Female sex-related stroke risk in atrial fibrillation: a thing from the past?

Abstract Background Whether female sex should be considered in stroke risk stratification among patients with atrial fibrillation (AF) is a subject of equipoise. Recent evidence suggests that stroke risk associated with female sex may be age-dependent and appears to have declined among patients diagnosed with AF in more recent years. We hypothesized that the observed decline in stroke risk associated with female sex and its age-dependency may reflect two sides of the same coin, potentially stemming from a cohort effect. Purpose To investigate whether sex differences in stroke risk are associated with the birth year of patients with AF. Methods The FinACAF retrospective, registry-based cohort study includes all patients diagnosed with AF in Finland between 2007 and 2018. We used a standard age-period-cohort approach and applied Poisson regression model incorporating time scales with natural splines for age, calendar year of the observation period, and birth year, along with an interaction term between sex and birth year. Ischemic stroke rate ratios comparing women to men were computed across the range of birth years. Results We identified 229 565 patients with new-onset AF (50.0% women; mean birth year 1942, SD 13.5; mean age at AF diagnosis 72.7 years, SD 13.2; mean follow-up 4.0 years, SD 3.2 years), of whom 7.1% experienced an ischemic stroke during follow-up. A significant interaction between sex and birth year was observed (p 0.001). The association between female sex and stroke decreased across the range of birth years (Figure: panel A unadjusted analysis and panel B adjusted analysis). In those born before 1940, women had significantly higher stroke rates than men, but this association diminished and became non-significant around the 1940–1960 birth years, with even a trend toward lower stroke rates in women born in the 1980s. Conclusions In this nationwide cohort, the higher stroke rate traditionally associated with female sex in AF was mainly observed in those born before 1940. Female sex may not need to be considered for stroke risk stratification or oral anticoagulant decision-making in contemporary and potentially future patients with AF.Incidence rate ratio of ischemic stroke