In this study, we assessed for the first time the prevalence of BVDV infections in all four indigenous wild ruminant species on the whole territory of Switzerland. Results provide a general picture of the current BVDV infection status of these species, which is necessary for the future planning and successful achievement of the implemented BVD eradication program in livestock. Although the number of investigated animals was too low to provide strong data at local level, the total sample size (> 400 animals tested per species) fulfilled criteria both for detection of “disease” (virus positive animals) and prevalence estimation at country level.
This study also provides valuable information on the occurrence of interspecific interactions among wild ruminants and between wild and domestic ruminants in an alpine environment. These quantitative data are necessary to understand the epidemiology of infectious diseases in natural habitats used by different animal species and to perform risk assessments. Data gathered by means of questionnaire surveys rely on observations of a number of voluntary participants and may be biased in various ways. Furthermore, exact and detailed information is nearly impossible to collect because 1) participants have to rely on memorized events to answer the questions, and 2) the response rate is expected to be negatively influenced by extensive question lists. However, this method represents a relatively easy and efficient way to obtain general information on a large geographical area over a long period of time.
The BVDV screening test gave more positive results than the confirmation test: only 42% percent (32/77) of the samples that were found positive by ELISA were confirmed as such by SNT. Two factors may account for this difference: First, it may be due to poor serum quality and cytotoxicity, which limits performance of the SNT and hampers result interpretation. Results that could not be validated by SNT were classified as non-interpretable, which may have lead to an underestimation of the final seroprevalence. Second, cross-reactions with other potential circulating pestivirus may have occurred in the ELISA. Because only one virus strain was used for the SNT, it was not possible to assess the occurrence of such cross-reactions
. Other studies have indeed shown a high circulation of pestivirus in wild ungulates populations, e.g. the Border disease virus (BDV) in ibex and chamois in the Italian Alps
[21, 22] or in chamois in France and Spain
[23–25], and it is likely that this virus also circulates in Swiss wildlife. The presence of other pestivirus than BVDV could also explain the presence of seropositive wild animals in communities without documented PI cattle (e.g. unit South-West).
Overall, we observed low seroprevalences in red deer, chamois and ibex and no antibodies were detected in roe deer. For red deer, the seroprevalence calculated in the present study (2.7%) is higher than previously reported in Switzerland (1.7%, 95% CI 0.46-4.25
) but the difference is not significant (p = 0.602). Our results are in line with a study from Austria (2%
), a country that started an eradication program for BVDV in livestock at regional level in 1997 and at national level in 2004
. Other European countries have reported low seroprevalences (0% to 5%) in red deer populations
Similarly, the situation in Swiss chamois is comparable to previous studies in Europe, which documented BVDV seroprevalences ranging from 3.4% to 5.5%
[28, 31, 34]. In ibex, seroprevalence has significantly decreased
, both at local level (from 3.9% to 0% in the unit South-East, p = 0.003) and at national level (from 4.3% to 1.8%, p = 0.021). In the unit South-West, a non-significant decrease was also noted (from 8.8% to 6.7%, p = 0.613). If this decline was a consequence of the implementation of the BVD eradication program in 2008, the prevalence is expected to continue to decrease in wild populations.
Concerning roe deer, previous reports have documented low seroprevalences in Germany (0% to 0.7%
[28, 35, 36]) and in France (0.7%
). However, in Germany higher prevalences reaching up to 9.8% have also been recorded, pointing at existing regional differences
[31–33]. A distinct BVDV-like strain was found in free-ranging roe deer in Germany, indicating that specific BVDV strains might circulate in this species
. In Scandinavia, antibodies against BVDV were detected in 12.3% of 635 Norwegian roe deer and the authors also concluded that this situation might be due to the circulation of a wild BVDV-like strain
We report here the detection of bovine-related viral RNA in a seropositive chamois kid. Given the fact that cattle harbouring the same strain were born in the region where this chamois was sampled, and according to the widespread circulation of virus in the cattle population prior to the eradication program, a cattle to chamois transmission is more likely than the contrary. To our knowledge, this is the first report of a detection of BVDV in a chamois and in an Alpine ungulate. There are two possible explanations for the simultaneous presence of virus and antibodies in this chamois kid. On the one hand, it may have had circulating maternal antibodies like it has been described in cattle calves
. However, the antibody titer in this chamois (1:861) was much higher than the maternal antibody titer predicted for 3 month-old calves (1:32)
. The chamois was > 3 months old and we expect that transfer and persistence of maternal antibodies is comparable in wild and domestic ruminants. On the other hand, it may have undergone an acute infection but the simultaneous occurrence of detectable antibodies and virus in a transiently infected animal is unlikely
. To distinguish between a PI animal and a transient infection, two consecutive analyses are necessary, a procedure that is not possible when examining hunted animals.
The facts that all BVDV seropositive animals were located within the Alps, and that the only virus positive wild ruminant was infected with a cattle strain, suggest a role of interactions on alpine pastures as a risk factor for BVDV infection, as already described for cattle infections
. Domestic PI animals may have been the source of the infections detected in wildlife.
It is now widely recognized that the epidemiology of diseases is often characterized by multiple host systems
, and interspecific interactions, especially between wild and domestic species, have received an increasing attention during the past decade. However, although the epidemiological significance of interactions at the wildlife-livestock interface is often incriminated or suspected (e.g.
[11, 25, 43]), such interactions are poorly documented in the internationally available scientific literature. Published studies on interactions between different animal species mostly focus on interferences or influences at the population level - such as mutualism, commensalism, competition and predation - while reports on behavioral and social interactions at the individual level, which are most relevant when talking about pathogen transmission, are scarce
. Furthermore, most studies on behavioral interactions between sympatric species focus on only few species, which is likely insufficient when addressing epidemiological questions in a natural environment.
