Home > Cardiology > ACC 2019 > Arrhythmias and Clinical Electrophysiology > Smartwatch can detect atrial fibrillation with high degree of accuracy

Smartwatch can detect atrial fibrillation with high degree of accuracy

Presented By
Dr Mintu Turakhia, Stanford University, USA
ACC 2019
Apple Heart Study
New developments in technology keep finding their way into medicine. A recent application of the Apple Watch is the detection of atrial fibrillation (AF). Early results from the Apple Heart Study, which included over 400,000 participants, showed that the device was not only able to detect AF but is also highly accurate [1].

The overall goal of the study was to evaluate the ability of the irregular pulse notification algorithm to identify AF and guide subsequent clinical evaluation. Each participant in this study needed to have an Apple Watch (series 1, 2, or 3) and an iPhone (the newest Apple Watch model, i.e. series 4 with a built-in electrocardiogram [ECG] app, was not included in this study). Exclusion criteria were AF and the use of anticoagulation. Photoplethysmography was used to measure blood flow activity as well to detect small changes which might indicate an irregular heartbeat. Then a tachogram was created and analysed by an algorithm. When 5 of 6 repeat tachograms showed an irregular pulse within a 48-hour period, a notification to the user was sent via the app. Those who received a notification were advised to contact the study doctor through the app for a video consultation to determine if they should wear an ECG patch. They were using the patch for up to 7 days. The primary endpoints of the study were AF for >30 seconds on ECG patch and simultaneous AF on ECG patch and tachogram. A total of 419,297 people from the USA were enrolled in the study. Interestingly, this large number of participants was collected in 8 months.

The results, presented by Dr Mintu Turakhia (Stanford University, USA) showed that the number of users receiving an AF alert was 0.5% (0.49-0.54). Participants aged over 65 years received notification rates most frequently; it was lowest in those aged <40 years. A total of 658 participants were sent ECG monitor patches of which 450 were returned and included in the analysis. AF was identified in 34% of people who received a notification and wore the ECG patch. The positive predictive value was 71% for the tachogram and 84% for notification. Limitations of the study included lack of a randomised control trial, the reliance on self-reporting, and participants’ own initiative to take part as well as not achieving the pre-specified precision of co-primary endpoints. Also, the target enrolment of 500,000 participants of which with 75,000 should be 65 years was not achieved. However, this trial offers a new approach to doing clinical research, and its findings highlight the useful application of a relatively simple and widely used electronic device.

1. Turakhia M, et al. Abstract 401-07. ACC 2019, 16-18 March, New Orleans, USA.

Posted on