Cardiography, Impedance

A whole-body impedance cardiography device (CircMon^R, JR Medical Ltd, Tallinn, Estonia), which records the changes in body electrical impedance during cardiac cycles, was used to determine beat-to-beat HR, stroke index (stroke volume in proportion to body surface area, ml/m²), cardiac index (cardiac output/body surface area, l/min/m²), and PWV (m/s)
[29 (link)-31 (link)]. Left cardiac work index (kg*m/min/m²) was calculated by formula 0.0143*(MAP–PAOP)*cardiac index, which has been derived from the equation published by Gorlin et al.
[32 (link)]. MAP is mean radial arterial pressure measured by tonometric sensor, PAOP is pulmonary artery occlusion pressure which is assumed to be normal (default 6 mmHg), and 0.0143 is the factor for the conversion of pressure from mmHg to cmH₂O, volume to density of blood (kg/L), and centimetre to metre. Systemic vascular resistance index (systemic vascular resistance/body surface area, dyn*s/cm⁵/m²) was calculated from the signal of the tonometric BP sensor and cardiac index measured by CircMon^R.
To calculate the PWV, the CircMon software measures the time difference between the onset of the decrease in impedance in the whole-body impedance signal and the popliteal artery signal. From the time difference and the distance between the electrodes, PWV can be determined. As the whole-body impedance cardiography slightly overestimates PWV when compared with Doppler ultrasound method, a validated equation was utilized to calculate values that correspond to the ultrasound method (PWV = (PWV_impedance*0.696) + 0.864)
[30 (link)]. PWV was determined only in the supine position because of less accurate timing of left ventricular ejection during head-up tilt
[30 (link)]. A detailed description of the method and electrode configuration has been previously reported
[31 (link)]. PWV was also recorded after the head-up tilt in all subjects, and the average difference between the mean PWV before and after the head-up tilt was 0.024 ± 0.388 m/s (mean ± standard deviation), showing the good repeatability of the method (repeatability index R 98%, Bland-Altman repeatability index 0.8)
[33 ]. The cardiac output values measured with CircMon^R are in good agreement with the values measured by the thermodilution method
[31 (link)], and the repeatability and reproducibility of the measurements (including PWV recordings) have been shown to be good
[34 (link),35 (link)].

A whole-body impedance cardiography device (CircMon^R, JR Medical Ltd, Tallinn, Estonia), which records the changes in body electrical impedance during cardiac cycles, was used to determine beat-to-beat HR, stroke volume, cardiac output, and aortic-popliteal PWV [27 (link)]. The mechanism of function, electrode placement, and processing of impedance cardiography data have been previously described [21 (link), 27 (link)–29 (link)]. Briefly, the impedance cardiography method calculates PWV between the level of the aortic root and the popliteal artery by the use of the whole-body impedance signal and the signal measured from the popliteal artery region [27 (link), 28 (link)]. The PWV results obtained using CircMon^R show good repeatability [29 (link)], and normal values for PWV in 799 individuals (age 25–76 years) have been previously published [30 (link)]. We have also shown that the determination of stroke volume using impedance cardiography versus 3-dimensional echocardiography show good correlation [29 (link)]. SVR was calculated from the tonometric radial BP signal and cardiac output measured by the CircMon^R device by subtracting average normal central venous pressure (4 mmHg) from mean arterial pressure and dividing it by cardiac output. Mean arterial pressure was calculated by using the formula: [(systolic BP)/3 + 2 × (diastolic BP)/3]. Cardiac output, stroke volume and SVR were indexed to body surface area (abbreviated as CI, SVI and SVRI, respectively).

We performed impedance-cardiography and cuff-based BP measurement at rest and under mental and physical load in 71 young and healthy adults. After welcoming subjects and receiving their written consent, we proceeded with the experiment. We asked the participants to sit on a chair behind a desk with all devices attached. After 2 min at total rest, we initiated the impedance cardiography and performed the first BP measurement. This measurement was taken to record at rest (baseline) measurements of BP and impedance cardiography results which we needed to show intraindividual changes in the following analysis. Subsequently, the subjects each went through two experimental phases (Figure 2).

