Validation of the new Complior sensor to record pressure signals non-invasively

Value and Variability of Pulse Shape Indicator for Estimating Mean Arterial Pressure in the Radial and Femoral Arteries

BACKGROUND

The form factor (FF) is a pulse shape indicator that corresponds to the fraction of pulse pressure added to diastolic blood pressure to estimate the time-averaged mean arterial pressure (MAP). Our invasive study assessed the FF value and variability at the radial and femoral artery levels and evaluated the recommended fixed FF value of 0.33.

METHODS AND RESULTS

Hemodynamically stable patients were prospectively included in 2 intensive care units. FF was documented at baseline and during dynamic maneuvers. A total of 632 patients (64±16 years of age, 66% men, MAP=81±14 mm Hg) were included. Among them, 355 (56%) had a radial catheter and 277 (44%) had a femoral catheter. The FF was 0.34±0.06. In multiple linear regression, FF was influenced by biological sex (P<0.0001) and heart rate (P=0.04) but not by height, weight, or catheter location. The radial FF was 0.35±0.06, whereas the femoral FF was 0.34±0.05 (P=0.08). Both radial and femoral FF were higher in women than in men (P<0.05). When using the 0.33 FF value to estimate MAP, the error was -0.4±4.0 mm Hg and -0.1±2.9 mm Hg at the radial and femoral level, respectively, and the MAP estimate still demonstrated high accuracy and good precision even after changes in norepinephrine dose, increase in positive end-expiratory pressure level, fluid administration, or prone positioning (n=218).

CONCLUSIONS

Despite higher FF in women and despite interindividual variability in FF, using a fixed FF value of 0.33 yielded accurate and precise estimations of MAP. This finding has potential implications for blood pressure monitoring devices and the study of pulse wave amplification.

High Fidelity Pressure Wires Provide Accurate Validation of Non-Invasive Central Blood Pressure and Pulse Wave Velocity Measurements

Central blood pressure (cBP) is known to be a better predictor of the damage caused by hypertension in comparison with peripheral blood pressure. During cardiac catheterization, we measured cBP in the ascending aorta with a fluid-filled guiding catheter (FF) in 75 patients and with a high-fidelity micromanometer tipped wire (FFR) in 20 patients. The wire was withdrawn into the brachial artery and aorto-brachial pulse wave velocity (abPWV) was calculated from the length of the pullback and the time delay between the ascending aorta and the brachial artery pulse waves by gating to the R-wave of the ECG for both measurements. In 23 patients, a cuff was inflated around the calf and an aorta-tibial pulse wave velocity (atPWV) was calculated from the distance between the cuff around the leg and the axillary notch and the time delay between the ascending aorta and the tibial pulse waves. Brachial BP was measured non-invasively and cBP was estimated using a new suprasystolic oscillometric technology. The mean differences between invasively measured cBP by FFR and non-invasive estimation were -0.4 ± 5.7 mmHg and by FF 5.4 ± 9.4 mmHg in 52 patients. Diastolic and mean cBP were both overestimated by oscillometry, with mean differences of -8.9 ± 5.5 mmHg and -6.4 ± 5.1 mmHg compared with the FFR and -10.6 ± 6.3 mmHg and -5.9 ± 6.2 mmHg with the FF. Non-invasive systolic cBP compared accurately with the high-fidelity FFR measurements, demonstrating a low bias (≤5 mmHg) and high precision (SD ≤ 8 mmHg). These criteria were not met when using the FF measurements. Invasively derived average Ao-brachial abPWV was 7.0 ± 1.4 m/s and that of Ao-tibial atPWV was 9.1 ± 1.8 m/s. Non-invasively estimated PWV based on the reflected wave transit time did not correlate with abPWV or with atPWV. In conclusion, we demonstrate the advantages of a novel method of validation for non-invasive cBP monitoring devices using acknowledged gold standard FFR wire transducers and the possibility to easily measure PWV during coronary angiography with the impact of cardiovascular risk factors.

New Method to Estimate Central Systolic Blood Pressure From Peripheral Pressure: A Proof of Concept and Validation Study

Objective: The non-invasive estimation of central systolic blood pressure (cSBP) is increasingly performed using new devices based on various pulse acquisition techniques and mathematical analyses. These devices are most often calibrated assuming that mean (MBP) and diastolic (DBP) BP are essentially unchanged when pressure wave travels from aorta to peripheral artery, an assumption which is evidence-based. We tested a new empirical formula for the direct central blood pressure estimation of cSBP using MBP and DBP only (DCBP = MBP²/DBP).

