The histopathologic alterations of hepatitis C virus (HCV) infection of the liver overlap with those of other diseases, making interpretation of liver biopsy specimens in some cases insufficient to render a diagnosis. Although HCV infection can be confirmed by detection of circulating anti-HCV antibodies, immunocompromised liver transplant recipients are often unable to mount an immunologic response to the virus, resulting in false-negative serologic testing. We describe the comparison of reverse transcription-polymerase chain reaction (RT-PCR) with histopathology, serology, and immunohistochemistry for the diagnosis of HCV. Sixty-three formalin-fixed, paraffin-embedded tissue samples (40 needle biopsy specimens and 23 native liver resection specimens) from 35 transplant patients were analyzed by use of a novel method of RNA extraction followed by nested PCR for HCV as well as albumin mRNA as an internal control. HCV was detected by RT-PCR in 50 of 51 (98%) paraffin sections of liver from transplant patients with circulating anti-HCV antibodies, 15 of which lacked characteristic histologic features of HCV infection. Overall, there were no false-negative results in 36 needle biopsy specimens from patients with hepatitis C infection, but three negative results were seen in end-stage cirrhotic native livers resected from HCV-infected patients. No false-positive test results were seen among 21 negative controls (10 liver samples from immunocompetent patients with abnormalities unrelated to hepatitis C and 11 liver biopsies from immunocompetent patients without histologic evidence of liver disease). In comparison, immunohistochemistry using antibody TORDJI-22 was positive for HCV in only 15 of 32 (47%) needle biopsies positive by RT-PCR. Our results indicate that RT-PCR is a more sensitive and specific method of detecting hepatitis C in routinely processed paraffin sections of formalin-fixed liver biopsy specimens than histopathologic examination or immunohistochemistry.