In a randomized, controlled trial of recombinant interferon alfa-2b with or without prednisone priming in Chinese adults with chronic hepatitis B virus infection, stratified randomization for pretreatment serum alanine aminotransferase levels was done. Partial or complete antiviral responses were achieved in 17 (21.5%) of 79 treated patients and 3 (8.3%) of 36 controls (P = 0.14). The response to interferon treatment was significantly better in those who had elevated pretreatment transaminase levels and comparable to that reported in white patients [15 (38.5%) of 39 patients compared with 2 (5%) of 40 who had normal pretreatment transaminase levels (P = 0.0005)]. The spontaneous seroconversion rate was also higher among the controls with elevated transaminase levels [3 (18.8%) of 16 compared with 0 of 20 with normal transaminase levels], but this difference was not statistically significant (P = 0.16). Among the interferon-treated patients, prednisone priming appeared to have a marginal benefit over treatment with interferon alone in patients with elevated transaminase levels (43% vs. 33%), but not in those with normal transaminase levels (0% vs. 9.5%). It was confirmed that Chinese patients with normal transaminase levels respond very poorly to interferon alfa therapy. However, the response was significantly better in patients with elevated transaminase levels.