Comparison between smartphone 6-lead ECGs and 12-lead ECGs in athletes

Abstract

Abstract Introduction Athletes occasionally experience transient arrythmias that are difficult to capture using traditional methods. Smartphone electrocardiogram (ECG) technology presents an innovative method of capturing transient arrythmias. A 6-lead (6L) smartphone device that measures leads I, II, III, aVR, aVL, and aVF, has been shown to be accurate in some populations but has limited evidence in athletes. A previous pilot study comparing 6L with 12L readings in athletes showed good correlations. Purpose To assess the accuracy of the 6L device in athletes by comparing the 6L with a 12-lead (12L) ECG in a larger sample size under real-world conditions. Methods From 2020-2023, athletes (n=217, mean age 18.0±3.8 years, 40% female) had a resting supine 12L ECG as part of cardiac screening and a seated 30s 6L reading within 2 hours of the 12L. Manual measurements of RR, QT, PR intervals, and QRS duration in the 6L and 12L ECGs were completed by a cardiologist using EPS Digital Calipers in lead II. A subset (n=30) was measured by 4 expert cardiologists and the mean was taken. Each measurement was calculated as an average of at least 6 consecutive beats in the 6L and 12L. Bland-Altman analysis was used to assess the quantitative agreement between heart rate (HR), QTc interval (Bazett), PR interval and QRS duration in the 6L and 12L. ECGs were analysed for agreement of QRS axis (normal, left and right axis deviation defined by the International Criteria for Athlete ECG Interpretation[1]). Results The 6L readings for HR were 5.0±11.0bpm faster than the 12L; 6L readings for other measurements were shorter than the 12L. QTc had the largest difference, with the 6L on average 14.2±27.4ms shorter than the 12L reading. For QTc, 5.5% of readings were >60ms different, mostly due to baseline interference in the 6L traces. PR intervals were 13.9±16.3ms shorter in the 6L than the 12L. QRS duration measurements were similar, with 6L readings being 0.8±10.2ms shorter than the 12L reading. There was 96.8% agreement on QRS axis classification between 6L and 12L. Conclusion The 6L readings had fair agreement with the 12L ECG, although the standard deviations were higher than measured in the pilot study, suggesting less consistent accuracy of the 6L. Most 6L measures (excluding HR) were slightly shorter than the 12L, consistent with studies in other populations. QRS axis categories were very similar. Given the change from supine to seated, it is unsurprising that HR was higher in the 6L. Our previous work showed good correlation between 6L and 12L readings and excellent utility of the portable 6L device for detecting transient arrhythmias in athletes. This larger study confirms most of those findings but adds caution that baseline interference is not uncommon in real-world 6L readings, meaning that QTc may be difficult to accurately measure.