Abstract WP37: Real-World Applicability of Thrombectomy in Anterior Circulation Large Vessel Occlusion Strokes Treated in the Extended Window: Analysis of the Prospective Trevo Registry

Abstract

Background: Two recent trials showed a strong benefit of thrombectomy in the extended window. However, these studies were performed in selected centers and utilized strict inclusion criteria. We aim to evaluate the outcomes of thrombectomy in a large prospective cohort treated outside the rigid clinical trial setting. Methods: Trevo Registry patients with ICA, MCA-M1 or M2 occlusions and pre-morbid mRS0-2 were categorized according to time-from-last-seen-well (TLSW) to puncture as early (0-6hours) vs. late (6-24hours). Uni- and multivariate analyzes were performed to identify good outcome (90-day mRS0-2) predictors. Subgroup analyses were performed for the basic DAWN (age >=18, NIHSS >=10, ICA or M1 occlusion, pre-morbid mRS 0-1) and DEFUSE 3 (age 18-90, NIHSS >=6, ICA or MCA-M1 occlusion, mRS 0-2) trial criteria. Results: As compared to the late (n=430), early patients (n=1173) were older (70 vs 68, p=0.011) and had higher IV tPA use (69 vs 25%, p<0.001), lower smoking frequency (33 vs 40%, p=0.011), larger baseline infarcts (21.2 vs 15.6 cc, p=0.045), less frequent ICA occlusions (18 vs 24%, p=0.015), and a trend towards higher admission NIHSS (16 vs 15, p=0.09). Despite significantly longer TLSW to puncture (3.5 vs 9.6 h, p<0.001), late patients had similar rates of mTICI2b-3 (92 vs 94%, p=0.20), good outcomes (60 vs 56%, p=0.128), symptomatic ICH (1.5 vs.1.4%, p=0.84), and 90-day-mortality (10.9 vs.11.4%, p=0.79). Age (OR 0.96, 95%CI [0.96-0.97]), admission-NIHSS (0.91 [0.89-0.93]), baseline mRS (0.49 [0.40-0.60]), ASPECTS >= 6 (1.37 [1.07-1.75]), DM (0.58 [0.44-0.77]), and time to treatment (0.98 [0.97-1.00]) were independent predictors of good outcomes. Imaging modality did not predict outcomes. Similar findings were observed in the early versus late DAWN-like (n=709 vs 257) and DEFUSE 3-like (n=855 vs 273) cohorts. There was great similarity between the outcomes of the Trevo Registry subsets vs their analogous RCTs: early DAWN-like vs SWIFT Prime (90-day mRS 0-2: 57.5 vs 60%; 90-day mRS 6: 11.% vs 9%), Late DAWN-like vs DAWN (50.2 vs 48.6%; 10.6 vs 18%), and Late DEFUSE 3-like vs DEFUSE 3 (52 vs 45%; 10.3 vs 14%). Conclusions: Our study provides favorable data for the generalizability of the safety and efficacy of thrombectomy in the “real-world” setting.