Abstract
Objective: We aimed to assess the impact of time to endovascular thrombectomy (EVT) on clinical outcomes in the DAWN trial, while also exploring the potential effect modification of mode of stroke onset on this relationship. Methods: The association between every 1-h treatment delay with 90-day functional independence (modified Rankin Scale [mRS] score 0-2), symptomatic intracranial hemorrhage, and 90-day mortality was explored in the overall population and in three modes of onset subgroups (wake-up vs. witnessed vs. unwitnessed). Results: Out of the 205 patients, 98 (47.8%) and 107 (52.2%) presented in the 6 to 12 hours and 12 to 24 hours time window, respectively. Considering all three modes of onset together, there was no statistically significant association between time last seen well to randomization with either functional independence or mortality at 90 days in either the endovascular thrombectomy (mRS 0-2 1-hour delay OR 1.07; 95% CI 0.93-1.24; mRS 6 OR 0.84; 95% CI 0.65-1.03) or medical management (mRS 0-2 1-hour delay OR 0.98; 95% CI 0.80-1.14; mRS 6 1-hour delay OR 0.94; 95% CI 0.79-1.09) groups. Moreover, there was no significant interaction between treatment effect and time (p = 0.439 and p = 0.421 for mRS 0-2 and 6, respectively). However, within the thrombectomy group, the models that tested the association between time last seen well to successful reperfusion (modified Treatment in Cerebral Infarction >= 2b) and 90-day functional independence showed a significant interaction with mode of presentation (p = 0.013). This appeared to be driven by a nominally positive slope for both witnessed and unwitnessed strokes versus a significantly (p = 0.018) negative slope in wake-up patients. There was no association between treatment times and symptomatic intracranial hemorrhage. Interpretation: Mode of onset modifies the effect of time to reperfusion on thrombectomy outcomes, and should be considered when exploring different treatment paradigms in the extended window.