The Week in Blood Pressure & Wearables — November 6-12, 2026

By Yair Lurie, Founder and CEO, Cardiowell. LinkedIn Newsletter edition, November 13, 2026.

The American Heart Association's Scientific Sessions ran November 6-9 in Chicago, and most of this week's news came from there. I spent Saturday morning at the poster session on blood pressure devices and measurement science. This week's deep dive is one abstract from that session: a head-to-head test of laser Doppler flowmetry against the optical sensor used in most wearables.

Also this week: two Brazilian trials of a systolic target below 120 reached different answers in diabetes and stroke; a team-based trial built around home monitors and single-pill combinations nearly doubled blood pressure control; and a Technion team described a cuffless optoacoustic sensor. Lorundrostat held a 17.8 mmHg drop for 48 weeks ahead of an FDA decision in December, and I compare the sensitivity of Google's and Apple's hypertension notifications. Oura agreed to buy Counsel Health for $625 million.

Deep dive: Laser Doppler beat PPG in a small cuffless blood pressure test

What the abstract tested

Abstract MDP0116 at AHA Scientific Sessions 2026 is titled "Laser Doppler Flowmetry Outperforms Photoplethysmography for Machine Learning-Based Cuffless Blood Pressure Estimation - Validation Against Oscillometry." It was listed in session HN.MDP.01, "Novel Blood Pressure Devices and Blood Pressure Measurement Science," on Saturday, November 7. The authors are Anna Florax, Libera Fresiello, and Dirk Donker of the University of Twente, and Yakup Kilic and Alwin Fransen of Sonion Nederland B.V. in Hoofddorp. The abstract was listed in the AHA program; no one presented the poster.

The design, from the abstract:

  • 49 adults, median age 40, range 21 to 77.

  • Two sessions up to three weeks apart, at rest and during low-intensity exercise.

  • PPG, laser Doppler, and oscillometric cuff readings taken at the same time on the upper arm. The cuff was the reference.

  • The same approach for both sensors: pulse wave analysis plus machine learning on more than 99 time- and frequency-domain features, with per-subject calibration.

  • 477 paired measurements for laser Doppler and 455 for PPG.

Two different optical signals

PPG, short for photoplethysmography, shines light into the skin and measures how much comes back. The signal changes as blood volume in the tissue rises and falls with each heartbeat, as described in Cureus. It's the sensor in most smartwatches and rings.

Laser Doppler flowmetry (LDF) uses laser light that scatters off moving red blood cells. The movement shifts the light's frequency. The output is a flux that is the product of red blood cell concentration and velocity, per the same Cureus paper by Harvey Mayrovitz, which puts the laser's reach at about 1 to 2 mm into the skin. A Scientific Reports study comparing the two methods says PPG measures "volumetric variations" and LDF tracks "the velocity and concentration of moving erythrocytes," with the depth depending on wavelength and probe design.

So the two sensors see different things. PPG sees how much blood is there. LDF sees how fast it moves.

What the abstract found

Measure Laser Doppler PPG Correlation with cuff, systolic / diastolic 0.885 / 0.7650.841 / 0.721 Mean error, systolic / diastolic (mmHg)-0.61 / -0.62-0.78 / -0.91SD of error, systolic / diastolic (mmHg) 7.07 / 5.648.92 / 5.99 Four-quadrant trend concordance, systolic 93.9% 94.4% Four-quadrant trend concordance, diastolic 90.5% 71.2% Polar angular concordance, systolic / diastolic 76.6% / 67.6%58.7% / 43.2%

As the authors applied the AAMI/ISO/ESH criteria, laser Doppler met them for systolic and diastolic. PPG exceeded the precision threshold for systolic, meaning its 8.92 mmHg spread was above the 8 mmHg limit.

Systolic trend tracking was a tie. PPG was slightly higher, 94.4% against 93.9%. The gap showed up in diastolic trends, 90.5% against 71.2%. The polar analysis favored laser Doppler on both, but it used a subset of readings.

Who is behind it

The abstract doesn't name a device. Sonion, a Danish components maker, announced Sonion Health on March 13, 2025. Its Sonion Health page describes the "Sonion Health Band," which "uses Laser Doppler Flowmetry (LDF) to optically measure arterial blood flow velocity," with machine learning software it calls Health-AI. The page says: "Investigational device. Limited by Federal law to investigational use. Not available for sale or commercial use." It lists seven filed US patents.

Two of the five authors work at Sonion Nederland. A link between the abstract and the Sonion Health Band is likely, but I couldn't confirm it.

Sonion Health pitched at MedTech World North America in West Palm Beach on May 13, 2026, and the video was posted June 9. The slides call the system "ViDo." These are the company's claims from that pitch, none of them published:

  • A study it calls GoLDFish: 54 subjects, 864 data points, 26 people over 60 and 10 over 70, described as "ISO compliant" and passing the ESH exercise test. This is a different dataset from the AHA abstract.

