Patients with heart failure are at increased risk for ischemic stroke. We aim to develop a more accurate stroke risk prediction tools identify high-risk patients with heart failure with reduced ejection fraction (HFrEF). Patient data were extracted retrospectively from the electronic medical database between January 2009 and February 2019. Univariate and multivariate Cox regression analysis were performed to identify independent predictors, which were utilized to construct a nomogram for predicting ischemic stroke. AUROC analysis was used to compare the prognostic value between the new risk score and CHADS2/CHA2DS2-VASc scores. In 6087 patients with HFrEF, the risk of first-ever ischemic stroke was 5.8% events/pts-years (n=468) during 8007.2 person-years follow-up. A nomogram constructed by integrating 6 variables, including age, atrial fibrillation (AF), deep vein thrombosis (DVT), d-dimer, anticoagulant use and spontaneous echocardiographic contrast (SEC)/left ventricular thrombus (LVT), exhibited a greater area under the curve of 0.727, 0.728 and 0.714 than that by CHADS2 score (0.515, 0.522 and 0.540), and by CHA2DS2-VASc score (0.547, 0.553 and 0.562) for predicting first-ever ischemic stroke at hospitalization, 30-day and 6-month follow-up (all p<0.001). This novel stroke risk score performed better than existing CHADS2/ CHA2DS2-VASc scores and showed improvement in predicting first-ever ischemic stroke in HFrEF patients.
Keywords: heart failure with reduced ejection fraction; ischemic stroke; predictors; score model.