Leptin, the ob gene product, has been proposed as a mediator of inflammatory cytokine-dependent decreased food intake and cachexia in rodents. In humans, leptin serum levels increase after administration of tumor necrosis factor-alpha (TNF-alpha) or interleukin-2 or during septicemia. However, the effect of human chronic inflammatory disease on serum leptin is unknown. We therefore determined the serum leptin level (radioimmunoassay), body mass index (BMI), percent body fat ([%BF] bioelectrical impedance analysis), and disease activity (Disease Activity Score [DAS]) in 58 patients with rheumatoid arthritis (RA) and 16 controls. The BMI, %BF, serum leptin, and ratio of leptin to %BF (leptin/%BF) did not differ significantly in 25 patients with moderate RA activity (DAS, 3.6 +/- 0.5), 33 patients with low RA activity (DAS, 1.8 +/- 0.5), and controls. A positive correlation for serum leptin and %BF was detected in all groups. Our data indicate that in RA, a human chronic cytokine-mediated inflammatory disease, the serum leptin level is directly related to %BF but not to disease activity.