Implementing a probabilistic definition of freedom from infection to facilitate trade of livestock: putting theory into praxis for the example of bovine herpes virus-1.

International trade of livestock and livestock products poses a significant potential threat for spread of diseases, and importing countries therefore often require that imported animals and products are free from certain pathogens. However, absolute freedom from infection cannot be documented, since all test protocols are imperfect and can lead to false-negative results. It is possible instead to estimate the "probability of freedom from infection" and its opposite, the probability of infection despite having a negative test result. These probabilities can be estimated based on a pre-defined target prevalence, known surveillance efforts in the target population and known test characteristics of any pre-export test. Here, calculations are demonstrated using the example of bovine herpes virus-1 (BoHV-1). In a population that recently became free of BoHV-1 without using vaccination, the probability of being infected of an animal randomly selected for trade is 800 per 1 million and this probability is reduced to 64 (95% probability interval PI 6-161) per 1 million when this animal is tested negatively prior to export with a gB-ELISA. In a population that recently became free of BoHV-1 using vaccination, the probability of being infected of an animal randomly selected for trade is 200 per 1 million, and this probability can be reduced to 63 (95% PI 42-87) when this animal is tested negatively prior to export with a gE-ELISA. Similar estimations can be made on a herd level when assumptions are made about the herd size and the intensity of the surveillance efforts. Subsequently, the overall probability for an importing country of importing at least 1 infected animal can be assessed by taking into account the trade volume. Definition of the acceptable level of risk, including the probability of false-negative results to occur, is part of risk management. Internationally harmonized target prevalence levels for the declaration of freedom from infection from selected pathogens provide a significant contribution to the facilitation of international trade of livestock and livestock products by allowing exporting countries to design tailor-made output-based surveillance programs, while providing equivalent guarantees regarding the probability of freedom from infection of the population. Combining this with an approach to assess the overall probability of introducing at least 1 infected animal into an importing country during a defined time interval will help importing countries to achieve their desired level of acceptable risk and will help to assess the equivalence of animal health and food safety standards between trading partners.

Evaluation of an indirect ELISA for detection of antibodies in bulk milk against bluetongue virus infections in the Netherlands.

After the introduction of bluetongue virus serotype 8 (BTV-8) in western Europe in 2006, an indirect ELISA for detection of serogroup-specific antibodies against BTV in serum samples was validated for individual milk samples by the Central Veterinary Institute and the Animal Health Service in the Netherlands (Kramps et al., 2008). In order to develop a cost-effective monitoring tool, we now have evaluated this ELISA also for use in bulk milk. Therefore, bulk milk samples and individual milk samples were collected from 92 herds in the affected southern region in the Netherlands in 2007, before the start of the vaccination campaign. In addition, bulk milk samples collected from 88 herds before the bluetongue introduction in 2006 ("historically negative" samples) have been tested. With these results ROC analyses were performed and herd specificity and herd sensitivity of the bulk milk ELISA were estimated. All "historically negative" bulk milk samples were negative in the ELISA, with a mean S/P ratio of 10 ± 0.8%. The herd sensitivity and herd specificity of the ELISA in bulk milk samples depend on the cut-off that is chosen. In order to detect a within-herd-prevalence of 1%, the optimal cut-off S/P ratio 13% was found. A few herds with one or two milk-positive animals would then be missed. The specificity will be 100%. A within-herd-prevalence of 10% can be detected with 100% sensitivity at a cut-off S/P ratio of 96%. In conclusion, the indirect ELISA in bulk milk samples is a very specific and sensitive test which can be implemented in monitoring and surveillance systems in unvaccinated populations.

Public health risk analysis of European bat lyssavirus infection in The Netherlands.

We present the frequency and the nature of contact incidents of the Serotine bat, Eptesicus serotinus, with humans and with companion animals (specifically cats and dogs), in The Netherlands between 2000 and 2005. Out of 17 bats in bite contact with humans, five tested positive for European bat lyssavirus (EBLV) type 1a. Cats had the most numerous contacts with bats (49 times) but a relatively low number of these bats were EBLV positive (six times). We estimated that the average incidence of human bat rabies infection might be between once per year and once per 700 years, depending mainly on the number of infectious viral particles in bat saliva. The risk of bat rabies is higher between April and October, and in the northern half of the country. This is the first study in Europe describing the risk of human bat rabies after bat contact incidents.

Enzyme-linked immunosorbent assay using a virus type-specific peptide based on a subdomain of envelope protein E(rns) for serologic diagnosis of pestivirus infections in swine.

