10 Years Before Symptoms Appear... New Technology Predicts Heart Disease During Sleep

Your heartbeat during sleep may carry signals about health risks you could face years later. Researchers have found that artificial intelligence can analyze the heart’s electrical activity during sleep to help predict the likelihood of cardiovascular disease and death over the next ten years.

The study, published in the journal SLEEP with support from the U.S. National Institutes of Health, suggests that data from electrocardiograms (ECGs) already recorded during sleep tests could be leveraged to improve early detection of health risks.

Researchers later tested the model on two independent groups: approximately 10,000 patients from Emory University Hospital and 13,000 patients from a medical center in Boston. The system relied on ECG data recorded during sleep to classify participants according to their long-term cardiovascular risk levels.

Researchers explained that relying on a single ECG signal could allow extraction of predictive information from existing tests, without necessarily requiring a separate test for this purpose.

Results showed that higher risk scores assigned by the model were associated with increased probabilities of developing several conditions over the following ten years.

For each one-standard-deviation increase in the risk score, the likelihood of developing atrial fibrillation, a heart rhythm disorder, rose by 103% to 172%. This was also associated with a 19% to 40% increased risk of stroke, a 69% to 133% increased risk of heart failure, and a 65% to 82% increased risk of death from any cause.

However, no significant increase in the future risk of myocardial infarction was observed except in one of the study groups. Researchers emphasized that adding the model’s risk score to known risk factors improved the ability to predict cardiac outcomes, potentially helping to identify individuals who would benefit from more proactive health monitoring.

Sleep tests, particularly polysomnography (PSG), also record the heart’s electrical activity, although this data is not always used to estimate future cardiovascular risks.

Researchers found that heart signals during the early and light stages of sleep may carry important information; heart rate patterns and electrical conduction change with the rapid shifts the nervous system undergoes as it transitions between sleep stages.

When ECG data extracted from sleep was added to current risk assessment models, the ability to predict atrial fibrillation improved by up to 6%, heart failure by 9%, and death by 7%.

Although these improvements appear modest, they may help better distinguish between individuals at higher risk of complications and those at lower risk.

Researchers hope to integrate this technology into routine sleep tests in the future, but note that its contribution to improving risk prediction compared to traditional methods must be evaluated before widespread adoption in clinical practice.