A network meta-analysis of therapeutic and prophylactic management of vasospasm on aneurysmal subarachnoid hemorrhage outcomes

Abstract

BackgroundVasospasm and cerebral ischemia after aneurysmal subarachnoid hemorrhage are associated with mortality and poor neurological outcomes. We studied the efficacy of all available strategies targeting vasospasm and cerebral ischemia on outcomes in a network meta-analysis.MethodsWe searched EMBASE and MEDLINE databases from 1 January 1990 and 28 November 2021 according to PRISMA guidelines. Randomized controlled trials and longitudinal studies were included. All curative or preventive strategies targeting vasospasm and/or cerebral ischemia were eligible. A network meta-analysis was performed to compare all interventions with one another in a primary (randomized controlled trials only) and a secondary analysis (both trials and longitudinal studies). Mortality by 3 months was the primary outcome. Secondary outcomes were vasospasm, neurological outcome by 3 months, and dichotomized as “good” or “poor” recovery according to each study definition.ResultsA total of 2,382 studies were screened which resulted in the selection of 192 clinical trials (92 (47.9%) and 100 cohorts (52.1%) and the inclusion of 41,299 patients. In randomized controlled studies, no strategy decreased mortality by 3 months. Statins (0.79 [0.62–1]), tirilazad (0.82 [0.69–0.97]), CSF drainage (0.47 [0.29–0.77]), and clazosentan (0.51 [0.36–0.71]) significantly decreased the incidence of vasospasm. Cilostazol was the only treatment associated with improved neurological outcomes by 3 months in the primary (OR 1.16, 95% CI [1.05–1.28]) and secondary analyses (OR 2.97, 95% CI [1.39–6.32]).DiscussionIn the modern era of subarachnoid hemorrhage, all strategies targeting vasospasm failed to decrease mortality. Cilostazol should be confirmed as a treatment to improve neurological outcomes. The link between vasospasm and neurological outcome appears questionable.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=116073, identifier: PROSPERO CRD42018116073.