Peptides deduced from the C-terminal end (residues 191 to 227) of pestivirus envelope protein E(rns) were used to develop enzyme-linked immunosorbent assays (ELISAs) to measure specifically antibodies against different types of pestiviruses. The choice of the peptide was based on the modular structure of the E(rns) protein, and the peptide was selected for its probable independent folding and good exposure, which would make it a good candidate for an antigenic peptide to be used in a diagnostic test. A solid-phase peptide ELISA which was cross-reactive for several types of pestivirus antibodies and which can be used for the general detection of pestivirus antibodies was developed. To identify type-specific pestivirus antibodies, a liquid-phase peptide ELISA, with a labeled, specific classical swine fever virus (CSFV) peptide and an unlabeled bovine viral diarrhea virus peptide to block cross-reactivity, was developed. Specificity and sensitivity of the liquid-phase peptide ELISA for CSFV were 98 and 100%, respectively. Because the peptide is a fragment of the E(rns) protein, it can be used to differentiate between infected and vaccinated animals when a vaccine based on the E2 protein, which is another pestivirus envelope protein, is used.

Development of a classical swine fever subunit marker vaccine and companion diagnostic test.

The development of a classical swine fever (CSF) subunit marker vaccine, based on viral envelope glycoprotein E2, and a companion diagnostic test, based on a second viral envelope glycoprotein E(RNS), will be described. Important properties of the vaccine, such as onset and duration of immunity, and prevention of horizontal and vertical transmission of virus were evaluated. A single dose of the vaccine protected pigs against clinical signs of CSF, following intranasal challenge with 100LD(50) of virulent classical swine fever virus (CSFV) at 2 weeks after vaccination. However, challenge virus transmission to unvaccinated sentinels was not always completely inhibited at this time point. From 3 weeks up to 6 months after vaccination, pigs were protected against clinical signs of CSF, and no longer transmitted challenge virus to unvaccinated sentinels. In contrast, unvaccinated control pigs died within 2 weeks after challenge. We also evaluated transmission of challenge virus in a setup enabling determination of the reproduction ratio (R value) of the virus. In such an experiment, transmission of challenge virus is determined in a fully vaccinated population at different time points after vaccination. Pigs challenged at 1 week after immunization died of CSF, whereas the vaccinated sentinels became infected, seroconverted for E(RNS) antibodies, but survived. At 2 weeks after vaccination, the challenged pigs seroconverted for E(RNS) antibodies, but none of the vaccinated sentinels did. Thus, at 1 week after vaccination, R1, and at 2 weeks, R=0, implying no control or control of an outbreak, respectively. Vertical transmission of CSFV to the immune-incompetent fetus may lead to the birth of highly viraemic, persistently infected piglets which are one of the major sources of virus spread. Protection against transplacental transmission of CSFV in vaccinated sows was, therefore, tested in once and twice vaccinated sows. Only one out of nine once-vaccinated sows transmitted challenge virus to the fetus, whereas none of the nine twice-vaccinated sows did. Finally, our data show that the E(RNS) test detects CSFV-specific antibodies in vaccinated or unvaccinated pigs as early as 14 days after infection with a virulent CSF strain. This indicates that the E2 vaccine and companion test fully comply with the marker vaccine concept. This concept implies the possibility of detecting infected animals within a vaccinated population.

A simple, rapid and reliable enzyme-linked immunosorbent assay for the detection of bovine virus diarrhoea virus (BVDV) specific antibodies in cattle serum, plasma and bulk milk.

To detect Bovine Virus Diarrhoea Virus (BVDV)-specific antibodies in cattle serum, plasma and bulk milk, a simple, reliable and rapid blocking ELISA ("Ceditest") has been developed using two monoclonal antibodies ("WB112" and "WB103") directed to different highly conserved epitopes on the non-structural peptide NS3 of pestiviruses. The test can be performed at high reproducibility using undiluted samples. In testing 1000 field serum samples, the ELISA showed a specificity and a sensitivity relative to the virus neutralization test of 99% and 98%, respectively. The blocking ELISA is able to detect specific antibodies in serum obtained 12 days after an acute infection and in serum of vaccinated and challenged animals. A frequency distribution diagram, obtained by testing almost 1800 random Dutch field serum samples, showed a clear separation between a negative population (maximum frequency of the % inhibition at -5%) and a positive population (maximum frequency of the % inhibition at 95%). Based on these data, the prevalence of seropositive animals was 65% (95% confidence interval: 63%-67%). By exchanging plasma- and bulk milk samples between two laboratories (The Netherlands and Denmark), a good overall agreement was found between results obtained with the Ceditest and those obtained with the Danish blocking ELISA as used in the Danish BVDV eradication programme. The results indicate that BVDV infections can reliably be diagnosed by the Ceditest ELISA and that the test is suitable for use in large scale screening and eradication programmes.

