International proficiency trial for bovine viral diarrhea virus (BVDV) antibody detection: limitations of milk serology

BACKGROUND: Control programs were implemented in several countries against bovine viral diarrhea (BVD), one of the most significant cattle diseases worldwide. Most of the programs rely on serological diagnostics in any phase of the program. For the detection of antibodies against BVD virus (BVDV), neutralization tests as well as a variety of (commercially available) ELISAs are used. Here, test systems applied in various laboratories were evaluated in the context of an international interlaboratory proficiency trial. A panel of standardized samples comprising five sera and five milk samples was sent to veterinary diagnostic laboratories (n=51) and test kit manufacturers (n=3). RESULTS: The ring trial sample panel was investigated by nine commercially available antibody ELISAs as well as by neutralization tests against diverse BVDV-1, BVDV-2 and/or border disease virus (BDV) strains. The negative serum and milk sample as well as a serum collected after BVDV-2 infection were mostly correctly tested regardless of the applied test system. A serum sample obtained from an animal immunized with an inactivated BVDV-1 vaccine tested positive by neutralization tests or by total antibody or E(rns)-based ELISAs, while all applied NS3-based ELISAs gave negative results. A further serum, containing antibodies against the ovine BDV, reacted positive in all applied BVDV ELISAs, a differentiation between anti-BDV and anti-BVDV antibodies was only enabled by parallel application of neutralization tests against BVDV and BDV isolates. For the BVDV antibody-positive milk samples (n=4), which mimicked prevalences of 20% (n=2) or 50% (n=2), considerable differences in the number of positive results were observed, which mainly depended on the ELISA kit and the sample incubation protocols used. These 4 milk samples tested negative in 43.6%, 50.9%, 3.6% and 56.4%, respectively, of all investigations. Overall, negative results occurred more often, when a short sample incubation protocol instead of an over-night protocol was applied. CONCLUSIONS: While the seronegative samples were correctly evaluated in most cases, there were considerable differences in the number of correct evaluations for the seropositive samples, most notably when pooled milk samples were tested. Hence, thorough validation and careful selection of ELISA tests are necessary, especially when applied during surveillance programs in BVD-free regions.