Seventy patients with rheumatoid arthritis (RA) completing a 12-week multicenter double blind trial comparing recombinant human interferon-gamma (r-IFN-gamma) with placebo were enrolled in a longterm prospective protocol evaluating r-IFN-gamma in RA. Forty (57%) patients after 1 year and 26 (37%) patients after 2 years continued the drug with sustained clinical benefit. Over 2 years, r-IFN-gamma was discontinued in 44 patients (lack of efficacy--25, withdrawn consent--7, noncompliant--4, suspected adverse drug reactions--2, concurrent illness--6). Two years of treatment with r-IFN-gamma were well tolerated with sustained clinical benefit in some patients with few significant adverse drug reactions.