A prospective controlled, double-blind multicenter trial compared placebo, auranofin (an orally administered gold complex), and parenteral gold sodium thiomalate (GST) in patients with active rheumatoid arthritis (RA). Of 193 patients who received any treatment, the only important improvement identified for either auranofin or GST was for pain/tenderness scores. When 161 patients who completed 20 weeks of treatment were examined, both auranofin and GST treatments were superior to placebo as measured by improvement in number of painful and/or tender joints, joint pain/tenderness scores, physician's assessment of disease activity, and decrease in erythrocyte sedimentation rate when elevated at entry. GST was superior to placebo in improvement of joint swelling scores, anemia, thrombocytosis, and rheumatoid factor. No drug-related remissions were observed. The only statistically significant advantages of GST over auranofin for efficacy were an increase in hemoglobin concentration and decrease of thrombocytosis with GST. Withdrawals for adverse effects were 5 times more frequent with GST treatment. Thrombocytopenia, proteinuria, elevated liver enzymes, "nitritoid" reactions, and "gold pneumonitis" were observed only in the GST treatment group. These results confirm that both parenteral and oral gold may be effective for the treatment of RA, that GST tends to show greater efficacy than auranofin, and that auranofin has fewer significant adverse effects than GST. However, long-term benefits, tolerability, and safety cannot be inferred from this study.