One hundred and fifty patients treated for an acute migraine attack all received a standard regimen consisting of metoclopramide 10 mg i.m. and—20 min later—paracetamol 1 g orally. Simultaneous trial medication was given. This consisted of placebo, diazepam 5 mg as tablets, or chlormezanone 400 mg as capsules given in a double dummy fashion. A successful treatment result was defined as a decrease of two or three steps on a four-point verbal pain report scale, and no need for further treatment. Patients with severe pain (all three treatment groups together) showed significantly better treatment results (p < 0.001) than patients with pain of medium severity. In the group with severe pain there was no significant difference between diazepam, chlormezanone or placebo. In the group with pain of medium severity there was a significant (p < 0.01) effect of chlormezanone, but no significant effect of diazepam. Chlormezanone may possibly be a valuable addition to antiemetic and analgesic therapy of migraine attacks.