Diagnostic performance of a wearing dynamic ECG recorder for Atrial Fibrillation screening: The HUAMI heart study

Abstract Background: Atrial fibrillation (AF) is the most prevalent cardiac dysrhythmia with a significant morbidity and mortality rate. Notably, one out of three patients with AF is asymptomatic. Given the asymptomatic and paroxysmal nature of AF, AF's timely detection with traditional instruments is somewhat unsatisfactory and delayed. Thus, wearing a dynamic electrocardiogram (ECG) recorder can help analyze, interpret, and distinguish AF from normal sinus rhythm accurately and safely, even in an upright position and after exercises, using an artificial intelligence (AI) algorithm.Methods: A total of 114 participants in the outpatient registry of our institution from June 24, 2020 to July 24, 2020, were enrolled. Participants were tested with a wearable dynamic ECG recorder and 12-lead ECG in a supine, an upright position and after exercises for 60seconds. Results: A total of 114 subjects (sixty-one with normal sinus rhythm, fifty-three with AF) were enrolled in the study. The number of cases unable to be determined by the dynamic ECG recorder wristband was two, one in each group. Case results not clinically objective were defined as false-negative or false-positive. The diagnostic accuracy, sensitivity and specificity using wearable dynamic ECG recorders in a supine position were 94.74% (95% CI% 88.76%-97.80%), 88.68% (95% CI 77.06%-95.07%) and100% (95% CI 92.91%-100%), respectively. Meanwhile, the diagnostic accuracy, sensitivity and specificity in an upright position were 97.37% (95% CI% 92.21%-99.44%), 94.34% (95% CI 84.03%-98.65%), and 100% (95% CI 92.91%-100%), respectively. The result after exercise was the same as the result of the upright position.Conclusion: AF can be detected using the widely accessible wearable dynamic ECG recorder with an AI algorithm after different postures and exercises. It may provide a useful and user-friendly screening tool, diagnosing AF early in at-risk individuals.