Epidemiological Study of Pestiviruses in South American Camelids in Switzerland
Corresponding author: Patrik Zanolari, Dr med vet FVH, Dip ECBHM, Clinic for Ruminants, Vetsuisse Faculty of Berne, Bremgartenstrasse 109a, P.O. Box 8466, 3001 Berne, Switzerland; e-mail: email@example.com.
Background: In the context of the ongoing eradication campaign for bovine viral diarrhea virus (BVDV) in cattle in Switzerland, the role of South American camelids (SAC) as a possible virus reservoir needed to be evaluated.
Objective: To assess and characterize the prevalence of pestivirus infections in SAC in Switzerland.
Animals: Serum samples collected from 348 animals (40 herds) in 2008 and from 248 animals (39 herds) in 2000 were examined for antibodies against pestiviruses and for the presence of BVDV viral RNA.
Methods: Cross-sectional study using stratified, representative herd sampling. An indirect BVDV-ELISA was used to analyze serum samples for pestivirus antibodies, and positive samples underwent a serum neutralization test (SNT). Real-time RT-PCR to detect pestiviral RNA was carried out in all animals from herds with at least 1 seropositive animal.
Results: In 2008, the overall prevalence of animals positive for antibodies (ELISA) and pestiviral RNA or was 5.75 and 0%, respectively. In 2000, the corresponding prevalences were 3.63 and 0%, respectively. The seroprevalences (SNT) for BVDV, border disease virus or undetermined pestiviruses were estimated to be 0, 1.73, and 4.02% in 2008, and 0.40, 1.21, and 2.02% in 2000, respectively.
Conclusions and Clinical Importance: At the present time, SAC appear to represent a negligible risk of re-infection for the BVDV eradication program in cattle in Switzerland.
border disease virus
bovine virus diarrhea virus
South American camelid
serum neutralization test
Pestivirus infection causes considerable economic loss in ruminants worldwide.1–4 The viruses of the genus Pestivirus, family Flaviviridae, which are pathogenic to cloven-hooved domestic animals, are divided into 6 well-characterized genotypes: bovine viral diarrhea virus types 1 and 2 (BVDV-1, -2), border disease viruses 1, 2, and 3 (BDV-1, -2, -3) and classical swine fever (hog cholera) virus (CSFV), and other less well-defined genotypes.2,5 Despite considerable genetic variability among these genotypes, there is antigenic relatedness among viruses of the entire genus with marked serological cross-reactions.6–8 Furthermore, pestiviruses readily cross species barriers.9–12 BVDV leads to substantial losses in cattle.13,14 BVDV-1 is distributed worldwide, whereas BVDV-2 has so far been found in only a few countries.15 BDV also is distributed worldwide, and it primarily affects sheep,4,10,16–18 although transmission of the virus from sheep to cattle and vice versa has been documented.9,19,20 Bovine viral diarrhea and BD seropositive South American camelids (SAC) have been documented.21–27
Cattle persistently infected (PI) with BVDV are viremic and permanently shed virus in all secretions and excretions. In contrast, the degree of viral shedding by transiently infected cattle is much lower and excretion is limited to a few days.28 Community pasturing during the summer months, animal shows, and purchasing animals that are pregnant with a PI fetus are important risk factors for the transmission of the virus within a population. In SAC, a similar mechanism of intrauterine infection with BVDV as observed in cattle has been proposed.27 The mean duration of gestation is 345 days in SAC and 284 days in cattle.29 In the latter species, infection of the fetus with BVDV between days 40 and 125 of gestation can lead to PI offspring.15 Based on the assumption that the ontogenesis of the immune system of SAC is analogous to that of cattle, a fetal infection before approximately 145 days of gestation in the former would be expected to produce PI offspring.27 Like PI cattle, PI SAC also may constitute a major potential source of pestivirus transmission to other susceptible farm animals including cattle.
In 2008, a BVDV eradication program based on detection of PI animals by screening the whole cattle population for the presence of virus was initiated in Switzerland.30,31 The initial seroprevalence was estimated at 58% for first-calf heifers and up to 80% for pluriparious cattle,1 the prevalence of PI animals was estimated to be 0.80%.13 There is growing concern that once eradication of BVDV is complete in the cattle population, BVDV reservoirs in other species may gain importance as a potential source of reinfection for cattle.5,32
The goal of this study was to determine the prevalence of pestivirus infection in SAC and to investigate the risk of transmission of these viruses between cattle and SAC. The underlying question was whether SAC could hamper the eradication of BVDV from the cattle population by acting as a potential reservoir for the virus.
