Magnesium and Hematoma Expansion in Intracerebral Hemorrhage: A FAST-MAG Randomized Trial Analysis.

ABSTRACT

BACKGROUND: Observational studies suggest that magnesium may have hemostatic effects. FAST-MAG (Field Administration of Stroke Therapy-Magnesium) was a pragmatic clinical trial of magnesium sulfate administered prehospital for acute clinical stroke syndromes and included patients with intracerebral hemorrhage. Exploratory secondary analysis by the treatment group found no reduction in hematoma expansion (HE) associated with magnesium treatment in intracerebral hemorrhage but did not consider serum magnesium levels achieved. We analyzed FAST-MAG intracerebral hemorrhage data for associations between serum magnesium level, HE, and early neurological deterioration, accounting for groupwise biases. METHODS: HE was defined as hematoma volume increase ≥3 mL within 24 hours and early neurological deterioration as ≥1-point Glasgow Coma Scale decline from arrival to hospital day 4. Comparing treatment and placebo groups confirmed biased availability of neuroimaging data. Therefore, HE and neurological deterioration were analyzed and stratified by treatment and placebo groups using univariate tests and adjusted logistic regression. RESULTS: Spontaneous intracerebral hemorrhage was present in 381 patients. Placebo patients had fewer serial neuroimaging studies available (123 [65.4%] versus 145 [75.1%]; P=0.038). Necessary data were available in 104 magnesium- and 85 placebo-treated patients (age, 64.9 [13.0] years; 67.7% male). In the magnesium group, higher magnesium level was associated with less HE (adjusted odds ratio, 0.64 per mg/dL [95% CI, 0.42-0.93]) and less neurological deterioration (adjusted odds ratio, 0.54 per mg/dL [95% CI, 0.33-0.82]). In the placebo group, magnesium level was not associated with either HE or neurological deterioration. CONCLUSIONS: Magnesium may exhibit a hemostatic effect that was only observable in the FAST-MAG magnesium treatment group. Equipoise should be maintained, and specific trials are needed. REGISTRATION URL https//www.clinicaltrials.gov; Unique identifier NCT00059332.