Electrical impedance plethysmography versus tonometry to measure the pulse wave velocity in peripheral arteries in young healthy volunteers: a pilot study

The leading cause of health loss and deaths worldwide are cardiovascular diseases (1). A predictor of cardiovascular diseases and events is the arterial stiffness, especially in central arteries such as the aorta (2,3). Previous studies have demonstrated that pulse wave velocity (PWV) is a valid marker for arterial stiffness that can be measured non-invasively (4). The PWV is the traveling speed of the pulse wave propagating through the arterial system. A high PWV is associated with a stiffer artery (5). The PWV between the carotid and femoral arteries has been considered as the “gold-standard” measurement of arterial stiffness in the aorta (3). In addition to aortic PWV having clinical value, measurements of peripheral PWV may also have a clinical value. In a study by Lee et al. it was demonstrated that in patients with chronic kidney disease that have both a high PWV in the central and peripheral arteries were associated with a rapid decline in kidney function (6). Also studies with diabetes patients show the clinical relevance of PWV in the legs. Diabetes patients have an increased risk of developing peripheral artery disease (PAD). Different studies with diabetes patients with and without any symptoms of PAD have shown to have stiffer femoral and popliteal arteries compared to a healthy control group (2,7, 8, 9).

Moreover, another study with diabetes patients found that peripheral neuropathy is associated with a high brachial-ankle PWV (10). As a high PWV in the legs is associated with different diseases, it is of clinical relevance that the PWV is measured with a good method.

All methods to estimate PWV rely on the measurement of a pulse wave, which is commonly acquired by pressure, distension, or flow waves. Different methods and techniques have been used to measure PWV, i.e. the PWV are the PulsePen® (tonometry and electrocardiogram (ECG)), Complior® (piezoelectric pressure mechanotransducers), SphygmoCor® (tonometry and ECG), Arteriograph® (oscillometric pressure curve analysis), photoplethysmography (infrared probes), ultrasound (Doppler waveforms simultaneously or with ECG), magnetic resonance imagining (MRI, aortic systolic flow wave at two different sites) and invasive measurement (catheterization from peripheral artery with angiography). Among these, pressure sensors such as the tonometer is considered the gold standard for measuring PWV (11). This approach needs holding or mechanical fixation of a probe above the artery during measurement, where operating skills affect the measurement and a trained operator is needed. Typically, these probes are directly placed on the patient’s tissue at two different arteries on a site where the pulsation is easily felt. This gives an average PWV for that segment of the arterial system, but could cause distortion the waveform and might be uncomfortable in some patients (11,12).

Another method is electrical impedance plethysmography (IPG), which is a kind of bioimpedance measurement. Instead of relying on mechanical contact to pick up the pressure waveform, the IPG technique relies on electrical contact through electrodes attached to the skin, using an applied weak electrical current and voltage sensing. The IPG method relies on changes in the ability of the tissue to conduct alternating current and is sensitive to blood volume changes (plethysmography), blood flow and blood composition (haematocrit). Either by sensing the IPG waveform at two arterial sites simultaneously, or together with an electrocardiography (ECG) the PWV can be estimated by measuring the time delay (pulse arrival time (PAT)) and distance between the two sites (13). This could be a promising alternative for the already existing methods, as the IPG is a simple, non-invasive and a low-cost method. The instrumentation and electronics required for IPG measurement is relatively simple, enabling development of a miniaturized device. Another advantage of the IPG is the possibility to observe a pulse in every artery, despite the depth of the artery (14). This could mean that bioimpedance could be used on places where the artery is difficult to detect for other devices.

Aria et al. showed in 2019 that using IPG for PWV measurement is a suitable and practical method, but that comparison with other techniques is still needed (14). A suitable comparison is a tonometer, regarded as one of the golden standard for PWV measurement (11). It is suitable as it can be used for every superficial palpable artery (5). The aim of this study was therefore to explore if the bioimpedance IPG can be a suitable technique to measure pulse wave velocity in legs as an alternative for the tonometer technique. The hypothesis is that the IPG is comparable to tonometry for the measurement PWV.

The aim of this study was to investigate if bioimpedance can be a suitable technique to measure the pulse wave velocity (PWV) in the legs as an alternative for the tonometer. We estimated with both methods, PWV in the a. fermoralis, a. popliteal, a. tibialis posterior and a. dorsalis pedis in nine young healthy participants. The bioimpedance and the tonometer obtained the pulse waves from the measured arteries. Gated by the R-wave in the ECG signal, the pulse arrival time (PAT) was estimated using a self-written script in Matlab. The PWV was estimated by dividing the distance from the heart to the femoral artery and between the measured arteries by the PAT. The results indicate that IPG may be a suitable technique to measure PWV. The estimated PWV results from bioimpedance and from the tonometer seem fairly in line with each other, see figure 6. The results of individual subjects seem to be in agreement with each other as well when they are illustrated in a scatter plot where both devices are compared with an x=y line (figure 4). These results support the hypothesis of the present study.

