Early Response to Endovascular Thrombectomy after Stroke: Early, Late, and Very Late Time Windows.

Background And Purpose: Endovascular thrombectomy (EVT) has benefits in selected patients 6-24 h after stroke onset. However, the response to EVT >24 h after stroke onset is still unclear. We compared the early response to EVT in patients with different time windows.

Methods: Patients who underwent EVT in an emergency setting were enrolled and categorized according to when EVT was performed: within 6 (early), 6-24 (late), and >24 h (very late) after stroke onset. Early neurological improvement (ENI) and deterioration (END) were defined as improvement and worsening, respectively, of National Institutes of Health Stroke Scale (NIHSS) score by ≥4 points after EVT. The three groups' clinical characteristics and response to EVT were compared. We also investigated factors associated with ENI and END.

Results: During study period, 274 patients underwent EVT (109 early, 104 late, and 61 very late). Patients who underwent EVT very late were younger (p = 0.007), had smaller ischemic cores, and had lower initial NIHSS scores (8 ± 5) than those who underwent EVT early (14 ± 6) and late (13 ± 7; p < 0.001). Stroke mechanisms also differed according to the time window (p < 0.001): cardioembolism was more common after early EVT, whereas large-artery atherosclerosis was more prevalent among patients who underwent EVT very late. ENI was significantly more common after early (60.6%) and late EVT (51.0%) than after very late EVT (29.5%; p = 0.001); however, rates of END did not differ (11.0%, 13.5%, and 4.9%, respectively). ENI was independently associated with male, higher NIHSS score, and early and late EVT. END was associated with failure of recanalization.

Conclusions: ENI was more observed and associated with early and late EVT. Highly selected patients receiving very late EVT may not benefit from ENI but may still have a chance to prevent END. The occurrence of END was associated not with time window but with failure of recanalization.