Informative value of an indirect enzyme-linked immunosorbent assay (ELISA) for the detection of bovine viral diarrhoea virus (BVDV) antibodies in milk.

Bulk and individual milk samples from 117 herds located in Brittany (west France) were used to assess: (i) the performance characteristics of an indirect enzyme-linked immunosorbent assay (ELISA) applied to individual milk for the detection of antibodies to bovine viral diarrhoea virus (BVDV); and (ii) the relationship between the bulk milk result obtained from this test and the within-herd prevalence of antibody-positive lactating cows. This ELISA test was based on a monoclonal antibody directed against non-structural protein NS2-3 of pestiviruses. At the individual level, based on 1113 matched milk/serum samples, the sensitivity and specificity of this test applied to milk, compared with the virus neutralization test on serum, were 95.0 and 97.7%, respectively. At the herd level, the relationship between the optical density percentage (OD%) of bulk milk and the within-herd prevalence of antibody-positive lactating cows was assessed using the receiver operating characteristics (ROC) analysis. Classes of OD% of bulk milk were determined so that they were associated with minimum intraclass and maximum between-class variances of within-herd prevalence of antibody-positive cows. The ROC analysis resulted in two classes of bulk milk results corresponding to different expected levels of within-herd prevalence. Herds with an OD% of bulk milk < 75% and > or = 75% had a mean observed prevalence of antibody-positive cows of 8.9 and 60.6%, respectively. Herds with a bulk milk result < 75% were expected to be BVDV free, whereas large variations in prevalence of antibody-positive cows existed in the herds with OD% > or = 75%. The test described in this study is suitable to identify herds likely to have a low prevalence of BVDV antibody-positive cows.