Time-trends of sex differences in stroke risk in atrial fibrillation: Are female patients still at higher risk today? A nationwide cohort study

Abstract Background Female sex is associated with higher risk of stroke in patients with atrial fibrillation (AF), particularly in the presence of other stroke risk factors. This has implications for oral anticoagulant treatment (OAC) decision-making, based on points of the CHA2DS2-VASc score. However, stroke rates have declined over recent years, and management has evolved over time towards a holistic approach that may have diminished the sex differences in AF-related stroke. Purpose To examine the changes in stroke risk over time in relation to risk differences between males and females not treated with oral anticoagulants in a nationwide cohort study. Methods This was an observational cohort study including patients with a hospital diagnosis of AF between 2007 and 2021. Data were obtained from the Danish nationwide registries. We identified a cohort of adult patients who were not receiving OAC treatment for up to 30 days after their AF diagnosis. Patients were followed up for thromboembolic events within one year after the AF diagnosis. Survival analysis was applied to describe the yearly rate of events and to estimate the risk ratio for females vs males of thromboembolism according to baseline risk levels, i.e., the CHA2DS2-VA score level (leaving out the ‘Sc’ sex criterion). Results A total of 188,704 AF patients (median age 75 years (IQR: 64 to 84); 49.7% female) who did not receive OAC treatment were included. The risk of thromboembolism during the entire study period was overall higher in females (3.5% in males vs. 4.7% among females). The overall one-year thromboembolic risk declined over time from 4.9% between 1997-2000 to 3.3% between 2017-2020, with male-female risk differences less evident in recent years (Figure 1a). The relative risk of stroke diminished from 1.22 (95%CI 1.10-1.34) in years 1997-2000 to 1.02 (95%CI 0.85-1.19) in years 2017-2020 for patients with a CHA2DS2-VA score of ≥2 (Figure 1b). For patients with a score of 2, the relative risk was lowest between 2005-2008 (1.02 (95%CI 0.63-1.42)) and was 1.18 (95%CI 0.54-1.83) in the latest time period. Formal statistical test for time trends according to score level revealed non-significant results (p-value=0.10 for score 0-1; p-value=0.17 for score levels of ≥2). Conclusion The risk of stroke among patients with AF has been declining over the last two decades. The excess stroke risk observed among females relative to males has also diminished, most evident among AF patients with ≥2 prevalent stroke risk factors.