  • A 170-subject study under way in Colorado.

  • FDA feedback through a Q-Sub meeting on prescription and over-the-counter 510(k) pathways, with a 510(k) submission targeted for the first half of 2027.

  • 10 or more patents filed and one granted, per the slides.

The slides also call it "first-in-kind" and the only accurate cuffless platform. Several cuffless blood pressure devices already hold FDA clearance, including Aktiia's bracelet system and Biozen's fingertip BP1000, which is covered below. I'd treat those slide lines as marketing.

Prof. George Stergiou is a Sonion Health advisor, per the company. He also spoke at Sky Labs' symposium at Hypertension Seoul last week.

One open technical question: Sonion says its sensor measures arterial flow velocity. The LDF literature above describes skin perfusion within a millimeter or two of the surface. Those may not be the same signal, and I'd like to see the company explain which it is.

Others working with laser light

Laser flow sensing for blood pressure isn't unique to Sonion.

  • Boston University and Meta Reality Labs. The Boas and Roblyer labs used speckle contrast optical spectroscopy, a related laser method, at the wrist or finger in 30 volunteers. Adding blood flow to blood volume improved estimates by up to 31% over PPG alone, with an average systolic error of 2.26 mmHg, per the Optica release on EurekAlert!. The paper is in Biomedical Optics Express (2025).

  • A*STAR, Singapore. A medRxiv preprint posted June 22, 2026 used diffuse speckle pulsatile flowmetry with an 850 nm laser at the thumb in 133 outpatients, against an Omron cuff. It reports mean absolute errors of 4.17 mmHg systolic and 2.42 mmHg diastolic. PPG-only features didn't meet the AAMI/ESH/ISO or BHS criteria in their analysis, while blood flow features did. It isn't peer-reviewed.

  • CSEM, Switzerland, has research combining laser Doppler and speckle sensors with PPG and ECG. It's research, not a product.

  • Patents. Covidien and Apple hold patents on laser Doppler or self-mixing laser sensors that mention blood pressure. Patents aren't products.

What I think

I like this abstract because it changes one variable. Same people, same arm, same cuff reference, same modeling approach, two sensors. Most cuffless claims compare a whole product against a cuff, so you can't tell whether the sensor or the software made the difference.

The result is also modest, which makes it more believable. Systolic trending was a tie. The clearest difference was in diastolic trends and in systolic spread. Those are the numbers I'd want to see hold up.

The limits are real. It's one abstract, not peer-reviewed, with 49 people and a median age of 40. Every reading was calibrated per person. Two of the five authors work at Sonion Nederland. A full ISO 81060-2 validation uses a larger sample than this, as I understand it. The 2026 ISH guidelines, covered here on October 30, still don't recommend cuffless devices for diagnosing or managing hypertension.

Here's where we sit. A clinic running a hypertension program titrates medication from validated cuff readings. We use Cardiowell’s ‘always-connected’ blood pressure monitor because each reading goes over cellular to the care team with no pairing, which reduces the points where data gets lost. A cuffless device earns a place in that program when it shows the same accuracy across older patients, patients on medication, and patients with stiff arteries, without a cuff. In the meantime, It's useful to surface trends and nighttime readings.

That's the group I'd worry about most with PPG. If laser flow sensing helps there, it matters. I hope we'll see a cleared laser-based wearable within a few years, and I'm glad the field is testing more than one kind of light. I wrote a shorter post on this abstract on LinkedIn.

What to watch

  • Whether the authors or Sonion Health name the device used in the abstract.

  • Publication of the GoLDFish data and the Colorado study, with the reference method and calibration rules stated.

  • Results in people over 60 and people on blood pressure medication.

  • The 510(k) submission, and whether it is for spot checks or continuous readings.

From the AHA 2026 blood pressure measurement session

Session HN.MDP.01, "Novel Blood Pressure Devices and Blood Pressure Measurement Science," was moderated by Timothy Plante, MD, on Saturday, November 7. Abstract text below comes from the AHA program. I found public links only where noted.

A fingertip device that measures in 23 seconds

"Validating the Accuracy of an Ultra-Compact, Portable, Blood Pressure Measurement Device Based on Fingertip Oscillometry with Shortened Measurement Duration." Goktug Ozmen, Michael Chan, John Berkebile and colleagues at Biozen, with last author Omer Inan of Georgia Tech, who discloses ownership in Biozen.

Biozen's BP1000 received FDA 510(k) clearance with a decision date of May 6, 2026, per FDA's 510(k) database, and Biozen announced it May 11-12. It's cleared for spot checks at home, over the counter, in adults 22 to 59 without arrhythmia, peripheral artery disease, or pregnancy. You press a fingertip on the sensor. Pressure and optical sensors run the oscillometric method at the finger, with no cuff calibration.