Methods and Results: First, we performed a post-hoc analysis of our prospective invasive high-fidelity aortic pressure database (n = 139, age 49 ± 12 years, 78% men). The cSBP was 146.0 ± 31.1 mmHg. The error between aortic DCBP and cSBP was −0.9 ± 7.4 mmHg, and there was no bias across the cSBP range (82.5–204.0 mmHg). Second, we analyzed 64 patients from two studies of the literature in whom invasive high-fidelity pressures were simultaneously obtained in the aorta and brachial artery. The weighed mean error between brachial DCBP and cSBP was 1.1 mmHg. Finally, 30 intensive care unit patients equipped with fluid-filled catheter in the radial artery were prospectively studied. The cSBP (115.7 ± 18.2 mmHg) was estimated by carotid tonometry. The error between radial DCBP and cSBP was −0.4 ± 5.8 mmHg, and there was no bias across the range.

Conclusion: Our study shows that cSBP could be reliably estimated from MBP and DBP only, provided BP measurement errors are minimized. DCBP may have implications for assessing cardiovascular risk associated with cSBP on large BP databases, a point that deserves further studies.

Heartbeat Detection by Laser Doppler Vibrometry and Machine Learning

Background: Heartbeat detection is a crucial step in several clinical fields. Laser Doppler Vibrometer (LDV) is a promising non-contact measurement for heartbeat detection. The aim of this work is to assess whether machine learning can be used for detecting heartbeat from the carotid LDV signal. Methods: The performances of Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF) and K-Nearest Neighbor (KNN) were compared using the leave-one-subject-out cross-validation as the testing protocol in an LDV dataset collected from 28 subjects. The classification was conducted on LDV signal windows, which were labeled as beat, if containing a beat, or no-beat, otherwise. The labeling procedure was performed using electrocardiography as the gold standard. Results: For the beat class, the f1-score (f1) values were 0.93, 0.93, 0.95, 0.96 for RF, DT, KNN and SVM, respectively. No statistical differences were found between the classifiers. When testing the SVM on the full-length (10 min long) LDV signals, to simulate a real-world application, we achieved a median macro-f1 of 0.76. Conclusions: Using machine learning for heartbeat detection from carotid LDV signals showed encouraging results, representing a promising step in the field of contactless cardiovascular signal analysis.

Estimation of central arterial pressure from the radial artery in patients undergoing invasive neuroradiological procedures

Backgrounds

Central arterial pressure can be derived from analysis of the peripheral artery waveform. The aim of this study was to compare central arterial pressures measured from an intra-aortic catheter with peripheral radial arterial pressures and with central arterial pressures estimated from the peripheral pressure wave using a pressure recording analytical method (PRAM).

Methods

We studied 21 patients undergoing digital subtraction cerebral angiography under local or general anesthesia and equipped with a radial arterial catheter. A second catheter was placed in the ascending aorta for central pressure wave acquisition. Central (AO) and peripheral (RA) arterial waveforms were recorded simultaneously by PRAM for 90–180 s. During an off-line analysis, AO pressures were reconstructed (AO_rec) from the RA trace using a mathematical model obtained by multi-linear regression analysis. The AO_rec values obtained by PRAM were compared with the true central pressure value obtained from the catheter placed in the ascending aorta.

Results

Systolic, diastolic and mean pressures ranged from 79 to 180 mmHg, 47 to 102 mmHg, and 58 to 128 mmHg, respectively, for AO, and 83 to 174 mmHg, 47 to 107 mmHg, and 60 to 129 mmHg, respectively, for RA. The correlation coefficients between AO and RA were 0.86 (p < 0.01), 0.83 (p < 0.01) and 0.86 (p < 0.01) for systolic, diastolic and mean pressures, respectively, and the mean differences − 0.3 mmHg, 2.4 mmHg and 1.5 mmHg. The correlation coefficients between AO and AO_rec were 0.92 (p < 0.001), 0.87 (p < 0.001) and 0.92 (p < 0.001), for systolic, diastolic and mean pressures, respectively, and the mean differences 0.01 mmHg, 1.8 mmHg and 1.2 mmHg.

Conclusions

PRAM can provide reliable estimates of central arterial pressure.

Electronic supplementary material

The online version of this article (10.1186/s12871-019-0844-1) contains supplementary material, which is available to authorized users.

