• atrial fibrillation;
  • epidemiology;
  • female sex;
  • stroke;
  • thromboembolism

Summary.  Background:  Female sex has been suggested as a risk factor for stroke/thromboembolism in patients with non-valvular atrial fibrillation (AF) and has therefore been included within risk scores, e.g. the CHA2DS2-VASc score, and guidelines.

Objectives:  To investigate the risk of stroke/thromboembolism associated with female sex in non-valvular AF patients.

Patients/Methods:  Using the national Danish registers, we identified non-anticoagulated patients discharged with non-valvular AF (1997–2008), and subdivided the population into three age intervals: < 65, 65–74 and ≥ 75 years. We calculated stroke rates according to sex, and assessed the stroke risk associated with female sex by using Cox regression analysis.

Results:  We included 87 202 AF patients, and 44 744 (51.3%) were female. The rate of stroke/thromboembolism for females aged < 65 and 65–74 years was not increased as compared with men, whereas the rate for females aged ≥ 75 years was increased. At both 1-year and 12-year follow-up, female sex did not increase the risk of stroke for patients aged < 75 years. At 1-year follow-up, the hazard ratios associated with female sex were 0.89 (95% confidence interval [CI] 0.70–1.13) and 0.91 (95 CI 0.79–1.05) for patients aged < 65 and 65–74 years, respectively, and being female and aged ≥ 75 years was associated with an increased risk of stroke of 1.20 (95 CI 1.12–1.28).

Conclusion:  Female sex was only associated with an increased risk of stroke for AF patients aged ≥ 75 years. Our study suggests that female sex should not be automatically included as an independent stroke/thromboembolic risk factor in guidelines or in the CHA2DS2-VASc score, without careful prior consideration of the ‘age < 65 and lone AF’ criterion.