Background: Some patients require retreatment after MMA embolization for subacute/chronic subdural hematoma (SDH). We evaluated clinical, imaging, and angiographic features associated with retreatment. Methods: Post-hoc analysis of EMBOLISE Surgery Cohort subjects treated with Onyx. Groups were defined by retreatment ≤90 days (Retreated, N=8) vs no retreatment (N=189). Variables included demographics, antithrombotic use, SDH morphology, core-lab–read hematoma metrics, embolization extent, Onyx distribution, and late contrast filling after embolization. Fisher’s exact and unpaired t-tests were used. Results: Age (73.9±9.1 vs 72.9±11.1 years, p=0.81), SDH thickness (19.8±7.7 vs 21.3±5.9 mm, p=0.48), midline shift (5.6±2.0 vs 7.1±3.7 mm, p=0.23), Onyx volume (0.88±0.47 vs 0.85±0.66 mL, p=0.93), fluoroscopy time (31.0±15.8 vs 28.3±13.5 min, p=0.58), and procedure time (69.0±29.6 vs 63.7±31.0 min, p=0.64) did not differ. Antiplatelet/anticoagulant use at symptom onset trended higher in the Retreated group (75.0% vs 36.5%, p=0.056). Core-lab angiography showed an association with retreatment: late contrast filling of proximal MMA branches after embolization occurred in 100% of Retreated vs 35.8% of non-Retreated (p=0.0026), whereas a “proximal occlusion only” late-filling pattern was absent in Retreated vs present in 53.6% of non-Retreated (p=0.011). SDH location and morphology and subdural membrane filling change were not different, though “no change” in membrane filling trended higher in Retreated (16.7% vs 0.56%, p=0.064). Conclusions: Retreatment rates after Onyx MMA embolization in the EMBOLISE study were quite low to make meaningful conclusions. Demographics, hematoma thickness, midline shift, and procedural metrics were similar between the retreated vs non-retreated groups. There was a trend toward retreatment being associated with persistent late filling of proximal MMA branches. However, distribution of Onyx was proximal in 50% and distal in 50% of the retreated subjects. Thus, the role of distal embolization is unclear. Further pooling of large studies is needed to understand factors influencing retreatments.
Hassan et al. (2026) studied this question.