Noninvasive Estimation of Aortic Stiffness Through Different Approaches

Aortic pulse wave velocity is a worldwide accepted index to evaluate aortic stiffness and can be assessed noninvasively by several methods. This study sought to determine if commonly used noninvasive devices can all accurately estimate aortic pulse wave velocity. Pulse wave velocity was estimated in 102 patients (aged 65±13 years) undergoing diagnostic coronary angiography with 7 noninvasive devices and compared with invasive aortic pulse wave velocity. Devices evaluating carotid-femoral pulse wave velocity (Complior Analyse, PulsePen ET, PulsePen ETT, and SphygmoCor) showed a strong agreement between each other ( r>0.83) and with invasive aortic pulse wave velocity. The mean difference ±SD with the invasive pulse wave velocity was -0.73±2.83 m/s ( r=0.64) for Complior-Analyse: 0.20±2.54 m/s ( r=0.71) for PulsePen-ETT: -0.04±2.33 m/s ( r=0.78) for PulsePen ET; and -0.61±2.57 m/s ( r=0.70) for SphygmoCor. The finger-toe pulse wave velocity, evaluated by pOpmètre, showed only a weak relationship with invasive aortic recording (mean difference ±SD =-0.44±4.44 m/s; r=0.41), and with noninvasive carotid-femoral pulse wave velocity measurements ( r<0.33). Pulse wave velocity estimated through a proprietary algorithm by BPLab (v.5.03 and v.6.02) and Mobil-O-Graph showed a weaker agreement with invasive pulse wave velocity compared with carotid-femoral pulse wave velocity (mean difference ±SD =-0.71±3.55 m/s, r=0.23; 1.04±2.27 m/s, r=0.77; and -1.01±2.54 m/s, r=0.71, respectively), revealing a negative proportional bias at Bland-Altman plot. Aortic pulse wave velocity values provided by BPLab and Mobil-O-Graph were entirely dependent on age-squared and peripheral systolic blood pressure (cumulative r²=0.98 and 0.99, respectively). Thus, among the methods evaluated, only those assessing carotid-femoral pulse wave velocity (Complior Analyse, PulsePen ETT, PulsePen ET, and SphygmoCor) appear to be reliable approaches for estimation of aortic stiffness.

Measuring Blood Pulse Wave Velocity with Bioimpedance in Different Age Groups

In this project, we have studied the use of electrical impedance cardiography as a possible method for measuring blood pulse wave velocity, and hence be an aid in the assessment of the degree of arteriosclerosis. Using two different four-electrode setups, we measured the timing of the systolic pulse at two locations, the upper arm and the thorax, and found that the pulse wave velocity was in general higher in older volunteers and furthermore that it was also more heart rate dependent for older subjects. We attribute this to the fact that the degree of arteriosclerosis typically increases with age and that stiffening of the arterial wall will make the arteries less able to comply with increased heart rate (and corresponding blood pressure), without leading to increased pulse wave velocity. In view of these findings, we conclude that impedance cardiography seems to be well suited and practical for pulse wave velocity measurements and possibly for the assessment of the degree of arteriosclerosis. However, further studies are needed for comparison between this approach and reference methods for pulse wave velocity and assessment of arteriosclerosis before any firm conclusions can be drawn.

Carotid pulse wave analysis: left or right, does it matter?

INTRODUCTION

Pulse wave analysis (PWA) over the carotid artery is one of the available methodological options to obtain central arterial pressures and other important hemodynamic parameters. However, limited data exist relating the PWA estimates of one carotid artery over the contralateral one as the majority of the available data rely on measures over the right carotid artery.

OBJECTIVE

To evaluate the agreement of the PWA estimates between the right carotid artery and the left carotid artery.

PATIENTS AND METHODS

A cross-sectional study of 38 patients, with a mean age of 28.85±1.70 years, was carried out. Brachial blood pressure was evaluated using a sphygmomanometer. Subsequently, PWA was obtained randomly over the left or the right carotid arteries with the Complior Analyse. All the evaluations were performed by the same experienced operator.

RESULTS

The overall mean differences observed by comparing the obtained parameters in each carotid territory were -1.50±8.06 mmHg for central systolic blood pressure (SBP), -1.63±7.98 mmHg for central pulse pressure, -3.37±27.80% for the augmentation index, and 1.50±8.06 mmHg for the SBP amplification, and were not statistically significant for all the parameters. The Bland-Altman analysis showed distinct correlations and concordance levels for different parameters: central SBP and central pulse pressure showed a very strong agreement (intraclass correlation of 0.926 and 0.886, respectively). In contrast, the concordance for the augmentation index and the SBP amplification was moderate (intraclass correlations between 0.5 and 0.8).

