Monitoring of cardiorespiratory vagal desynchrony using novel biomarkers derived from smartwatch electrocardiograms in a patient recovering from long COVID: case report

Abstract

Background

Long COVID and cardiovascular autonomic dysfunction, including postural orthostatic tachycardia syndrome (POTS), present significant healthcare challenges. Long-term monitoring is challenging due to the evolving nature of symptoms and the limited availability of objective diagnostic tools. With over 200 million electrocardiogram (ECG)-enabled smartwatches sold worldwide, these devices offer a promising solution for at-home diagnostics and disease tracking.

Methods and results

This study examines a 35-year-old male with long COVID, POTS, and chronic fatigue syndrome (CFS), who recorded 328 ECGs over using a Samsung smartwatch. The protocol required ECG recordings to be taken first in a sitting posture, followed by a standing position, with slow, controlled breathing. For testing, the patient used a Samsung smartwatch to perform a 30-s hand-to-hand single-lead ECG while engaging in 0.1 Hz diaphragmatic controlled breathing, consisting of 5 s of inhalation followed by 5 s of exhalation (Appendix 1). S-/R-peak amplitude ratios, heart rhythm changes, and other biomarkers were analysed to assess autonomic function. Fatigue levels were self-reported via the BREATHE FLOW app using a three-grade scale, and health status was tracked monthly with the EQ-5D-5L model. Initially, the patient experienced severe fatigue and heart rhythm changes consistent with POTS. Electrocardiogram analysis revealed an increased S-wave amplitude and higher S/R ratio in standing posture, along with worsening respiratory sinus arrhythmia (RSA), indicating cardiorespiratory desynchrony. Over time, as symptoms improved, heart rate responses between sitting and standing normalized, and S/R ratio and RSA index followed self-reported fatigue levels, including fluctuations due to post-exercise fatigue.

Conclusion

Smartwatch-derived S-/R-wave amplitude ratio may serve as an accessible biomarker for tracking disease progression in long COVID. Given the widespread availability of smartwatches, standardized at-home protocols could improve diagnostics and monitoring for autonomic dysfunction.

Learning points.

• By utilizing posture-based electrocardiogram (ECG) recordings with a smartwatch, the patient and healthcare providers can get more information about the patient’s health.

• Posture-based protocols can provide new types of cardiac information that previously has not been available to healthcare providers.

• Remote monitoring with smartwatches is a viable alternative for ‘at-home’ monitoring of cardiovascular diseases, especially such which are connected to the parasympathetic and sympathetic signalling to the heart.

Introduction

Long COVID and one of its principal manifestations, postural orthostatic tachycardia syndrome (POTS), have become significant health challenges in the post-COVID-19 pandemic era. Although effective treatment is not available, general measures, such as physiotherapy, off-label drug prescription, and spontaneous recovery, may lead to substantial improvement over time. Objective monitoring of long COVID patients is limited by a lack of reliable tests and the need for frequent evaluation, especially in the home environment.

Electrocardiogram (ECG)-enabled smartwatches record and transmit a single-lead ECG equivalent to Lead I on a standard 12-lead ECG. In healthy individuals, the R-wave is generally considerably larger than the S-wave in Lead I. The Samsung smartwatch ECG has received CE marking as a medical device for detecting atrial fibrillation. In this study, we used only raw ECG signal data, excluding Samsung or other third-party analytics or diagnostics (Figure 1).

Figure 1.

Lead I and smartwatch-based ECG PQRST and amplitude description.

Cardiorespiratory vagal desynchrony disrupts the coordination between heart rate and respiration, a process regulated by the vagus nerve, leading to imbalances in autonomic function and cardiac performance.^1,2 Common biomarkers used to monitor this desynchrony include heart rate variability (HRV), respiratory sinus arrhythmia (RSA), and vagal tone indexes.³ These biomarkers reflect the autonomic regulation of heart rate in relation to respiratory patterns (Figure 2).⁴

Figure 2.

Summary timeline of biomarkers and health status.

