PO-614-06 SMART WATCH SYNDROME IN PEDIATRICS - HOW WELL DOES IT PREDICT ARRHYTHMIA?

Abstract

Smart watches have the capability of assessing heart rate (HR) and rhythm; some can produce a single lead ECG tracing. These features may enhance detection of atrial fibrillation in adults based on a recent study. Tachyarrhythmias in children such as supraventricular tachycardia (SVT) may also be detectable with a smart watch. Alternatively, misunderstood or inaccurate smart watch data may lead patients to seek unnecessary evaluation. Assess the likelihood of a true arrhythmia in pediatric patients presenting with concerns about smart watch cardiac data. Single center retrospective review of children aged 10-18 years who had ever presented to the pediatric cardiology clinic with concerns related to smart watch cardiac data. The primary study outcome was diagnosis of arrhythmia based on clinical evaluation or documentation of arrhythmia by clinical testing. There were 126 patients (mean age 15.6 ± 2.4 yrs) that presented with a smart watch based rhythm concern - tachycardia in 89%. Symptoms were present in 95 (75%); with palpitations accounting for 78% of those. Smart watch measured HRs were available in 121/126 (96%) with 45 (37%) reporting HR ≥ 190 bpm. Presenting smart watch data was sufficient to diagnose SVT in 3. Additional testing was used to confirm or rule out arrhythmia in 72 (57%). The majority, 83 (66%), were discharged after a single visit +/- testing. In all, 19 / 126 (15%) patients were diagnosed with true arrhythmia: 13 SVT, 3 Wolff Parkinson White, 2 atrial tachycardia, 1 ventricular ectopy. The odds of a true arrhythmia diagnosis with symptoms vs no symptoms was 3.2 (95%CI 0.7 - 14.5); and with HR ≥ 190 bpm vs HR < 190 was 14.3 (95%CI 3.8 - 52.8). The positive predictive value of HR ≥ 190 AND symptoms together to predict arrhythmia was only 39% (95%CI 28 - 52). The negative predictive value for arrhythmia having neither symptoms nor HR > 190 was 95% (95%CI 75 - 99). The likelihood of a true arrhythmia in pediatric patients presenting with a smart watch based HR concern was low. Symptoms and HR > 190 improved but did not optimize the predictability of an arrhythmia. The absence of symptoms or HR > 190bpm predicted no arrhythmia in 95% of patients. Rarely, smart watch EGMs or trend data was sufficient for arrhythmia diagnosis.