Interleukin 1 (IL-1), IL-6, and tumour necrosis factor (TNF) alpha are pleiotropic cytokines produced predominantly by macrophages which have been implicated in the pathogenesis of rheumatoid arthritis (RA). Sulphasalazine has been shown to have disease modifying properties and to inhibit the production of cytokines in vitro. To evaluate the effect of sulphasalazine on cytokine production in vivo, serum cytokine levels were measured in a group of patients with RA entered into a randomised controlled trial. Serum levels of IL-1 alpha, IL-1 beta, IL-6, and TNF alpha were measured at baseline and at two monthly intervals for six months in 17 patients receiving sulphasalazine and in 22 patients treated with placebo. The two groups of patients had a similar age and sex distribution, had had RA for less than a year, had no joint erosions, and had not been treated previously with any other disease modifying drugs. In the 39 patients studied IL-1 alpha was detected (> 0.1 ng/ml) at baseline in 14 patients (median 0.24 ng/ml), IL-1 beta in 25 patients (median 1.0 ng/ml), TNF alpha in 27 patients (median 1.2 ng/ml), and IL-6 in 33 patients (median 0.44 ng/ml). In the group treated with sulphasalazine there was a progressive and significant decline in serum IL-1 alpha, IL-1 beta, and TNF alpha levels over the six month period (median levels at six months were < 0.1, 0.12, and 0.44 ng/ml respectively). Interleukin 6 levels were significantly reduced only at the four month time point (median level of 0.23 ng/ml). These reductions were associated with improvements in clinical and laboratory measures of disease activity. In contrast patients receiving the placebo showed no changes in serum cytokine levels and no improvement in clinical and laboratory indices of disease activity. These results suggest that sulphasalazine may exert its disease modifying effect partly by suppressing cytokine production in vivo.