The abstract cut measurement time from 38.09 to 22.80 seconds. The new algorithm was tested in 234 subjects (643 measurements) using leave-one-subject-out cross-validation against dual-observer auscultation: systolic error 0.51 ± 7.06 mmHg, diastolic 0.72 ± 6.59 mmHg. Mean age was 35.8, and 59.4% had Fitzpatrick skin types V or VI. The faster version isn't cleared yet.

In the Q&A, the team estimated an end price near $174 depending on form factor and said a spot check could later calibrate continuous wearables. Treat both as estimates. The label stops at 59, and most Medicare remote monitoring patients are older. I covered this one on LinkedIn.

Three of five common measurement mistakes lowered the reading

"Effects of Common Blood Pressure Measurement Deviations on Office Blood Pressure: A Randomized Repeated-Measures Study," presented by Pablo Reyes-Guzmán, MD, with Roberto Anjed Velazco Llinas and colleagues at Hospital Universitario "Dr. José Eleuterio González" in Monterrey, Mexico.

265 adults, 133 with hypertension, were measured with one validated automated device under six conditions in random order. Systolic difference from a standard reading:

  • Active conversation: +3.03 mmHg

  • Crossed legs: +1.04

  • Unsupported arm: -2.31

  • Unsupported back: -1.70

  • Cuff over clothing: -1.19

The unsupported-arm effect was -3.20 in people without hypertension and -1.43 in people with it. Three of the five errors lowered the reading. We usually assume measurement errors push numbers up. Some of them hide hypertension.

Most landmark trials didn't report a validated device

"Validation Status of Blood Pressure Measuring Devices and Reporting Quality of Blood Pressure Measurement in Landmark Randomized Controlled Trials," presented by Saby Belen Alderete Quispe of Ica, Peru. First author Fabian Chavez; co-authors include Eugene Yang (University of Washington), Pedro Ordunez (PAHO), Carlos Vergara-Sanchez (Mayo Clinic), and Carlos Toro-Huamanchumo (Edith Cowan University).

The team reviewed 87 trials from 2013 to 2025 in 13 journals, with 305,844 participants, and checked 92 devices against the STRIDE BP, ValidateBP, and Medaval listings.

  • 32 of 87 trials (36.8%) used only validated devices.

  • Of the devices, 39.1% were validated in a relevant population, 13.0% weren't validated, and 41.3% couldn't be classified from what the trial reported.

  • The device model was reported in 59.8% of trials.

  • Reporting didn't improve over the 13 years.

SPRINT's Omron HEM-907XL is a useful example. STRIDE BP lists the HEM-907 as validated under older protocols (AAMI 1992 and ESH 2002), and a 2010 study by Ostchega and colleagues found the XL met AAMI criteria except for teen diastolic readings. None of the reviewed devices used ISO 81060-2:2018. So it was validated under older protocols, and I see no reason here to doubt SPRINT's results. The bigger point is reporting. If a trial doesn't name its device, readers can't check it.

Update on last week's crisis-prediction study

Last week's edition covered abstract MDP0117, on wearables and AI predicting hypertensive crisis up to three days early. We followed the AHA press release, which says a portable cuff-based monitor took the blood pressure readings and a wrist device tracked heart rate and HRV. We wrote flatly that it wasn't a cuffless study.

The abstract text says something different. It says blood pressure came from "a validated wrist-worn PPG device (BIHS grade A)." The release and the abstract disagree on how blood pressure was measured, and I can't resolve which is right. The October 30-November 5 edition should have said the method was unclear.

The abstract adds numbers the release didn't give: AUC 0.87 (0.84-0.90), sensitivity 82.1%, specificity 79.4%, positive predictive value 31.6%, negative predictive value 97.4%, using gradient boosting. It describes a "geographically distinct 20% hold-out," while the release describes one hospital.

Remote monitoring programs: what the posters showed

Buffalo, majority African American primary care. "Remote Hypertension Monitoring with Clinician-Led Medication Titration in a Majority-African American Urban Primary Care Population: Longitudinal Outcomes and Racial Disparities," presented by Ryan Mackey, MD, University at Buffalo, senior author Ellen Rich. 503 patients, 63.6% African American, sent readings from a Bluetooth home cuff, with monthly clinician review and titration. Control rose from 10.1% to 37.1% at nine months, and systolic fell from 145 to 132. Each month in the program was linked to lower odds of uncontrolled blood pressure (adjusted odds ratio 0.82). Compared with African American patients, White patients had lower odds of uncontrolled blood pressure (aOR 0.57 at the moderate threshold, 0.40 at the strict one). It's retrospective with no comparison group. A related preprint is on medRxiv and isn't peer-reviewed.

