Serum samples from 100 patients with non-A, non-B hepatitis-related chronic liver disease and 100 patients with hepatitis B-related chronic liver disease were tested by first-generation and second-generation enzyme-linked immunosorbent assays, a second-generation recombinant immunoblot assay and the nested polymerase chain reaction. In non-A, non-B hepatitis-related chronic liver disease, second-generation enzyme-linked immunosorbent assay (anti-c22 and/or c200) and second-generation recombinant immunoblot assay showed 98% positivity, whereas first-generation enzyme-linked immunosorbent assay (anti-c100-3) showed 89% positivity. The two second-generation recombinant immunoblot assay-negative samples were positive by nested polymerase chain reaction, but one second-generation recombinant immunoblot assay-positive sample was polymerase chain reaction negative. However, when this second-generation recombinant immunoblot assay-positive sample was tested by polymerase chain reaction using another set of primers, it was polymerase chain reaction positive. Therefore, 100% of the non-A, non-B hepatitis-related chronic liver disease serum samples were hepatitis C virus RNA positive by polymerase chain reaction. Nine hepatitis B-related chronic liver disease samples were first-generation enzyme-linked immunosorbent assay positive. Of the eight second-generation enzyme-linked immunosorbent assay-positive hepatitis B-related chronic liver disease samples, six were first-generation enzyme-linked immunosorbent assay positive and five were second-generation recombinant immunoblot assay positive and polymerase chain reaction positive. One indeterminate second-generation recombinant immunoblot assay sample was polymerase chain reaction negative. Therefore, second-generation recombinant immunoblot assay appears to be as useful as polymerase chain reaction for detecting a chronic hepatitis C virus infection, although some discrepancies were noted.(ABSTRACT TRUNCATED AT 250 WORDS)