We compared epidemiologic and clinical data from 310 patients with migraine and 30 patients with acute migrainous stroke to identify factors predictive of migrainous stroke and assess the risk of future stroke in these two populations. We found no significant differences in gender ratio or in the frequency of smoking, estrogen use, hypertension, mitral valve prolapse, or family history of migraine between the two groups. A history of migraine with aura was significantly more common in the migrainous stroke group (24 of 30 [80%] versus 142 of 310 [46%]; p < 0.001), as was a history of prior stroke (nine of 30 [30%] versus four of 310 [1.3%]; p < 0.001). We followed 173 of the migraine patients for at least 1 year and a mean of 35.8 months, and no strokes occurred in this group. We followed 28 of the migrainous stroke patients for a mean of 25.3 months, and there were six recurrent strokes in that group, all again migraine-associated. Migrainous stroke is more common in individuals with aura than in those who are aura-free, but this association is of little value in attempting to distinguish patients destined for migrainous stroke from the migraine population at large. Patients with a history of migraine-associated stroke are at significantly increased risk for recurrent stroke.