AI Analysis of Sleep ECGs Predicts 10-Year Heart Failure Risk
A new AI model can predict long-term cardiovascular risks, including heart failure and atrial fibrillation, using single-lead ECG data recorded during standard sleep studies. Published in the journal Sleep, the research demonstrates that analyzing heart activity during the lightest stage of sleep offers predictive value beyond traditional risk factors. The model, trained on data from nearly 16,000 patients, showed significant improvements in predicting outcomes for atrial fibrillation, heart failure, and mortality, though results for stroke and heart attack were less consistent. Researchers suggest this could integrate into routine polysomnography tests for early detection.
