B19 virus can be transmitted by contaminated blood or blood products. Recent observations, in healthy volunteers, suggest that active B19 infection can follow the administration of plasma pools with a concentration > or =10(7) genome equivalents/ml (geq/ml) of B19 DNA. However, patients receiving batches with levels of virus DNA lower than 10(4) geq/ml do not show any evidence of transmission of the virus. The aim of the study was to show, by in vitro assays, a threshold of viral load in B19 contaminated plasma pools over which the infection can be transmitted. Twenty plasma pools, each containing 960 single donations, were tested to correlate the viral load and the level of antibodies anti-B19 with the in vitro infectivity and expression of B19 virus. All the plasma pools, titrated for B19 viral load by competitive PCR, were inoculated into KU812Ep6 erythroid human cell line. Five of the nine contaminated plasma pools, with a B19 DNA concentration > or =3.60 x 10(6) geq/ml, were able to infect KU812Ep6 cells. In vitro infectivity was shown in KU812Ep6 cells at 24 h post-infection by in situ hybridisation and amplification assays for viral DNA and RNAs. Plasma pools with a viral load in the range of 6.00 x 10(3)-8.96 x 10(4) geq/ml did not show infectivity when inoculated into KU812Ep6 cells. Medium-high titres of IgG antibodies anti-B19 were detectable in all the plasma pools and the neutralising activity associated with specific IgG anti-B19 may explain the lack of infectivity of plasma pools contaminated with a low viral load. In conclusion, in situ hybridisation and amplification assays for viral DNA and RNAs in KU812Ep6 cells inoculated with plasma pools can be valid assays to test for the presence of infectious virus in the production of B19-safe material.