20.11.2023
Diagnostic accuracy of Apple Watch Series 6 recorded single-lead ECGs for identifying supraventricular tachyarrhythmias: a comparative analysis with invasive electrophysiological study
verfasst von:
Kivanc Yalin, Ali Ugur Soysal, Baris Ikitimur, Beyza Irem Yabaci, Sukriye Ebru Onder, Adem Atici, Hasan Tokdil, Gunduz Incesu, Hakan Yalman, Murat Cimci, Hakan Karpuz
Erschienen in:
Journal of Interventional Cardiac Electrophysiology
|
Ausgabe 5/2024
Einloggen, um Zugang zu erhalten
Abstract
Background
The advancements in wearable technology have made the detection of arrhythmias more accessible. While smartwatches are commonly used to detect patients with atrial fibrillation, their effectiveness in the differential diagnosis of supraventricular tachycardias (SVT) lacks consensus.
Methods
A study was conducted on 47 patients with documented SVTs on a 12-lead ECG. All patients in the cohort underwent electrophysiology study with induction of SVT. A 6th generation Apple Watch was used to record ECG tracings during baseline sinus rhythm and during induced SVT. Cardiology residents and attending cardiologists evaluated these recordings to diagnose the differential diagnosis of SVT.
Results
The evaluation revealed 27 cases of typical atrioventricular nodal reentrant tachycardia (AVNRT), 11 cases of atrioventricular reentrant tachycardia (AVRT), and 9 cases of atrial tachycardia/atrial flutter (AT/AFL) among the induced tachycardias. Attending physicians achieved an accuracy of 66.0 to 76.6%, and residents demonstrated accuracy rates between 68.1 and 74.5%. Interrater reliability was assessed using Fleiss’s Kappa method, resulting in a moderate level of agreement between residents (Kappa = 0.465, p < 0.001, 95% CI 0.30–0.63) and attendings (Kappa = 0.519, p < 0.001, 95% CI 0.35–0.68). The overall Kappa value was 0.417 (p < 0.001, 95% CI 0.34–0.49).
Conclusions
Smartwatch recordings demonstrate moderate feasibility in diagnosing SVT when following a pre-specified algorithm. However, this diagnostic performance was lower than the accuracy obtained from 12-lead ECG tracings when blinded to procedure outcomes.