Canada has not introduced the non-A, non-B (NANB) surrogate marker tests (antibodies to hepatitis B core antigen and alanine aminotransferase) to screen donated blood. We evaluated the effect of NANB surrogate markers in reducing post-transfusion hepatitis in a prospective randomised intervention study. From 1988 to 1992, 4588 subjects were enrolled into two study groups that received allogeneic blood from which units positive for NANB surrogate markers were either withheld (n = 2311) or not withheld (n = 2277). We also assessed a simultaneous non-randomised cohort (n = 650) of subjects who received only syngeneic blood. All subjects were followed up for 6 months and assessed for the presence of post-transfusion hepatitis due to hepatitis A, B, C, non ABC, Epstein-Barr virus (EBV) and cytomegalovirus (CMV). Withholding of blood containing NANB surrogate positive units reduced the overall post-transfusion hepatitis rate by 40% (p = 0.065) and the hepatitis C rate by 70% (p = 0.05). Most of the benefit of NANB surrogate testing was due to reduced frequency of hepatitis C virus after transfusion before all donor blood was screened for anti-HCV. During this time the overall post-transfusion hepatitis rate per 1000 transfusion recipients was 20.2 in the no-withhold group compared with 5.0 in the withhold group (p = 0.05), and the HCV hepatitis rate was 12.6 and 0 respectively (p = 0.06). After the introduction of HCV screening, the overall post-transfusion hepatitis rates were 8.6 and 6.8 per 1000 (p = 0.55) respectively.(ABSTRACT TRUNCATED AT 250 WORDS)