CHA₂DS₂-VASc score performance to predict stroke after acute decompensated heart failure with and without reduced ejection fraction

Aims: To validate the utility of CHA₂DS₂-VASc score to predict the annual risk of stroke in patients admitted with acute heart failure, comparing those with preserved ejection fraction (HF-presEF) and reduced ejection fraction (HF-redEF). Methods and results: We investigated 2922 patients with known atrial fibrillation who were admitted to the Sheba Medical Center for acute decompensated heart failure (HF). Anticoagulation therapy was prescribed based on CHA₂DS₂-VASc score or physician's discretion. Subjects were divided into four pre-specified groups based on HF type and median CHA₂DS₂-VASc score: HF-presEF with CHA₂DS₂-VASc <5(N = 731), HF-presEF with CHA₂DS₂-VASc ≥5 (N = 1102), HF-redEF with CHA₂DS₂-VASc <5 (N = 563), and HF-redEF with CHADS₂-VASc ≥5 (N = 526). The primary endpoint was an ischaemic stroke at 1 year. Mean age of the study population was 79 ± 11 years, of whom more than half were women. The median CHA₂DS₂-VASc score for the entire study population was 5.0 (interquartile range 25-75%: 4-6). Stroke rate for the entire study population was 6.6%. Multivariate Cox regression proportional hazards regression analysis revealed that in both HF-redEF and HF-presEF patients, each one-point increment in CHA₂DS₂-VASc was associated with a corresponding 28% increase in stroke risk (P < 0.001). The Kaplan-Meier's survival analysis revealed that in the same CHADS₂-VASc category (high vs. low), no difference was found between HF-redEF and HF-presEF with regards to the risk of stroke. Conclusion: Our key finding is that the CHA₂DS₂-VASc score is a valid and powerful predictor of subsequent stroke among patients admitted with acute heart failure decompensation regardless of heart failure type.