Background: 
Cryptogenic stroke (CS) patients have a wide spectrum of risk of subsequent stroke, cardiovascular events and death. To date, the is no simple and convenient scoring system to predict the risk of adverse outcome in CS patients. We investigated whether CHADS2 and CHA2DS2-VASc scores were useful predictive tools to assess the risk for adverse events in CS patients.
Methods and results: This nation-wise hospital-based prospective study was conducted at 59 hospitals between August 2006 and May 2016. Totally 3,775 CS patients with ACS were enrolled, and CHADS2 and CHA2DS2-VASc scores were calculated. The primary endpoint was occurrence of major adverse cardiovascular events (MACEs), including all cause death, recurrent stroke or myocardial infarction.  CHADS2 and CHA2DS2-VASc scores were significant predictors of MACEs among CS patients in separate multivariate regression analyses. A Kaplan–Meier analysis of CHADS2 and CHA2DS2-VASc scores of ≥2 showed a higher rate of MACEs as compared with scores of <2 (P<0.001; log-rank test). CHA2DS2-VASc score was better than CHADS2 score in predicting MACEs; the area under the receiver operating characteristic curve increased from 0.58 to 0.61 (p<0.001). Patients with CHADS2 scores of 0 or 1 were further classified according to CHA2DS2-VASc score, using a cutoff value of 2. The rate of MACEs significantly differed between those with a score of <2 and those with a score of ≥2 (5.50 % vs.7.96%, P=0.046).
Conclusions: CHADS2 and CHA2DS2-VASc scores were useful predictors of subsequent MACEs in CS patients.

Prognostic Impact of CHADS2 and CHA2DS2-VASc scores on Cryptogenic Stroke: The Taiwan Stroke Registry