We evaluated the efficacy of long-term interferon (IFN) therapy in patients with e-antigen-positive chronic hepatitis B. The study design was a prospective, randomized controlled clinical trial. Fifty-three patients were randomly assigned into one of two groups, treated with 3 million units (MU) of IFN (low dose group; n=27) or 6 MU IFN (high dose group; n=26), administered twice weekly for 52 weeks. Responders were defined as patients whose hepatitis B virus (HBV)-DNA level, determined by branched DNA signal amplification and hepatitis B e-antigen (enzyme immunoassay) were negative and whose serum alanine transaminase (ALT) levels fell to within the normal range (ALT<50 IU/l) at 6 month after termination of IFN therapy. One patient in high dose group was dropped out because of transfer. Remainder 52 patients were examined by intention-to-treat (ITT) analysis. The response rates by ITT analysis were 40.7% (11/27) in low dose group and 20% (5/25) in high dose group. The difference between low and high dose group was not statistically significant. Univariate analysis of clinical factors that contribute to the response demonstrated that IFN therapy had a significant effect when, (1) the serum HBV-DNA level was <200 Meq/ml prior to the commencement of IFN therapy (P=0.0327). (2) Transient acute exacerbation of ALT was present during or after IFN therapy (P=0.0311). Multivariate analysis showed that the risk ratio for the development of response in patients with serum HBV-DNA level of less than 200 Meq/ml was 3.60 compared with patients with that of 200 Meq/ml or more than 200 Meq/ml (95% CI, 1.012-12.81). In conclusion, the results of this trial show that: (1) long-term twice weekly IFN therapy could be a worthwhile strategy for e-antigen-positive chronic hepatitis B patients with serum HBV-DNA level of less than 200 Meq/ml and (2) patients with transient acute exacerbation of ALT during or after IFN therapy could often respond well after exacerbation of ALT.