- Open Access
Noninvasive measurement of cardiac stroke volume using pulse wave velocity and aortic dimensions: a simulation study
© Babbs; licensee BioMed Central Ltd. 2014
- Received: 9 June 2014
- Accepted: 16 September 2014
- Published: 19 September 2014
Concerns about the cost-effectiveness of invasive hemodynamic monitoring in critically ill patients using pulmonary artery catheters motivate a renewed search for effective noninvasive methods to measure stroke volume. This paper explores a new approach based on noninvasively measured pulse wave velocity, pulse contour, and ultrasonically determined aortic cross sectional area.
The Bramwell-Hill equation relating pulse wave velocity to aortic compliance is applied. At the time point on the noninvasively measured pulse contour, denoted th, when pulse amplitude has fallen midway between systolic and diastolic values, the portion of stroke volume remaining in the aorta, and in turn the entire stroke volume, can be estimated from the compliance and the pulse waveform. This approach is tested and refined using a numerical model of the systemic circulation including the effects of blood inertia, nonlinear compliance, aortic tapering, varying heart rate, and varying myocardial contractility, in which noninvasively estimated stroke volumes were compared with known stroke volumes in the model.
The Bramwell-Hill approach correctly allows accurate calculation of known, constant aortic compliance in the numerical model. When nonlinear compliance is present the proposed noninvasive technique overestimates true aortic compliance when pulse pressure is large. However, a reasonable correction for nonlinearity can be derived and applied to restore accuracy for normal and for fast heart rates (correlation coefficient > 0.98).
Accurate estimates of cardiac stroke volume based on pulse wave velocity are theoretically possible and feasible. The precision of the method may be less than desired, owing to the dependence of the final result on the square of measured pulse wave velocity and the first power of ultrasonically measured aortic cross sectional area. However, classical formulas for propagation of random errors suggest that the method may still have sufficient precision for clinical applications. It remains as a challenge for experimentalists to explore further the potential of noninvasive measurement of stroke volume using pulse wave velocity. The technique is non-proprietary and open access in full detail, allowing future users to modify and refine the method as guided by practical experience.
- Stroke Volume
- Pulse Pressure
- Pulse Wave Velocity
- Pulmonary Artery Catheter
- Pulse Contour
Monitoring cardiac pump function is extremely useful in critical care medicine and has been the standard of care in order to ensure tissue oxygenation . Clinical measurement of cardiac output is traditionally done using a flow-directed Swan-Ganz catheter, placed percutaneously into the pulmonary artery. Typically the pulmonary artery catheter is equipped with an injection port and a downstream temperature sensor to measure total pulmonary artery blood flow by thermodiluiton . In recent years, however, the cost effectiveness of this invasive procedure has been called into question [3–5], especially given complications that occur in about 10 percent of cases . In the SUPPORT trial the 30 day survival critically ill patients was greater in 3552 patients managed without a pulmonary artery catheter than in 2184 similar patients managed with a pulmonary artery catheter . Moreover, in 1008 matched pairs of patients managed either with or without pulmonary artery catheters, the total cost was 38% ($13,600) greater when pulmonary artery catheters were used to monitor heart function. Hence it is timely to revisit the issue of clinical monitoring of stroke volume and cardiac output with an eye toward less invasive and less expensive options.
According to Geerts, Aarts, and Jansen  the ideal technique for the measurement of cardiac output would be one that is accurate, precise, operator independent, fast responding, non-invasive, continuous, easy to use, inexpensive, and safe. The present paper proposes a new technique for measurement of stroke volume and cardiac output by combining standard cuff-based brachial artery blood pressure data, noninvasive measurements of aortic pulse wave velocity , and ultrasonic measurements of aortic luminal cross sectional area . The theoretical feasibility of this alternative technique is explored in three stages. The first stage is an analytical description of the underlying theory of the proposed method, including consideration of tapering of the aortic diameter, and axial gradients in local aortic compliance. The second stage is a test of the validity and accuracy of the method using a numerical model of the aorta and systemic circulation to explore effects of potential confounding variables, including simulated atherosclerosis, nonlinear compliance, and fast heart rates. The third stage is a discussion of the precision of the method based on the propagation of errors from the multiple separate measurements required in the calculation of stroke volume and cardiac output.
