To determine the optimal conditions for preparation of serum specimens for quantitative hepatitis C virus RNA determination, patient samples were processed such that differences in time from clot formation to centrifugation, centrifugation to separation of serum and collection of serum until freezing could be independently assessed. The effects of multiple cycles of freezing and thawing were also determined. There was progressive and significant loss of hepatitis C virus RNA activity when the time from the formation of the clot until centrifugation was longer than 2 hr. This reduction reached 32% after 6 hr and 49% after 24 hr. If centrifugation was performed immediately after formation of the clot, loss of hepatitis C virus RNA activity was reduced to less than 10% even though the serum remained unseparated from the clot for up to 6 hr. Centrifugation of blood through a paraffin plug (serum separator tube) prevented loss of hepatitis C virus RNA activity for up to 24 hr. There was no loss of hepatitis C virus RNA activity with up to three freeze-thaw cycles. When patient specimens were prepared under these optimal conditions, the sensitivity of the quantitative branched DNA signal amplification assay in patients with hepatitis C virus infection was 83% and the specificity in patients with liver disease was 100%. Fluctuations in hepatitis C virus RNA levels were shown to correlate with biochemical changes observed in patients treated with recombinant inter-feron-α2b. These data demonstrate that improper or inconsistent methods of serum preparation may result in falsely low and unreliable levels of hepatitis C virus RNA. Clinical applications of quantitative hepatitis C virus RNA assays will only be possible when sera are properly prepared. (HEPATOLOGY 1994;19:1337–1341.)