Smartwatch Algorithm Highly Effective for Detection of Atrial Fibrillation

This article originally appeared here.
Share this content:
The KardiaBand allows patients to collect an electrocardiogram (ECG) reading using their Apple Watch. <i>Image Credit: AliveCor</i>
The KardiaBand allows patients to collect an electrocardiogram (ECG) reading using their Apple Watch. Image Credit: AliveCor

The KardiaBand allows patients to collect an electrocardiogram (ECG) reading using their Apple Watch. According to study data published in the Journal of the American College of Cardiology,1 the KardiaBand algorithm accurately detects atrial fibrillation with high sensitivity and specificity comparable to readings performed by electrophysiologists.

Investigators performed a prospective, nonrandomized, and adjudicator-blinded study at a hospital-based electrical cardioversion laboratory. Study enrollees were individuals with atrial fibrillation who were seen at the laboratory for scheduled elective cardioversion recording (n=100). Among initial enrollees, 85% subsequently underwent scheduled cardioversion; 15% either cancelled the procedure or were excluded following presentation in sinus rhythm. From the remaining participants, 169 simultaneous 12-lead ECG and KardiaBand recordings were obtained. Patients were of mean age 68.2±10.86 years, and 17% were women.

Among the readings for which the KardiaBand algorithm provided an interpretation, atrial fibrillation was correctly identified with 93% sensitivity, 84% specificity, and a K coefficient of 0.77 (95% CI, 0.65-0.89) compared with electrophysiologist-interpreted 12-lead ECG. The KardiaBand algorithm classified 57 readings as noninterpretable; among these, electrophysiologists corroborated with 18. Reasons for noninterpretable results included short records (<30 seconds), low-amplitude P waves, and baseline artifacts.

When KardiaBand tracings produced by the smartwatch were interpreted by electrophysiologists instead of the algorithm, results demonstrated 99% sensitivity, 83% specificity, and a K coefficient of 0.83 (95% CI, 0.74-0.92). These data solidify the fidelity of KardiaBand tracings and the efficacy of the associated algorithm in detecting atrial fibrillation and differentiating it from sinus rhythm.

Published in response to this research,2 Shaun K. Giancaterino, MD, and Jonathan C. Hsu, MD, MAS, of the Cardiac Electrophysiology Section at the University of California, San Diego, offered follow-up commentary on the implications of these data. Although controversial, they wrote, mass screening for atrial fibrillation by means of smartphone technology represents a remarkable development in medical device technology. Smartwatch screening could become an element of standard care, given its high accessibility and capacity to detect subclinical fibrillation. Additionally, data synchronization from smartwatch monitoring could serve as a data source for large-scale epidemiologic studies on heart disease. Although issues of “bias, privacy, and data ownership” are implicit in the implementation of smartphone screening, many medical professionals agree on the utility of the KardiaBand in detecting atrial fibrillation, particularly for older patients.

References

  1. Bumgarner JM, Lambert CT, Hussein AA, et al. Smartwatch algorithm for automatic detection of atrial fibrillation. J Am Coll Cardiol. 2018;71(21):2381-2388.
  2. Giancaterino SK, Hsu JC. The smartwatch will see you now: implications of mass screening for atrial fibrillation. J Am Coll Cardiol. 2018;72(12):1433-1434.
You must be a registered member of The Cardiology Advisor to post a comment.

Sign Up for Free e-Newsletters