We measured thyrotropin receptor antibodies in serum obtained from 2 groups of patients participating in clinical trials of recombinant interferon-alpha 2b for viral hepatitis. Group I: Patients with hepatitis B (N = 8), received interferon 5 x 10(6) units thrice weekly for 4 months. Group II: Patients with non-A, non-B hepatitis (N = 16) were randomized to receive interferon in a dose of either 0.25 x 10(6) or 3 x 10(6) U thrice weekly for 6 months and then crossed over to receive the other dosage schedule for a further 6 months. None of the patients developed thyrotoxicosis. Thyrotropin receptor antibody activity was detectable within the "normal range" (less than 10 U/l) in 6 patients prior to treatment. In Group I, thyrotropin receptor antibodies became detectable in 6 patients on treatment, in 4 of whom it was 10 U/l. In Group II, thyrotropin receptor antibody activity was unchanged on low-dose interferon, but on the higher dose became detectable in 9 patients, in 7 of whom it was greater than 10 U/l. We conclude that treatment with interferon is associated with the development of thyrotropin receptor antibodies in a large proportion of patients. It is possible that in some patients treated with higher doses of interferon the increase in thyrotropin receptor antibody activity may be sufficient to induce hyperthyroidism.