accordance with predictions from mathematical models: when an epidemic occurs in a spatially structured population, several models predict that incidence should be the highest near the point of emergence and should decrease when the 40 S. Rossi et al distance to the point of emergence increases [4,29].…”

“…Thus, one must note that Figure 5 does not represent an exact prediction of incidence but rather a qualitative image of the evolution of incidence over space and time. Non-spatial factors, such as local population dynamics, structure of the habitat and local hunting habits, likely explain part of the residual variance in the data [33]. The HosmerLemeshow test did not show inadequacy in the goodness-of-fit and residuals revealed no particular structure, thus we had no indication of specific points to improve in the model.…”

“…Polynomial modelling supports that incidence peaked near the supposed point of emergence, where the first virus isolation had been performed [3,31,33]. This observation suggests that CSF emerged in that part of the forest.…”

“…Seroprevalence certainly underestimated the true incidence in juveniles, because they are particularly susceptible to CSFV, so that most of them would have died before developing antibodies [13,39]. We suppose that this underestimation mainly affected the results in 1992, when virus-induced mortality was the highest [33]. Consequently, we also supposed that we overestimated incidence in yearlings in 1993.…”

“…CSF is expected to persist in large populations [2,24]. It is also expected to persist in a nonhomogeneous way, depending on factors favouring CSF transmission, such as local population dynamics or management measures [2,24,33,37]. However, the detailed spatial pattern of CSF persistence has been poorly documented and we lack a theoretical understanding of the persistence of the disease [2].…”