Using a smartwatch electrocardiogram to detect abnormalities associated with sudden cardiac arrest in young adults

Abstract

Abstract Aims  Smartwatch electrocardiograms (ECGs) could facilitate the detection of sudden cardiac arrest (SCA)-associated abnormalities. We evaluated the feasibility of using smartwatch-derived ECGs for detecting SCA-associated abnormalities in young adults and its agreement with 12-lead ECGs. Methods and results  Twelve-lead and Apple Watch ECGs were registered in 155 healthy volunteers and 67 patients aged 18–45 years with diagnosis and ECG signs of long-QT syndrome (n = 10), Brugada syndrome (n = 12), ventricular pre-excitation (n = 19), hypertrophic cardiomyopathy (HCM, n = 13), and arrhythmogenic right ventricular dysplasia/cardiomyopathy (ARVC/D, n = 13). Cardiologists separately analysed 12-lead ECGs and the smartwatch ECGs taken from the left wrist (AW-I) and then from chest positions V1, V3, and V6 (AW-4). Compared with AW-I, AW-4 improved the classification of ECGs as ‘abnormal’, increasing the sensitivity from 64% to 89% (P < 0.01). Pre-excitation was detected in most cases using AW-I (sensitivity 89%) and in all cases using AW-4 (sensitivity 100%, P = 0.48 compared with AW-I, specificity 100% for both). Brugada was missed using AW-I but was detected in 11/12 patients using AW-4 (sensitivity 92%, specificity 100%, P = 0.003). Long QT was detected in 8/10 cases using AW-I (sensitivity 80%, specificity 100%) and in 9 patients using AW-4 (sensitivity 90%, specificity 100%, P > 0.99). Hypertrophic cardiomyopathy was correctly suspected using AW-I and AW-4 (sensitivity 92% and 85%, specificity 85%, and 100%, P > 0.99). AW-I was mostly (62%) considered normal in ARVC/D whereas AW-4 was useful in suspecting ARVC/D (100% sensitivity, 99% specificity, P = 0.004). Conclusions  Detection of SCA-associated ECG abnormalities (pre-excitation, Brugada patterns, long QT, and signs suggestive of HCM and ARVC/D) is possible with an ECG smartwatch.