Outcomes of endovascular thrombectomy for acute ischemic stroke with concurrent intracranial hemorrhage: multicenter registry study.

Abstract

Patients with large vessel occlusion (LVO) and concurrent intracranial hemorrhage are systematically excluded from thrombectomy trials, leaving a critical evidence gap. We examined the association between endovascular thrombectomy (EVT) and clinical outcomes in this population, and derived a preliminary hemorrhage volume threshold to inform treatment selection. This multicenter, retrospective cohort study included 85 patients with acute LVO and intracranial hemorrhage (EVT: n=49; best medical management (BMM): n=36) across 18 stroke centers. The primary endpoint was 90 day functional independence (modified Rankin Scale score of 0-2). Multivariable logistic regression and inverse probability of treatment weighting (IPTW) were applied to partially mitigate selection bias in these two groups. A receiver operating characteristic derived hemorrhage threshold was internally validated via bootstrap iterations. Baseline groups (EVT vs BMM) differed in hemorrhage volume (median 5.0 vs 3.0 mL; P=0.019), reflecting confounding by treatment preference. Despite this, EVT was consistently associated with greater 90 day functional independence across all analytical approaches (53.1% vs 19.4%; IPTW adjusted OR 4.2, 95% CI 1.49 to 11.85; P=0.007). Mortality difference did not reach significance after adjustment (IPTW OR 0.51; P=0.219). Hemorrhage expansion and symptomatic intracranial hemorrhage rates were comparable. A preliminary threshold of 4.5 mL (AUC 0.70) identified patients with higher functional independence (66.7% vs 25.0%; P=0.015) and lower mortality (3.0% vs 31.2%; P=0.011). In this hypothesis generating cohort, EVT was associated with favorable outcomes in patients with LVO and concurrent intracranial hemorrhage, particularly those with hemorrhage ≤4.5 mL. The non-randomized design and modest sample preclude definitive conclusions. Prospective validation is needed.