In a double blind noncrossover study, dietary supplementation with fish oil (18 g/day), was compared with an olive oil supplement over a 12-week period in patients with rheumatoid arthritis receiving established conventional therapies. An improvement in tender joint score and grip strength was seen at 12 weeks in the fish oil treated group but not in the olive oil treated group. The more subjective measures of mean duration of morning stiffness and analogue pain score improved to a similar extent in both groups, although statistical significance was only achieved in paired analyses in the olive oil treated group. Production of leukotriene B4 by isolated neutrophils stimulated in vitro was reduced by 30% in the fish oil treated group and unchanged in the olive oil treated group.