A split-sample study was conducted to evaluate the clinical performance of an enzyme immunoassay that detects the human parvovirus B19 virus (B19V) immunoglobulin M (IgM) or IgG in the sera of pregnant women. The initial study compared a baculovirus-expressed VP2 enzyme immunoassay (BVP2 EIA) (Biotrin International Inc., Dublin, Ireland) with the currently available and commonly used Escherichia coli-expressed VP1 enzyme immunoassay (EVP1 EIA) (MRL Diagnostics, Cypress, Calif.). There was a high degree of agreement between the two assays in the detection of IgM antibodies (283 of 307 [92.2%]) or IgG antibodies (279 of 311 [89. 7%]), with the majority of discrepancies (IgM, 17 of 24 [71%]; IgG, 16 of 31 [50%]) being due to equivocal data obtained with the EVP1 EIA. Specimens with discordant BVP2 EIA and EVP1 EIA results (23 of 24 IgM and 32 of 32 IgG results) were analyzed further by baculovirus-based VP1 immunofluorescence assays (BVP1 IFAs) (Biotrin International). The BVP2 EIA and BVP1 IFA results for 20 of 23 and 28 of 32 specimens for IgM and IgG, respectively, were concordant. In contrast, the EVP1 EIA and BVP1 IFA data for only 3 of 23 and 4 of 32 specimens for IgM and IgG, respectively, were in agreement, despite the fact that the same capsid antigen was used. Both the BVP2 EIAs and BVP1 IFAs utilize a conformational viral capsid antigen, while the EVP1 EIA uses a denatured viral capsid antigen. In conclusion, the BVP2 EIAs produced far fewer equivocal results for IgM and IgG, correlating more closely to the confirmatory BVP IFAs, than did the EVP1 EIAs and proved to be more accurate for detecting B19V antibodies in the sera of pregnant women.