A bovine respiratory syncytial virus strain with mutations in subgroup-specific antigenic domains of the G protein induces partial heterologous protection in cattle.

Bovine respiratory syncytial virus (BRSV) strains are tentatively divided in subgroups A, AB and B, based on antigenic differences of the G protein. A Dutch BRSV strain (Waiboerhoeve: WBH), could not be assigned to one of the subgroups, because the strain did not react with any monoclonal antibody against the G protein. We describe here that the WBH strain has accumulated critical mutations in subgroup-specific domains of the G protein gene, which also occur but then independently in G protein genes of BRSV subgroup A or B strains. Although the comparison of nucleotide residues 256-792 of the G gene of the WBH strain with those of subgroup A and B strains showed that the G gene of the WBH strain is different from that of BRSV subgroup A and B strains, the sequence divergence was not more than observed within the G genes of human respiratory syncytial virus subgroup A or B strains. The WBH strain did not induce severe disease after experimental infection of calves, and induced partial protection against a heterologous challenge. Despite the dissimilarity of the conserved central regions of the G protein of the WBH strain and that of the challenge strain, a secondary antibody response against this region was induced in WBH-infected calves after challenge. We conclude that complete BRSV virus can partially protect against a BRSV infection with a strain that contains an antigenic dissimilar G protein.

Validation of a monoclonal antibody-based ELISA to detect antibodies directed against swine vesicular disease virus.

A simple, rapid and sensitive competitive monoclonal antibody-based ELISA for the detection of antibodies directed against swine vesicular disease virus (SVDV) was developed. The ELISA was validated using field sera originating from SVDV-infected and non-infected Dutch pig herds, reference sera obtained from the Community Reference Laboratory for Swine Vesicular Disease at the Institute for Animal Health, Pirbright Laboratory, UK, and sera from animals infected experimentally. When testing 4277 sera originating from non-infected Dutch pig herds and collected as part of the national screening program, this ELISA had only 0.6% false positive results, whereas approximately 2% of false positive results were obtained with a conventional blocking ELISA used until recently. A sensitivity relative to the virus neutralisation test of > 97% was achieved when testing sera collected from Dutch pig farms where an outbreak of SVDV had occurred. All international reference sera scored consistently correct. Sera collected sequentially from pigs experimentally infected with SVDV isolates representing all currently recognized antigenic groups, were scored positive slightly earlier by the ELISA compared to the virus neutralisation test. This monoclonal antibody-based competitive ELISA for SVDV antibodies designated the Ceditest ELISA for SVDV-Ab, is as sensitive but more specific than the ELISA used until recently. Because sera are tested at a single dilution (1:5), incubations are carried out at room temperature and test results are available within 3 h, this ELISA is simple, easy to automate and therefore very suitable for screening large numbers of serum samples.

Immune response of calves experimentally infected with non-cell-culture-passaged bovine respiratory syncytial virus.

The immune response of calves was studied following infection with non-cell-passaged Bovine respiratory syncytial virus (BRSV). Two groups of 6 specific pathogen free (SPF) calves were housed in separate isolation rooms. One group was inoculated intranasally with a non-cell-passaged BRSV strain and the control group was mock-infected. A BRSV specific antibody response was observed for all the BRSV infected calves. These antibodies were shown to have neutralizing activity. No lymphocyte proliferation response was detected in the mock-infected group whereas three animals in the infected group were positive three weeks after the infection. All BRSV-infected calves, except one, produced interferon-gamma (IFN-gamma) one week post-infection and IFN-gamma was observed in all six infected calves after three weeks. The control group showed no IFN-gamma synthesis. In spite of the limits of the BRSV infection model, humoral and cellular immune responses were actively developed by all the calves against this pathogen.

Critical factors affecting the diagnostic reliability of enzyme-linked immunosorbent assay formats.

