by Vincent Jacobbi,Mayo Clinic

Credit: Mayo Clinic

Mayo Clinic researchers have developed an artificial intelligence (AI) algorithm that can identify obstructive sleep apnea (OSA) using the results from an electrocardiogram (ECG)—a common heart test. The innovation could make it faster, cheaper, and easier to spot sleep apnea, particularly in women, who are often underdiagnosed.

OSA affects more than 936 million adults ages 30–69 worldwide and poses significant cardiovascular risks. People with OSA experience repeated episodes of upper airway collapse or blockage during sleep. This collapse causes breathing to stop or become shallow repeatedly, which often leads to loud snoring and gasping. Despite its prevalence, it often goes undiagnosed.

"Obstructivesleep apneaor OSA is a highly prevalent disease with important cardiovascular consequences," says Virend Somers, M.D., Ph.D., Alice Sheets Marriott Professor of Cardiovascular Medicine and senior author of the studypublishedinJACC: Advances.

"OSA affects the heart to the point where AI algorithms can detect the OSA signature from the ECG, which in essence is a representation of the electrical activity of the heart muscle cells," Dr. Somers adds.

In the study, the researchers used AI algorithms to review the 12-lead ECG test results of 11,299 patients at Mayo Clinic who had undergone the test along with sleep evaluations. More than 7,000 of them had a known diagnosis of OSA, and 4,000 were controls.

"The most surprising finding was the increased visibility on the ECG of OSA in the females compared to the males, even though the OSA severity was less in the females," says Dr. Somers.

"This is relevant since emerging data consistently suggest that females have a greater relative likelihood of suffering the cardiovascular consequences of OSA, even if their OSA may be considered 'milder' by standard diagnostic criteria," he adds.

The test also strongly suggests women may suffer more damage to their heart muscle cells from OSA, Dr. Somers says.

Dr. Somers underscores that this approach may have the potential to evaluate whether a given OSA treatment may be able to reduce a patient's cardiovascular risk.

More information: Naima Covassin et al, Deep Neural Network Algorithm Using the Electrocardiogram for Detection of Obstructive Sleep Apnea, JACC: Advances (2025). DOI: 10.1016/j.jacadv.2025.102139

Provided by Mayo Clinic