Stroke risk prediction in patients with atrial fibrillation with and without rheumatic heart disease

Aims Patients with atrial fibrillation (AF) and rheumatic heart disease (RHD), especially mitral stenosis, are assumed to be at high risk of stroke, irrespective of other factors. We aimed to re-evaluate stroke risk factors in a contemporary cohort of AF patients. Methods and results We analysed data of 15 400 AF patients presenting to an emergency department and who were enrolled in the global RE-LY AF registry, representing 47 countries from all inhabited continents. Follow-up occurred at 1 year after enrolment. A total of 1788 (11.6%) patients had RHD. These patients were younger (51.4 +/- 15.7 vs. 67.8 +/- 13.6 years), more likely to be female (66.2% vs. 44.7%) and had a lower mean CHA(2)DS(2)-VASc score (2.1 +/- 1.7 vs. 3.7 +/- 2.2) as compared to patients without RHD (all P<0.001). Significant mitral stenosis (average mean transmitral gradient 11.5 +/- 6.5 mmHg) was the predominant valve lesion in those with RHD (59.6%). Patients with RHD had a higher baseline rate of anticoagulation use (60.4% vs. 45.2%, P<0.001). Unadjusted stroke rates at 1 year were 2.8% and 4.1% for patients with and without RHD, respectively. The performance of the CHA(2)DS(2)-VASc score was modest in both groups [stroke at 1 year, c-statistics 0.69, 95% confidence interval (CI) 0.60-0.78 and 0.63, 95% CI 0.61-0.66, respectively]. In the overall cohort, advanced age, female sex, prior stroke, tobacco use, and non-use of anticoagulation were predictors for stroke (all P<0.05). Mitral stenosis was not associated with stroke risk (adjusted odds ratio 1.07, 95% CI 0.67-1.72, P=0.764). Conclusion The performance of the CHA(2)DS(2)-VASc score was modest in AF patients both with and without RHD. In this cohort, moderate-to-severe mitral stenosis was not an independent risk factor for stroke.