FDA Clears AI Wearable That Detects Heart Attacks Before Symptoms

The U.S. Food and Drug Administration has cleared a wrist-worn wearable that uses a deep-learning model to detect myocardial infarction up to 20 minutes before the wearer reports symptoms - the first device authorized for preemptive cardiac-event alerting.

What the Model Sees

The device continuously samples a multi-lead ECG, skin temperature, and micro-accelerometry. Its model, trained on 4 million cardiac events, detects subtle electrophysiological patterns that precede infarction but are invisible to conventional monitors.

"A 20-minute head start is the difference between a stent in the cath lab and irreversible muscle damage," said the cardiologist who led the trial.

Caveats and Rollout

The clearance requires a prescription and a paired physician dashboard. The model carries a documented false-alert rate of roughly 3%, which the FDA deemed acceptable given the consequence of a missed event.

Critics warn that preemptive alerting could drive unnecessary ER visits. The manufacturer has committed to a post-market study tracking outcomes across 50,000 users over two years.