SmartWATCHes for Detection of Atrial Fibrillation

Atrial Fibrillation

Atrial fibrillation, cardiovascular App, Rhythm monitoring, Stroke prevention

In this trial the Preventives Heartbeats algorithm will be tested in two wearable devices for its specificity and sensitivity to distinguish between AF and SR.

Introduction Particularly in stroke prevention, the detection of atrial fibrillation is of pivotal importance. Recent studies have impressively confirmed the merits of long-term monitoring. Nevertheless, and despite very mature technology, the currently available methods like long-term ECG with conductive adhesive electrodes or implantable loop recorders are burdened with disadvantages in that they are inconvenient, costly and/or invasive. A predecessor study gave us the possibility to test an app that employs photoplethysmographic signals on a smartphone camera-similar to pulse oximetry technology-to distinguish between atrial fibrillation and sinus rhythm. That retrospective study achieved a sensitivity and specificity of 95% (Krivoshei et al., Europace 2016). The prospective follow-up study DETECT AF pro is currently being conducted in cooperation with Prof. Dörr in Greifswald, Germany. Our research group has been involved with testing the app's quality in clinical settings before implementing it for use. This procedure differs markedly from the otherwise common practice of marketing qualitatively inferior health apps without any clinical testing. There is no app equipped with comparable technology worldwide. The WATCH AF study is the first study on sophisticated rhythm analysis using a smartwatch worldwide. Study design Prospective, blinded, multicenter study. The blinded data will be evaluated and monitored externally. An interim analysis to assess signal quality will be performed after inclusion of 50% of the planned subjects. If there are \<10% non-evaluable signals, the number of subjects will be adjusted accordingly. The aim is to enroll 600 evaluable subjects. Methods The subjects will be asked to place a smartwatch on each arm to allow their pulse curve to be recorded for five minutes. At the same time, an ambulatory ECG system will take a synchronous ECG for reference. The ECG system employed is equipped with a validated automatic atrial fibrillation detector, used as a reference for assignment to groups. Data will be collected on person-related information, comorbidities and medication (estimated overall duration/patient \<20 min). No follow-up is planned. Analogous to predecessor studies, the pulse wave curve data are coded with the patient's identification number (ID) and externally analyzed at Preventicus; based on the results, the patients will be assigned to SR/AF groups. After evaluation of all files, the grouped results will be aggregated, unblinded and evaluated under monitoring. Primary target parameters are the app's sensitivity and specificity in correctly detecting atrial fibrillation compared to an automatically interpreted ECG. Secondary target parameters include the proportion of non-evaluable recordings in the overall study, differences between the two smartwatches and the app's sensitivity and specificity in patients with \>2% premature beats.

Germany, Switzerland