Ithaca, rural residency clinic. "Implementing a Remote Blood Pressure Monitoring Program at Rural Resident Primary Care Clinic," Sameer Acharya and colleagues, Cayuga Medical Center, with Yashashwi Pokharel (Wake Forest). 72 Medicare patients 65 and older. Prevounce supplied monitors, platform, and coaching, and residents titrated every six weeks. Patients sent about 7.3 readings a week. Systolic fell 11.4 mmHg and diastolic 3.4, with most of the drop in the first two weeks and a plateau after week 13. 63.9% had a medication change, and retention was 100%. Single arm, no control.

HealthSnap, 12 months of readings. "Proportion of Home Blood Pressure Readings in Target Range During 12 Months of Remote Patient Monitoring," presented by Craig Flanagan with Wesley Smith, both of HealthSnap. 2,987 patients at eight US cardiovascular practices, mean age 77 to 79. Of 1,935 whose month-12 average systolic was under 130, only 55.3% had at least 80% of their readings under 130. The share of readings under 130 rose most in patients who started high; in the 150-159 group, it went from 17.0% to 41.3%. It's retrospective; it only includes patients who stayed 12 months with readings on 16 or more days a month, and the authors work for HealthSnap. The 80% goal hasn't been tested against outcomes. A SPRINT post hoc analysis by Fatani and colleagues in JACC (2021) linked more time in target range with fewer cardiovascular events.

An average under 130 can hide many readings above it. That's an argument for looking at the distribution of home readings, which only frequent home measurement gives you.

Clinical & Research

Two trials of a systolic target under 120: different answers in diabetes and stroke

Two open-label trials from Brazil tested the same strategy—a systolic target under 120 versus under 140—and were published in NEJM during the meeting.

OPTIMAL-DIABETES randomized 9,476 adults 50 and older with type 2 diabetes and added cardiovascular risk at 33 sites, per the AHA. It achieved an average systolic blood pressure of 122 vs 135 over about four years. Major cardiovascular events were 6.0% vs 6.7%, about 10% lower, but not statistically significant. Kidney events were higher with intensive treatment (5.7% vs 3.7%), mostly declines in kidney function.

OPTIMAL-Stroke randomized 4,368 adults with a prior ischemic stroke or TIA at 30 centers, per the AHA. Major cardiovascular events were 10.9% against 13.5%, a hazard ratio of 0.79 (0.67-0.94), per HCPLive. Recurrent stroke was 4.7% against 6.4%. Kidney events were higher here too, 3.9% against 2.2%.

Both trials were open-label and in one country. The AHA releases don't say how blood pressure was measured. For clinics, the takeaway is that a lower target needs kidney monitoring alongside the blood pressure readings.

Single-pill combinations plus home monitors nearly doubled control

The CATCH trial randomized 40 primary care clinics in Colombia and Jamaica, enrolling 1,706 adults with uncontrolled blood pressure; it was published in JAMA and presented as an AHA late-breaker. UT Southwestern led the NIH-funded study.

The intervention combined several pieces:

  • Single-pill combinations, mostly valsartan/amlodipine or valsartan/amlodipine/hydrochlorothiazide, funded by the study.

  • A home blood pressure monitor for every participant, with readings at least two days a week.

  • Nurses who reviewed clinic and home readings and relayed them to physicians, who adjusted medication.

  • Community health workers coaching on adherence and lifestyle.

At 18 months, systolic fell 19.1 mmHg against 9.0 in usual care, an adjusted difference of 9.7 mmHg. Control under 130/80 was 69.5% against 29.2%. The gap narrowed by 24 months (55.8% against 34.2%). It was unblinded, and nearly everyone was already on treatment.

This is the trial I'd send to a clinic manager. It shows the parts working together: fewer pills, home readings, a nurse who looks at them, and a physician who acts.

Spironolactone after stroke, measured at home

The REDUCE trial randomized 160 adults with uncontrolled blood pressure after stroke at five US centers to add daily spironolactone or continue usual care, per the AHA. The outcome was measured with home digital monitors. At three months, home systolic was about 126 against 133, and 52% against 30% were under 130/80. About 4% had high potassium. Enrollment ended early; the trial was open-label, and the spironolactone group used more medicines.

Lorundrostat held its effect for 48 weeks, with an FDA decision due December 22

Lorundrostat is an aldosterone synthase inhibitor from Mineralys Therapeutics. It blocks production of aldosterone, a hormone that raises blood pressure by making the body retain salt and water. Manish Saxena of Barts Health, the global chief investigator, presented the TRANSFORM-HTN results in a late-breaking session, per Barts Health on LinkedIn.

