In a blind, placebo-controlled study of dietary manipulation therapy in outpatients with rheumatoid arthritis there was significant objective improvement during periods of dietary therapy compared with periods of placebo treatment, particularly among "good responders". Possible explanations for improvement include reduced food intolerance, reduced gastrointestinal permeability, and benefit from weight loss and from altered intake of substrates for prostaglandin production. A proportion of the improvement was due to a placebo response, but this was not sufficient to explain the whole improvement.