Serum substance P was assayed in rheumatoid arthritis patients and healthy controls to evaluate whether neurogenic inflammation with substance P release is significant in rheumatoid arthritis. A very sensitive competitive immunoenzymetric assay was used. Mean serum substance P level was significantly higher in rheumatoid arthritis patients than in controls and was not correlated with disease duration, morning stiffness duration, Thompson's articular index, Larsen's radiographic score, or the following laboratory indices of inflammation: erythrocyte sedimentation rate, C-reactive protein, and alpha 1-acid glycoprotein. Neurogenic inflammation with substance P release may contribute significantly to the pathogenesis of rheumatoid arthritis. The absence of correlations between serum substance P and clinical or laboratory indices of inflammation may reflect complex interactions between neurogenic inflammation and other pathogenic mechanisms, which may influence clinical features and laboratory tests in rheumatoid arthritis patients.