Wearable technology to screen for atrial fibrillation: Does it raise more questions than it answers?

When the roughly 52,000-member American College of Cardiology (ACC) meets — as it did this past weekend in Orlando for its 67th Annual Scientific Session — it’s a big deal.

Not only does it generate significant media buzz, but some of the studies presented can lead to major changes in how heart disease is diagnosed and treated.

One area generating buzz is “mobile health” approaches; this includes smartphones, watches, and even wearable patches that can now measure much more than simply your pulse or blood pressure. This year’s meeting featured two devices — the Kardia Band for the Apple Watch, and the iRhythm Zio Patch — that provide electrocardiogram (ECG) readings. Both technologies aim to help doctors screen for atrial fibrillation — an irregular beating of the heart’s two upper chambers — that is considered a risk factor for stroke. They seek to find atrial fibrillation in individuals who show no symptoms of the problem (i.e. are asymptomatic) and might not otherwise receive medical care for this condition.

Is screening for atrial fibrillation a good thing?

Will finding more atrial fibrillation, and treating it, lead to better outcomes?

John Mandrola, MD

News coverage of these devices generally suggests that the answer is a resounding “yes.” But John Mandrola, MD, a cardiologist practicing in Kentucky who specializes in the heart’s electrical system, isn’t so sure. Although he has nothing against the “mobile health” approach, he’s not convinced detecting more atrial fibrillation — our most common abnormal heart rhythm (or, “arrhythmia”)  — will necessarily improve health outcomes like preventing stroke.

In an article for Medscape three months ago — ‘Overdiagnosis Only a Matter of Time With ECG Watches‘ — he cites about a half a dozen studies which “cast doubt on the value of AF (atrial fibrillation, or ‘Afib’) as a surrogate for stroke risk.” He adds:

The scariest part of enhanced AF detection is overdiagnosis and overtreatment

Because Afib can make blood pool in the heart and lead to clots, a mainstay of treatment is anticoagulant drugs. These “blood thinner” drugs carry significant risks, costs, and a wide range of worrisome drug interactions. People diagnosed with Afib are usually prescribed other drugs as well and put through a variety of tests, some of which can be quite invasive. All those downsides must be factored into the risk-benefit calculation of finding more cases of Afib. 

Mandrola’s concern is backed up by a December 2017 report  by the US Preventive Services Task Force (USPSTF) which found insufficient evidence to assess the balance of benefits and harms to support screening asymptomatic adults over age 65 for atrial fibrillation by ECG. They cited one randomized control trial (RCT) documenting harms, and no RCTs supporting benefits.

But not everyone agrees. This viewpoint published in JAMA in November, 2015 argued that screening may reduce strokes in asymptomatic, elderly adults.

However, the USPSTF recommendations alone might justify approaching wearable devices for Afib detection with some degree of skepticism. For example, let’s take a look at the data presented this past weekend from a study involving the Zio Patch — a wearable, wristwatch-sized heart monitor. Although the study findings did not capture much mainstream media attention, the news releases from the ACC are instructive, and may help readers critically evaluate the coverage that will inevitably come as these devices gain more widespread acceptance.

Is a wearable patch better for diagnosing atrial fibrillation?

Here are the broad outlines of the study:

• 1,732 people over age 55, with no known heart rhythm issues, but at moderate risk of Afib, were compared with a matched control group of 3,646 to see if the patch could detect Afib better than standard outpatient testing.
• Primary endpoint was newly diagnosed Afib at one year: the patch group had a nearly 3 times increased likelihood of Afib diagnosis than the control group (6.3 vs 2.3 %)
• The patch group had a higher rate of starting anticoagulant drugs (5.4 vs. 3.4 %), receiving  anti-arrhythmic therapy (0.8 vs. 0.3%), and pacemaker placement (0.7 vs. 0 %)

And yet, despite a three-fold higher likelihood of being diagnosed with Afib, and a higher rate of treatment with anticoagulant and anti-arrhythmic drugs, the patch group showed no reduction in the risk of outcomes that actually matter, such stroke, heart attack, or thromboembolism (blood clots traveling from the heart to other parts of body) during 1-year follow-up. The researchers will keep following these patients for two more years to see if benefits begin to accrue with longer duration of treatment.   

