Objective: To determine if CHA2DS2-VASc score, a score commonly used to assess risk of cerebrovascular events among adults with atrial fibrillation, predicts ischemic stroke, thromboembolism, and death in a cohort of patients with heart failure with and without atrial fibrillation. Design: Prospective cohort study. Setting and participants: Patients in Denmark aged 50 years or older discharged with a primary diagnosis of incident heart failure between 1 Jan 2000 and 31 December 2012. Patients with atrial fibrillation were identified by a hospital diagnosis of atrial fibrillation or atrial flutter from 1994 onwards. The study excluded patients treated with vitamin K antagonist within 6 months prior to heart failure diagnosis and patients with a diagnosis of cancer or chronic obstructive pulmonary disease. The study utilized 3 national Danish registries: the National Patient Registry (which records all hospital admissions and diagnoses using ICD-10), the National Prescription Registry (prescription data), and the Civil Registry System (demographics and vital statistics). The registries were linked and have been well validated. Main outcome measure: The primary outcome measure was defined as a hospital diagnosis of ischemic stroke or thromboembolic events, transient ischemic attack, systemic embolism, pulmonary embolism or myocardial infarction within 1 year after heart failure diagnosis. A secondary outcome measure was all-cause death at 1 year. Analysis: Patients were risk stratified using the CHA2DS2-VASc score. Patients were given 1 point for congestive heart failure, hypertension, age 65 to 74 years, diabetes mellitus, vascular disease, and female sex and 2 points for age 75 years or older and previous thromboembolic events. The authors conducted a time-to-event analysis to examine the relationship between CHA2DS2-VA S c s c o r e and the risk of ischemic stroke, thromboembolic event, and death separately among those with atrial fibrillation and without. Patients were censored if they began anti-coagulation therapy during follow-up. The properties of CHA2DS2-VASc score in predicting the risk of outcomes were quantified using C statistics. Multiple sensitivity analyses were conducted to account for patients who had a diagnosis of atrial fibrillation shortly after diagnosis of heart failure, to include patients with chronic obstructive pulmonary disease, and split sample analysis by date of heart failure diagnosis was conducted. Main results: A total of 42,987 patients with incident heart failure during 2000-2012 were included in the cohort, with 21.9% of these having atrial fibrillation at baseline. The median follow-up period was 1.8 years. For patients with heart failure with or without a diagnosis of atrial fibrillation, the 1-year absolute risk for all outcomes were high and increased with increasing CHA2DS2-VASc score. For ischemic stroke and death, absolute risks were higher among patient with heart failure and atrial fibrillation when compared with patients without atrial fibrillation. At high CHA2DS2-VASc score, the risk of thromboembolism was higher among patients without atrial fibrillation when compared with those with atrial fibrillation. CHA2DS2-VASc score predicted the end point of ischemic stroke at 1 and 5 years modestly with C statistics 0.67 and 0.69 among those without atrial fibrillation and 0.64 and 0.71 among those with atrial fibrillation. The negative predictive value for all events at 1 year was around 90% when using a cutoff score of 1 for patients without atrial fibrillation, but only around 75% at 5 years. Conclusions: Although the CHA2DS2-VASc score was developed to predict ischemic stroke among patients with atrial fibrillation, it also has modest predictive accuracy when applied to patients with heart failure without atrial fibrillation. Among patients with heart failure with a high CHA2DS2-VASc score, the risks of all adverse outcomes were high regardless of whether concomitant atrial fibrillation was present, and the risk of thromboembo-lism was higher among those without atrial fibrillation than those with concomitant atrial fibrillation. Because of the modest predictive accuracy, the clinical utility of CHA2DS2-VASc among patients with heart failure needs to be further determined.