AI Analysis of Smartwatch ECGs May Help Detect Weakened Heart Function

Left ventricular systolic dysfunction (LVSD) occurs when the heart’s main pumping chamber does not contract effectively. The condition can remain asymptomatic for some time but may progress to symptomatic heart failure without treatment. Diagnosis typically relies on echocardiography, which may not be performed until clinical suspicion arises. To support earlier identification, researchers have evaluated whether artificial intelligence analysis of smartwatch electrocardiogram data can help flag patients who may benefit from further cardiac screening.

The novel artificial intelligence model was developed by Medical AI Co., Ltd. (Seoul, South Korea) and evaluated in adults enrolled at Incheon Sejong Hospital in South Korea. The system analyzes 30-second, single-lead electrocardiogram recordings captured by smartwatches for signs of weakened heart function. LVSD was defined as a left ventricular ejection fraction of 40% or less and was confirmed clinically using transthoracic echocardiography. Smartwatch electrocardiograms were recorded within 14 days before or after the ultrasound examination.

The study included two patient groups. The first consisted of 2,211 adults scheduled for echocardiography at an outpatient cardiology clinic between July 2023 and May 2024. Each participant underwent back-to-back electrocardiogram recordings using both an Apple Watch and a Samsung Galaxy Watch under staff supervision. The model distinguished participants with and without LVSD with similar accuracy across the two devices, although 200 participants, or 9%, were classified differently when the same AI score cutoff was applied.

The second group included 146 outpatients enrolled between July 2024 and December 2025 who recorded electrocardiograms at home using their own smartwatches. Among 108 participants who recorded on at least five days during the first week, requiring positive AI results on at least two of three consecutive recording days reduced false-positive LVSD classifications from 18 to nine among 70 participants without LVSD. Using this rule, the model identified 31 of 38 participants with LVSD, compared with 33 using a single-positive-day analysis. Across all 146 home-monitoring participants, 81% of those with consistently positive results had LVSD confirmed by echocardiography, compared with about 6% of those with consistently negative results.

The findings were presented as preliminary, independent late-breaking science at the American Heart Association’s Scientific Sessions 2026 in Chicago. The study was conducted at a single hospital and included patients already referred for echocardiography, which may limit how well the results apply to the general population. The researchers also noted that the study did not determine whether using the AI results

“Some people have reduced heart pumping function without noticeable symptoms. Our study explored whether AI analysis of smartwatch ECGs could help identify people who may benefit from additional screening for LVSD or other heart conditions,” said Hak Seung Lee, chief medical officer of Medical AI Co., Ltd., and cardiologist at the Digital Healthcare Institute, Sejong Medical Research Institute.

“A single smartwatch ECG is a snapshot, and repeated recordings can give different results. Requiring repeated positive results reduced false positives in our study, yet it also missed some additional cases that should have been positive. Larger studies are needed to determine how best to use these results to guide further evaluation,” said Lee.

Related Links
Medical AI Co., Ltd.
American Heart Association