Bleeding and stroke risk in a real-world prospective primary prevention cohort of patients with atrial fibrillation

BACKGROUND: All stroke risk stratification schemes categorize a history of stroke as a "truly high" risk factor. Therefore, stratifying stroke risk in atrial fibrillation (AF) should perhaps concentrate on primary prevention. However, the risk factors for stroke also lead to an increase in the risk of bleeding. Our objective was to evaluate the agreement among the currently used stroke risk stratification schemes in "real-world" patients with AF in the primary prevention setting, their correlation with adverse events recorded during warfarin treatment, and the relationship between stroke and bleeding risk. METHODS: We prospectively followed up 3,302 patients with AF taking warfarin for primary prevention. Stroke risk was assessed using the CHADS(2) (congestive heart failure, hypertension, age ≥ 75 years, diabetes, previous stroke or transient ischemic attack), Atrial Fibrillation Investigators, American College of Chest Physicians Conference on Antithrombotic and Thrombolytic Therapy, American College of Cardiology/American Heart Association/European Society of Cardiology, and National Institute for Health and Clinical Excellence schemas, and for bleeding risk, the outpatient bleeding risk index was calculated. Bleeding and thrombotic events occurring during follow-up were recorded. RESULTS: Patients classified into various stroke risk categories differed widely for different schemes, especially for the moderate- and high-risk categories. The rates of bleeding and thrombotic events during follow-up were 1.24 and 0.76 per 100 patient-years, respectively. All stroke stratification schemes correlated closely to bleeding risk. Stroke rate increased progressively from low- to moderate- to high-risk patients. CONCLUSIONS: Stroke risk stratification models differed widely when categorizing subjects into the moderate- and high-stroke-risk categories. Bleeding and stroke risk were closely correlated and both were low among low-risk patients and were similarly high among moderate/high-risk groups.