Assessment of markers of hepatitis C virus infection in a Japanese adult population.

Abstract

Latent-class analysis was used to evaluate the usefulness of markers of hepatitis C virus (HCV) infection in characterizing the true, underlying infection in a community-based Japanese population. Antibodies to HCV were detected in 24%, HCV RNA in 22%, and HCV core protein in 19% of stored serum samples from 372 adults. A 2-class model suggested that positive results for any 2 virus markers defined the current HCV infection class, with an estimated prevalence of 22% (95% confidence interval, 18%-26%). The sensitivity for detection of current HCV infection was highest for anti-HCV (97%) and was more moderate for HCV RNA (91%) and HCV core protein (85%). The specificity for each marker was > or =96%. In general, the association between demographic factors and current HCV infection status was strengthened by use of latent-class analysis that combined data for markers of HCV infection, when compared with results of logistic regression analysis for each marker separately.