The purpose of this study was to investigate determinants of endemic hepatitis C virus (HCV) infection within communities in Taiwan. A two-phase study, including a seroprevalence survey and a prevalent case-control study at the first phase, which has been published previously, and a follow-up seroconversion determination and an incident case-control study during the second phase, was carried out to evaluate correlates of persistent endemic HCV infection. At the first phase, a total of 12,021 men and 1,819 women who were 30-64 years old and living in seven townships in Taiwan were tested for the seroprevalence of antibodies to HCV (anti-HCV). In addition, a prevalent case-control study involving 272 HCV-positive cases and 282 seronegative controls identified from the anti-HCV testing was conducted to investigate risk factors associated with HCV prevalence. During the second phase, a total of 2,728 men and 834 women who were seronegative at recruitment participated in the 1-year prospective study on anti-HCV seroconversion. Subsequently, an incident case-control study based on 39 seroconverters and 81 persistently seronegative controls were carried out to elucidate determinants of HCV seroconvertion. Antibodies to HCV were tested by the second-generation enzyme immunoassay. Information on risk factors of HCV infection was collected from subject interviews. The prevalence of anti-HCV consistently increased with age (range 2.9-5.4%), whereas no apparent age trend was observed for anti-HCV seroconversion rate (range 0.9-1.7%). A striking geographical variation in seroprevalence and seroconversion rates of anti-HCV was observed in the study townships. Furthermore, a significant geographical correlation between HCV seroprevalence and seroconversion rates was noted (r = 0.962, P = 0.001). From the results of both prevalent and incident case-control comparisons, medical injections were found to be the main mode to sustain the persistent endemic state of HCV infection within a community (odds ratios for prevalent and incident case-control studies were 2.5 (95% CI = 1.7-3.6) and 3.1 (95% CI = 1.4-7.1), respectively. The data indicate that the basis for HCV transmission has already been existed in study areas and the iatrogenic risk factor tended to be the major determinant for sustaining persistent endemicity within a community.