Based on the Vagus Health Ltd. database findings of >50 000 ECG recordings (www.vagus.co), the R-peak amplitude is usually about three times larger than the S-peak. The reference data are from 2883 generally healthy users who recorded 41 350 ECGs between March 2020 and June 2024 using Vagus Health Ltd. apps and ECG-enabled smartwatches. The average S/R ratio for this reference data set was −0.305. The average R-wave peak value was 76.0 mV and the S-wave peak was −23.2 mV. Five hundred thirteen users have at least one ECG recording where the average S-/R-peak ratio was <−1. 21.7% of all tests that have an S/R ratio of <−0.5. A total of 6.95% has an S/R ratio of is < −1. The average RSA index is 78.

Vagus Health Ltd. is a Cambridge, UK-based wearable health analytics and biomarkers company. Its analytics and biomarkers were primarily developed during the pandemic to enhance remote diagnostics using smartwatch ECGs. The users and patients of smartwatches have agreed to the app’s terms, which state that the data can be used for scientific research by Vagus Health and its partner research institutions. The application, ‘BREATHE FLOW’, which was used to collect test data in this study, has been commercially available to users at no cost since the onset of the pandemic.

Summary figure

Case presentation

The patient, a 35-year-old male, was previously healthy and active with no medical conditions. He experienced his first COVID-19 infection in June 2020. After recovering from acute illness, in the following weeks, he developed long COVID symptoms with fatigue as the primary manifestation. Still, he improved after 2 months and returned to regular activity. After a new COVID-19 infection in February 2021, he developed typical long COVID symptoms of dysautonomia, post-exertional malaise (PEM), and symptoms compatible with POTS. These symptoms continued with some variability until the monitoring started in January 2024.

At the beginning of the monitoring in January 2024, he experienced fatigue, which worsened after physical activity. During these periods, the heart rate difference between sitting and standing posture was comparable to clinical POTS symptoms. The patient has self-reported his health according to the EQ-5D-5L model published by the EUROQOL group (Figure 3).⁵

Figure 3.

The patient’s self-reported health.

The health improvement was accompanied by a reduced heart rate difference between sitting and standing. At the end of the observation period, his supine-to-standing pulse increase during the tests did not indicate POTS-like symptoms. He also experienced far fewer and milder episodes of post-exertional fatigue.

During the testing period, the patient was taking the following medications: 10 mg prednisone every other day, 60 mg pyridostigmine bromide three times a day, 40 mg famotidine twice a day, and 162 mg tocilizumab once a week, which was tapered in June 2024.

The patient used a Nurosym^TM auricular vagus nerve stimulation device during the first month of the study. Then he switched to an experimental cymba concha vagus nerve-stimulating device during the rest of the data collection period, i.e. until the end of July 2024. Both devices used a 20 Hz stimulation programme.

Changes in S/R ratio and heart rate variability during the test period

In January 2024, the patient began using a Samsung ECG-enabled smartwatch and the BREATHE FLOW app. He received instruction on the test protocol on 19 January 2024. For 6 months, he performed 328 ECG recordings, which were analysed in real time on the Vagus Health Ltd. data platform, creating more than 40 biomarkers per test. This analysis makes more than 40 biomarkers per test through detailed ECG and QRST pattern analysis.

Patient electrocardiograms and data

See Figures 4 and 5.

Figure 4.

First month of testing electrocardiograms.

Figure 5.

Sixth month of testing electrocardiograms.

Vagal tone (heart rate variability)

Heart rate variability is the most commonly used biomarker for vagal tone (higher HRV indicates better vagal tone). When the patient’s health improved, the HRV in sitting posture increased. This increase was even more significant during standing posture. Both posture-based HRV indicators reached similar levels when the patient was feeling better (Table 1).

Table 1.

Biomarker data summary

Month	Health status	Avg S/R ratio (sitting)	Avg S/R ratio (standing)	Avg HRV (SDNN, ms), all tests	Avg RSA index, all tests
January	Fatigued	−0.35	−1.22	95	76
February	Severe fatigue	−0.30	−0.92	101	77
March	Fatigued	−0.32	−0.77	103	82
April	Improving	−0.31	−0.71	106	75
May	Improved	−0.27	−0.42	128	80
June	Improved	−0.24	−0.44	129	83
July	Improved	−0.24	−0.34	135	88
August	Occasional Relapse	−0.25	−0.49	133	84

During January to March, the patient self-reported as either fatigued or severe fatigue (n = 145 tests). During May–August, the patient reported improved health and no POTS-like heart rate changes. Occasional periods of fatigue relapses occurred. (n = 73 tests). The average heart rate changes between sitting and standing posture were January 28 b.p.m., February 19 b.p.m., March 14 b.p.m., April 17 b.p.m., May 11 b.p.m., June 11 b.p.m., July 8 b.p.m., and August 10 b.p.m.