This paper aims to evaluate different formats of the enzyme-linked immunosorbent assays (ELISAs) for detection of virus-specific antibodies and focuses on factors that may influence the diagnostic reliability of such tests. Newly developed and well-established ELISAs for detection of infections of bovine herpesvirus 1 (BHV1), bovine respiratory syncytial virus (BRSV), classical swine fever virus (CSFV), pseudorabies virus (PRV) and bovine viral diarrhoea virus (BVDV) are used as examples. Differences between competitive and non-competitive ELISAs are described, with special reference to the influence of the antigen, the conjugated antibody and the test sample on the test results. Attention is drawn to interference, which may result in false positive or false negative test results, with special emphasis on the 'bridging' phenomenon. The use of monoclonal antibodies and discriminatory tests are briefly discussed. Diagnostic reliability is described for tests that are used in monitoring or eradication programmes, emphasising the consequences of false negative and false positive test results. Finally, reducing assay-time and functional quality control for such tests are discussed.

Antigenic and molecular analyses of the variability of bovine respiratory syncytial virus G glycoprotein.

Antigenic variation among eight bovine respiratory syncytial virus (BRSV) isolates was determined using monoclonal antibodies (MAbs) specific for the attachment (G) protein. Two major (and one intermediate) subgroups were identified, as well as one strain that did not fit any pattern. The subgroups could also be differentiated on the basis of the Mr of the F protein cleavage product, F2. The nucleotide sequence of the G gene of seven of the BRSV strains was determined and compared with published G gene sequences. Subgroups A and A/B were more closely related in protein sequence than subgroups A and B or subgroups A/B and B. These results could not be correlated with those obtained by the determination of the Mr of the F2 polypeptide. Multiple sequence alignments showed a high level of amino acid identity at the inter-subgroup level (85% identity between subgroup A and subgroup B strains), similar to the intra-subgroup human (H)RSV identity, suggesting that the BRSV isolates form a continuum rather than distinct subgroups. However, unusual variability was observed within the immunodominant domain (amino acids 174-188) in contrast with the situation in HRSV strains belonging to the same subgroup.

Probability of detecting antibodies to bovine herpesvirus 1 in bulk milk after the introduction of a positive animal on to a negative farm.

The purpose of this study was to assess the probability that the introduction of one or more bovine herpesvirus 1 (BHV-1)-seropositive animals would result in the bulk milk of a clean herd becoming BHV-1-positive. Probability calculations (stochastic and deterministic) were based on the distribution of the log(titre) of 828 positive animals and the daily milk production of the herds and of the individual cows. They showed that the probability in average sized herds of 45 dairy cows is only between 10 and 25 per cent and that even in small herds of 25 cows the introduction of a positive animal would go undetected in the majority of cases. It is concluded that if the bulk milk has become BHV-1-positive it is most likely that the infection has spread.

Antibody responses against epitopes on the F protein of bovine respiratory syncytial virus differ in infected or vaccinated cattle.

The fusion protein F of bovine respiratory syncytial virus (BRSV) is an important target for humoral and cellular immune responses, and antibodies against the F protein have been associated with protection. However, the F protein can induce antibodies with different biological activity, possibly related to distinct antigenic regions on the protein. Therefore, epitopes were mapped on the F protein using monoclonal antibodies. Two epitopes (A and B) were identified that induced neutralizing antibodies, and one epitope (C) that did not elicit neutralizing antibodies. Subsequently, antibody responses were analysed against these epitopes in cattle sera after natural infection, experimental infection or vaccination. After natural infection or reinfection, the antibody titres against epitope A were significantly higher than those against epitope B or C. After experimental infection and after vaccination with an inactivated vaccine, antibody titres against epitope B and C were significantly higher than after natural infection. Conversely, virus neutralizing antibody titres were significantly lower in these animals with higher antibody titres against epitopes B and C than in naturally infected cattle. Because after natural infection the epitope-specific-antibody titres against epitope A, B or C differed markedly between the cattle, the magnitude of the antibody titres against epitope A, B or C in relation to the major histocompatibility complex (MHC) genes of cattle (BoLA) was studied. The magnitude of the antibody responses against epitope A of the F protein, but not against the G protein, appeared to be associated with the bovine lymphocyte antigen (BoLA) haplotype.

Serological indication for persistence of bovine respiratory syncytial virus in cattle and attempts to detect the virus.

