Previous clinical trials demonstrated 90-day benefits for endovascular therapy in acute ischemic stroke with large-core infarcts, but protocols relied on advanced multimodal perfusion imaging.1–4 Selection based on noncontrast computed tomography (NCCT) is an alternative that could broaden access to mechanical thrombectomy.5 In the Thrombectomy for Emergent Salvage of Large Anterior Circulation Ischemic Stroke (TESLA) randomized clinical trial, which used an NCCT-alone selection strategy and compared endovascular therapy with medical management (MM) alone for patients with large-core infarcts, the primary outcome assessed at 90 days did not meet the criterion for statistical superiority.6 Because recovery or decline after large-core stroke may extend beyond 90 days, this article presents the prespecified 1-year secondary efficacy and safety outcomes.
TESLA was a multicenter, open-label, blinded end point randomized clinical trial conducted at 47 US stroke centers.6 Eligible patients were aged 18 to 85 years, presenting within 24 hours of last known well with a National Institutes of Health Stroke Scale score of 6 or higher, internal carotid artery or middle cerebral artery occlusion, Alberta Stroke Program Early Computed Tomography Score (ASPECTS) of 2 to 5 on baseline NCCT, and premorbid modified Rankin Scale (mRS) score of 0 to 1. Participants were randomized 1:1 to undergo intra-arterial thrombectomy (IAT) plus best MM vs MM alone. All patients provided informed consent. The trial protocol was previously published.6 TESLA followed the Consolidated Standards of Reporting Trials (CONSORT) guidelines.
The primary 1-year end point was the mean utility-weighted mRS score (range, 0-10; higher scores indicate preferable functional states), calculated via standard utility multipliers.7 Prespecified secondary end points assessed at 1 year included functional independence (mRS 0-2), independent ambulation (mRS 0-3), mRS ordinal shift, European Quality of Life 5 Dimensions, 5 Levels (EuroQol EQ-5D-5L) index score (range, 0-100; higher scores indicate better quality of life),8 and all-cause mortality.
Analyses evaluated the intention-to-treat population with 1-year data and a per-protocol cohort. The main outcome used a baseline ASPECTS-adjusted bayesian model; a posterior probability of .975 or greater defined statistical superiority.
Exploratory frequentist secondary analyses used χ2 or Fisher exact tests for categorical variables, Wilcoxon rank-sum tests for continuous distributions, and a proportional odds model for ordinal shift. Sensitivity analyses (eMethods in Supplement 1) included per protocol; core-lab ASPECTS adjudication; multivariable adjustment for age, diabetes, and site-adjudicated ASPECTS; last observation carried forward; and tipping-point analysis. A 2-sided P < .05 indicated statistical significance. Analyses were performed using SAS version 9.4 (SAS Institute).
Of 302 randomized patients, 300 were included in the intention-to-treat cohort (152 IAT; 148 MM).6 Complete 1-year functional data were available for 144 patients in the IAT group and 133 in the MM group; 23 individuals were lost to follow-up or withdrew consent. Among the 277 participants, baseline profiles were balanced, although the IAT group was slightly younger (median [IQR] age, 66 [54-74] vs 68 [59.5-76.5] years) and had a higher proportion of diabetes (28.5% vs 16.8%).
The primary 1-year end point, mean utility-weighted mRS score, was higher in the IAT vs MM group (3.65 vs 2.78), a bayesian adjusted mean difference of 1.18 points (95% credible interval, 0.42-1.93; posterior probability of superiority, .999) (Table).
