Intravenous glibenclamide for cerebral oedema after large hemispheric stroke (CHARM): a phase 3, double-blind, placebo-controlled, randomised trial

Background: No treatment is available to prevent brain oedema, which can occur after a large hemispheric infarction. Glibenclamide has previously been shown to improve functional outcome and reduce neurological or oedema-related death in patients younger than 70 years who were at risk of brain oedema after an acute ischaemic stroke. We aimed to assess whether intravenous glibenclamide could improve functional outcome at 90 days in patients with large hemispheric infarction. Methods: CHARM was a phase 3, double-blind, placebo-controlled, randomised trial conducted across 143 acute stroke centres in 21 countries. We included patients aged 18–85 years with a large stroke, defined either by an Alberta Stroke Program Early CT Score (ASPECTS) of 1–5 or by an ischaemic core lesion volume of 80–300 mL on CT perfusion or MRI diffusion-weighted imaging. Patients were randomly assigned in a 1:1 ratio to either intravenous glibenclamide (8·6 mg over 72 h) or placebo. The study drug was started within 10 h of stroke onset. The primary efficacy outcome was the shift in the distribution of scores on the modified Rankin Scale at day 90, as a measure of functional outcome. The primary efficacy outcome was analysed in a modified intention-to-treat population, which included all randomly assigned patients aged 18–70 years. The safety population comprised all randomly assigned patients who received a dose. This trial is registered with ClinicalTrials.gov (NCT02864953). The trial was stopped early by the sponsor for strategic and operational reasons (slow enrolment because of COVID-19), before any unblinding or knowledge of the trial results. Findings: Between Aug 29, 2018, and May 23, 2023, 535 patients were enrolled and randomly assigned, of whom 518 received a dose (safety population) and 431 were aged 18–70 years and comprised the modified intention-to-treat population (217 were assigned glibenclamide and 214 placebo). The mean age of patients was 58·7 (SD 9·0) years in the placebo group and 58·0 (9·5) years in the glibenclamide group; the median US National Institutes of Health Stroke Scale (NIHSS) score was 19 (IQR 16–23) in the placebo group and 19 (IQR 16–22) in the glibenclamide group; and the mean time from stroke onset to study drug start was 8·9 h (SD 2·1) in the placebo group and 9·2 h (2·1) in the glibenclamide group. Intravenous glibenclamide was not associated with a favourable shift in the modified Rankin scale at 90 days (common odds ratio [OR] 1·17 [95% CI 0·80–1·71], p=0·42). 90-day mortality was 29% (61 of 214) in the placebo group and 32% (70 of 217) in the glibenclamide group (hazard ratio 1·20 [0·85–1·70]; p=0·30). Serious adverse events in the prespecified safety population were consistent with the known safety profile of glibenclamide and included hypoglycaemia in 15 (6%) of 259 patients in the glibenclamide group and in four (2%) of 259 patients in the placebo group, leading to dose interruption or reduction in seven (3%) patients in the glibenclamide group and in one (<1%) in the placebo group. Interpretation: Intravenous glibenclamide did not improve functional outcome in patients aged 18–70 years after large hemispheric infarction, although the trial was underpowered to make definitive conclusions because it was stopped early. Future prospective evaluation could be warranted to identify a possible benefit of intravenous glibenclamide in specific subgroups. Funding: Biogen.