A novel risk model to predict first-ever ischemic stroke in heart failure with reduced ejection fraction

Area of Science:

• Cardiology
• Neurology
• Medical Informatics

Background:

• Patients with heart failure, particularly heart failure with reduced ejection fraction (HFrEF), face a substantially elevated risk of ischemic stroke.
• Existing stroke risk prediction models, such as CHADS ₂ and CHA ₂DS ₂-VASc, may not accurately identify all high-risk HFrEF patients.

Purpose of the Study:

• To develop and validate a novel, more accurate risk prediction tool for ischemic stroke specifically in patients with HFrEF.
• To compare the performance of the new risk score against established CHADS ₂ and CHA ₂DS ₂-VASc scores.

Main Methods:

• Retrospective analysis of electronic medical records from 6087 HFrEF patients between January 2009 and February 2019.
• Development of a nomogram-based stroke risk score integrating key predictors identified through Cox regression analysis.
• Comparison of the novel score's predictive accuracy using Area Under the Receiver Operating Characteristic (AUROC) curves against CHADS ₂ and CHA ₂DS ₂-VASc scores.

Main Results:

• The study identified 468 first-ever ischemic strokes over 8007.2 person-years of follow-up (incidence 5.8% events/pts-years).
• A novel nomogram incorporating age, atrial fibrillation, deep vein thrombosis, d-dimer, anticoagulant use, and spontaneous echocardiographic contrast/left ventricular thrombus demonstrated superior predictive performance.
• The new risk score achieved higher AUROC values (0.727-0.728) compared to CHADS ₂ (0.515-0.540) and CHA ₂DS ₂-VASc (0.547-0.562) for predicting stroke at hospitalization, 30 days, and 6 months (all p<0.001).

Conclusions:

• The developed stroke risk score significantly improves the prediction of first-ever ischemic stroke in HFrEF patients.
• This novel tool offers a more accurate and refined method for identifying high-risk HFrEF individuals compared to current standard scores.
• Implementation of this score can aid in better risk stratification and potentially guide more personalized stroke prevention strategies in HFrEF management.