Some unanswered questions

The study — part of a larger study called mHealth Screening To Prevent Strokes (mSToPS) trial — was presented as an abstract at the meeting. It has not been published yet in a peer-reviewed journal. As is so often the case with major medical meetings, journalists either learned about the preliminary results by attending the conference, or via a news release.

Zio Patch (from irhythmtech.com)

In this case there were two news releases from the ACC – a longer one and a shorter one:

Self-applied chest patch catches common irrregular heartbeat more quickly than usual care (long)

Self-applied ECG patch leads to increased Afib diagnosis versus routine care in mSToPS trial (short)

The shorter news release predominantly highlighted the study design and results, but little else. Both news releases framed the findings as an unqualified success, and buried the lack of difference on clinical outcomes in a brief acknowledgment at the bottom of the release. Neither explained what the current (“routine” or “usual”) approach to diagnosing Afib is, and how the patch compares in terms of accuracy, costs, risks, benefits, or availability.

The potential harms of using the patch — such as a false-positive result, or the risk of a major bleeding event due to subsequent anticoagulant therapy — were not mentioned.

Let’s not forget: a key finding of the report is that those who wore the patch — although they clearly used more healthcare resources — had no difference in the rate of stroke or other important health outcomes as those who did not. It’s possible those benefits will materialize with further study, but the uncertainty surrounding this result is certainly cause for tempered enthusiasm.  

The issue of mass screening

Christopher Labos, MD

Christopher Labos MD, a cardiologist and regular contributor at HealthNewsReview.org, says these outcomes are important when taken in the context of screening.

“The data for screening asymptomatic patients is weak and there’s a fair amount of uncertainty about whether it’s cost effective,” said Labos.

“Right now screening asymptomatic patients beyond checking their pulse at clinic visits is not recommended. It’s unclear if the excess costs from mass screening would be offset by disease prevention, especially in lower risk populations. This study showed that their device seems to work. But a different kind of study is needed to show the benefit of mass screening the population.”

This relevant background information was not brought up in either of the two news releases from the ACC, or in any of the interviews of the lead author I came across. Which makes me wonder: who IS the target audience of news releases from medical meetings?

Another concern for me that was not highlighted in the shorter press release is that this study was at least partially funded by Janssen, a company that sells an anticoagulant (Rivaroxaban; brand name = Xarelto). Janssen would potentially stand to benefit financially from an increase in the diagnosis of atrial fibrillation among asymptomatic individuals.

A nod to novel study methods

It’s worth noting that the authors believe this study may be the first completely digital, nationwide, direct-to-participant clinical research program that didn’t require the participants to physically show up at a medical facility. The subjects were recruited by email and the wearable devices were returned by mail for analysis. The authors argue this approach, when appropriate, could lead to larger study enrollment, more diversity of subjects, and reach previously inaccessible subjects who may not be located near major medical research facilities.

Wearable technology that provides participant-generated data is a burgeoning area of research. And a booming business. As more studies of this kind generate news coverage, I’d recommend paying special attention to two things at the very least.

First, because tech companies are inevitably involved, follow the money closely. Who is funding the studies and do the authors have financial conflicts of interest? Inevitably many will, which should focus your attention on the data and tune your radar for biased spin.

Second, just because a new test or intervention features new technology, doesn’t necessarily mean it’s automatically better than existing methods. We still need to know how it measures up in terms of risks, benefits, costs … and improving outcomes that actually matter.

Here’s some of our other coverage of the American College of Cardiology’s 67th Annual Scientific Session:

LifeVest study results: PR spin machine in full effect at American College of Cardiology meeting

Did a pricey cholesterol-lowering drug really reduce deaths, as headlines claim?