Abstract Body: Introduction: Prior attempts to assess risk of symptomatic intracranial hemorrhage (sICH) after endovascular thrombectomy have been limited by reliance on information obtained after thrombectomy (such as TICI reperfusion) and lack of generalizability to routine clinical practice.
Methods: Patients presenting to GWTG-Stroke participating hospitals between July 2021 and June 2023 with last known well within six hours prior to presentation, who received endovascular thrombectomy were included. The primary outcome was sICH; secondary outcomes included in-hospital mortality, mRS at discharge, and length of stay. The study population was divided into a derivation and validation cohort with 70:30 partition. According to a pre-specified statistical analysis plan, a full model of 31 candidate variables and subsequently a highly parsimonious model including only variables measured before EVT deployment was fit for each endpoint, with variable retention guided by multiple factor analysis (MFA). Models were then externally validated in the HERMES clinical trial population.
Results: 31,668 patients (median age 71 [Q1: 61, Q3: 81]) were included, of whom 1,799 (5.7%) developed sICH. In the validation cohort, the area under the receiver operating characteristics curve (AUC) for the full model was 0.649 (Table 1), and the AUC for the simplified points score was 0.589 (Table 2). At the conference, we will present results of external validation and secondary endpoints, details of model calibration, and direct comparisons to existing risk scores.
Conclusions: A risk score for sICH after thrombectomy for acute stroke devised using routinely collected data known prior to intervention had good performance compared to existing approaches.