Methods	Multicentre, randomised, double-blind, parallel-group. Single dose to treat single attack Medication administered when migraine headache pain was of moderate or severe intensity Assessments at 0.5, 1, 2, 2,5, 3, 4, and 24 h after dosing Second dose of study medication available after 2 h for those who had not obtained relief Rescue medication (excluding ergotamine, dihydroergotamine, sumatriptan, or steroids) available 1 h after second injection if relief was still inadequate At the 1 h evaluation, intramuscular prochlorperazine edisylate (10 mg) or metoclopramide hydrochloride (10 mg) could be given for vomiting
Participants	Aged 18 to 65 years, meeting IHS criteria for migraine (1988) with or without aura. At least 1-year history of migraine (untreated severity ≥ moderate) with 1 to 6 attacks per month Prophylactic medication for migraine was permitted providing there were no changes in the medication for at least 2 weeks before study dosing Participants were excluded if experienced aura phase with a duration longer than 1 h, were currently using serotonin reuptake inhibitors, or if they used opioid or other analgesics for more than 3 days per week The use of any form of ergot alkaloid or sumatriptan was prohibited in the 72 h preceding study drug administration, as well as use of antiemetics and narcotic analgesics in the 24 h preceding administration N = 310 M 38, F 272 (88%) Mean age 41 years Migraine without aura was the principal headache diagnosis Although all participants had moderate or severe baseline pain intensity, there was a difference in the distribution of moderate and severe pain between groups, therefore the authors adjusted pain ratings for baseline values (no further details)
Interventions	Sumatriptan 6 mg, n = 158 (150 for efficacy) Subcutaneous dihydroergotamine (DHE) mesylate 1 mg, n = 152 (145 for efficacy)
Outcomes	Headache relief (at 1 and 2 h) Improvement in nausea and vomiting at 1 h Adverse events Withdrawals
Notes	Oxford Quality Score: R1, DB1, W1. Total = 3.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not reported
Study size	Unclear risk	Treatment groups 50 to 200 participants