ABSTRACT: The performance of scoring systems for risk stratification in patients with atrial fibrillation (AF) was not validated well in patients with stroke. The purpose of this study was to evaluate whether the risk scoring systems predict vascular outcomes in stroke patients with AF.Data were obtained from a nationwide multicenter registry for acute stroke with AF from January 1, 2013, to December 31, 2015. We investigated the predictive power of the CHADS2, CHA2DS2-VASc, ATRIA, and Essen stroke scores in stroke patients with AF. The subjects were further stratified into groups according to treatment with or without oral anticoagulants (OACs).A total of 3112 stroke with AF subjects were included. The rate of recurrent ischemic stroke and any stroke were not associated with the CHADS2, CHA2DS2-VASc, ATRIA, and Essen stroke risk scores. The risks of death and major adverse cerebrovascular and cardiovascular events (MACEs) increased sequentially with the increase of each risk score in OAC group. (the range of C-index 0.544-0.558 for recurrent ischemic stroke; 0.523-0.537 for any stroke; 0.580-0.597 for death; 0.564-0.583 for MACEs). However, in the group treated with OACs, all risk scores were significantly associated with the risk of MACEs. The C-statistics of the 4 scoring systems were 0.544 to 0.558, 0.523 to 0.537, 0.580 to 0.597, 0.564 to 0.583, respectively, for recurrent ischemic stroke, any stroke, death, and MACEs.The performance of the CHADS2, CHA2DS2-VASc, ATRIA, and Essen stroke risk scores for the prediction of recurrent stroke was unsatisfactory in stroke patients with AF whereas the performance for the prediction of recurrent stroke was not MACEs or death was good. A new risk stratification scheme that is specific for secondary stroke prevention in the AF population is needed.