To identify putative persistent bovine respiratory syncytial virus (BRSV) infections in cattle, seven cattle that had experienced BRSV infections were treated with corticosteroids for two periods of 5 days. During the 5-day periods and the 3 weeks after treatment, attempts were made to isolate BRSV from lung lavage fluid and nasal swab specimens. Fluorescent antibody tests were used to detect BRSV antigen in lung lavage cells. A BRSV specific polymerase chain reaction (PCR) assay was developed, and was performed on lung lavage samples of all seven cattle as well as on various tissues of five of the cattle. In addition, nasal swabs of 74 over-one-year-old cattle, in a closed dairy herd were also assayed by PCR. The virus or its RNA was not detected in putative carriers, by any of the methods used, whereas all positive controls were positive. After corticosteroid treatment, three of the seven cattle showed a fourfold rise in antibody titre, suggesting induction of virus replication. BRSV-seronegative sentinel calves, that were housed together with each corticosteroid-treated animal, did not develop antibodies to BRSV indicating that BRSV was not shed by corticosteroid-treated cattle, or was shed at a very low level. In addition BRSV was not detected in seropositive cattle in a closed farm in summer. Although we consider the rises in antibody titres against BRSV an indication for persistence of BRSV in cattle, BRSV or its RNA was not detected in infected cattle.

Subgrouping of bovine respiratory syncytial virus strains detected in lung tissue.

Bovine respiratory syncytial virus is an important respiratory pathogen in cattle. Recently, subgroups of BRSV have been identified, based on antigenic differences. However, little is known about subgroups of BRSV that circulate in the cattle population. Therefore, we determined the reactivity of monoclonal antibodies (mAbs), directed against the G, F, or P protein of BRSV, with lung tissue from 47 calves, that suffered from severe respiratory disease. Fourteen animals (30%) proved to be infected with BRSV, because they all reacted with mAbs against the P or F protein, as detected by fluorescent antibody tests. Monoclonal antibodies against the G protein were able to discriminate between the BRSV-positive specimens: 7 strains were identified as subgroup A strains, and 5 strains as subgroup AB, which is introduced as BRSV subgroup in this paper. Two strains could not be identified unambiguously. It is concluded that BRSV subgroup A and AB were associated with severe respiratory disease, and that strains belonging to either subgroup circulated concurrently in the cattle population in the Netherlands.

A European inter-laboratory trial to evaluate the reliability of serological diagnosis of bovine herpesvirus 1 infections.

The detection of cattle latently infected with bovine herpesvirus 1 (BHV1) is of importance in control programs and in international trade activities. Therefore, tests to detect specific antibodies in serum must be highly sensitive. To evaluate the reliability of serological diagnosis of BHV1 infections in Europe, seventeen laboratories in 15 European countries were asked to determine BHV1-specific antibodies in a panel of bovine serum samples using the serological tests available in their laboratory. Laboratory tests included virus neutralisation tests (1, 2 and 24 h), indirect immunofluorescence tests and ELISAs. The serum panel consisted of 12 duplicate lyophilised samples which were randomly coded from 1 to 24 and included negative, weak- and strong-positive samples as well as international reference sera. Virus neutralisation tests and ELISAs showed a high specificity. All participants using neutralisation tests (n = 13) scored the negative samples correctly. Twenty three of 25 ELISAs showed 100% specificity. A serum sample obtained at day 7 after infection was scored as negative by all tests except one home-made blocking ELISA. Samples obtained at day 9 and at day 11 after infection were scored as positive by approximately half and by all tests, respectively. To score the weak-positive European reference standard (EU2) correctly, 24 h neutralisation tests (positive by 8 of 9 laboratories) and home-made blocking ELISAs (positive by 5 of 6 laboratories) were the most reliable. The results indicate that standardisation is urgently needed to ensure that BHV1-infected animals with low antibody titres are recognised.

Antibody responses against the G and F proteins of bovine respiratory syncytial virus after experimental and natural infections.

Antibodies against the two major surface glycoproteins of bovine respiratory syncytial virus (BRSV), G and F, play a role in protection against BRSV-associated disease, but only the antibody response against the F protein has been well described. Therefore, we used a novel peptide-based enzyme-linked immunosorbent assay (G peptide-ELISA) to compare immunoglobulin G (IgG) and IgG subclass antibody responses against the G protein with the antibody response against the F protein, as measured by a conventional BRSV ELISA (F-ELISA). Experimental infection of cattle induced significantly lower antibody titers than did natural infection. After natural primary infection, G peptide-specific antibodies declined more rapidly and to lower levels than the F protein-specific antibodies. As a consequence, the G peptide-ELISA detected more reinfections than did the F-ELISA. Ratios of G- and F-specific IgG1/IgG2 antibody titers did not differ markedly after infection or vaccination. Interestingly, after natural infection calves did not develop an IgG2 response to the complete G protein. In contrast, adult cattle had high IgG2 titers against this protein. Vaccination with a live vaccine induced low antibody titers, similar to the titers after experimental infection, whereas vaccination with an inactivated vaccine induced high titers. The results indicate that the kinetics of the G- and F-specific antibody responses differ. Furthermore, the IgG subclass response against the unglycosylated central region of the G protein is similar to the IgG subclass response to the F protein, but the IgG subclass response differs from the response to the complete G protein.

