If you look hard to find people who have atrial fibrillation (AF) you will in fact find people who have atrial fibrillation, a new paper published in JAMA shows. But the paper offers no evidence whatsoever that the new diagnosis improves outcomes in these people, though it does find that the diagnosis leads to increased use of drugs and procedures.
Steven Steinhubl (Scripps Translational Science Institute) and colleagues performed a randomized controlled trial (RCT) and an observational study to examine the impact of a broad AF screening strategy in people at high risk for AF.
In the RCT portion of the study 2,659 people were randomized to immediate or delayed (at 4 months) screening for AF using the Zio iRhythm self-applied wearable electrocardiogram (ECG) patch. At 4 months (before the control group underwent screening) new AF was identified in 53 out of 1,366 patients (3.9%) compared with 12 out of 1,293 patients (0.9%) in the control group, an absolute difference of 3%.
In the observational portion of the study the investigators compared the patients (in either group of the RCT) who received screening with a matched cohort who did not receive screening. After one year the rate of new AF diagnosis was more than doubled in the monitored group (from 2.6 to 6.7 per 100 person-years). People in the active monitoring group were more likely to start taking anticoagulants, more likely to visit a cardiologist, and more likely to receive a pacemaker or defibrillator. The investigators say that they will analyze outcomes after 3 years.
The argument in favor of monitoring is simple. People with AF have an increased risk of stroke, and identification of AF can lead to initiation of anticoagulant therapy which has been shown to prevent strokes.
But it is important to recognize that people identified by this broad screening method are different from AF patients studied in trials, most of whom had typical AF symptoms. By contrast, people diagnosed with AF in this study did not have typical AF symptoms and had, as the authors reported, a “relatively low burden of paroxysmal AF.” “Silent AF” may now become the new name for this condition, and you shouldn’t be surprised if drug and device companies and industry supported health organizations seek to raise awareness of it. And you probably won’t be surprised to learn that the study was financed by Janssen (Johnson & Johnson), which makes the anticoagulant Xarelto (rivaroxaban), which along with Pfizer’s and Bristol-Myers Squibb’s Eliquis (apixaban) and Boehringer Engelheim’s Pradaxa (dabigatran), are the drugs that stand to benefit from the larger AF patient pool. The big problem, though, is that no one knows how the risk-benefit equation works out in this group.
In an accompanying editorial, Benjamin Steinberg and Jonathan Piccini write that before the findings of the study “can be incorporated into clinical practice, 2 major questions must be considered with regard to structured AF screening: (1) does earlier or more sensitive detection of AF improve clinical outcomes? (2) And if so, is it cost-effective?”
Brave New World?
In a second editorial Eric Peterson and Robert Harrington write that the study “represents a brave new world [italics added] for clinical research: an innovative, highly commendable, contemporary pragmatic health care IT study that tested an important question and yielded significant clinical findings.”
But Peterson and Harrington ignore the fact that the brave new worlds may turn out to be a dystopia. James Stein (University of Wisconsin) outlines in thoughtful detail precisely why we need to be extremely careful before rushing to bring about this brave new world.
It’s all well and good to diagnose treatable diseases early. This study, like many others using a wide range of techniques (community ECG screening, smart watch and smart phones apps, implantable monitors, etc.) shows that if you screen for afib, you find afib. The conclusion “further research is needed regarding clinical implications” is correct, because AF screening (as for many other screening techniques) leads to more upfront costs, inconvenience, exposure to medications with risks for uncertain benefits, and possibly labeling healthy people as sick people for unknown benefits. In this study, 1313 people needed to be screened to find 53 cases of afib – basically 25 screened to find one case. Not to prevent one stroke or death, but just to find a case of AF. We don’t know, however, the rate of strokes or systemic embolism in patients discovered this way or the effectiveness of oral anticoagulation. The ease of screening will, however, lead more and more people at lower risk than those selected for this study to be screened, leading to more costs for unclear benefit, anxiety, and stigmatization.
The study was well conducted and has a pragmatic design for which I commend the authors, but the praise heaped upon them by the editorialists seems excessive. After all, something as important as making a lifelong disease diagnosis and starting a medicine with potential for great benefit, great harms, and great costs, and great profits deserves rigorous study and clear demonstration of efficacy. This is a pretty small step in the right direction to understanding all the pieces of the puzzle of AF screening. I worry that several small pieces will be created by those who profit from this adventure– those who make the monitors, those who make the anticoagulants, and those who promote monitoring and treating on behalf of the companies– before we really understand its role. I further worry that we’ll stop conducting rigorous research once a few pieces point in a direction and the story of screening and treating is spun well enough by thought leaders that it becomes a forgone conclusion. I know I sound like a nihilist but I am not– I just want stronger research before I start exposing my patients to potential harms. The clinical research on AF and stroke prevention is not strong. Even the predictive value of the CHA2DS2-VASc score, which his accepted as gospel and is the basis of our guidelines, is quite poor. So this next step makes me even more concerned.
Stein’s concerns are warranted. Even prior to the publication of the paper in JAMA I received a message from a public relations executive on behalf of iRhythm Technology, the manufacturer of the Zio, describing the study as “a groundbreaking discovery.” No doubt the study may be “groundbreaking” from the perspective of industry and others who benefit from the increased medicalization of the population. Whether or not it will be groundbreaking from the perspective of patients remains to be seen.