TRANSFORM-HTN is the open-label extension of Launch-HTN, a 12-week randomized, placebo-controlled trial in people whose blood pressure stayed high on two or more medications. Per HCPLive and RTTNews:

  • 632 people entered the extension on lorundrostat 50 mg, and 583 (92%) completed 48 weeks of follow-up.

  • Automated office systolic fell 17.8 mmHg from a baseline of 149/87, and diastolic fell 9.2.

  • 56% had systolic under 130 at week 48.

  • Results were similar for people on two background medications and on three or more.

  • 37 people (5.9%) stopped because of side effects, 10 (1.6%) of them for high potassium. High potassium led to a dose change, pause or stop in 32 (5.1%), and potassium above 6 mmol/L occurred in 2.5%.

About a quarter of participants were Black or African American and almost half were women, per Barts Health. FDA set a target action date of December 22, 2026, per Mineralys, for use with other blood pressure medications.

The limits: the extension had no control group, and everyone in it had already tolerated 12 weeks of the drug. The 48-week durability is the new part.

I think this is the best hypertension news of the year. Resistant hypertension is where clinics spend the most effort for the least progress, and a drug that holds a 17.8 mmHg drop for nearly a year changes that conversation. It also adds work. A practice prescribing it will need potassium checks alongside the home blood pressure readings, so the lab results and the readings belong in the same review.

Primary aldosteronism stayed in remission after treatment stopped

A letter in NEJM from a Queen Mary University of London team followed 14 patients with primary aldosteronism who took an aldosterone synthase inhibitor for 6 to 12 months, per News-Medical. Up to two years after stopping, the aldosterone marker stayed 99% lower and blood pressure averaged 25 points lower. Almost a quarter had normal blood pressure without medication. It's 14 people with no control group, and the authors call themselves "duly cautious." A larger trial is under way.

Sacubitril/valsartan and atrial fibrillation after ablation

In 320 adults 55 and older with atrial fibrillation and hypertension after ablation at three hospitals in China, sacubitril/valsartan cut 12-month arrhythmia recurrence to 18% from 28.3% with valsartan, per the AHA. Clinic systolic fell about 10 mmHg in both arms. Open-label, with intermittent rhythm monitoring, and not yet peer-reviewed.

Devices & Wearables

Google's research model vs Apple's notification: what the sensitivity numbers mean

Cardiologist Kenneth Civello MD, MPH compared the two on LinkedIn this week. Apple's hypertension notification caught 41.2% of people with hypertension in its validation study. Google's model caught 65.8%. He noted it wasn't a head-to-head comparison.

Radiologist Sumer Sethi made a related point about the Google study: the model flags signs of hypertension from several days of wear, with no cuff calibration, and it doesn't give a blood pressure reading. The same is true of Apple's feature.

The primary sources:

Google (AI-PPG-ACC-HTN)Apple (Hypertension Notifications) Status Research model. Google calls the methods "purely exploratory." FDA-cleared feature on Apple Watch Series 9 or later and Ultra 2 or later. Signals PPG plus accelerometer, plus height PPG, 60-second segments about every two hours while sitting still. People analyzed: 196 at five US sites; 76 with hypertension1, 863,585 without hypertension. Population: no known heart disease; not on BP medication; no prior hypertension diagnosis. Reference: 24-hour ambulatory monitoring, average of 130/80 or higher; 30-day average of twice-daily home upper-arm cuff readings; 2017 ACC/AHA categories. Watch wear: mean 7.3 days, about 15 hours a day; 30 days, at least 12 hours a day. Sensitivity (95% CI): 65.8% (54.0-76.3); 41.2% (37.2-45.3). Stage 1: 29.6%. Stage 2: 53.7%. Specificity (95% CI): 90.0% (83.2-94.7), 92.3% (90.6-93.7)

On these numbers, Google's model caught more cases at similar specificity, and the confidence intervals don't overlap. That's a real difference between these two studies. It doesn't tell us which watch is better, for three reasons:

  • Different reference tests. Google compared against 24-hour ambulatory monitoring. Apple compared against a month of home cuff readings. These two methods don't always classify the same person the same way.

  • Different sample sizes. Google's estimate rests on 76 people with hypertension. Apple's rests on 585. Google's interval runs from 54% to 76%.

  • Different products. Apple's 41.2% describes a cleared feature people can turn on today. Google's 65.8% describes a research model. Google says the work "helped inform" the blood pressure trends feature in Health Guardian, and also that the methods "are not indicative of any existing or upcoming commercial product feature." Google calls that trends feature general wellness, as covered here in October. I found no published sensitivity for what Pixel Watch owners actually get.

What I take from it. Both companies chose high specificity, so an alert is usually right, and most misses are silent. Civello put it plainly: "Silence does not mean you are OK." The more useful number for clinics may be in Google's own confirmation data, where multi-day home monitoring had the highest sensitivity, 89.2%. A watch alert is a reason to start home readings with a validated cuff. People without an alert still need a cuff check.

