Background This study aimed to evaluate and compare the predictive performance of the National Institutes of Health Stroke Scale (NIHSS) assessed at baseline, at 24 h, and derived change metrics for 90-day unfavorable functional outcome (modified Rankin Scale 3–6) in cardioembolic stroke patients with Anterior Circulation Occlusion (ACO) post-endovascular thrombectomy (EVT). Methods A retrospective analysis of 103 eligible patients was performed. Univariate and multivariate logistic regression identified predictors. Receiver operating characteristic (ROC) curve analysis and DeLong’s test compared the predictive performance of baseline NIHSS, 24-h NIHSS, ΔNIHSS (baseline NIHSS − 24-h NIHSS) and the percent ΔNIHSS (ΔNIHSS × 100/baseline NIHSS). Results Multivariate analysis confirmed 24-h NIHSS, baseline NIHSS, ΔNIHSS, and percent ΔNIHSS as independent predictors. ROC analysis showed that 24-h NIHSS had the highest predictive power (AUC = 0.850), significantly outperforming baseline NIHSS (AUC = 0.702), ΔNIHSS (AUC = 0.735), and percent ΔNIHSS (AUC = 0.780). The optimal cut-off value was ≥12, with 82.2% sensitivity and 75.6% specificity. Combining 24-h NIHSS with other NIHSS-based metrics did not improve predictive performance compared to 24-h NIHSS alone. Conclusion The 24-h NIHSS score is the strongest prognostic predictor of 90-day unfavorable functional outcome in cardioembolic stroke patients post-EVT, superior to baseline scores, ΔNIHSS and percent ΔNIHSS. It serves as an early and effective tool for prognostic stratification.
Yang et al. (Wed,) studied this question.