Mobile health technology to detect atrial fibrillation results in a high rate of false positives and inconclusive results in some patients with certain heart conditions, researchers report in the Canadian Journal of Cardiology
Cardiovascular electronic devices can increase the detection of atrial fibrillation (AF), but have limitations including short battery life and lack of immediate feedback.
Can new smartphone tools that can record an electrocardiogram (ECG) strip and perform automated diagnosis overcome these limitations and facilitate timely diagnosis?
The largest study to date, in the Canadian Journal of Cardiology , finds that the use of these devices is challenging in patients with abnormal ECGs. Researchers say better algorithms and machine learning can help these tools provide more accurate diagnoses. "Previous studies have validated the accuracy of the Apple Watch for diagnosing AF in a limited number of patients with similar clinical profiles," explained lead investigator Marc Strik, MD, PhD, LIRYC Institute, Bordeaux University Hospital, Bordeaux, France.
"We tested the accuracy of the Apple Watch ECG app in detecting atrial fibrillation (AF) in patients with a variety of coexisting ECG abnormalities." The study included 734 consecutive hospitalized patients. Each patient underwent a 12-lead ECG, immediately followed by a 30-second recording on the Apple Watch. Smartwatch automated single-lead ECG AF detections were classified as “no signs of atrial fibrillation,” “atrial fibrillation,” or “inconclusive reading . ”
The smartwatch recordings were given to an electrophysiologist, who performed a blinded interpretation , assigning each trace a diagnosis of "AF" , "absence of AF", or "unclear diagnosis" . A second blinded electrophysiologist interpreted 100 randomly selected tracings to determine the extent to which observers agreed.
In about one in five patients, the smartwatch ECG failed to produce an automatic diagnosis. The risk of having a false positive in automated AF detection was higher in patients with premature atrial and ventricular contractions (PAC/PVC), sinus node dysfunction, and second- or third-degree atrioventricular block.
For patients with atrial fibrillation (AF), the risk of having a false-negative tracing (missed AF) was higher for patients with ventricular conduction abnormalities (interventricular conduction delay) or rhythms controlled by an implanted pacemaker.
Cardiac electrophysiologists had a high level of agreement for the differentiation between AF and non-AF.
The smartphone app correctly identified 78% of patients who had AF and 81% who did not have AF. Electrophysiologists identified 97% of patients who had AF and 89% who did not.
Figure : Example of electrocardiograms (ECG) false positive (a normal ECG could not be diagnosed in a patient without AF) caused by the appearance of premature complexes or bradycardia and false negative (AF could not be detected in a patient with AF) ECG caused by wide QRS complexes and ventricular pacing (Credit: Canadian Journal of Cardiology) .
Patients with ventricular premature beats (EVs) were three times more likely to have false-positive AF diagnoses on the smartwatch ECG, and identification of patients with atrial tachycardia (AT) and atrial flutter (AFL) was very poor .
"These observations are not surprising, since smartwatch automated detection algorithms rely solely on cycle variability ," noted Dr. Strik, explaining that ventricular extrasystoles (VEs) cause short and long cycles, which increases cycle variability. “Ideally, an algorithm would better discriminate between ventricular extrasystoles (VE) and AF. Any algorithm limited to cycle variability analysis will have poor detection accuracy. “Machine learning approaches may increase the accuracy of smartwatch AF detection in these patients.”
In an accompanying editorial, Andrés F. Miranda-Arboleda, MD, and Adrian Baranchuk, MD, Division of Cardiology, Kingston Health Science Center, Kingston, ON, Canada, noted that this is the first “ real world” study to focus on the use of the Apple Watch as a AF diagnostic tool. “It is of notable importance because it allowed us to know that the performance of the Apple Watch in diagnosing AF is significantly affected by the presence of underlying ECG abnormalities. In some ways, smartwatch algorithms for detecting AF in patients with cardiovascular diseases are still not smart enough. But they may soon be,” Dr. Miranda-Arboleda and Dr. Baranchuk said. "With the increasing use of smartwatches in medicine, it is important to know what medical conditions and ECG abnormalities could affect and alter AF detection using the smartwatch to optimize care for our patients," Dr. Strik said. d. "AF detection with smart watches has great potential, but is more difficult in patients with pre-existing heart disease."
