Application and Comparison of the CHADS2 and CHA2DS2-VASc Risk Scores in a Population with Atrial Fibrillation

Abstract

IntroductionThe CHADS2 score and the CHA2DS2-VASc score recently adopted by the medicalcommunity have been developed with international registry data and are widely usedin clinical practice. However, they have not been evaluated in national registries.ObjectivesThe aims of this study were first to evaluate the predictive power of the CHADS2and CHA2DS2-VASc stroke risk scores in the Atrial Fibrillation Registry conductedby the Argentine Society of Cardiology Research Area and second to compare bothscoring systems.MethodsThe Atrial Fibrillation Registry of 2001 was a multicenter, prospective study of allconsecutive patients with chronic atrial fibrillation (permanent and persistent)treated in 70 medical centers in Argentina. Demographic data, socioeconomic characteristics,background and clinical features were obtained. A 2-year follow-up wasperformed to assess stroke rate. For the present analysis patients without anticoagulanttreatment were selected. In this population, the two risk score systems wereassessed; a ROC curve was built for each score (reported as C statistic) and bothscoring systems were compared.ResultsThe study population consisted of 303 patients (49.3 %) not receiving anticoagulanttherapy. The stroke rate in the selected population was 9.5%. Both scoring systemspredicted significant stroke risk.The stroke rate increased as the CHADS2 and the CHA2DS2-VASc scores were higher,and were similar in both risk scales.The CHADS2 and CHA2DS2-VASc scores had C statistic values of 0.67 (0.55-0.78)and 0.69 (0.59 to 0.78), respectively, without significant differences between them.The score analyses divided into three risk profiles -low, moderate and high- revealedthat the predictive power decreased markedly. The C statistic of the CHADS2 was0.63 (95% CI 0.57-0.68) and that of the CHA2DS2-VASc score was 0.57 (95% CI 0.51-0.62, with a slightly better predictive trend for the CHADS2 score but without statisticalsignificance.ConclusionsThe two scoring systems used to predict stroke in an Argentine population of patientswith persistent and permanent atrial fibrillation have a similar predictivepower comparable to results reported in the literature.