Smartphone Detection of Atrial Fibrillation Challenging with Abnormal ECGs - MD Magazine

Mobile wellness technologies to observe atrial fibrillation (AF) are linked to a precocious complaint of mendacious positives and inconclusive results successful patients with definite cardiac conditions, according to caller findings.

The largest real-world survey to day indicated that the usage of these devices is peculiarly challenging successful patients with abnormal electrocardiograms (ECGs). Both improved algorithms and instrumentality learning could assistance the tools supply much close diagnoses, according to investigators.

“With the increasing usage of smartwatches successful medicine, it is important to cognize which aesculapian conditions and ECG abnormalities could interaction and change the detection of AF by the smartwatch successful bid to optimize the attraction of our patients,” said pb survey writer Marc Strik, MD, PhD, LIRYC institute, Bordeaux University Hospital. “Smartwatch detection of AF has large potential, but it is much challenging successful patients with pre-existing cardiac disease.”

In theory, the usage of extended cardiac monitoring successful patients and the usage of implantable cardiovascular physics devices whitethorn summation detection of AF. However, limitations with the devices see a abbreviated artillery beingness and a deficiency of contiguous feedback.

New smartphone tools that person the quality to grounds an ECG portion and marque an automated diagnosis whitethorn frankincense flooded the supra limitations and pb to a timely diagnosis. Strik noted that anterior studies person validated the accuracy of the Apple Watch for the diagnosis of AF successful a “limited fig of patients with akin objective profiles.”

The investigators performed a trial connected the accuracy of the Apple Watch ECG app successful the detection of AF successful patients with a assortment of coexisting ECG abnormalities.

Their survey included a full of 734 consecutive hospitalized patients. Each underwent a 12-lead ECG, with contiguous follow-up by a 30-second Apple Watch recording.

Investigators reported that each smartwatch’s automated single-lead ECG AF detections were classified arsenic “no signs of atrial fibrillation,” “atrial fibrillation,” oregon “inconclusive reading.” The recordings were fixed to an electrophysiologist who performed a blinded reading, assigning each tracing a diagnosis of “AF,” “absence of AF,” oregon “diagnosis unclear.” A 2nd blinded electrophysiologist analyzed 100 randomly selected traces to find “the grade to which the observers agreed.”

The findings indicated that the smartwatch ECG failed to nutrient an automatic diagnosis successful astir 1 successful each 5 patients.

Additionally, investigators reported the hazard of having a mendacious affirmative automated AF detection was higher for patients with premature atrial and ventricular contracts (PACs/PVCs), sinus node dysfunction, and second- oregon third-degree atrioventricular-block.

For those successful AF, the hazard of having a mendacious antagonistic tracing (missed AF) was reported arsenic higher for patients with ventricular conduction abnormalities (intraventricular conduction delay) oregon rhythms controlled by an implanted pacemaker. Moreover, the cardiac electrophysiologists had a precocious level of statement for differentiation betwixt AF and non-AF.

Data bespeak the smartphone app correctly identified 78% of patients successful AF and 81% who were not successful AF. Meanwhile, the electrophysiologists identified 97% of patients who were successful AF and 89% who were not successful AF.

Those with PVCs were 3 times much apt to person mendacious affirmative AF diagnoses from the smartwatch ECG according to the data, and the recognition of patients with atrial tachycardia (AT) and atrial flutter (AFL) was considered precise poor.

“These observations are not surprising, arsenic smartwatch automated detection algorithms are based solely connected rhythm variability,” Strik added. “Ideally, an algorithm would amended discriminate betwixt PVCs and AF. Any algorithm constricted to the investigation of rhythm variability volition person mediocre accuracy successful detecting AT/AFL. Machine learning approaches whitethorn summation smartwatch AF detection accuracy successful these patients.”

The article, “Role of Coexisting ECG Anomalies successful the Accuracy of Smartwatch ECG Detection of Atrial Fibrillation,” was published successful the Canadian Journal of Cardiology.