A well known clinical rule of thumb is that stroke volume is related directly to arterial pulse pressure (that is, systolic minus diastolic pressure) in a given patient, when aortic compliance is constant. The compliance of an elastic tube is the ratio of volume change ΔV to pressure change ΔP when an incremental volume ΔV is introduced. In symbols C = ΔV/ΔP. The dynamic compliance is the instantaneous slope of the volume versus pressure curve. For an aorta of given compliance the stroke volume, SV, is related to pulse pressure PP directly: SV ~ C⋅PP [9–12] if runoff of blood from the aorta is taken into account. Because the compliance is strongly dependent on the diameter of the vessel, the aorta itself and its largest branches are responsible for most arterial compliance seen by blood ejected from the left ventricle. Classically [11, 13] in typical human patients the stroke volume in ml is about 1.5 times pulse pressure in mmHg. That is, in a normal adult human the effective aortic compliance is about 1.5 ml/mmHg. However, in patients with significant abnormalities in arterial compliance, such as atherosclerosis and hypertension, the ratio may be substantially different . The present paper shows theoretically that (1) aortic compliance can be determined from aortic pulse wave velocity and aortic volume, both of which can be measured noninvasively and (2) stroke volume can be determined from systolic and diastolic blood pressure and compliance, even in the presence of pulse reflections and continuous runoff of blood from the aorta, also using completely noninvasive measurements.
Estimation of aortic wall compliance
in units of , where V is the instantaneous volume of fluid in the tube.
Equations (3) state that compliance is simply related to the volume of the vessel, the blood density, and the square of the pulse wave velocity. This relationship is identical to the Bramwell-Hill equation , first published in 1922, given the definition of compliance as ΔV/ΔP. Importantly in the present context, aortic volume can be estimated noninvasively using external anatomic landmarks to establish the effective length of the aorta and using ultrasound to measure the mid-level cross sectional area of the aortic lumen [8, 16]. Hence, in principle, aortic compliance can be estimated non-invasively.
From compliance to stroke volume
Using the relationship, C = ΔV/ΔP, for the whole aorta, the volume change can be determined from the compliance and the pressure change, under steady state conditions in which the pressure and volume changes are allowed to stabilize. If this state of affairs were strictly true for the aorta, then given a reasonable estimate of compliance, we would have SV ≈ C⋅ΔP, where ΔP is pulse pressure. However, in a real aorta the situation is more complicated for three reasons. First, there is continual runoff of blood from the aorta during and following ejection of blood from the left ventricle, so that the aorta is a leaky compliance. Second, a finite time is required for the pulse wave to travel along the aorta from aortic valve to iliac arteries, hence the pressure may not be the same in all parts of the aorta at the same time, especially during the early part of the pulse wave. Third, there are damped reflections of the pulse wave, so that pressure waveforms in proximal and distal aortic sites differ. Nonetheless, techniques for a more refined estimate of compliance-based stroke volume can be derived as follows.
The small animal method
The half time method
To ensure more steady state conditions one can consider the half time, th, defined as the time after the onset of ejection when blood pressure falls to a point halfway between systolic and diastolic pressures. This time is substantially longer than the pulse propagation delay in humans. Further, by this time the oscillations in pulse pressure have been damped out to a large extent, and pressures at all points along the length of the aorta are similar. At time th the total aortic compliance then comes into play. The halftime can be determined from the pulse contour, as indicated in Figure 2, which can be recorded noninvasively with external pulse pickups [17–19], including those used to measure pulse wave velocity.
where , with CVP denoting central venous pressure.
The CVP can be either assumed to be negligible, or if elevated, estimated noninvasively from physical examination of the jugular veins.