I expect both companies to improve sensitivity as their models train on more data, making the watch a better front door to a home blood pressure program.

Technion describes a cuffless optoacoustic blood pressure sensor

Technion researchers Amir Rosenthal, Daniel Lange and Gil Gelbert described SPADE, a silicon-photonics sensor, in Nature Communications, per Ynet and ICT&health. Short light pulses create ultrasound in an artery. The sensor measures the artery's absolute diameter and its changes with each beat, 100 times a second, and an arterial-wall model rebuilds the pressure waveform.

It's a third optical approach next to PPG and laser Doppler, and it measures the artery itself. It's also a proof of concept. The coverage gives no sample size, accuracy figures or reference comparison. The paper was published in September; this week was the press push.

Omron validates a lower-pressure, faster cuff method

A study in Hypertension Research validated the Omron HEM-7290T, an upper-arm oscillometric monitor using a low-pressure, rapid inflation method, against ISO 81060-2 in 85 people, per the journal. Mean differences were 0.5 ± 7.09 mmHg systolic and 0.6 ± 5.50 diastolic, within the standard. Omron funded the study, and three authors are Omron employees. The comfort benefit wasn't measured.

Oura agrees to buy Counsel Health for $625 million

Oura signed a deal to buy Counsel Health, an AI primary care company that pairs AI with board-certified physicians, for $625 million in cash, per Becker's. Oura users would get diagnoses, prescriptions, and referrals in the app, with plans that include chronic and cardiometabolic care. Christina Farr's Second Opinion covered it. Closing is expected before the end of 2026.

A ring company that can prescribe changes what a wearable alert can lead to. I'll watch whether the physicians on the other end start asking for validated cuff readings.

Fitbit Air gets manual blood pressure logging

Google Health 5.12 lets Fitbit Air owners log readings from a separate cuff, per Android Police. The Air's blood pressure trends feature, which doesn't measure blood pressure, isn't available on it yet. Rollout is staged.

Smartwatch ECGs flag weak heart pumping

A single-center South Korean study used AI on 30-second single-lead ECGs from Apple Watch and Galaxy Watch to detect an ejection fraction of 40% or less, confirmed by echocardiogram, per the AHA. At home, requiring two of three positive days cut false positives from 18 to 9 of 70 people without the condition while catching 31 of 38 cases. No regulatory status was stated.

Policy & Regulation

Still no final RPM rates or "direct employee" definition

Nothing new from CMS this week on the 2027 remote monitoring rules covered in last week's deep dive. Nixon Peabody confirms CMS didn't define "direct employee" and promised guidance before January 1, 2029. Practice-expense changes for the device codes phase in over at least two years, and the AMA's RUC reviews all 19 RPM and RTM codes in January 2027. I didn't find the final payment table (Addendum B) with confirmed RPM rates.

STAT reports that vendors, hospitals, and insurers plan to use the two years to push back. The article is paywalled.

States get a playbook for self-measured blood pressure programs

ASTHO and ACPM published a CDC-funded brief with five self-measured blood pressure program case studies, including Henry Ford in Detroit, Johns Hopkins in Baltimore and the Oklahoma City Indian Clinic, per ASTHO. At the Oklahoma City clinic, 77% of patients who completed the program reached control, with drops of 14/11.6 mmHg. The brief recommends states work with Medicaid and payers on sustainable funding, without naming a specific policy.

Also worth a glance

  • DEA telemedicine prescribing. The final special registration rule has been under White House review since August 25, and current flexibilities expire December 31, 2026, per Frier Levitt. The text isn't public.

  • A wristband for veterans. Dallas startup Optimus OS is launching a 15-veteran pilot of its Cadence band, which tracks heart rate, HRV, respiration, and temperature using hardware from Sensor Bio, per Dallas Innovates. It makes no blood pressure claim and says it doesn't diagnose.

  • What patients say about remote monitoring. "Beyond Blood Pressure Control: A Qualitative Study of Engagement and Perceptions of Remote Monitoring for Hypertension" (MDP0124), presented by Herlyne Das of the University of Chicago, senior author Amber E. Johnson, interviewed 8 patients. Themes were accountability ("pushes me to do the right thing"), care-team support ("someone's watching"), motivation, and lasting habits. AHA abstract; no public link found.

  • Home cuffs in an underserved clinic. "Impact of Home Blood Pressure Monitoring on Hypertension Outcomes and Patient Engagement in an Underserved Primary Care Clinic," presented by Micaela MacKay, MD, Washington University in St. Louis, senior author Angela L. Brown, AHA-funded. In 45 patients, clinic blood pressure was 151/87 at cuff handout, 142/81 at 0-3 months and 145/83 at 3-6 months, and 84% reported a positive experience. Small sample. AHA abstract; no public link found.

