Value of CHADS-VASc score in predicting stroke recurrence in first-ever ischemic stroke survivors without atrial fibrillation / 南方医科大学学报

ABSTRACT

OBJECTIVE@#To test the performance of CHADS-VASc score in predicting stroke recurrence in first-ever ischemic stroke survivors without atrial fibrillation (AF).@*METHODS@#A total of 768 patients were included in this study, including 475 male (61.85%) and 293 female patients (38.15%) with a mean age of 61.52±12.59 years (17-90 years). The baseline information of the patients was collected by face-to-face questionnaire survey and electronic medical record review, and their follow-up information was collected by telephone follow-up once every 3 months. Chi-square test and Wilcoxon rank sum test were used to compare the baseline characteristics between the patients regularly followed up and those withdrawn from the study. The area under the ROC curve (AUC), sensitivity, specificity, positive likelihood ratio and other indicators of CHADS-VASc score were determined, and C-index based on Cox proportional hazards model was used to evaluate the performance of CHADS- VASc score in predicting the risk of stroke recurrence in patients at different time points during the follow-up.@*RESULTS@#The 1-year, 3-year, and 5-year recurrence rate of stroke was 10.59%, 20.45%, and 29.46% in these patients, respectively. The AUC value, Optimal Operating Point (OOP) and the corresponding positive likelihood ratios (LR+) for predicting stroke recurrence were 0.558 (95% 0.492-0.624), 4.5, and 1.256 at 1 year; 0.574 (95% 0.517-0.630), 4.5, and 1.397 at 3 years; and 0.604 (95% 0.548-0.661), 4.5, and 1.655 at 5 years, respectively. Sensitivity analysis of congestive heart failure showed that the predictive effect of CHADS-VASc score was basically unchanged.@*CONCLUSIONS@#CHADS-VASc score can be used to predict the 3-year and 5-year risk of stroke recurrence in first-ever ischemic stroke survivors without AF. The score has a better prediction ability for long-term stroke recurrence, but the prediction value remains low, suggesting the need to further improve the score or establish a new score for predicting stroke recurrence.