Dr. Chris Notte and Dr. Neil Skolnik
In an ad for one of these products, KardiaMobile, a cardiologist says this device “detects atrial fibrillation, one of the major causes of stroke.” You might also have heard that the Apple Watch has an opt-in feature that constantly screens for atrial fibrillation without any effort being made by the patient, or can check on-demand for AFib if a wearer experiences palpitations or an abnormal heart beat. Both of these devices generate a standard limb–lead ECG (essentially lead I) by connecting the device to both arms and producing a 30-second rhythm strip.
KardiaMobile recently introduced a newer device. When you place this device on a bare knee and touch one electrode with fingers from the right hand and another electrode with fingers from the left hand, the device produces a six-lead ECG. These small devices send an image of the ECG to a patient’s smartphone over Bluetooth, and the results can be easily read, printed out, or sent to the doctor for further analysis. Additionally, both of KardiaMobile’s devices utilize artificial intelligence to analyze a rhythm strip in real time and let the patient know if the ECG is normal, shows AFib, or is unable to be analyzed.
The electrocardiographic technology was formerly only available in a medical setting. It required an expensive machine and could only be interpreted by someone with expertise developed through years of training. Now it is readily available to patients in their homes. But how accurate is the technology and how are we going to use it?
How effective is KardiaMobile at detecting AFib?
Studies have looked at both KardiaMobile and the Apple Watch. One study of KardiaMobile in patients with Afib who were admitted for antiarrhythmic drug initiation showed that about a quarter of readings could not be classified because of artifact and other reasons. After exclusion of unclassified recordings, the KardiaMobile interpretation had 97% sensitivity and 94% specificity for AFib detection when compared with physician-interpreted ECGs.1 In a large review of the device’s accuracy, there was about 85% sensitivity and specificity of the automated readings.2
How does the Apple Watch find AFib?
Like the KardiaMobile device, the Apple Watch can be used whenever patients notice symptoms or whenever they and their physicians decide the device would be useful. In addition, though, the Apple Watch has a function where the wearer can opt in to have the watch screen for AFib in the background whenever the watch is worn.
The watch monitors heart rate using photoplethysmography, where light-sensitive photodiodes detect blood pulses to assess heart rate variability. When an irregular heart rate is detected, the AW alerts the user of possible AFib. Once alerted, the wearer can then utilize a second function to obtain a single-lead ECG. Heart rate, rhythm, and a 30-second ECG tracing are saved in the Bluetooth-linked iPhone’s health app and can be exported for review by a physician.
In a study of over 400,000 participants, among participants notified of an irregular pulse through screening there was a positive predictive value of 84%.3 Single-lead EKGs initiated by watch wearers had a specificity for AFib of 99.6% among tracings with good wave forms, indicating very few false positives. Only 1 individual of the 263 individuals who had normal sinus rhythm on 12-lead ECG was classified as having AFib, though in 7% sinus rhythm could not be confirmed because of poor tracings.4,5
What should we do with the results?
It’s impressive that these devices deliver accurate information with very good specificity. Our hope is that detecting AFib with one of these devices will lead to an intervention being made that will decrease a patient’s risk of stroke. But it is not clear if routine screening in asymptomatic adults will accomplish this.
While more data is needed, we must acknowledge that our patients will soon be bringing us results from home. Regardless of what we think of this technology, we need to decide what to do when patients call us with results from these devices.
Dr. Notte is a family physician and chief medical officer of Abington (Pa.) Hospital–Jefferson Health. Follow him on Twitter (@doctornotte). Dr. Skolnik is professor of family and community medicine at Sidney Kimmel Medical College, Philadelphia, and associate director of the family medicine residency program at Abington Hospital–Jefferson Health. They have no conflicts related to the content of this piece.
References
1. William A et al. Heart Rhythm. 2018 Oct;15(10):1561-5.
2. KardiaMobile for the ambulatory detection of atrial fibrillation. NICE Medtech innovation briefing. 29 October 2020 Oct 29. www.nice.org.uk/guidance/mib232.
3. Perez MV et al. N Engl J Med. 2019; 381:1909-17.
4. Using Apple Watch for Arrhythmia Detection, December 2018. Apple. https://www.apple.com/healthcare/site/docs/Apple_Watch_Arrhythmia_Detection.pdf. Accessed 2019 Apr 5.
5. De Novo Classification Request for ECG App. https://www.accessdata.fda.gov/cdrh_docs/reviews/DEN180044.pdf. Accessed 2019 Apr 29.