Abstract
Transient ischemic attack (TIA) carries a substantial risk of early ischemic stroke, particularly within the first 48 hours. Although the ABCD2 scoring system is widely used for short-term risk stratification, its exclusion of laboratory markers may limit predictive accuracy. To investigate whether selected biochemical markers - thyroid function tests, lipid profiles, and Hemoglobin A1c - hold prognostic value for early post-TIA stroke, and whether their integration enhances the predictive performance of the ABCD2 scoring system. This retrospective cohort included 130 first-ever TIA patients admitted between March 2022 and 2024. ABCD2 scores and laboratory parameters were recorded at admission. Patients were followed for 30 days and categorized based on stroke timing: early (≤48 hours), intermediate (3-7 days), and late (8-30 days). Receiver operating characteristic (ROC) curve analysis and binary logistic regression were utilized to assess predictive performance and determine independent risk factors. Ischemic stroke occurred in 52 patients (40%) within 30 days, with 65% of events arising in the first 48 hours. Patients who developed stroke had significantly elevated triglyceride (TG) levels (P = .038), while those with early stroke exhibited significantly lower free thyroxine (fT4) levels (P < .001). ROC analysis identified TG ≥ 152 mg/dL (area under the curve (AUC): 0.607) and fT4 ≤ 1.17 ng/dL (AUC: 0.708) as critical thresholds. Integration of these biomarkers into the ABCD2 model significantly improved its performance: the AUC rose from 0.647 (ABCD2 alone) to 0.751 (with TG) and 0.775 (with fT4). Notably, fT4 demonstrated a high negative predictive value (88.6%) for ruling out early stroke. Elevated TG and reduced fT4 levels were independently associated with increased stroke risk after TIA. Incorporating these biomarkers into the ABCD2 scoring system significantly enhanced early stroke prediction. In particular, low fT4 may serve as a robust marker to exclude high-risk cases. These findings support a multimodal risk stratification strategy incorporating clinical, biochemical, and imaging data. Validation through prospective multicenter studies is recommended.