As of 2004, the European Pharmacopoeia demands that plasma pools for production of anti-D immunoglobulin should not contain more than 104 IU per mL of parvovirus B19 (B19V) DNA. Hence, before pooling, highly viremic donations have to be identified, and after pooling the level of B19V DNA must be determined. The performance of a new real-time B19V DNA PCR test (Roche, Mannheim, Germany) was studied, using a DNA extractor (NucliSens, bioMerieux, Boxtel, the Netherlands) for isolation of nucleic acid, and using a DNA quantification test (LightCycler apparatus, Roche, Mannheim, Germany) for amplification and detection. Dilutions of the international B19V DNA standard and reference preparations were tested to determine the precision, linear range, and accuracy of the assay and to calculate the factor for conversion of B19V DNA copies to IUs. The internal control signals, invalid test results, and the effect of cryo-poor plasma were studied as a measure for robustness. Routine performance was assessed by testing 164 manufacturing pools (not screened for B19V) and 1048 test pools of 480 donations each. The copies-to-IU conversion factor was calculated to be 3.34 (95% CI, 3.07-3.63). The assay appears linear between 10(3) and 10(7) IU per mL. Between 10(3) and 10(5) IU per mL, the test can discriminate samples differing a factor two in B19V DNA content. Overall, 0.78 percent of the test results were invalid. Of 127 B19V DNA negative control plasma samples, 7 were contaminated with low levels of B19V DNA. Of 164 nonscreened manufacturing plasma pools, 92 contained B19V DNA (56%); 13 contained more than 10(4) IU per mL. Of 503,040 donations, 29 contained more than 5 x 10(6) IU per mL B19V DNA (1:17,346). The B19V DNA quantification test (LightCycler, Roche ) is suitable for quantitative, routine, in-process measurement of B19V DNA levels in plasma pools, using the DNA extractor (NucliSens, bioMerieux) for nucleic acid isolation.