The aims of the present study were to determine (i) the long-term immunogenicity and the decay rate of hepatitis B virus (HBV) surface antibody (anti-HBs) from universal hepatitis B vaccination at infancy for a healthy population in an area of hyperendemicity and (ii) whether the anti-HBs levels measured by enzyme immunoassay (EIA) were closely correlated with those assayed by radioimmunoassay (RIA) methods during long-term monitoring. A total of 1,337 apparently healthy children (696 boys and 641 girls) who were vaccinated against HBV at infancy and monitored for anti-HBs annually from 7 to 16 years of age entered the study. Serum samples were analyzed for anti-HBs by RIA at 7 to 15 years of age and were also analyzed by EIA at 13 to 16 years of age. Antibody titers were quantified in mIU/ml by EIA as well as by the ratio of the count in the sample to the count for a negative control (S/N) by RIA. In non-boosted children, the average decay of anti-HBs from 7 to 16 years of ages indicated that approximately 20% of the geometric mean titer decays per year. There was a good correlation between serum anti-HBs levels measured by the RIA and the EIA methods (r=0.91; P<0.0001). An equation for RIA to EIA level conversion was established: log EIA titer=-0.12+ (1.31 . log RIA S/N). The anti-HBs titers measured by EIA correlate well with the S/N assayed by RIA. The annual decay rate of the log anti-HBs level may help in planning booster immunizations for hypo-responders or individuals at risk in adolescence.