Wearables plus AI flagged high blood pressure crises up to three days early in a preliminary study
In more than 4,000 people with hypertension followed for a year, the system's alerts preceded about 32% of crises and 97.4% of people without an alert stayed crisis-free; the abstract is for the AHA's Scientific Sessions and has not been peer-reviewed.
Why it matters
A hypertensive crisis is a sudden, severe rise in blood pressure that can lead to stroke or heart attack, and today it is usually caught only once it happens. The study used continuous readings from wearable devices and found two warning patterns: blood pressure that did not fall during sleep, linked to nearly five times the odds of a crisis, and a morning rise of more than 35 mm Hg, linked to nearly four times the odds. When medication was adjusted within 24 hours of an alert, crises fell by about 61%. The work is from one hospital and is preliminary.
Key facts
- More than 4,000 adults with hypertension followed for 12 months
- 312 participants had 418 crises
- Predictions up to 72 hours ahead
- About 32% of alerts were followed by a crisis; 97.4% without an alert stayed crisis-free
- Medication adjusted within 24 hours of an alert: about 61% fewer crises
- To be presented at AHA Scientific Sessions, 6–9 November, Chicago