This study was performed to evaluate the long-term effects of interferon-alpha 2a (IFN-alpha 2a) vs no treatment in patients with chronic hepatitis B and to determine whether viral clearance, following therapy or occurring spontaneously, was sustained. Patients originating from three previously published multicentre, randomized, controlled trials were analysed. Information about survival and response during long-term follow-up was available in 340 (73%) and 308 (66%) of 469 randomized patients respectively. Response to therapy (viral clearance) was defined as: loss of hepatitis B virus (HBV) DNA and loss of hepatitis B e antigen (HBeAg) and improvement in alanine aminotransferase level. Scheduled treatment-free follow-up was 12 months in all studies. Median long-term follow-up time after inclusion in the individual studies was 4.7 years (range: 0.2-7.5 years). Viral clearance after IFN-alpha 2a, or occurring spontaneously, was sustained in 70 out of 80 evaluable patients (88%) who were responders at the end of the original trials and 21 (30%) lost hepatitis B surface antigen (HBsAg). A total of 80 patients received (re)treatment during the long-term follow-up period and 33% of them responded, irrespective of previous treatment category. Overall response rate was not significantly affected by gender, sexual inclination or ethnic origin. Durability of response did not depend upon ethnic origin or presence of cirrhosis. At the end of the original trial periods, 253 patients were histologically evaluated and 22 (9%) had histologically confirmed progression to cirrhosis. During long-term follow-up an additional five patients developed cirrhosis. Hepatocellular carcinoma developed in three patients (1%): in one patient during the follow-up period of the original trial and in two patients (one untreated) during the long-term follow-up period. Ten of 25 deaths were liver-related (hepatocellular carcinoma in three, gastrointestinal bleeding in two and liver failure in five). The distribution of clinical events (progression to cirrhosis, hepatocellular carcinoma and liver-related deaths) was unrelated to original treatment category and response to treatment. Hence, 90% of responding patients will, irrespective of treatment category, have a sustained response. At least 30% of responding patients will eventually lose HBsAg. For a number of reasons, the present patient population and observation period are insufficient to establish a presumed beneficial effect of IFN-alpha 2a on disease progression and survival.