Novel composite scoring system to predict unknown atrial fibrillation in acute ischemic stroke patients.

OBJECTIVES

The aim of the study is to ascertain whether a new risk-stratification scheme that combination of CHADS2 or CHA2DS2-VASc score and neuroimaging features could reliably predict unknown AF in patients with acute ischemic stroke.

METHODS

From October 2013 to June 2015, patients from 20 hospitals throughout China were selected from the study of Atrial Fibrillation Screening in the patients with acute ischemic stroke and transient ischemic attack (TIA). Brain imaging characteristics in prediction of AF detection were analyzed with multivariate logistic regressions. The clinical prediction of the models was examined using the C-statistic.

RESULTS

Among the 661 patients, previously undetected AF was in 53 patients (8.02%). Age, history of congestive heart failure, acute cortical involvement, acute insular cortex involvement, and prior cortical infarction were independently associated with newly detected AF. Using CHADS2 and CHA2DS2-VASc scores, C statistics for newly detected AF was 0.54 (OR 1.11, 95% CI: 0.88-1.38) and 0.58 (OR 1.17, 95% CI: 0.99-1.38), respectively; adding newly identified AF-associated imaging characteristics to the two scores, the value of C statistic was significantly improved to 0.74 (P<0.001) and 0.75 (P<0.001), respectively.

CONCLUSION

Both CHADS2 and CHA2DS2-VASc scores are useful but not optimal for predicting newly detected AF in patients with acute ischemic stroke. A newly identified embolic features combining with the CHADS2 or CHA2DS2-VASc score for predicting AF in patients with stroke is superior to the CHADS2 or CHA2DS2-VASc score alone.