Objective.- To reexamine the efficacy of terminating migraine headache by administration of sumatriptan during the visual-aura phase of the attack. Background.- Although the antimigraine action of triptans is most effective soon after onset of the headache, treatment during the aura phase has been found to be ineffective. Methods.- Nineteen subjects having migraine with aura were studied using a 4-way crossover, open-label design. Each patient was asked to treat 8 consecutive attacks with 100 mg of sumatriptan RT: 3 attacks treated at a timing of the patient's discretion (baseline); 1 attack treated 4 hours after onset of pain (late); 2 attacks treated within 1 hour of onset of pain (early); 2 attacks treated during the aura phase - before the onset of pain (aura). Pain level and cutaneous allodynia were reported by the patients at the onset of pain, at the time of treatment, and 2 and 24 hours after treatment. Results.- Sumatriptan treatment during the aura preempted the development of headache in 34/38 (89%) attacks. The same patients were rendered pain-free in 30/38 (79%) of attacks treated within 1 hour of pain onset, and in 4/19 (21%) of attacks treated 4 hours after the onset of pain. The incidence of allodynia at the time of treatment was 2/38 (5%) in attacks treated during aura, 8/38 (21%) in attacks treated early, and 14/19 (74%) in attacks treated late. Conclusion.- Considering the discrepancy between the present and previous clinical studies, it is worthwhile revisiting the efficacy of preemptive triptan therapy during the aura phase of migraine attacks, using larger-scale, 3-way, crossover, placebo-controlled studies.