A randomized, controlled clinical trial was conducted to compare the use of epirubicin (EPI) and doxorubicin (DOX) in Lipiodol (Laboratoire Guerbet, Roissy-Charles-de-Gaulle Cedex, France)-transcatheter arterial chemoembolization as a treatment of hepatocellular carcinoma. One hundred ninety-two hospitals participated, and 415 patients were enrolled in the study during the period between October 1989 and December 1990. The patients were randomly allocated to group A (EPI) or group B (DOX) by a centralized telephone registration. The actual doses of EPI and DOX were 72 mg/body and 48 mg/body, respectively. The 1-, 2-, and 3-year survival rates were, respectively, 69%, 44%, and 33% for group A and 73%, 54%, and 37% for group B. There were no statistically significant differences (P=. 2296, log-rank test). When each group of patients was classified retrospectively into high-risk and low-risk subgroups based on the severity index calculated by the Cox regression model from the significant prognostic factors (the pretreatment tumor size, the pretreatment serum alpha-fetoprotein level, tumor encroachment, and Child's classification), the survival curve of the low-risk DOX subgroup was significantly superior to that of the low-risk EPI subgroup (P=. 0182). However, there was no significant difference between the high-risk subgroups (P=. 4606). The change in the serum alpha-fetoprotein level, the extent of Lipiodol accumulation in the tumor, and the extent of tumor reduction after the treatment did not show any significant differences between the groups. The white blood cell count in group B showed a tendency to decrease slightly more than in group A at 3 weeks after Lipiodol-transcatheter arterial chemoembolization. In conclusion, there was no statistically significant difference between the survival curves of the EPI and DOX groups in Lipiodol-transcatheter arterial embolization treatment of hepatocellular carcinoma.