Discussion

We present here the first case report on using smartwatch-ECG–based custom-made analysis for objective monitoring of patients suffering from long COVID. The focus of the study was on monitoring the autonomic integrity and parasympathetic control of RSA using a smartwatch-based application and dedicated, custom-made software. Initially, the patient’s cardiac-respiratory synchronization and HRV suggested autonomic dysregulation when standing. Conversely, most recordings taken while sitting were classified as usual. Notably, during the initial months, the S-wave in the ECG recordings became significantly higher when standing, a pattern that was reduced when health improved. The S/R ratio decreased consistently over the last months, correlating with the patient’s overall health improvement, as measured by both orthostatic response and subjective symptom scores.

The patient’s 328 ECG recordings revealed that during periods of poor health, the patient’s heart rate generally increased to more than 30 b.p.m. when moving from a supine to a standing posture. It is important to note that the smartwatch posture-based test protocol (Appendix 1) does not directly correspond to a clinical POTS test protocol. The ECG recordings last only 30 s. The patient has slightly varying times between supine and standing postures. The clinical turntable posture-based pulse change measurement is the only established medical biomarker of POTS.

In this study, we did not aim to evaluate the therapeutic effect of various treatment options as they were applied on pragmatic and experimental grounds, and different therapies overlapped. Rather, the focus was on monitoring autonomic integrity and the parasympathetic control of RSA using a novel smartwatch-based application and dedicated, custom-made software.

We found that an ECG-based S-/R-peak biomarker with a trigger threshold set at −0.5 correlated well with the patient’s self-reported health, as reported in terms of fatigue and according to self-reported monthly health (Figure 4). A strong correlation was also seen between the S/R ratio, pulse increase at standing compared to sitting, and low HRV (SDNN).

The ECG-based S/R amplitude rate and respiratory desynchronization biomarkers such as RSA and HRV, together with a posture test protocol, represent a novel, non-invasive method and easy ‘at-home’ triage and monitoring of long COVID, POTS, and fatigue. It offers a new smartwatch ECG-derived biomarker to be used in combination with other established biomarkers. This approach aligns with findings from other studies that emphasize the importance of monitoring autonomic functions in cardiovascular dysautonomia.¹

Possible mechanisms for S-wave amplitude posture changes

The S-wave represents the downward deflection following the R-wave in the QRS complex of an ECG, corresponding to the depolarization of the ventricles.⁶ S-wave changes in amplitude and morphology can be influenced by several physiological factors, including alterations in autonomic tone, blood volume distribution, and myocardial contractility, which are notably affected in POTS and fatigue conditions.^7,8

Autonomic nervous system dysfunction in postural orthostatic tachycardia syndrome

Postural orthostatic tachycardia syndrome is characterized by an exaggerated autonomic nervous system response, particularly an overactive sympathetic nervous system, and impaired parasympathetic function.⁸ When standing, patients with POTS experience excessive sympathetic stimulation, leading to increased heart rate and altered myocardial contractility.⁸ This heightened sympathetic tone can increase the amplitude of the S-wave due to changes in ventricular depolarization dynamics.

Conclusion

In the beginning, during the first and second months of testing, when the patient’s health was considered ‘bad’, the patient’s S-/R-peak amplitude ratio was consistently <−0.50 when he was doing the test while in a standing posture. The S/R ratio during standing posture was statistically significantly lower during January–April compared to May–July (P < 0.0001). This is evident in the ECG graphs of data measured during the first month and sixth month of testing (Figure 4).

During the last month of testing (when the patient reported better health, Graph 4), this change in the S/R ratio pattern virtually disappeared. In most tests, the patient’s QRST ratio, dysautonomia, and pulse changes are considered normal during standing posture. During the last month of testing, the S/R ratio rarely exceeded the benchmark of ‘−0.5’.

