Validation of contemporary stroke and bleeding risk stratification scores in non-anticoagulated Chinese patients with atrial fibrillation

BackgroundRisk stratification schemes assessing stroke and thromboembolism (stroke/TE) and bleeding relating to atrial fibrillation (AF) have largely been derived and validated in Western populations. We assessed risk factors that constitute scores for assessing stroke/TE (CHADS2, CHA2DS2-VASc) and bleeding (HAS-BLED), and the predictive value of these scores in a large cohort of Chinese patients with AF.Methods and resultsWe studied 1034 AF patients (27.1% female, median age 75; 85.6% non-anticoagulated) with mean follow-up of 1.9 years. On multivariate analysis, vascular disease was independently associated with stroke/TE in non-anticoagulated patients (p = 0.04). In patients with a CHADS2 or CHA2DS2-VASc score = 1, the rate of stroke/TE was 2.9% and 0.9% respectively, but in patients at “high risk” (scores ≥ 2), this rate was 4.6% and 4.5%, respectively. The c-statistics for predicting stroke/TE with CHADS2 and CHA2DS2-VASc were 0.58 (p = 0.109) and 0.72 (p < 0.001), respectively. Compared to CHADS2, the use of CHA2DS2-VASc would result in a Net Reclassification Improvement (NRI) of 16.6% (p = 0.009) and an Integrated Discrimination Improvement (IDI) of 1.1% (p = 0.002). Cumulative survival of the patients with a CHA2DS2-VASc score ≥ 2 was decreased compared to those with a CHA2DS2-VASc score 0-1 (p < 0.001), but the CHADS2 was not predictive of mortality. There was an increased risk of major bleeding with increasing HAS-BLED score (c-statistic 0.61, 95% CI: 0.51-0.71, p = 0.042).ConclusionsVascular disease was a strong independent predictor of stroke/TE in Chinese patients with AF. The CHA2DS2-VASc score performed better than CHADS2 in predicting stroke/TE in this Chinese AF population. Cumulative survival of the patients at high risk with the CHA2DS2-VASc score (but not using CHADS2) was significantly decreased.