AI and Wearable Trackers Predict Hypertensive Crisis Days Early

Predicting Severe Blood Pressure Surges up to 72 Hours in Advance

Artificial intelligence (AI) paired with portable blood pressure monitors and wrist devices can detect impending hypertensive crises up to three days before they occur, according to research scheduled for presentation at the American Heart Association (AHA) Scientific Sessions 2026. The preliminary study evaluated whether automated pattern analysis could warn physicians before patients experienced dangerously high blood pressure spikes.

A hypertensive crisis involves blood pressure exceeding 180/120 millimeters of mercury (mm Hg) accompanied by acute end-organ damage, including heart attack, stroke, or kidney damage. In the 12-month study of more than 4,000 adults with stage 2 high blood pressure, 312 participants experienced 418 hypertensive crises. The AI tool differentiated higher-risk patients from lower-risk individuals up to 72 hours ahead of the event.

Specific Nighttime and Morning Blood Pressure Patterns Raised Crisis Odds

Researchers gathered continuous measurements across daily routines and sleep using a portable cuff alongside a wrist fitness tracker that tracked heart rate and heart rate variability (HRV). The analysis revealed specific cardiovascular signals that signaled elevated short-term vulnerability.

Among participants flagged by an alert, approximately 32% experienced a hypertensive crisis, whereas 97.4% of participants without an alert remained crisis-free. Two specific blood pressure behaviors showed strong associations with upcoming emergencies.

• Participants whose blood pressure did not show the expected decrease during sleep had nearly five times the odds of an upcoming hypertensive crisis.
• Participants who experienced a blood pressure surge of more than 35 mm Hg after waking, compared to nighttime levels, had nearly four times the odds of an event.

Rapid Medication Adjustments Reduced Emergency Department Visits

Early alerts enabled clinicians to intervene before organ damage occurred. When the AI system issued an alert, physicians contacted patients, reviewed current symptoms, and modified antihypertensive medications if necessary.

Patients whose physicians adjusted treatment within 24 hours of an alert experienced approximately 61% fewer hypertensive crises compared to individuals who did not receive timely intervention. In addition, hospital emergency department visits were about 45% lower in the intervention group.

Current Limitations of the Wearable Monitoring Evidence

The findings represent preliminary observational research led by investigators at the University of Maryland Capital Region Medical Center. While the data show clear predictive utility, the protocol required participants to wear both a cuff-based monitor and a wrist device throughout the day and night.

About 68% of participants who received a positive alert did not develop a full hypertensive crisis, meaning false positives occurred in approximately two out of three alerts. Individuals tracking their cardiovascular recovery or hypertension should speak with a clinician about medication management rather than altering prescriptions based solely on consumer tracker metrics.