CONCLUSION

PWA provides similar measures of central blood pressure, whether measured over the right or the left carotid arteries, even though the morphological features of the pulse waves showed moderate agreement. The advantages of PWA over each arterial territory and the requirements that should mediate the choice of one of the both sides require further investigation.

Clinical evaluation of an optical fiber-based probe for the assessment of central arterial pulse waves

The aim of this study, enrolling 118 patients, was to clinically evaluate the accuracy of carotid pulse waveform acquisition with a new non-invasive optical fiber probe using invasive and non-invasive pressure readings as references. Pulse waves were acquired simultaneously in the ascending aorta and right common carotid; for the non-invasive study, the pulses were sequentially acquired using the optical fiber device and the Complior Analyse® (Alam Medical, France) device in the right carotid artery. For all subjects, the pulse waveforms assessed using the optical fiber sensor and the references were superimposed to analyze the deviation and point-by-point correlation. Augmentation index and central pressure were compared using intraclass correlation and Bland-Altman analyses with a confidence interval of 95%. For the invasive study, the acquired waves presented a mean deviation of 11 ± 3% and a mean intraclass correlation of 0.97 ± 0.02. Concerning the augmentation index and central systolic pressure, correlations of 0.79 (p < 0.001) and 0.94 (p < 0.001) were found, respectively. In the non-invasive comparison, the assessed mean deviation between the morphologies of the waves was 13 ± 5%, with correlation coefficients of 0.91 (p < 0.001) for the augmentation index and 0.98 (p < 0.001) for central systolic pressure. The results show that the optical fiber probe results were highly correlated with those obtained using the reference techniques in terms of the pulse waveforms, central systolic pressure (cSP), and augmentation index assessment.

Cardiorespiratory interactions: Noncontact assessment using laser Doppler vibrometry

The application of a noncontact physiological recording technique, based on the method of laser Doppler vibrometry (LDV), is described. The effectiveness of the LDV method as a physiological recording modality lies in the ability to detect very small movements of the skin, associated with internal mechanophysiological activities. The method is validated for a range of cardiovascular variables, extracted from the contour of the carotid pulse waveform as a function of phase of the respiration cycle. Data were obtained from 32 young healthy participants, while resting and breathing spontaneously. Individual beats were assigned to four segments, corresponding with inspiration and expiration peaks and transitional periods. Measures relating to cardiac and vascular dynamics are shown to agree with the pattern of effects seen in the substantial body of literature based on human and animal experiments, and with selected signals recorded simultaneously with conventional sensors. These effects include changes in heart rate, systolic time intervals, and stroke volume. There was also some evidence for vascular adjustments over the respiration cycle. The effectiveness of custom algorithmic approaches for extracting the key signal features was confirmed. The advantages of the LDV method are discussed in terms of the metrological properties and utility in psychophysiological research. Although used here within a suite of conventional sensors and electrodes, the LDV method can be used on a stand-alone, noncontact basis, with no requirement for skin preparation, and can be used in harsh environments including the MR scanner.

Pulse wave analysis with two tonometric devices: a comparison study

Pulse wave analysis is a pivotal tool to estimate central haemodynamic parameters. Available commercial devices use applanation tonometry and have been validated against invasive catheterism. We previously observed differences on a radial second systolic peak (rSPB2) between two commonly used devices: SphygmoCor (AtCor, Australia) and PulsePen (DiaTecne, Italy). The aim of our study was to further quantify differences in radial and carotid signals from the two devices.We measured radial and carotid waveforms in 38 patients with minimal changes between systolic, diastolic blood pressure and heart rate.rSBP2, mean pressure, form factor and augmentation index were different with SphygmoCor providing lower values (mean differences: 2.2 ± 3.8 mmHg; 1.5 ± 1.7 mmHg; 3.2 ± 3.5%; 4.2 ± 8.4%, respectively). Carotid augmentation index and form factor were similar. However carotid systolic pressure (cSBP) from PulsePen was higher than cSBP from SphygmoCor (2.7 ± 4.4 mmHg, p < 0.001). For both carotid and radial signals, harmonics moduli were similar across the spectrum with the exception of the 1st harmonic.PulsePen and SphygmoCor sensors are not equivalent and provide different wave shapes despite similar harmonics content with more discrepancy on radial derived parameters than on carotid derived parameters. Further studies are required to compare invasive pressure parameters to indices derived from these two devices.