Liquid Embolic Volume Does Not Predict Outcome After Middle Meningeal Artery Embolization for Chronic Subdural Hematoma

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Purpose: Middle meningeal artery embolization (MMAE) is increasingly used for chronic subdural hematoma (CSDH), but the relationship between injected liquid-embolic volume and outcomes is unclear. We evaluated dose–outcome associations and searched for a minimum effective dose. Methods: This retrospective multicenter cohort included 313 patients treated with liquid embolic agents at 12 MESH Registry centers from 2016 to 2024. MMAE was performed alone in 172 patients (55.0%) and adjunctively with surgical evacuation in 141 (45.0%). Dose was assessed as total injected volume per patient and normalized to middle meningeal artery diameter, baseline hematoma thickness, and estimated hematoma volume. The primary outcome was a 180-day composite of retreatment, residual or recurrent hematoma > 10 mm, or death. Logistic regression with multiple imputation, restricted cubic splines, and sensitivity analyses evaluated associations and nonlinearity. Results: Median injected volume was 0.60 mL (interquartile range, 0.50–0.90). Center accounted for 44.5% of variation in log volume, whereas hematoma burden accounted for 0.3%. No dose metric was significantly associated with the primary outcome or 90-day recurrence/reoperation. For absolute volume, the odds ratio per interquartile-range increment was 0.80 (95% confidence interval, 0.56–1.16; p = 0.24). No reliable minimum effective dose was identified. Findings remained consistent across sensitivity analyses and treatment indications (interaction p = 0.30). Procedural complications occurred in 6.4% and were not dose-associated. Conclusion: Within the observed range, injected volume alone was not a consistent determinant of outcome and should not be a procedural target after a satisfactory angiographic endpoint. Future studies should evaluate embolic distribution, distal penetration, and vascular territory coverage.

This publication has 20 references indexed in Scilit: