Randomized clinical trial of intravenous magnesium sulfate as an adjunctive medication for emergency department treatment of migraine headache

Study objective: We test the hypothesis that intravenous magnesium sulfate is an effective adjunctive medication for treatment of acute migraine. Methods: In this randomized, double-blind, placebo-controlled trial, adults presenting to 2 urban emergency departments with headache meeting International Headache Society criteria for acute migraine received either 20 mg of intravenous metoclopramide plus 2 g of intravenous magnesium sulfate or 20 mg of intravenous metoclopramide plus a placebo of intravenous saline solution at 15-minute intervals for a maximum of 3 doses or until pain relief occurred. At 0, 15, 30, and 45 minutes, patients recorded pain intensity using a standard visual analog scale (VAS). The primary study end point was the between-group difference in pain improvement when initial and final VAS scores were compared. Results: 0f 44 patients enrolled (21 randomized to metoclopramide plus magnesium and 23 to metoclopramide plus placebo), 42 (95%) were women. Baseline features were comparable in both groups. Each group experienced a more than 50-mm improvement in VAS score during the study. However, this improvement was smaller in the magnesium group for the primary end point (16-mm difference favoring placebo [95% confidence interval (Cl)-2 to 34 mm]), as was the proportion with normal functional status at their final rating (36% absolute difference also favoring placebo [95% Cl 7% to 65%]). Using a 50% reduction in pain to dichotomize VAS scores, the number needed to harm with magnesium plus metoclopramide versus metoclopramide alone is 4 patients (95% Cl 2 to 36). Conclusion: Although this result was unexpected, our data suggest that the addition of magnesium to metoclopramide may attenuate the effectiveness of metoclopramide in relieving migraine. Countertherapeutic cerebral vasodilatation caused by magnesium is a plausible, although unproven, explanation for this finding. Because of the preponderance of women in our trial, these data may not be generalizable to men.