Apple Watch hypertension notifications: how PPG spots possible high blood pressure

Short answer: Apple Watch hypertension notifications don't measure blood pressure. The watch reads its optical heart sensor (PPG) in the background for 30 days and, if the pulse data looks like what it learned from people with high blood pressure, it sends a "possible hypertension" alert. In the FDA-reviewed validation study, the feature caught 41.2% of people with hypertension and correctly cleared 92.3% of people without it (FDA 510(k) K250507).

So an alert is worth acting on, and silence tells you little. The rest of this page explains how the feature works, what the numbers mean, and why a pulse waveform carries blood pressure information at all.

How the feature works

Apple's validation paper describes a passive feature. It runs during waking hours, needs no calibration, and evaluates data in separate 30-day windows (Apple, Hypertension Notifications Validation Paper, Sept 2025). There is no number on screen. You either get a notification or you don't.

According to the same paper, the feature is available on Apple Watch Series 9 or later and Apple Watch Ultra 2 or later, not Apple Watch SE. It is meant for people 22 and older who have not been diagnosed with hypertension, and it is not intended for use during pregnancy.

The FDA cleared it on September 11, 2025 under a new device type, "Hypertension Machine Learning-Based Notification Software" (FDA 510(k) database, K250507).

What the validation study found

The FDA summary describes the pivotal study:

Source: FDA 510(k) summary K250507.

Apple tuned the feature to avoid false alarms. If you have normal blood pressure, there is about a 1-in-20 chance of a notification over a 30-day window. If you have hypertension, the odds of being flagged are a bit worse than a coin flip, a little better if it is stage 2.

That trade-off fits a consumer screening tool worn by millions of people who never asked to be screened. It also means the feature can't rule anything out.

Why a pulse waveform says anything about blood pressure

Each heartbeat sends a pressure wave down the arteries. A PPG sensor sees that wave as a change in light absorption at the wrist. Stiffer, higher-pressure arteries change the shape of the wave: how fast it rises, where the reflected wave lands, and where the small dip called the dicrotic notch sits. Our PPG blood pressure estimation methods guide covers these features in depth.

Apple hasn't published its model in detail, but academic work shows the same idea can screen for hypertension without estimating a number. A 2026 study in European Heart Journal – Digital Health by Hong, Nandi, Zheng and Alastruey tested a simple hypothesis: people with normal blood pressure show bigger changes in their PPG between waking and sleep than people with unmanaged hypertension (Hong et al., 2026, DOI 10.1093/ehjdh/ztag131).

They extracted 20 PPG indices from the Aurora-BP dataset (180 people: 138 normotensive, 42 hypertensive) and compared each person's awake and sleep periods. A support vector machine reached 81.1% accuracy on the internal test set. On an independent dataset of 26 people (CUHK-BP), a version using only timing-based indices reached 84.6% accuracy. Timing features tied to the dicrotic notch held up best across datasets.

That study is small, and it measures something different from Apple's feature, which runs during waking hours. It does show why a month of passive PPG data can carry a hypertension signal. The nightly dip in blood pressure is itself a clinical marker; see our piece on nocturnal blood pressure dipping. You can browse the paper's record in the ChatPPG research hub.

What a notification means, and what it doesn't

If you get an alert:

• Treat it as a reason to measure. Use a validated upper-arm cuff, take readings on several days, and bring them to a clinician.
• Don't start or change medication based on the watch.
• Remember that specificity of 92.3% still leaves some false alerts across a large population.

If you never get one:

• That isn't a clean bill of health. In the study, most people with hypertension were not flagged.
• Masked hypertension, where office readings look fine but daily readings run high, is common. Our guides on masked hypertension detection with wearables and white-coat vs masked hypertension explain why home readings matter.

How it compares with cuffless blood pressure watches

Some wearables try to show an actual blood pressure number from PPG, usually after calibration with a cuff. Those devices face a stricter validation question, which we cover in cuffless blood pressure validation and the blood pressure watch guide. Apple's approach avoids that problem by answering a narrower yes-or-no question over a month.

For the wider research picture, see PPG hypertension monitoring and the ChatPPG research hub, which tracks new PPG blood pressure papers, datasets and models as they come out.

This article is educational and is not medical advice. Talk to a clinician about blood pressure readings or symptoms.

References

1. U.S. Food and Drug Administration. 510(k) summary K250507, Hypertension Notification Feature (Apple Inc.), decision date September 11, 2025. https://www.accessdata.fda.gov/cdrh_docs/pdf25/K250507.pdf
2. Apple Inc. Hypertension Notifications Validation Paper, September 2025. https://www.apple.com/health/pdf/Hypertension_Notifications_Validation_Paper_September_2025.pdf
3. Hong J, Nandi M, Zheng Y, Alastruey J. Hypertension screening via awake–sleep differences in photoplethysmogram signals. European Heart Journal – Digital Health. 2026. DOI 10.1093/ehjdh/ztag131