Our data on frequency of interactions originate from a retrospective questionnaire survey and do therefore not present the accuracy of a prospective field study carried out by ethologists. However, they arise from a particularly long observation period (mean of 17 years per district) that also reflects the experience level of the participating game wardens, two factors that are expected to positively influence result reliability. Because the concerned wild species are most active at dawn or dusk
[45–48], it is likely that only a small part of the occurring interactions were observed, i.e., their reported occurrence is probably underestimated.
Compared to former questionnaire surveys and prospective observational studies on interactions between domestic sheep, ibex and chamois in the Swiss Alps
[14, 15, 49], we report as many or more observations of 1) physical contacts and short distance encounters between chamois and ibex, and between wild caprids and sheep; 2) mixed herds of chamois and ibex; and 3) encounters at salt licks between chamois and ibex. This supports the expected reliability of our questionnaire data and also shows that rarely observed events like physical contacts are more likely to be recorded by questionnaires than by direct observations on selected study sites. Reports on interactions between domestic goats and wild ruminants had not yet been documented, except for the occasional occurrence of hybrids between goats and ibex
. We show here that these interactions are not rare, particularly between goat and ibex.
Interactions among deer species (encounters at < 50 m between roe and fallow deer) were reported to be frequent in an Italian Mediterranean national park
. Here we also report frequent encounters of < 50 m and other types of interactions between roe deer and red deer. Interactions between red deer and ibex have already been observed before in the Swiss Alps
. However, we report for the first time observations of physical contacts between cervids (red deer) and caprids (chamois). Our data suggest that red deer interact more often with chamois than ibex, which can be explained by the larger overlap of deer and chamois habitat. Similarly, direct interactions between red deer and cattle have already been documented in the Alps
 but we newly report physical contacts between these two species. Also, we show that roe deer can get close to domestic species, especially cattle, with which almost all types of interactions were reported significantly more often than with small ruminant species. Since the questions on proximity and type of observed interactions were not linked with each other in our questionnaire, it is difficult to conclude from the obtained data which circumstances more likely favor close interspecific contacts. However, additional comments of the participating game wardens and common sense suggest that the simultaneous use of common food sources bear the highest risk for close interactions including physical contacts.
Wildlife feeding has never been performed by the hunting authorities in the canton of Graubünden and was officially abolished 25 years ago by the cantonal hunters’ association (G. Brosi, pers. comm.). However, a number of game wardens reported that intentional feeding of wildlife in wintertime is still common practice, and that interspecific interactions occur at feeding sites, especially between roe deer and red deer. Furthermore, both cervid species but especially red deer were observed near cattle farms - outside or inside livestock enclosures - during winter, profiting of livestock food resources like bale silage. In Switzerland, wildlife management is largely defined at cantonal level and official restrictions regarding wildlife feeding in Graubünden may not apply to other cantons. This means that active wildlife feeding may be even more common in other regions. An aggregation of animals of the same or different species either at intentional or non-intentional feeding places is a recognized risk factor for pathogen transmission
[53, 54] that needs to be taken into consideration in the frame of disease prevention or control programs.
The risk of pathogen transmission between sympatric species depends on the occurrence, frequency and intensity (i.e., proximity and duration) of interactions. Furthermore, two major transmission pathways have to be considered: 1) by direct contact, i.e., physical contact with an infected host or by contact with the latter’s infected discharges, including contaminated aerosols, feed, water or environment (= direct transmission); or 2) by indirect contact via a living or inanimate intermediate vehicle that transmits pathogens between infected and susceptible hosts, such as arthropod vectors (= indirect transmission)
[42, 55]. For BVDV, both direct and indirect transmissions are possible
 and we have shown that all kinds of interactions representing a risk for transmission, including physical contact, short distance encounters and use of the same food resources, are commonly observed between wild and domestic ruminants in the Swiss Alps. Overall, we documented a regular occurrence of interspecific interactions allowing virus transmission.
We showed that BVDV infections are significantly less frequent in wildlife than in Swiss cattle and that seropositive wild ruminants were mostly found in areas with PI cattle. Furthermore, the virus positive chamois was infected with a cattle strain, and seroprevalence in wildlife is apparently decreasing now that the eradication program in livestock is ongoing. It has been previously suggested that conditions to maintain certain pathogens are indeed more favorable in domestic livestock than in wild ruminants
: Swiss wildlife populations are much smaller than cattle and sheep populations; furthermore, in livestock intraspecific contacts and mixing of herds (for grazing, shows and markets) and movements throughout the country (commercial exchanges, alpine summering) are more frequent and intense than among wild populations, which is expected to favor pathogen maintenance.
All tested wild animals from the bioregion Jura were negative despite the presence of infected domestic livestock on pastures. Interactions between wild and domestic species are also known to occur in this region but may be less frequent than in the Alps due to the lower number of domestic ruminants on summer grazing pastures. However, data from the Alps suggest that spill-over is rare and our sample size does not allow excluding the infection of single individuals in the Jura.
Although interactions between cattle and roe deer are regularly observed, BVDV antibodies were not detected in this wild species. Earlier studies have shown that roe deer are susceptible to BVDV infection
[31–33] and that a distinct BVDV strain exists in roe deer populations from Germany
. However, due to the high variability of data obtained for this species across Europe, it seems difficult to draw further conclusions on the susceptibility of roe deer to BVDV cattle strains. The frequent interactions between wild and domestic caprids, and the occurrence of BDV in sheep and goats in Switzerland
, further underline the need to consider the occurrence of infections with other pestiviruses than BVDV in wild ruminants.