A double-blind, cross-over, randomized study of acute migraine attack compared treatment results of naproxen with that of placebo. Each treatment period continued for either three months or six migraine attacks, whichever occurred first. The initial dose of naproxen was 750 mg, with additional 250-500 mg doses taken if and when required, to a maximum of five 250 mg tablets within a period of 24 h in each migraine attack. Forty-one patients were enrolled in the study; they had all experienced at least two but not more than eight migraine attacks a month during the preceding year. Thirty-two patients completed the two treatment periods. Naproxen was statistically significantly superior to placebo in reducing the severity of head pain, nausea, and photophobia; in shortening the duration of head pain, nausea, vomiting, photophobia, and lightheadedness; in diminishing the frequency of vomiting; and in decreasing the need for escape medication. Both patient and physician treatment preferences significantly favoured naproxen. Nine side effects were experienced by seven patients while receiving placebo and seven by five patients during naproxen treatment. Mild gastrointestinal discomfort was the main complaint. Only one patient withdrew from treatment because of a side effect, which occurred while receiving placebo.