Materials and Methods
Nearly all SAC owners in Switzerland are members of 1 or both Swiss SAC breeding organizations (Swiss Association of Llama and Alpaca breeders, Swiss Alpaca Association). The address lists of these 2 breeding organizations were merged with addresses of SAC owners obtained from the Swiss Extension and Health Service for Small Ruminants and Camelids to obtain a combined list of all registered SAC owners in Switzerland. All 350 SAC owners on this list were contacted by letter and invited to participate in the study. One hundred (28.6%) were willing to participate in the study. There was no difference in the geographic distribution of farms that volunteered to participate and farms that did not respond. Of the 100 respondents, a stratified random sample of 40 SAC owners was selected to participate in the study. The sample was stratified by geographical region (canton) to ensure that the lowland and mountain regions and the French, German, and Italian speaking areas of Switzerland were equally represented in the sample. Herd size varied from 2 to 145 animals. In small farms with up to 26 animals, all animals of the herd were sampled. In larger farms, 26 animals were randomly selected for sampling. This sample size was sufficient to detect at least 1 animal seropositive for a pestivirus in the herd with at least 95% confidence if the seroprevalence within the herd was at least 10%. Samples were collected from 348 animals. This represented approximately 10% of the Swiss SAC population, based on 2,884 animals registered with the Swiss Association of Llama and Alpaca breeders at that time which includes approximately 75% of the entire population (Steiner, personal communication). The SAC included in this study consisted of 184 alpacas (52.9%) and 164 llamas (47.1%); 194 were females (55.7%) and 154 were males (44.3%), of which 73 (47.4%) were castrated. The animals ranged in age from 9 days to 17 years. In the majority of animals, the age was recorded only as “adult” because the exact date of birth was not known. Experimental protocols followed the current Swiss Act on Animal Protection and were approved by the Swiss Committee for Animal Care and Protection.
A detailed questionnaire was completed for each herd with the owner's support. The variables of interest related to individual animals as well as to the herd and comprised the type of farm (size, other animals species with a focus on contact with cattle), animal movements, general management, feeding, prophylactic health measures, disease incidence, and BVD disease awareness.
Jugular blood samples were collected into plastic tubesa from 348 animals during the summer of 2008 just before the start of the BVDV eradication program. Between 2 and 26 animals were sampled in each of the 40 farms. The samples were cooled to 4°C and centrifuged at 1,400 g on the same day, and serum was stored at −20°C until analysis.
An additional 248 frozen serum samples from 39 SAC herds were available for analysis from a previous study performed in 2000.33 The animals of that study were representative of the Swiss SAC population at that time. Two of these farms also were involved in the 2nd study.
Detection of Pestiviral Antibodies and RNA
An indirect pestiviral ELISA was used to examine the 596 serum samples for BVDV antibodies. ELISA microtiter plates were coated with BVDV antigen derived from cell cultures that had been infected with the R1935/72 strain (Oregon C24V, Sub-Genotyp BVDV-1a). Sera at a 1 : 10 dilution then were added to the plates. This dilution promotes the binding of antibodies directed to the well-conserved nonstructural NS23 protein,34 allowing detection of a wide range of pestiviruses.35–38 Protein G peroxidase was used as the conjugate.39 Samples that had positive test results subsequently underwent a serum neutralization test (SNT) that used a strain of BVDV-1a (R1935/72) and a strain of BDV-1 (Moredun) as challenge viruses.40,41 SNT titer ratios of a given serum sample in the 2 assays of ≥4 were considered positive.42 A higher titer with 1 of the challenge viruses (BVDV or BDV) compared with the other 1 was considered to indicate that a virus of this genotype was the source of infection.6,7,38,43 In case of titer ratios <4, the source of infection was considered to be a virus of another genotype than BVDV-1 and BDV-1 used in the SNT (ie, an undetermined pestivirus [UPD]).44 Finally, if a herd had at least 1 animal positive in the SNT, all serum samples from that herd underwent real-time RT-PCR to test for the presence of BVDV RNA.2,44–47 Qiagen BioRobotb was used for RNA extraction and the cador BVDV RT-PCR Kitc for real-time RT-PCR, following the manufacturer's instructions using the ABI 7300 Real Time PCR Systems.d
