Wearable Devices for Remote Monitoring of Heart Rate and Heart Rate Variability-What We Know and What Is Coming

Abstract

Heart rate at rest and exercise may predict cardiovascular risk. Heart rate variability is a measure of variation in time between each heartbeat, representing the balance between the parasympathetic and sympathetic nervous system and may predict adverse cardiovascular events. With advances in technology and increasing commercial interest, the scope of remote monitoring health systems has expanded. In this review, we discuss the concepts behind cardiac signal generation and recording, wearable devices, pros and cons focusing on accuracy, ease of application of commercial and medical grade diagnostic devices, which showed promising results in terms of reliability and value. Incorporation of artificial intelligence and cloud based remote monitoring have been evolving to facilitate timely data processing, improve patient convenience and ensure data security.

Keywords: cardiovascular risk; heart rate; heart rate variability; remote monitoring.

Figure 1

Electrocardiographic recordings of heart activity. (a) consumer grade heart rate monitor (AliveCor Kardia). The algorithm detects the QRS intervals and transforms them into markers. An inappropriately marked artifact is seen (non-physiological signal, no repolarization). These findings are easily detected and rejected by trained human interpreters but remain challenging for automated rhythm analysis. (b) Implantable cardiac rhythm monitor (Medtronic Reveal Linq). These devices benefit from improved signal/noise ratio due to the subcutaneous position of the device, close to the heart. However, sensing issues are still not infrequent, such as in this patient with atrial fibrillation and an undersensed premature ventricular contraction (PVC).

Figure 2

Electrocardiographic recording of heart activity with a commercial heart rate monitor (AliveCor Kardia). Inappropriate detection of high frequency noise led to oversensing and incorrect determination of rhythm (atrial fibrillation)—the patient is in normal sinus rhythm, with frequent PACs (premature atrial contraction—a benign cardiac rhythm abnormality).

Figure 3

Typical graphical and textual presentation of a medical grade 48 h wearable ambulatory monitor (Holter). The heart rate trends between 45–94 bpm, with longer and shorter cycle length variations due to diurnal changes and variable activity levels. Valid data was collected during 95% of the monitoring interval, the rest was rejected due to inadequate signals. Heart rate variability parameters are automatically calculated. The percentage of ectopic, abnormal rhythms is also shown (premature atrial and ventricular contraction).

Figure 4

Holter monitor heart rate tracings of patients with abnormal 48 h HR/HRV. (a) paroxysmal supraventricular tachycardia—on the first day of monitoring, the heart rate suddenly increased to 150 bpm for a few minutes, with abrupt termination. (b) paroxysmal atrial fibrillation. Normal sinus rhythm in the first half of the tracing, with sudden, sustained increase in the heart rate for the second half—the arrhythmia did not terminate during the monitoring period. (c) sick sinus syndrome. The heart rate is low (30–65 bpm). The diurnal variation is abnormal, lower heart rate during daytime and increased heart rate at night.

Figure 5

Long term heart rate variability in a patient with a Medtronic implanted biventricular defibrillator (cardiac resynchronization device implanted in patients with severe cardiomyopathy and ventricular dyssynchrony). The device continuously records heart rate and heart rate variability data. In addition, thoracic impedance is measured, which correlates with the degree of pulmonary congestion and severity of heart failure symptoms. Sudden decrease in heart rate variability may be observed at the onset of increasing pulmonary congestion (decreased thoracic impedance due to fluid buildup). As these findings may precede the onset of symptoms by several days, remote monitoring may identify high risk patients for targeted intervention.

Figure 6

Long-term ECG signal recording. (a–c) ECG signals measured after various amount of time using (a) porous PEDOT:PSS/PDMS electrodes; (b) planar PEDOT:PSS electrodes; and (c) commercial Ag/AgCl electrodes. (d) Comparison of the signal-noise ratio between the three different types of electrodes. Maintaining a high signal-to-noise ratio over longer time periods with novel electrode technologies may help to decrease the need for frequent electrode replacements when using ECG-based wearable devices. Reprinted from [36]. No changes were made to the image, which has been published under CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/legalcode).

Figure 7

Number of HRV analysis studies using machine or deep learning within one year, from 2010 to 2021. Reprinted with permission from [23].

Figure 8

Examples of wearable cardiac monitors. (a) iRhythm ZIO AT chest patch monitor (medical grade, https://www.irhythmtech.com, accessed on 17 August 2022), (b) Polar OH1 upper arm band monitor (https://www.polar.com/us-en, accessed on 17 August 2022), (c) Wahoo TICKR X chest strap (https://www.wahoofitness.com, accessed on 17 August 2022), (d) FitBit Versa smart watch (https://www.fitbit.com/global/us/home, accessed on 17 August 2022), (e) Garmin VivoSmart 4 wrist strap monitor (https://www.garmin.com/en-US, accessed on 17 August 2022).

Figure 9

Example of RR intervals recorded by ECG and a PPG-based device (Polar H7) during recovery after exercise in one subject. (a) RRE = RR interval series recorded by ECG; (b) RRP = RR intervals series recorded by the PolarH7. Compared to the ECG, the PPG method overestimates the HRV. Reprinted from [55]. No changes were made to the image, which has been published under CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/legalcode).