A randomised controlled trial of induced hypermagnesaemia following aneurysmal subarachnoid haemorrhage

Background: The effect of serum magnesium concentration on the incidence of cerebral arterial vasospasm following aneurysmal subarachnoid haemorrhage (SAH) is unclear. Objective: To test whether induced hypermagnesaemia reduces the incidence of cerebral arterial vasospasm following aneurysmal SAH. Methods: The study was conducted at two tertiary hospitals in Australia and patients were recruited between 1 April 2005 and 31 December 2009. Within 72 hours of aneurysmal SAH, patients were randomly assigned to a high or normal target for serum magnesium concentration (1.60-2.50 mmol/L or 0.65-1.05 mmol/L, respectively). The primary end point was cerebral arterial vasospasm diagnosed by blinded assessment of digital subtraction angiography. Secondary outcomes included severity of vasospasm and functional recovery at 90 days. Analysis was by intention to treat. Results: Of 162 patients, 81 were assigned to the normal range group and 81 were assigned to the high-range group; the primary outcome was available for 78 and 79 patients, respectively. The groups had similar baseline characteristics. Vasospasm occurred in 40 patients (50.6%) and 50 patients (64.1%) assigned to high-range and normal-range groups, respectively (adjusted OR, 0.51; 95% CI, 0.26-1.02; P = 0.06). At 90 days, neurological recovery between the groups was not significantly different (adjusted OR for worse outcome, 0.71; 95% CI, 0.39-1.32; P = 0.28). Patients in the high-range group were treated with more noradrenaline to support arterial blood pressure (79 [16-218] mg) v 59 [14-129] mg; P = 0.03) and had lower mean (SD) serum calcium concentration (1.9 [0.2] mmol/L v 2.1 [0.2] mmol/L, P < 0.001). Conclusion: Patients assigned a higher serum magnesium concentration had a reduced incidence of vasospasm as seen by angiography, but the difference was not statistically significant. Clinically significant outcomes were not different between groups. A firm recommendation for induced hypermagnesaemia cannot be made from this study.