Abstract 16810: Validation of an iPhone ECG Application Suitable for Community Screening for Silent Atrial Fibrillation: A Novel Way to Prevent Stroke
Background Atrial fibrillation (AF) is the most common sustained arrhythmia and carries a 5-fold increased risk of stroke. AF is common in the elderly, and previously unrecognised AF is estimated to occur in 1-2% of those over 65. A significant proportion of ischemic strokes are due to previously unrecognised AF, which is largely asymptomatic, and may therefore require community screening to detect the arrhythmia. Until now, community ECG screening has not been considered a feasible approach for AF detection.
Aim We tested a novel iPhone application suitable for community use (AliveCor), which records and sends a high quality single-lead ECG to a server for automated analysis, and subsequent interpretation by a physician to diagnose AF.
Methods A single-lead ECG was recorded using an iPhone in 109 patients (70 sinus rhythm and 39 AF) soon after a 12-lead ECG had been performed. iPhone ECGs were uploaded to the AliveCor server for subsequent interpretation by two cardiologists blinded to the rhythm diagnosis. The iPhone ECG was also processed on the server to provide an automated diagnosis of sinus rhythm or AF. Results were compared with the 12-lead ECG diagnosis by a third cardiologist.
Results Sensitivity and specificity of the iPhone single lead ECG for AF diagnosis, overall accuracy and kappa (κ) were 100%, 90%, 94%, κ 0.87 for cardiologist A; 95%, 94%, 95%, κ 0.88 for cardiologist B, and 87%, 97%, 94%, κ 0.86 for the automated algorithm respectively. After algorithm optimisation to enhance sensitivity, results for automated analysis were 100%, 96%, 97%, κ 0.94 respectively.
Conclusion A high quality single-lead ECG can be rapidly and simply recorded using an iPhone with the AliveCor application, to accurately diagnose AF, making this an ideal enabling technology for community screening programs to detect silent AF. Screening programs utilising this device could have a substantial impact on reducing ischemic stroke related to previously undiagnosed AF.
- © 2012 by American Heart Association, Inc.