Supraventricular tachycardias (SVT) such as AV-nodal reentry tachycardia (AVNRT) and AV-reentry tachycardia (AVRT) are common etiologies of symptomatic regular tachycardia and can lead to significant subjective distress [1]. Ablation during an electrophysiology (EP) study is the curative first-line treatment [2, 3]. Inappropriate sinus tachycardia (IST) is an important differential diagnosis not amenable to invasive treatment in the majority of cases. The recording of an ECG during a symptomatic episode, therefore, remains the cornerstone of a correct diagnosis in suspected SVT [4]. However, due to often short-lasting symptoms, patients frequently struggle to reach a health-care setting in time for the recording of an ECG. This leaves the treating physician with the choice to either conduct an empirical EP study, even in patients with IST, or to have the patient continue trying to record the symptomatic tachycardia. Smartphone-based one-lead ECG devices (sECG) might be useful to quickly record the underlying rhythm, removing the need for an EP study when IST is documented. This would eliminate possible risks of invasive EP studies [5] and could have a beneficial impact on overall health care expenditure with optimized access to appropriate treatment. We, therefore, initiated this unsponsored, investigator-driven study with the FDA-cleared AliveCor Kardia sECG to test the hypothesis that the quality of an sECG recording would be sufficient to differentiate SVT from IST.

Forty seven of the 75 patients (63%) were female, the mean age of the patient population was 50 ± 18 years and the mean BMI was 26 ± 5 kg/m². In 60 patients, a typical slow–fast AVNRT was the underlying SVT etiology. The remaining 15 patients exhibited an AVRT with only retrograde conduction properties of the accessory pathway. Accessory pathways were located posteriorly in ten patients, parahisian in two patients, midseptal in two patients and inferoseptal in one patient.

The present unsponsored, investigator-initiated study is the first to report on the diagnostic utility of a smartphone-based one-lead ECG recorder for the diagnosis of regular supraventricular tachycardia. We were able to demonstrate high accuracy of the sECG in the detection of SVT when compared to the gold-standard invasive EP procedure.

Our studied population of consecutive patients closely resembles a standard SVT population in gender and age distribution. While many patients present at a relatively young age, SVT can develop at any stage of life. We were able to show that sECG reliably record the underlying rhythm in paroxysmal tachycardia regardless of the age of the affected patient or other possible signal influencing factors such as gender or BMI.

Smartphone-based ECG recording devices have previously been evaluated for the diagnosis of atrial fibrillation and were shown to reliably document the underlying rhythm [6‐8]. Subsequently, one-lead ECG devices such as the AliveCor Kardia and the Apple Watch have been increasingly used by patients to document symptomatic tachycardias [10]. However, concerns seem to persist among cardiologists: in a recent European survey, only half of respondents would have performed an EP study when sECG showed (a) a narrow complex tachycardia in a patient with (b) typical on/off palpitations [11]. In this regard, the present study might close an important gap by establishing that regular SVTs can be differentiated from inappropriate sinus tachycardia in sECG with a high accuracy, which holds especially true in high-quality sECG.

With exponentially growing medical expenditures, there is urgent need to reduce global economic burden on our health care systems. One of the promising strategies is digital medicine to make processes more efficient. Out of necessity, during the current COVID-19 pandemic, digitalization of medical processes has been hugely intensified. While the 12-lead ECG remains the most accurate noninvasive diagnostic test to distinguish SVT from sinus tachycardia, our data may help clinicians integrate sECG into clinical practice in suspected SVT, thereby promoting the transfer to digital medicine.

Sensitivity (89%) and specificity (91%) for the detection of SVT were comparable with previously reported sensitivity (90–100%) and specificity (89–97%) for the detection of atrial fibrillation [6‐8]. The diagnostic utility and interobserver agreement considerably increased when sECG were high quality, highlighting the need for patient instruction on how to record episodes. Strategies to reduce artifact in recorded ECG may include constant pressure of the fingers on the sECG recorders, no movement of fingers while recording, the use of different fingers, small amounts of water to reduce electrical impedance or the use of alternative recording vectors [7, 12]. Unsurprisingly, specificity for SVT detection increased when two electrophysiologists agreed on the underlying rhythm.

This is the first study to compare sECG recordings to the gold-standard EP study. We were thus able to directly compare the diagnostic yield of the sECG to a 12-lead ECG, which showed an equivocal recording in rare cases (see Fig. 3). While the differentiation of atrial fibrillation from sinus rhythm can be achieved even when there is a considerable amount of artifact in the recording by analyzing the regularity of R–R intervals, distinguishing SVT from IST, at least for now, requires interpretation of the whole ECG by a human specialist (see Figs. 1, 2), as both SVT and IST typically present with a tachycardia with regular R–R intervals. We were able to demonstrate that sECG recordings can—with human interpretation—reliably be employed even in this higher demand setting and diagnostic accuracy does not relevantly diminish in comparison with the diagnosis of atrial fibrillation. Whether machine learning/artificial intelligence can reliably interpret narrow complex sECG in the near future remains to be seen.

In everyday electrophysiology practice, patients regularly present with a history of highly symptomatic paroxysmal tachycardias suggestive of SVT without ECG documentation. Our data show that smartphone-based one-lead ECG recordings can be utilized in suspected SVT and may, therefore, reduce the time to definitive diagnosis and a potential cure in the form of EP-guided ablation. Furthermore, should the underlying rhythm be inappropriate sinus tachycardia, a potentially unnecessary invasive EP study can be avoided, reducing possible distress and complications for the individual and decreasing overall economic healthcare burden.

The present results are not applicable to all SVT as it would be very difficult to distinguish atrial tachycardia from sinus tachycardia in a single-lead ECG. However, as atrial tachycardias comprise only about 5–10% of patients [13, 14], there would only be a small fraction of SVT patients, where a differentiation would not be possible. Recorded SVT had a shorter median cycle length than recorded sinus tachycardias. While this may have allowed the electrophysiologists to distinguish between the recordings more easily (the faster the narrow complex tachycardia, the more likely the diagnosis is SVT) the difference in cycle length did not affect the likelihood of a correct diagnosis, making confounding by cycle length unlikely. Additionally, by recording sinus tachycardia and SVT in the same patients we were able to considerably reduce the possible confounder of inter-individual sECG recording quality. Furthermore, we would like to stress the importance of patient instruction in how to record a high-quality sECG, since insufficient recording quality may remain as the predominant factor in preventing the widespread use of sECG.

A smartphone-based one-lead ECG can reliably differentiate SVT from IST, especially in high-quality recordings. This highlights the need to instruct patients how to optimally use wearables to increase the diagnostic yield. Our analyses also indicate that in difficult cases, consulting a colleague increases the likelihood of a correct ECG diagnosis. The results of this study hopefully encourage cardiologists to implement digital medicine in their workflow of diagnosing several cardiac arrhythmias, not only atrial fibrillation.