The purpose of this study was to determine whether selected antirheumatic drugs would suppress elevated circulating interleukin-6 (IL-6) levels in patients with rheumatoid arthritis (RA). The 267 patients who enrolled in a double-blind randomized protocol received placebo, naproxen (1500 mg/day), or prinomide (1500 mg/day) for up to 16 weeks. Serum samples from 143 of the patients completing the trial and from 135 normal donors were analyzed by quantitative sandwich enzyme-linked immunosorbent assay for IL-6 concentrations. A mean normal IL-6 value was determined to be 3 pg/ml (95th percentile value = 10 pg/ml). IL-6 levels at baseline for the patients with RA were significantly higher than those for control subjects (p < 0.0001). Elevated IL-6 levels (> 10 pg/ml) at baseline were found in 80% of subjects with RA (median = 36 pg/ml, range 12 to 403). For patients with elevated levels of either IL-6, C-reactive protein (CRP), or erythrocyte sedimentation rate (ESR) at baseline, initial median values of IL-6, CRP, and ESR were compared with those from the final visit for each treatment group. There was no significant decrease in IL-6 levels with treatment. Median CRP levels decreased significantly, from 1.9 to 0.8 mg/dl (p = 0.002), as did median ESR (37 to 34 mm/hr, p = 0.013), only in the prinomide-treated group.