Sample size and statistical power of the study were determined by commercial software packages.e,48 The sample size was calculated based on a power of 0.80 and on a confidence level for detection of seropositive animals of 0.95. The sample size of the 2008 study was sufficient to detect at least 1 seropositive animal assuming a BVDV seroprevalence of 1% and a population size of 3,600 animals. The sample size was large enough to estimate the seroprevalence of any pestivirus infection with an accuracy of ±5%. The samples collected in 2000 were representative of the Swiss SAC population at that time.33 Provided that the SAC population was free from BVDV in 2000 (seroprevalence ≤ 0.1%), the sample size was sufficient to reliably detect an increase in seroprevalence of 3.5% between 2000 and 2008. Pestivirus seroprevalences and prevalence of BVDV were calculated by commercial software packagese,48 with a confidence interval of 0.95, after correction for the sensitivity of the ELISA by the method described by Rogan and Gladen.49 Based on the results of an earlier study in SAC,44 a sensitivity and specificity of the ELISA of 96 and 97% were used, respectively. Because PCR and SNT are considered to be the gold standard methods for the detection of pestiviral RNA and specific antibodies, respectively, their sensitivities and specificities are difficult to assess accurately; however, they can be considered to be very high.36,50–58 Therefore, sensitivity and specificity values of 99.9% were used to calculate seroprevalences and BVDV prevalence.
The maximal prevalence of BVDV in the Swiss SAC population was determined by commercial software.59 Differences between prevalences in 2000 and 2008 and between animals with and without potential risk factors were statistically evaluated by Fisher's exact test. The risk factors evaluated included purchase of animals from other herds in Switzerland, import of animals, presence of other cloven-hooved animals particularly cattle and small ruminants on the farm, cattle or small ruminant farms in the neighborhood, and other potential contacts to ruminants. Association of pestivirus infection in SAC farms with the presence of PI cattle in the neighborhood was evaluated on the basis of the postal codes of PI cases detected and notified in the frame of the BVDV eradication campaign.
Twenty-eight of the 348 serum samples collected in 2008 were positive in the pestiviral ELISA. Twenty of these also were positive in the SNT for BVDV reflecting an overall seroprevalence of BVDV in SAC of 5.75%. Comparative serum neutralization using the 2 test viruses yielded a seroprevalence of 0% for BVDV, 1.73% for BDV, and 4.02% for UPV (Table 1). Eleven of the 248 serum samples collected in 2000 were positive in the pestiviral ELISA. Nine of these were positive in the SNT for BVDV, which corresponded to an overall seroprevalence of 3.63%. According to the ratio of SNT titers for BVDV-1 and BDV-1, the prevalences were 0.40% for BVDV, 1.21% for BDV, and 2.02% for UPV, respectively. The 2 ELISA-positive samples reacting negatively by SNT were considered to be false-positives. There were 12 herds in 2008 and 8 herds in 2000 with seropositive animals, corresponding to herd prevalences of 30.0% in 2008 (according to SNT titer ratios 0% for BVDV, 10% for BDV, and 27.5% for UPV) and 20.5% in 2000 (according to SNT titer ratios 2.6% for BVDV, 7.7% for BDV, and 12.8% for UPV), respectively (Table 2). There was no statistically significant difference between 2008 and 2000 with respect to seroprevalence values. The 2 herds that were included in both studies had positive animals in 2008 but not in 2000.
Table 1. Individual seroprevalence for pestiviruses and estimated seroprevalences for bovine viral diarrhea virus (BVDV), border disease virus (BDV), and undetermined pestiviruses (UPV) in South American camelids in Switzerland.
Table 2. Herd seroprevalence for pestiviruses and estimated seroprevalences for bovine viral diarrhea virus (BVDV), border disease virus (BDV), and undetermined pestiviruses (UPV) in South American camelids in Switzerland.