  • South Korea's telehealth limits. First-time telehealth prescriptions would be capped at 7 days, extendable to 30, under a proposed rule whose comment period ended November 9, per SBS News.

The thread this week

This week was about measurement.

A laser sensor did better than PPG in one small test. A fingertip device got faster. Three common measurement mistakes lowered office readings, and most landmark trials didn't say which device they used. The trial with the biggest gain in control, CATCH, gave every patient a home monitor and a nurse to review the readings. Google's and Apple's watches miss many people with hypertension, and Google's own data found home monitoring the most sensitive way to confirm it.

Better sensors are coming, and I'm optimistic about them. In the meantime, the readings we already have work best when they're taken correctly, and someone acts on them.

This edition is also published on the Cardiowell blog: [ARTICLE URL — cardiowell.com, not yet published]

If a colleague forwarded this to you, subscribe to The Week in Blood Pressure & Wearables to get it every Friday.

References

  1. AHA Scientific Sessions 2026 abstract MDP0116, "Laser Doppler Flowmetry Outperforms Photoplethysmography for Machine Learning-Based Cuffless Blood Pressure Estimation - Validation Against Oscillometry" (Florax, Fresiello, Kilic, Fransen, Donker). No public link found.

  2. Relationship Between Finger Photoplethysmographic Pulses and Skin Blood Perfusion. Cureus (Mayrovitz, 2024).

  3. Different lasers reveal different skin microcirculatory flowmotion - data from the wavelet transform analysis of human hindlimb perfusion. Scientific Reports.

  4. Announcement of Sonion Health. Sonion.

  5. Sonion Health. Sonion.

  6. Sonion Health pitch, MedTech World North America 2026 (video title not confirmed). YouTube.

  7. FDA clears over-the-counter, cuffless blood pressure monitor with the Hilo bracelet. Fierce Biotech.

  8. Biozen Receives FDA Clearance for Calibration-Free Cuffless Blood Pressure Technology. PR Newswire.

  9. Researchers take major step toward cuff-free blood pressure monitoring. EurekAlert! (Optica).

  10. Clinical-grade Cuffless Blood Pressure Monitoring via Deep-tissue Diffuse Speckle Pulsatile Flowmetry. medRxiv (preprint).

  11. A laser beat the standard optical sensor for cuffless BP.. Yair Lurie on LinkedIn.

  12. AHA Scientific Sessions 2026 abstract, "Validating the Accuracy of an Ultra-Compact, Portable, Blood Pressure Measurement Device Based on Fingertip Oscillometry with Shortened Measurement Duration" (Ozmen, Chan, Berkebile, Inan and colleagues). No public link found.

  13. FDA-cleared cuffless BP exists. It is labeled for ages 22-59.. Yair Lurie on LinkedIn.

  14. AHA Scientific Sessions 2026 abstract, "Effects of Common Blood Pressure Measurement Deviations on Office Blood Pressure: A Randomized Repeated-Measures Study" (Reyes-Guzmán, Velazco Llinas and colleagues). No public link found.

  15. AHA Scientific Sessions 2026 abstract, "Validation Status of Blood Pressure Measuring Devices and Reporting Quality of Blood Pressure Measurement in Landmark Randomized Controlled Trials" (Chavez, Alderete Quispe, Yang, Ordunez and colleagues). No public link found.

  16. Wearable devices + AI analysis may predict high blood pressure crisis up to 3 days early. American Heart Association Newsroom.

  17. AHA Scientific Sessions 2026 abstract MDP0117, "Wearable-Derived Continuous Blood Pressure Monitoring Predicts Hypertensive Crises with 72-Hour Lead Time in High-Risk Ambulatory Patients." No public link found.

  18. The Week in Blood Pressure & Wearables — October 30-November 5, 2026. Yair Lurie on LinkedIn.

  19. AHA Scientific Sessions 2026 abstract, "Remote Hypertension Monitoring with Clinician-Led Medication Titration in a Majority-African American Urban Primary Care Population: Longitudinal Outcomes and Racial Disparities" (Mackey, Rich and colleagues). No public link found.

  20. Clinician-Led Remote Hypertension Monitoring and Blood Pressure Control in a Majority-Minority Primary Care Cohort (full title not reproduced). medRxiv (preprint).

  21. AHA Scientific Sessions 2026 abstract MDP0120, "Implementing a Remote Blood Pressure Monitoring Program at Rural Resident Primary Care Clinic" (Acharya, Pokharel and colleagues). No public link found.

  22. AHA Scientific Sessions 2026 abstract, "Proportion of Home Blood Pressure Readings in Target Range During 12 Months of Remote Patient Monitoring" (Flanagan, Smith). No public link found.

