OBJECTIVE to determine the efficacy of fish-oil dietary supplements in active rheumatoid arthritis and their effect on neutrophil leukotriene levels. METHODS nonrandomized, double-blinded, placebo-controlled, crossover trial with 14-week treatment periods and 4-week washout periods. METHODS academic medical center, referral-based rheumatology clinic. METHODS forty volunteers with active, definite, or classical rheumatoid arthritis. Five patients dropped out, and two were removed for noncompliance. METHODS treatment with nonsteroidal anti-inflammatory drugs, slow-acting antirheumatic drugs, and prednisone was continued. Twenty-one patients began with a daily dosage of 2.7 g of eicosapaentanic acid and 1.8 g of docosahexenoic acid given in 15 MAX-EPA capsules (R.P. Scherer, Clearwater, Florida), and 19 began with identical-appearing placebos. The background diet was unchanged. RESULTS the following results favored fish oil placebo after 14 weeks: mean time to onset of fatigue improved by 156 minutes (95% confidence interval, 1.2 to 311.0 minutes), and number of tender joints decreased by 3.5 (95% Cl, -6.0 to -1.0). Other clinical measures favored fish oil as well but did reach statistical significance. Neutrophil leukotriene B4 production was correlated with the decrease in number of tender joints (Spearman rank correlation r=0.53; p less than 0.05). There were no statistically significant differences in hemoglobin level, sedimentation rate, or presence of rheumatoid factor or in patient-reported adverse effects. An effect from the fish oil persisted beyond the 4-week washout period. CONCLUSIONS fish-oil ingestion results in subjective alleviation of active rheumatoid arthritis and reduction in neutrophil leukotriene B4 production. Further studies are needed to elucidate mechanisms of action and optimal dose and duration of fish-oil supplementation.