Koivistoinen et al. (19) used whole-body impedance cardiography in the aortic arch and the popliteal artery in a healthy population (age 25-41 years). They reported the PWV in male and female was 7.7 ± 1.4 m/s and 7.2 ± 1.2 m/s respectively (18). Rossow et al. (20) measured the PWV in women with the SphygmoCor in the carotid-femoral and femoral-tibialis posterior segments. The PWV in the younger group (age 19-25 years (22 ± 2)) was in the carotid-femoral 5.8 m/s and in the femoral-tibialis posterior arteries 8.5 m/s (19). Hashimoto et al. (21) used the tonometer of the SphygmoCor to measure the PWV in a healthy group (mean age 55±14 years) and measured a PWV of 7.8 (6.8-9.3) m/s in the carotid-femoral arteries and in the femoral-dorsalis pedis a PWV of 9.2±1.5 m/s (21).

The results of the present study are mostly in agreement with the previous studies, except for the estimated PWV between the heart and the femoral artery. We measured the following median PWV with bioimpedance: heart-femoral segment 2.90 m/s; femoral-popliteal segment 7.34 m/s; popliteal-dorsalis pedis segment 9.12 m/s and poplitealtibialis posterior segment 9.18 m/s. For the tonometer the median estimated PWV was in the heart-femoral segment 3.19 m/s; femoral-popliteal segment 7.33 m/s; poplitealdorsalis pedis segment 9.36 m/s and popliteal-tibialis posterior segment 10.94 m/s. A possible explanation for this difference could be that the ECG was not correctly synchronised with the other devices, as this is the only ECG dependent segment. The other PWV segments were calculated based on the differences in PAT, cancelling out possible signal lag of the ECG sampling, provided that it is constant.

Just as seen in the previous studies, it should be noted that the PWV increases when the artery is more distally. Two reasons for this are the change in diameter and change in wall properties from the aorta to the peripheral arteries. When the distance from the heart increases, the amount of elastin decreases, and the amount of collagen increases in the vessel walls. Elastin causes the ability for an artery to stretch, while collagen causes stiffening of the vessel wall (4,21).

A remarkable result is that it seems that the tonometer estimates a slightly higher PWV than the bioimpedance (figure 3 and 4). However, when the mean and the confidence intervals from the estimated PWV are compared, it seems that there was no statistical difference detectable, probably due to the given sample size of the study.

Another interesting result is that it seems that the bioimpedance signal has a third wave which is not present in the tonometer signal (figure 2). A possible explanation for this is that the bioimpedance is more sensitive to various changes in the blood vessel compared to the tonometer. The bioimpedance is affected by changes in blood volume, blood velocity and blood composition (haematocrit, the ratio of volume of red blood cells to the total volume of the blood), while the tonometer is only dependent on the pressure difference (23,24). The signal of the tonometer shows two peaks, the first peak is a result from the ejection of the left ventricle and the second peak from the wave reflections from the peripheral arteries (25). The third peak in the bioimpedance signal could be the dicrotic wave, which is a result from the dicrotic notch. The dicrotic notch reflects the aortic valve closure, which is partly closed by the reflection wave. It represents the end of the systole and the beginning of the diastole (26). The dicrotic notch affects the pulse pressure and the blood flow (27,28). As the bioimpedance detects both change in volume and blood flow, it could be more sensitive for the dicrotic notch than the tonometer. The dicrotic wave could be of clinical relevance as it represents etiologic factors. For example aortic valve diseases or an infectious that causes arterial vasodilatation and a compliant aorta and arterial tree (29,30).

Another difference between the devices that should be noted is the ease of use of the bioimpedance compared to the tonometer. During the measurements we observed that finding the bioimpedance signal was easier than finding the signal of the tonometer. Moreover, when a participant moved the bioimpedance signal recovered, while often the signal of the tonometer had to be found again.

There are limitations of this study that should be considered. Although the analogue ECG and tonometer signals were streamed directly from the Sphygmocor device with a high sampling rate (838 Hz) into the Zurich Instruments device (with a <1 ms delay between channels), the devices were not synchronised using trigger signals before the measurements. There could be a delay in the ECG compared to the SphygmoCor and/or IPG that caused the low PWV between the heart and the femoral artery and probably incorrect results. However, we believe that this possible error would only affect the heart to femoral estimates (as the other estimates use differences in pulse arrival times) assuming that such device lag would be constant. The heart to femoral PWV is also the least relevant segment in this context of evaluating peripheral pulse wave velocities. Another limitation that could influence the results is the method of measuring distances between the measured sides. The measured distance is not the real distance of the artery tracts, as it is measured on the skin. A third limitation is the number of participants. Only nine volunteers participated in this pilot study, which results in a low statistical power.

Despite these limitations, our results do provide an indication for the use of bioimpedance for estimating PWV as the estimated PWVs seem valid because they are in line with each other and with other studies. Furthermore, the estimated PAT was reliable as there was a small variation between the beats. These variations were also comparable between the two methods (figure 5). However, more research will be necessary to validate this technique. Future studies should focus on a larger sample, using devices that are synchronised with each other and using simultaneous methods that use dual impedance measurement and pulse transit times for PWV measurement (without the need for an ECG).

Although further research is needed, the present study indicates that the use of IPG may be good alternative for estimating PWV in the legs and might have a potential to be implemented for clinical use in the future.