Document resource
Background Coronary heart disease (CHD) affects nearly one-quarter of patients undergoing middle meningeal artery embolization (MMAE) for chronic subdural hematoma (cSDH). Whether CHD impairs hematoma resolution or worsens treatment outcomes is unknown. CHD patients are frequently on antiplatelet and statin medications, which may independently influence outcomes. We sought to determine the independent effect of CHD on cSDH resolution and clinical outcomes after MMAE using causal inference methods.Methods We performed a retrospective cohort study using a multicenter international MMAE registry. The primary outcome was complete SDH resolution at last imaging follow-up. Secondary outcomes included mortality, reoperation, recurrence, good functional outcome (mRS 0-2), and SDH thickness change. Propensity scores for CHD status were estimated using logistic regression with 17 directed acyclic graph-informed covariates including age, sex, hypertension, diabetes, chronic kidney disease, prior stroke, dementia, baseline SDH thickness, laterality, midline shift, and anticoagulant use. Stabilized inverse probability of treatment weights (IPTW) were computed and trimmed at the 1st/99th percentiles. IPTW-weighted risk ratios (RR) with bootstrap 95% confidence intervals (2,000 resamples) constituted the primary analysis. Doubly-robust modified Poisson regression provided confirmatory estimates. Antiplatelet therapy and statins were deliberately excluded from the propensity model as potential mediators; their role was assessed through formal mediation analysis.Results Of 1,781 patients, 1,663 met inclusion criteria (401 CHD, 1,262 no CHD). The propensity score model achieved a C-statistic of 0.755. CHD patients were older (median 76 vs 74 years), more frequently hypertensive (83.8% vs 59.6%), and more often on antiplatelet therapy (63.1% vs 24.2%; all P<0.001). Baseline SDH thickness was similar (15.0 vs 15.0 mm; P=0.87). After IPTW adjustment, complete SDH resolution was virtually identical between groups (34.0% vs 34.1%; RR 1.00, 95% CI 0.82-1.18; P=0.95; figure 1). No secondary outcome differed significantly: mortality (RR 1.15, 0.81-1.57; P=0.44), reoperation (RR 1.09, 0.62-1.74; P=0.78), recurrence (RR 0.86, 0.41-1.48; P=0.61), mRS 0-2 (RR 1.02, 0.93-1.11; P=0.66), and thickness change (−0.2mm difference; P=0.66). The crude mortality signal (unadjusted RR 1.52; P=0.01) was fully attenuated after IPTW, indicating confounding by age and comorbidities. Doubly-robust estimates confirmed all findings. Mediation analysis found no significant indirect effect through the antiplatelet pathway.Conclusions In this large multicenter cohort, CHD had no independent effect on cSDH resolution or any clinical outcome after MMAE. The crude mortality association was attributable to confounding rather than a direct CHD effect. These findings suggest that CHD status should not influence the decision to proceed with MMAE for cSDH.Disclosures B. Demessie: None. A. Karandish: None. H. Salim: None. M. Collaborators: None.Abstract E-032 Figure 1