  23. Systolic Blood Pressure Time in Target Range and Cardiovascular Outcomes in Patients With Hypertension. Journal of the American College of Cardiology (Fatani et al., 2021).

  24. Intensive blood pressure therapy did not cut cardiovascular events in adults with diabetes. American Heart Association Newsroom.

  25. Intensive blood pressure control in stroke survivors lowered risk of cardiovascular events. American Heart Association Newsroom.

  26. AHA 2026: Intensive Blood Pressure Control Reduces Events After Ischemic Stroke. HCPLive.

  27. Implementation of Single-Pill Combination Therapy Through Team-Based Care for Hypertension Control in Resource-Constrained Settings: A Cluster Randomized Clinical Trial. JAMA.

  28. Single-pill combination helps more patients control hypertension. UT Southwestern Medical Center.

  29. Can a low-cost diuretic medication help to lower blood pressure after a stroke?. American Heart Association Newsroom.

  30. Transform-HTN: Lorundrostat Maintains Blood Pressure Reductions in Open-Label Extension. HCPLive.

  31. Mineralys Therapeutics Announces FDA Acceptance of NDA for Lorundrostat for Treatment of Adults with Hypertension and Topline Explore-OSA Trial Results. Mineralys Therapeutics via Barchart.

  32. New treatment may offer lasting cure for hormone-related hypertension. News-Medical.

  33. Blood pressure medication combo linked to fewer irregular heart rhythms after ablation. American Heart Association Newsroom.

  34. No pressure: Israeli researchers develop tiny cuffless blood pressure sensor. Ynet.

  35. Tiny sensor advances cuffless blood pressure monitoring. ICT&health.

  36. Toward patient-centered blood pressure monitoring: validation of a novel low-pressure and rapid oscillometric method according to the ISO 81060-2 (full title not confirmed). Hypertension Research.

  37. Oura to acquire AI primary care company for $625M. Becker's Hospital Review.

  38. Oura to buy Counsel Health. Second Opinion.

  39. Blood pressure logging is arriving for Fitbit Air owners. Android Police.

  40. AI-analysis of ECG data measured by smartwatch may help detect reduced heart pumping function. American Heart Association Newsroom.

  41. CMS Finalizes 2027 Medicare RPM/RTM Changes; Direct Employee Rule Delayed to 2029 (headline approximated from the URL; not confirmed). Nixon Peabody.

  42. After Medicare punts RPM overhaul to 2029, advocates see window to fight back. STAT.

  43. Strengthening Efforts to Reduce Hypertension Through Self-Measured Blood Pressure Monitoring Programs. Association of State and Territorial Health Officials.

  44. DEA telemedicine special registration rule under OIRA review (headline not confirmed). Frier Levitt.

  45. Dallas Startup Optimus OS Develops Wearable Tech to Combat Veterans' Health Issues. Dallas Innovates.

  46. AHA Scientific Sessions 2026 abstract MDP0124, "Beyond Blood Pressure Control: A Qualitative Study of Engagement and Perceptions of Remote Monitoring for Hypertension" (Das, Johnson). No public link found.

  47. AHA Scientific Sessions 2026 abstract, "Impact of Home Blood Pressure Monitoring on Hypertension Outcomes and Patient Engagement in an Underserved Primary Care Clinic" (MacKay, Brown). No public link found.

  48. First-Time Telehealth Restricted to Local Clinics; Prescriptions Capped at 7 Days (Max 30 Days). SBS News.

  49. My wife says I'm not sensitive enough. Wearables screening for hypertension have the same problem.. Kenneth Civello MD, MPH on LinkedIn.

  50. Opportunistically Detecting Signs of Hypertension on a Consumer Smartwatch. Nature (Di Achille, Poh et al., 2026; paywalled).

  51. Detecting signs of hypertension using wearable sensors. Google Research.

  52. Opportunistically Detecting Signs of Hypertension on a Consumer Smartwatch (preprint). medRxiv.

  53. Hypertension Notifications Validation Paper, September 2025. Apple.

  54. Apple receives FDA nod for hypertension notification feature. HLTH.

  55. Breaking trial news: a new drug has lowered blood pressure for nearly a year in people whose condition hasn't responded to several medications. Barts Health NHS Trust on LinkedIn (shared by Manish Saxena).

  56. Mineralys' Lorundrostat Sustains Blood Pressure Reduction Through 48 Weeks (headline approximated from the URL; not confirmed). RTTNews.

  57. Your smartwatch could prompt you to get checked for hypertension without ever inflating a cuff. Dr Sumer Sethi on LinkedIn.

Previous
Previous

The Week in Blood Pressure & Wearables — November 13-19, 2026

Next
Next

The Week in Blood Pressure & Wearables — October 30-November 5, 2026