Abstract TP278: Final infarct volume reduction may mediate EVT effects on mortality

Background: Recent trials of endovascular stroke therapy (EVT) have shown reduced rates of mortality compared to medical management, a finding in contrast with earlier studies. Because earlier studies had also suggested EVT-derived clinical benefits were largely not mediated by effects on final infarct volume (FIV), here we determine how FIV mediates mortality outcomes in a more modern cohort of EVT treatments. Methods: From our prospectively collected multi-center registry across four comprehensive stroke centers (CSC) in the Houston area, we identified patients from 2017 to 2023 with large vessel occlusion acute ischemic stroke (LVO AIS) who underwent thrombectomy. A large infarct core was defined by CTP as exceeding 70 ml or by CT ASPECTS < 6. This study conducted a mediation analysis to investigate the relationship between EVT (independent variable), 90-day mortality (outcome), and FIV (mediator). Bootstrapping was used to estimate indirect effect, adjusting for confounders, including age and sex. Separate mediation analyses were performed for severe and non-severe subgroups. Results: Among 616 patients with LVO AIS who underwent EVT, median NIHSS was 14 [8,20 IQR], age was 67 [57,77 IQR] and 47.2% were female. Median FIV was 20.38 [ 4.26, 60.20 IQR] and 19.2 % of patients suffered mortality. In the overall cohort, the total effect of EVT on 90-day mortality was not statistically significant (P = 0.46 unadjusted; P = 0.68 adjusted). EVT significantly influenced FIV with a P-value < 0.001, and FIV was a strong predictor of mortality with a P-value < 0.001. However, when accounting for FIV, the direct effect of EVT on mortality was not significant (P = 0.17 unadjusted; P = 0.31 adjusted). The impact of EVT on 90-day mortality was partially mediated via FIV. The indirect effect of EVT on mortality through FIV was significant (P < 0.001). Subgroup analyses indicated that in severe stroke patients (NIHSS ≥ 16), EVT did not significantly affect FIV (Path A, P = 0.21), precluding mediation analysis. In non-severe stroke patients (NIHSS < 16), EVT had a significant impact on FIV (P < 0.001), and FIV was a significant predictor of mortality (P = 0.026). In this group, 77.7% of EVT’s effect on mortality was mediated by FIV. Conclusions: Reductions in FIV partially mediate the effect of EVT on 90-day mortality, particularly in patients with NIHSS < 16. These findings support the results of recent clinical trials demonstrating mortality benefit with reperfusion.