This prospective unblinded 24-month-study compared the therapeutic value of oral gold with injectable gold to maintain rheumatoid arthritis (RA) patients in clinical remission and prevent the progression of erosive disease. Forty-six patients with definite RA in remission with injectable gold were randomized into two groups: a control group, continued on maintenance injectable gold (Solganal, aurothioglucose, 50-100 mg, intramuscularly, 2 to 4 weeks); and an experimental group, switched to oral gold (6-9 mg auranofin by mouth daily). Only 29% of the original 24 oral gold patients remained on assigned treatment at 24 months compared with 64% of the injectable gold group. By six months, over one-half of the oral gold patients had electively stopped their randomized therapy. Sixty-seven percent of the oral gold patients had adverse reactions, mostly gastrointestinal complaints, compared with one proteinuria in the injectable gold group. The oral gold group experienced significantly more deterioration in all the primary measures of treatment effect over the follow-up period. At the termination of the trial, 88% of the group had increases of 5 or more points in radiographic scores suggesting progression of erosive disease compared with only 29% of the control group. These data suggest that oral gold is not an effective substitute for injectable gold in maintaining remission in rheumatoid arthritis.