Background: Accurate risk stratification is considered the first and most important step in the management of patients with non-valvular atrial fibrillation (NVAF). We compared the performance of the widely used CHA2DS2-VASc and the recently developed R2CHADS2 and ATRIA scores, for predicting thromboembolic (TE) event in either non-anticoagulated or anticoagulated patients with NVAF.
Methods: The non-anticoagulated cohort was comprised of 154 patients, whereas 911 patients formed the cohort of patients on vitamin-K-antagonist. The scores were computed using the criteria mentioned in their developmental cohorts. Measures of performance for the risk scores were evaluated at predicting TE event.
Results: In the non-anticoagulated cohort, 9 TE events occurred during 11 ± 2.7 months. CHA2DS2-VASc showed significant association with TE occurrence: hazard ratio (HR) = 1.58 (95 % confidence interval [95 % IC] 1.01-2.46), but R2CHADS2 and ATRIA did not (HR = 1.23 (95 % CI 0.86-1.77) and 1.20 (95 % CI 0.93-1.56), respectively. In the anticoagulated cohort, after 10 ± 3 months of follow up, 18 TE events were developed. In that cohort, the three scores showed similar association with TE risk: HR = 1.49 (95 % CI 1.13-1.97), 1.41 (95 % CI 1.13-1.77) and 1.37 (95 % CI 1.12-1.66) for CHA2DS2-VASc, R2CHADS2 and ATRIA, respectively. In both cohorts, no TE event occurred in patients classified in the low risk category according to CHA2DS2-VASc or R2CHADS2.
Conclusions: In this study of NVAF patients, CHA2DS2-VASc has better association with TE events than the new R2CHADS2 and ATRIA risk scores in the non-anticoagulated cohort. CHA2DS2-VASc and R2CHADS2 can identify patients at truly low risk regardless of the anticoagulation status.