We used the CardioScreen^® 1000 (medis Medizinische Messtechnik GmbH, Ilmenau, Germany) device to perform impedance cardiography. The device has been validated (as bedside monitor, named niccomo™-monitor) (16 (link), 23 (link)–25 (link)). We recorded all parameters provided by the device, including the PEP, heart rate, left ventricular ejection time, and Heather index. The Heather index is calculated as a combination of acceleration and velocity of blood flow (both as index values, relative to body surface area) and is a parameter of impedance cardiography said to represent contractility and overall sympathetic tone (26 (link)). For BP measurement, we used the validated cuff-based OnTrak 90227 device (Spacelabs^® Healthcare) (27 (link)). For ensuring valid cuff-based measurement, we recorded the cuff’s pressure curves via a Y-connection, and recorded data via a SOMNOtouch™ NIBP (SOMNOmedics GmbH).
The devices were time synchronized in a two-step process. The CardioScreen^® and SOMNOtouch™ devices were both initiated to within second precision during device setup. As both devices recorded an ECG, we were then able to synchronize the signals with millisecond precision.
We used a Ergometrics 900 L (ergoline GmbH, Bitz, Germany) recumbent bike ergometer with 60° inclination to enable controllable physical load while minimizing upper body movement. This is of utmost importance for the 8-lead cardio impedance device and during cuff-based BP measurement. The experimental setup, combined with quality assessment of cuff pressure curves, allowed us to maximize the quality of measurements included in the final analysis. We time synchronized all devices.

We performed a quality check for all impedance cardiography datasets, excluding measurement errors defined as abnormal changes in heart rate (change > 30% in less than 3 s). In line with manufacturer recommendations, we averaged the beat-to-beat data over four heartbeats.
Blood pressure measurements were assessed for undisturbed pressure curves. We excluded measurements when there was an increase of cuff pressure of more than 8 mmHg during cuff deflation.
To decide whether to use the ECG’s Q-wave (physiological beginning of PEP = start of ventricular depolarization) or the R-wave (easier detection), we compared the PEP derived from both starting points using a correlation analysis and analyzed the changes in the Q-R-time and its dependency to mental or physical load.
We analyzed changes of PEP during the TSST and ergometer load separately by modeling a mixed linear model (IBM SPSS Statistics 27). For Regression analyses, we performed linear regression and non-linear, R²-calculation via Scikit-learn (30 ). To find differences in PEP’s behavior under different circumstances, we calculated a mixed linear model and adjusted for the heart rate as a covariate factor. Subsequently, we trained a k-nearest-neighbor classifier with patient mean values for rest, TSST and bike ergometer load. In accordance with best practices of Machine Learning we performed a strict separation of training and test data. We tested the classifier using a subject-dependent train-test split. This means that all three data-points of any given patient had to be either in the train or the test group, therefore reducing the change for data-leakage. We retrieved our results by evaluating the classifier in an 80/20 k-fold (equates to fivefold) evaluation scheme. We averaged the results (positive predictive value/sensitivity) over all three outcomes (rest, mental load, physical load) and over all k-fold iterations to provide an aggregate estimation of the classifiers performance.
To illustrate patient specific differences at rest and under mental and physical load, we visualized data at rest measurement, first TSST question and last ergometer step (maximum load) in a boxplot. We analyzed the differences via a one-way ANOVA. The statistical analysis was performed in close collaboration with Charité’s Institute of Biometry and Clinical Epidemiology. We used the conservative Bonferroni correction for all multi-group comparisons.
To estimate the effect of neglecting or estimating the PEP, we analyzed a previously published and often-cited relation between PWV and BP (6 (link)). We applied the relation for a standard human (1.80 m) and translated the relation to a pulse-arrival-time (PAT, time from ECG Q-wave to arrival of the pulse wave in the periphery) vs. BP relation. Thereafter, we used our data to estimate the PEP’s proportion of the PAT for different BP levels (regression) and subtracted the PEP to receive the pulse-transit-time (PTT = PAT–PEP). We then analyzed the PEP estimation uncertainty (intra- and interindividual PEP variability at similar BP levels) and calculated confidence intervals for one SD (67% of values) and two SDs (95% of values) for PTT. Applying this relation, we were able to determine the BP measurement estimation uncertainty caused by either neglecting or estimating the PEP for PWV based BP measurement.