The aim of the study was to evaluate the specificity and sensitivity of detection of hepatitis C virus (HCV)-RNA in formalin-fixed paraffin-embedded (FFPE) liver biopsies by polymerase chain reaction (PCR). Routinely processed FFPE diagnostic needle liver biopsies as well as stored serum samples from 43 patients with liver disease were tested for HCV-RNA by reverse transcription-nested PCR using the same sets of primers and following strict anticontamination measures. Twenty-nine cases were positive and 14 were negative for serum HCV-RNA. Tissue HCV-RNA was detected in 17 out of the 29 serum HCV-RNA-positive cases but not in any of the 14 serum HCV-RNA-negative cases. Compared to serum-PCR, tissue-PCR was 100% specific, 58.6% sensitive, and 72% efficient. HCV-RNA was detected more frequently in biopsies stored for less than 1 year, than in those stored for more than 1 year (P = 0.046). In biopsies stored for up to 1 year detection of HCV-RNA by PCR was 81.8% sensitive and 90.9% efficient. Short (< 0.5 cm) liver biopsies were as sufficient for nucleic acid extraction and amplification as long (> 0.5 cm) ones. It is concluded that following strict anticontamination measures, HCV-RNA detection by PCR in routinely fixed, processed, and stored diagnostic liver biopsies provides a valuable adjunct to diagnosis of HCV infection. In this study, this option was free from contamination problems, even though routine batch histological processing schedules were used.