Refining the CHA2DS2VASc risk stratification scheme: shall we drop the sex category criterion?

Abstract

Aims: The CHA2DS2VASc score is recommended for stroke risk stratification in patients with atrial fibrillation (AF). This score assigns one extra point to female sex based on evidence from the early 2000s, suggesting higher thromboembolic risk in women. This incremental risk of thromboembolism in women has decreased over time between 2007 and 2018, becoming non-significant in recent years. The objective of this study was to assess the impact of removing the sex category (Sc) from the CHA2DS2VASc score, thus validating a non-sex CHA2DS2VASc (i.e. CHA2DS2VA) score. Methods and results: We analysed UK primary and secondary care data comprising 195 719 patients with AF followed between 1998 and 2016 (mean age: 75.9 ± 12.3 years; 49.2% women). Among 126 428 non-anticoagulated patients, we compared the CHA2DS2VASc vs. CHA2DS2VA scores every calendar year. Throughout 413 007 patient-years, a total of 8742 events of ischaemic stroke or systemic embolism were recorded. Sex differences in thromboembolic risk were not observed in the lower-risk population, but higher stroke rates were consistently seen in female patients in the higher-risk category (i.e. CHA2DS2VA ≥2). C-statistics for both CHA2DS2VA and CHA2DS2VASc scores were similar over the years (ranging from 0.62 to 0.71). With CHA2DS2VA, no relevant differences were observed in integrated discrimination improvement, and net reclassification improvement (NRI) resulted in improved reclassification (11%) in lower thromboembolic risk groups. The NRI suggested misclassification in higher thromboembolic risk patients (-7%), but this did not affect their indication for anticoagulation (i.e. patients retained their high-risk status). Conclusion: Removing Sc from the CHA2DS2VASc score does not affect its ability to discriminate thromboembolic events in the population with AF. The use of CHA2DS2VA may simplify initial decision-making for thromboprophylaxis.