A novel risk prediction model for post-treatment ischemic stroke recurrence using plasma antithrombin III and thromboelastography.

BACKGROUND: The primary aim of this investigation was to build and test a novel risk prediction model incorporating plasma antithrombin III (ATIII) activity and thrombelastography (TEG)-derived parameters. This approach seeks to enhance the ability to stratify the risk of stroke recurrence among ischemic stroke (IS) patients receiving dual antiplatelet therapy (DAPT). METHODS: In this prospective cohort study, 200 consecutive patients diagnosed with non-cardiogenic IS were recruited during a one-year period (May 2024 to May 2025). All participants had their ATIII activity, TEG parameters, and coagulation function indicators tested within 24-72 hours of DAPT commencement. After initial variable screening via univariate analysis. A logistic regression-based risk model was built, with its ability to distinguish outcomes evaluated using the receiver operating characteristic (ROC) curve. RESULTS: Recurrent events occurred in 21.00% (42/200) of cases within 3 months post-DAPT Multivariate analysis established ATIII, TEG-LY30, and D-Dimer as independent risk factors and TEG-MA as protective. The resultant model exhibited superior predictive power (A U C = 0.9480, 95% C I= 0.9l48~ 0.9813; sensitivity 90.48% , specificity 86.71%). Internal validation, yielding AUCs of 0.9521 in the training set and 0.9437 in the validation set, verified the model's strong generalizability. Subgroup evaluations further revealed the model's robust performance in both large artery atherosclerosis (AU C= 0.9505) and small vessel occlusion (AUC=0.9395) subtypes. CONCLUSIONS: Integrating ATIII activity with TEG parameters enables the development of a novel model to predict the risk of IS relapse following DAPT.