We have noticed that suboptimal specimen processing and storage conditions may cause false-negative results in the detection of hepatitis C virus (HCV) RNA in plasma or serum. To establish the influence of specimen handling in a serological laboratory on the rate of detection of HCV RNA by the cDNA polymerase chain reaction (cDNA-PCR), we tested routine serum samples and fresh-frozen plasma samples from the same bleeding from confirmed anti-HCV-positive blood donors. When primers from the NS3/NS4 region were used, HCV RNA was detected in fresh-frozen plasma from 67% of the donors, whereas positive results were obtained with only 50% of the serum samples that had been subjected to routine serological procedures. Analysis of the same samples with primers from the highly conserved 5'-terminal region (5'-TR) revealed an HCV RNA detection rate of 92% for both the routine and the fresh-frozen samples. However, the yield of the amplification product in routine samples was strongly reduced compared with that in fresh-frozen plasma. Comparison of both primer sets for cDNA-PCR showed that the 5'-TR primer set was 10- to 100-fold more effective in detecting HCV RNA. We also analyzed the effect of storage of whole EDTA-blood and serum at room temperature and at 4 degrees C on the yield of the amplification product. A rapid decline in detectable HCV RNA of 3 to 4 log units was observed within 14 days when whole blood and serum were stored at room temperature. By contrast, no perceptible reduction in the cDNA-PCR signal was found in freshly prepared serum stored at 4 degrees C.