Background: Diagnosis of atrial fibrillation (AF) can be achieved by using an ECG recording of 30 seconds and is often diagnosed with delay in asymptomatic patients. Meanwhile, commercially available devices enable patients to record ECGs themselves, giving the opportunity to increase the likelihood of early AF detection. Smart watches allow ECG registration only in a single lead, usually equivalent to Einthoven lead I. However, best practice of ECG recording with these devices has not been studied so far. We hypothesized that a modified ECG registration, creating a lead equivalent to Einthoven lead II and thereby following the main electrical cardiac vector, could optimize diagnosis of AF with the automated algorithm.

Material and Methods: In total, 39 patients (27 male, median age 70.9 ± 10.2 years, median BMI 29.0 ± 5.5 kg/m², median left ventricular ejection fraction 52.1 ± 10.0 %) with persistent atrial fibrillation scheduled for electrical cardioversion were included in the present study. ECG recording was performed with the Apple Watch Series 5® (Apple^® Inc, Cupertino, CA, USA). All ECGs were recorded in two different leads: Einthoven lead I between the left wrist and right index finger and Einthoven lead II between the left lower abdominal region and right index finger [figure 1]. Comparisons were performed using the Mann-Whitney U test. A p-value < 0.05 was considered to be statistically significant.

Results:

Correct diagnosis of AF with the automated diagnostic algorithm of the smart watch using lead I could be achieved in 46.2% patients. However, in significantly more patients correct diagnosis of atrial fibrillation was achieved by recording the ECG between the left lower abdomen and right index finger (79.5%; p = 0.002). Amplitudes of the recorded QRS-complexes were significantly higher by recording ECGs between the left lower abdominal region and right index finger instead of using the left wrist and right index finger (0.8 ± 0.3 mV versus 0.6 ± 0.2 mV, respectively; p = 0.004).

Conclusion: When using a smart watch for documentation of atrial arrhythmias, patients should be instructed to perform ECG recording between the lower left abdomen and right index finger to optimize sensitivity of the automated diagnostic algorithm and to increase the quality of ECG documentation for interpretation by physicians.

Figure 1: Difference of ECG amplitude in two different leads recorded by a smart watch in the same patient with sinus rhythm (male, 39 y/o, BMI 21.1 kg/m², LVEF 60%).  

a) Registration of Einthoven lead I between the left wrist and right index finger. Diagnosis by the automated diagnostic algorithm was “unclassified” b) Recording of Einthoven lead II between the left lower abdominal region and right index finger. Note the higher amplitude of QRS-complexes and clearly visible p-waves, leading to the correct diagnosis of sinus rhythm by the automated algorithm.