Rizatriptan is an effective and fast acting drug for the acute treatment of migraine. As with any other acute treatment for migraine, headache recurrence may occur in up to one-third of responders. Combination with non-steroidal anti-inflammatory drugs (NSAIDs) seems to reduce the incidence of headache recurrence in clinical practice. Rofecoxib is a member of a new class of NSAIDs, which selectively inhibits the COX-2 enzyme and therefore is associated with a lower risk of gastrointestinal side-effects; the drug has a long plasma half-life (17 h). This open label study compared rizatriptan with rizatriptan plus rofecoxib in the acute treatment of migraine. Fifty-six triptan naive patients from a tertiary centre (37 women and 19 men, ages 16–55 years, mean 35 years) with International Headache Society migraine were randomized into two groups. They were instructed to treat three consecutive moderate or severe attacks with either 10 mg rizatriptan (group 1: 18 women and 10 men) or with 10 mg rizatriptan plus 25 mg rofecoxib (group 2: 19 women and 9 men). The presence of headache and nausea at 1, 2 and 4 h, and of side-effects, use of rescue medication and recurrence were compared. Fifty-four patients completed the study. Group 1 treated 76 attacks and group 2 treated 81 attacks. Absence of headache at 1 h was seen in 19 attacks (25%) in group 1 and in 34 attacks (42%) in group 2 (P=0.082); at 2 h absence of headache was seen in 60% of group 1 attacks and in 76% of group 2 attacks (P=0.115). At 4 h, 75% of group 1 attacks and 88% of group 2 attacks were pain free (P=0.122). With regard to nausea, of those who had nausea at baseline, 31% and 49% of attacks in groups 1 and 2, respectively, were nausea free at 1 h (P=0.091), 75% and 79% at 2 h (P=0.736) and 82% and 91% (P=0.479) at 4 h. Recurrence, based on all attacks of those patients who achieved pain free at 4 h, was observed in 53% of group 1 and 20% of group 2 attacks (P<0.001). Sustained pain-free rates (for the 4-h time point) were 45.6% of group 1 and 78.9% of group 2 attacks. There were no significant differences with regard to rescue medication consumption after 4 h and side-effects in both groups. There was a non-significant trend for the combination group to have a higher response rate. The group treated with rizatriptan and rofecoxib had a lower recurrence rate than the group treated with rizatriptan. This study demonstrated that combining a fast acting triptan such as rizatriptan with rofecoxib reduced headache recurrence rates, was well tolerated and may be more effective than the use of rizatriptan alone. Double-blind, placebo-controlled studies are necessary to confirm these observations.