By introducing an innovative smartwatch ECG-based biomarker for pre-medical diagnostic indicators of POTS, long COVID, and chronic fatigue syndrome (CFS), this study contributes to the ongoing research in autonomic dysfunction. It provides a practical tool for patient self-monitoring and management. The significant correlation between S/R rate and patient health status underscores the potential of this biomarker to aid in the diagnosis and management of POTS and related conditions. The S/R ratio and other heart rhythm pattern changes are individual, self-report. Patients must follow a posture-based ECG testing protocol and self-report using a smartphone.

A substantial number of patients with post-COVID autonomic dysfunction, including POTS and inappropriate cardiac autonomic responses, have placed a significant strain on healthcare resources. The implementation of remote monitoring strategies may help alleviate this burden by reducing the need for in-person clinical visits. In this context, wearable technologies such as smartwatch ECGs, combined with readily available data collection applications, enable standardized follow-up of patients with POTS, long COVID, and other autonomic nervous system-related cardiac disorders. The posture-based test protocol applied in this study is user-friendly and suitable for use in a home setting. We propose that integrating this accessible test methodology into routine care may improve the quality of patient follow-up and reduce healthcare costs, even for patients with severe autonomic dysfunction.

Hypothesis

Artificial intelligence analytics combined with this novel dysautonomia- and posture-based smartwatch ECG test protocol provides new biomarkers for wearable and out-of-hospital cardiology and health diagnostics. One such biomarker is the change in the S/R ratio between sitting and standing postures, as presented here.

This posture- and controlled breathing-based test protocol appears to be especially valid for patients experiencing cardiorespiratory vagal dysfunction, fatigue, POTS, and long COVID.

Lead author biography

Gustaf Kranck has worked full time since 2013 on developing technological solutions to treat and diagnose the vagus nerve. He has a patent for a vagus stimulation device. He has worked in Scandinavia, the UK, and the USA with leading specialists in the field of digital health diagnostics. His company is a leading smartwatch-based ECG analytics provider. When being a speaker on conferences, he is usually referred to as ‘the vagus man’. He invented the ‘controlled breathing’ test protocol to diagnose cardiac respiratory synchronization with smartwatch ECG. He is the founder and CEO of Vagus Health Ltd (based in Cambridge, UK).

Appendix 1 Postural test protocol with an ECG-enabled smartwatch

Hand-to-hand single lead ECGs are recorded with a smartwatch while performing 0.1 Hz diaphragmatic controlled breathing (5 s inhale followed by 5 s exhale, 12).

Devices: BREATHE FLOW apps for Apple Watch, Samsung Galaxy Watch, or Withings ECG Watch

The user performed the following two ECG recordings consecutively:

1. Sitting: Normal VAGUS ECG test while in sitting posture (with hands and elbows on the table) while performing controlled breathing (labelled ‘sitting’).

• (2) Standing: VAGUS ECG test immediately after standing up. Hands and elbows should be leaning on a flat surface that is on the level of the chest while remaining standing (labelled ‘standing’).

The users note posture and subjective health statements as a reply to the app’s chatbot’s first question.

After the test label, users can add additional text about how she/he is feeling. The users can also add fatigue levels, brain fog status, balance problems, pain, or other symptoms such as fainting and heart palpitations.

Samsung’s page describing how to best do an ECG with a Samsung smartwatch: https://www.samsung.com/us/apps/samsung-health-monitor/

Author contributions

Gustaf Kranck (Conceptualization [lead], Data curation [lead], Formal analysis [lead], Investigation [lead], Methodology [lead], Project administration [lead], Resources [lead], Software [lead], Validation [lead], Visualization [lead], Writing—original draft [lead]), Marcus Ståhlberg (Validation [supporting]), Ulf Andersson (Investigation [supporting]), Johan Lundin (Conceptualization [supporting]), and Artur Fedorowski (Writing—review & editing [supporting])

Consent: Written informed consent was obtained from the patient for publication of the clinical details and accompanying images in this case report.

Funding: Gustaf Kranck is the founder and CEO of Vagus Health Ltd, which provided the monitoring and analytics for free to the patient and the study. Neither the company nor Gustaf Kranck received any other funding related to this study. The other authors declare no conflicts of interest or financial contributions related to this study.

Data availability

Data is available on request. The data underlying this article will be shared on reasonable request to the corresponding author.