Based on information derived from the questionnaires, all 40 herds participating in the 2008 study had acquired SAC from other herds in Switzerland at least once, and 10 had imported animals from foreign countries including Chile (n = 5), Peru (n = 3), Great Britain (n = 3), Canada (n = 1), Germany (n = 1), and France (n = 1). Temporary housing of animals from other herds for breeding purposes was practiced by 25 herds (62.5%), and animals from 26 herds (65.0%) potentially could have had direct contact with cattle of neighboring farms. In 6 herds (15.0%), the SAC had contact with sheep and goats only, and animals from 14 herds (35.0%) regularly attended shows. Animals from 15 herds (37.5%) were used for trekking. These animals potentially came in contact with other animal species and were possibly also exposed to indirect transmission (of BVDV) by people or fomites. Transport of animals in trailers regularly took place in 29 herds (72.5%).
Of all the risk factors, only the import of animals from foreign countries had a significant effect on seroprevalence (P= .04) with an odds ratio (OR) of 5.44 (ie, farms that had imported animals were 5.44 times more likely to harbor seropositive SAC). There was no association between the occurrence of PI cattle in the neighborhood and seropositive SAC. Forty-seven samples from the 8 seropositive herds of 2000 and 136 samples from the 12 seropositive herds of 2008 that underwent RT-PCR all were negative, indicating 0% prevalence of pestivirus carriers. Maximal seroprevalences of 8.38% for pestiviruses (0.77% for BVDV, 3.35% for BDV, and 6.27% for UPV) were calculated and a maximum prevalence of virus-positive animals in the SAC population of 2.05% was calculated. For the samples collected in 2000, the corresponding calculated maximum seroprevalence for pestiviruses was 6.20% (1.76% for BVDV, 3.00% for BDV, and 4.10% for UPV) and the maximum prevalence of virus-positive animals in the SAC population was 6.02%.
Examination of serum samples from SAC in the year 2008 indicated an overall pestivirus seroprevalence of 5.75%. Based on antibody differentiation using cross-neutralization titer ratios, seroprevalences of 0% for BVDV, 1.73% for BDV, and 4.0% for UPV were calculated. The prevalence of BVDV carriers was 0%. These data are likely to be good estimates of the prevalence in the SAC population in Switzerland because nearly 10% of the population was sampled, and all regions of the country and all types of farms from very small to large herds were represented. SAC owners are not concerned about their BVDV status and do not test for this disease, it can therefore be assumed that owners of BVDV negative and positive animals would have been equally likely to be willing to participate in this study. Nevertheless, nonresponse bias cannot be excluded because less than one third of the SAC owners were willing to participate, and it was not possible to compare farm characteristics such as herd size and management of participating farms with the general SAC population. Prevalence estimates were not corrected for the fact that the sampled animals were clustered in herds. Instead, seroprevalence estimates were given separately for the individual animal and herd level. There was only weak clustering of seropositive animals within herds. On average, 3 animals per positive herd tested positive, so clustering did not bias the prevalence estimate substantially.
SACs belong to the suborder Tylopoda within the order Artiodactyla, the members of which are all susceptible to pestiviruses, including BVDV.60 There have been several studies investigating BVDV seroprevalence in SAC in North and South America with results ranging from 2.0 to 25.4%.22–24 A study from the USA, which involved 270 llamas from 21 herds, indicated a seroprevalence of 4.4%,22 and, in a study from Argentina involving 239 llamas, 8 (3%) were seropositive.23 The herd seroprevalence in a study that involved 323 alpacas from 63 herds in the United States was 25.4%.24 Our study yielded a similar herd seroprevalence for pestiviruses, but the application of cross-neutralization-based titer ratios allowed us to estimate more specifically the BVDV-induced herd seroprevalence, which was 0%. Similarly, the overall individual pestiviral seroprevalence of 5.75% found in this study was comparable to that found in other reports, whereas the estimation of the BVDV-induced individual seroprevalence was nil. Although only overall seroprevalences can be compared in different studies if cross-neutralization for classification of the source of infection is not applied, the present results clearly show the advantage of the procedure for purposes of differential risk assessment.
To date, there have been only 2 case reports of SAC PI with BVDV.25,26 A study from the United States determined a herd prevalence of BVDV-positive animals of 6.3% (4 of 63 herds) with a predominance of BVDV genotypes 1a and 1b.24 No PI animals were detected in the present study. In South and North America, herds tend to be larger than in Switzerland, which can be associated with increased animal movement and thus, increased risk of contact with BVDV. In South America, alpacas and llamas are commonly pastured together with cattle, a practice that also increases the risk of exposure to BVDV.23
According to the literature, antibodies to BDV are found mostly in sheep, but also can be detected in goats.4,17,18,61 In a recent study in Switzerland, an individual seroprevalence of 9% was observed in the individual sheep population,44 whereas in a previous study the seroprevalences at individual and herd levels were 20 and 67%, respectively.3 Despite these relatively high seroprevalence values, reports on sheep PI with BDV are rare.17,18 BDV has been isolated from cattle,19,20,62 and seroprevalence values of 0.9 and 11.0% have been reported in 2 studies in SAC.21,44 The sample analyzed in Danuser et al44 was only representative for sheep, but not for goats and SAC. Apart from the methods used, there is no crossover of samples and of data.