Experimental reproduction of respiratory disease in calves with non-cell-culture-passaged bovine respiratory syncytial virus.

To reproduce experimentally clinical bovine respiratory syncytial virus (BRSV) infections in cattle, we isolated BRSV from a calf in the field that suffered from acute respiratory disease. Cell culture passage of the virus was avoided to prevent any modification of the biological properties of the virus. The isolated BRSV was passaged in specific-pathogen-free (SPF) calves. Lung lavage fluids of these calves, which contained at least 10(3) TCID50/ml BRSV and which were found to be free of other known respiratory pathogens, were collected and pooled for experimental infection. To reproduce a clinical BRSV infection, two groups of six SPF calves were inoculated intranasally with 2 ml of 10(3.9) TCID50/ml BRSV of the obtained virus stock. Another five calves, which were persistently infected with bovine virus diarrhoea virus (BVDV), were given the same inoculum. One group of six calves served as mock-infected controls. Clinical signs were closely monitored from 1 week before until 16 days after inoculation. Reproducible clinical signs consisting of significantly (p < 0.05) increased respiratory rates and elevated body temperatures were recorded but not in all BRSV-inoculated calves. Although clinical signs were induced by experimental infection with non-cell-culture-passaged BRSV, the respiratory signs were not as serious as in the most severe cases in the field.

Type-specific serologic diagnosis of respiratory syncytial virus infection, based on a synthetic peptide of the attachment protein G.

Peptides deduced from the central hydrophobic region (residues 158-189) of the G protein of bovine and ovine respiratory syncytial virus (RSV) and of human RSV subtypes A and B were synthesized. These peptides were used to develop ELISAs to measure specifically antibodies against these types and subtypes of RSV. We have evaluated the bovine RSV-G peptide in both an indirect ELISA and in a blocking ELISA. Specificity and sensitivity, relative to a routine diagnostic ELISA that detects antibodies against the RSV F-protein in bovine sera, were 98% and 92% respectively for the indirect peptide-based ELISA, and 98% and 98% for the blocking peptide-based ELISA. In paired serum samples, rises in antibody titer were detected more frequently with the indirect peptide-based ELISA than with the routine F-ELISA. Furthermore, the peptide-based G-ELISAs were able to differentiate between antibodies against BRSV and HRSV, and between those against BRSV and ORSV. In addition, the indirect peptide-based ELISA was selective for HRSV subtype A and B antibodies. This study shows that peptides, corresponding to the central hydrophobic region of the attachment protein G of several RSVs, can be used successfully as antigens in highly specific and sensitive immunoassays.

Quantitative investigation of population persistence and recurrent outbreaks of bovine respiratory syncytial virus on dairy farms.

OBJECTIVE: To determine if transmission of virus among seropositive cattle is a plausible mechanism for the permanent presence of bovine respiratory syncytial virus (BRSV) in dairy herds, and how likely, with the scenario for persistence, there will be only 1 clinical outbreak of BRSV per year. DESIGN: Build a stochastic model, parameter estimation from serologic data on BRSV, and interpret the estimated parameter values from the model analysis. SAMPLE POPULATION: Monthly data on the prevalence of antibodies directed against BRSV in sera from all cattle in 6 dairy herds. PROCEDURE: Parameter estimation applying general linear models, model analysis using calculation of the reproduction ratio for simplified models, and Monte-Carlo simulation for the whole model. RESULTS: Persistence of BRSV by transmission among seropositive cattle given estimated parameter values would be accompanied by frequent extinctions (once every 10 to 50 years) and long infectious periods in seropositive cattle (100 days). Moreover, in the model, a single clinical outbreak among seronegative cattle only occurred with external forcing. CONCLUSIONS: From these data, transmission among seropositive cattle is not a plausible mechanism for persistence of BRSV in dairy herds.