Modified CHADS and CHADS-VASc scores to predict atrial fibrillation in acute ischemic stroke patients.

It is important to identify candidates who warrant extended cardiac monitoring after ischemic stroke. We investigated the predictive performance of the CHADS and CHADS-VASc scores for previously unknown atrial fibrillation during in-hospital electrocardiographic monitoring. Patients were selected from a prospective trial in China. The clinical prediction of the scores was examined using the C statistic. Multivariate logistic regressions were performed to analyze the relevant risk factors. Among 1315 patients enrolled in study, previously unknown atrial fibrillation was detected in 110 (8.4%). Age, heart failure, NIHSS on admission, creatinine, and triglycerides were independently associated with newly detected atrial fibrillation. For newly detected atrial fibrillation, the C statistic value was 0.55 (OR 1.14, 95% CI: 0.97-1.33) for CHADS and 0.62 (OR 1.26, 95% CI: 1.12-1.42) for CHADS-VASc; adding newly identified risk factors to these two scores, the value of C statistic was improved to 0.74 and 0.75, respectively. Age, heart failure, NIHSS on admission, creatinine and triglycerides were independent predictors of previously unknown atrial fibrillation. The CHADS and CHADS-VASc scores are useful but not optimal for atrial fibrillation prediction. Addition of newly identified risk factors to these two scores resulted in significant improvement of the predictive performance.