To analyse the incidence and risk factors of clinical rebound and hepatic decompensation during or upon withdrawal of prednisolone pretreatment before interferon (IFN) therapy, two series of Taiwanese patients with chronic viral hepatitis from two independent randomized controlled trails were compared. Group 1 included 41 patients with chronic hepatitis B who were pretreated with daily prednisolone (30 mg) for 3 weeks, 15 mg for 1 week and no prednisolone for 2 weeks prior to lymphoblastoid IFN therapy. Group 2 consisted of 59 patients with chronic hepatitis B who were pretreated with daily prednisolone (40 mg) for 2 weeks, 30 mg prednisolone for 2 weeks, 20 mg prednisolone for 2 weeks and no prednisolone for 2 weeks prior to INF alpha-2a therapy. Clinical rebound developed more frequently in group 2 (67.8%) than in group 1 patients (41.5%; P < 0.01). The peak serum transaminase levels of group 1 and 2 patients during clinical rebound were similar. Icteric and symptomatic clinical rebound occurred in four (one cirrhotic) group 2 patients. The incidence of hepatic decompensation was 3.4% in group 2 patients, or 5.0% in group 2 patients with clinical rebound. Patients pretreated with a higher dose (40 mg) of prednisolone (odds ratio 3.0; 95% CI 1.3-6.6; P < 0.01) and non-cirrhotic patients (odds ratio 6.2; 95% CI 1.2-32.1; P < 0.02) tended to suffer from clinical rebound more frequently. However, once clinical rebound develops in cirrhotic patients, the relative risk of decompensation is 16 times that of non-cirrhotic patients. These results suggest that clinicians should be cautious in prescribing a short course of corticosteroids for patients with chronic viral hepatitis, because hepatic decompensation might occur in Oriental people with or without cirrhosis.