The authors compared the diagnostic performance of a second-generation recombinant immunoblot assay (RIBA) (RIBA HCV 2.0 SIA) and the recently introduced third-generation RIBA (RIBA HCV 3.0 SIA) with that of hepatitis C virus (HCV) RNA by the polymerase chain reaction (PCR) in 55 patients on chronic hemodialysis. Compared with HCV RNA by PCR, RIBA 3.0 increased the sensitivity of HCV detection to 72% as compared with 56% of RIBA 2.0. Both assays underestimated the prevalence of HCV infection as determined by PCR. However, RIBA HCV 3.0 outperformed RIBA HCV 2.0, detecting all of the RIBA 2.0-positive patients plus an additional eight (8 of 22 RIBA 2.0 negative; confidence interval [CI] = [17.2%, 59.3%]). Forty-three of 51 patients with positive RIBA 3.0 or positive HCV RNA by PCR underwent a liver biopsy. Thirty (70%) had chronic hepatitis (three with cirrhosis), 10 (23%) had nonspecific changes, and three (7%) had normal liver histology. Thirty of 37 patients (81%) with hepatitis C viremia and positive anti-HCV had chronic hepatitis, whereas none of the viremic patients with negative anti-HCV had chronic hepatitis. Among the reactive antigens on RIBA 3.0, c33c was found to be most predictive of chronic hepatitis (P = 0.0002). Detection of HCV RNA continues to be the method of choice in the early phase of HCV infection. In places where a validated HCV RNA assay is not available, RIBA HCV 3.0 (soon to be commercially available) is a better alternative. Early detection of HCV infection and the implementation of an isolation strategy might be important in preventing the spread of HCV infection among hemodialysis patients.(ABSTRACT TRUNCATED AT 250 WORDS)