The relatively large proportion of infections that could not be attributed to either BVDV-1 or BDV-1 viral genotypes (UPV) in the present study indicated the occurrence of pestiviruses from genotypes other than BVDV-1 and BDV-1 used in SNT for cross-neutralization. The use of viral strains of other genotypes as the 2 challenge viruses, although limited due to technical reasons, might help decrease the proportion of undetermined infection sources substantially.8,38,43 However, the presence of viruses of different subgenotypes (eg, BVDV-1b) should not influence genotype differentiation in cross-neutralization assays.6,7 In Switzerland, subgenotype 1e predominates over the subgenotypes 1h, 1k, and 1b.63
Like cattle, SAC may be infected with BVDV via transiently or PI animals. Even considering the calculated maximum seroprevalence in Switzerland, the likelihood of cattle as well as SAC becoming transiently infected with BVDV after successful completion of the BVDV eradication program, which is projected for the year 2010, is very small. According to the present study, PI SAC which would represent the main source of BVDV infection are not present in Switzerland. The prevalence of newborn PI calves was 0.43% in September 2009 (http://www.admin.bvet.ch). On the basis of the calculated maximum seroprevalence of BVDV in SAC of 0.77% in the present study and considering that cattle, the principal host of BVDV, had a prevalence of BVDV antibody of 58–80% before eradication, indicating broad distribution of the virus in the bovine population, we conclude that SAC have a very limited susceptibility to BVDV. In an experimental study, 4 of 4 seronegative female llamas that were 67–102 days pregnant seroconverted after infection with BVDV, but none of their crias was PI.58 However, recent reports of SAC PI with BVDV have challenged the general assumption that SAC are relatively resistant to BVDV infection and that cattle are the sole source of infection.24–27,64,65 These reports and a recent study that analyzed various BVDV genotypes isolated from SAC in North America suggest that certain BVDV types may have adapted to SAC and be directly transmitted among alpacas and llamas.66 However, the low seroprevalence found in the present study does not indicate the circulation of easily transmissible BVDV strains within the SAC population in Switzerland. Thus, the risk of transmission of BVDV among SAC and from SAC to cattle can be considered to be very small in Switzerland. Furthermore, BVDV has limited persistence in the environment,67 being susceptible to pH fluctuations, high temperatures, and UV irradiation. It is readily inactivated by many disinfectants, which further reduces the likelihood of indirect transmission of the virus by people or fomites.28,68 Although cattle are highly susceptible to BVDV, their susceptibility to other pestiviruses has not been assessed in detail.4,9,10,17,61,68
In conclusion, the results of the present study indicate that SAC represent a very low risk of reinfection of cattle with BVDV, at least under the present conditions in Switzerland, representing a negligible challenge to the ongoing BVDV eradication program in cattle. However, in view of reports of relatively high prevalences of virus carriers in SAC in other parts of the world, it is advisable that countries with successfully completed or ongoing BVDV eradication programs limit the import of SAC to animals with negative BVDV test results to minimize the risk of reintroducing the virus into a naive population.
aMonovette, Sarstedt, Nümbrecht, Germany
bQIAamp DNA Blood BioRobot MDxKit Handbook; Qiagen AG, Hombrechtikon, Switzerland
cQiagen AG, Hombrechtikon, Switzerland
dApplied Biosystems Inc, Foster City, CA
eWinEpiscope 2.0, Software for Quantitative Veterinary Epidemiology, Facultad de Veterinaria, Zaragoza, Spain, Agricultural University Wageningen, the Netherlands, University of Edinburgh, UK
The authors thank the Swiss Llama and Alpaca breeders, and Libby Nebel, Elisabeth Santschi, and Daniel Kunz from the Institute of Veterinary Virology for their assistance. The project was financed in part by the Federal Veterinary Office (project number 1.08.20).