Axial tapering and compliance gradients
An obvious limitation of the forgoing analysis is the assumption of uniform diameter and wall thickness for the entire aorta that is implied in Equation (1). Anatomically, it is well known that the aorta tapers slightly in both its internal and external diameters from the level of the aortic root to the aortic bifurcation. Also wall thickness is greater in the thoracic than in the abdominal portions of the aorta.
Note that if the total taper, ϵL, is less than r0, then for this model, the compliance of the tapered aorta is only slightly greater than that of an un-tapered aorta having the middle value of radius. From the data of Voges et al.  we can estimate that ϵL/r0 is approximately 6 mm/8 mm, giving about a 4% difference in total compliance with tapering vs. a model of uniform thickness. This is an acceptable error for clinical purposes. If desired, the tapering correction term in parentheses could be taken as a constant for humans, approximately equal to 1.04. Accurate determination of the mid-level aortic cross sectional area, , however, remains important. Also, the issue of nonlinear compliance must be addressed, as shown subsequently.
Noninvasive data acquisition
Pulse wave velocity
At least three different commercial systems are available for measuring aortic pulse wave velocity from external sensors placed over the carotid and femoral arteries [7, 21]. These devices detect the carotid and femoral pulses using external pulse pickups [18, 19] and employ sophisticated algorithms such as cross correlation [22, 23] to determine the time delay between pulses. The time delay is divided into an estimate of effective aortic length to obtain aortic pulse wave velocity. Commercial systems give pulse wave velocities between about 8 and 10 m/sec for typical adult patients having some degree of atherosclerosis .
Conventional ultrasound sector scanning can be used to determine aortic cross sectional area in the high abdominal or lower thoracic aorta near its midpoint . One can use external landmarks to estimate the effective length of the aorta, for example as the distance from supra-sternal notch to either anterior-superior iliac spine. For use in Equation (3) the effective length, L, is defined as the length of an un-branched tube having the same total compliance as the natural aorta and its largest (brachiocephalic, carotid, and iliac) branches. The effective length, L, is slightly longer than the anatomic distance from the aortic valve in the chest to the aortic bifurcation in the mid abdomen, including the curvature of the aortic arch. Blood density is essentially constant at 1.03 g/ml. Combining these values in Equation (3) gives a noninvasive estimate of total aortic compliance.
Pressures and time intervals
Systolic and diastolic aortic pressures can be determined in the usual way using an external arm cuff or an indwelling arterial catheter. Characteristic time, th, and cycle length, T, can be determined from a noninvasive pulse pickup that records the pulse contour.
In this way cardiac output and stroke volume can be estimated quantitatively from a suite of noninvasive measurements, based on the underlying physics of the aorta. These measurements can be combined to provide estimates of stroke volume by the halftime method. A simple spreadsheet on a laptop computer or application on a smart phone can be programmed to do the calculations automatically, given the input data.
Testing the proposed methods in a computer model of the circulation
The accuracy of the proposed method can be tested and refined using a computational model of the human systemic circulation. In such a model one can compute pulse wave velocity from proximal and distal pulse waveforms and utilize instantaneous aortic cross sectional area to estimate aortic compliance. In turn, one can compute pulse contour parameters and stroke volume using the halftime method, comparing the results vs. “actual” values in the model. This test system can mimic the details of an idealized aorta in a way that makes it straightforward to investigate possible confounding effects such as, varying stiffness, nonlinear compliance, varying myocardial contractility, and varying heart rate, including both normal and shock-like states.
Model parameters for a normal adult human
Rs1 -- Rs9
Local systemic resistance
Pump input resistance
Pump output resistance
Ra0, Ra2 -- Ra8
Aortic segment resistance
Carotid and femoral resistance
Ca0 -- Ca9
Compliance of an aortic segment
Compliance of lumped venous reservoir
Pump compliance in diastole
Aortic segment length
Initial aortic segment inner radius
Initial volume of lumped veins
Initial pump volume
Cardiac frequency (heart rate)
Maximal external pump pressure
Time step for numerical integration
Time at which output data start being plotted
Time at which output data stop being plotted
Number of calculated points skipped between plotted points
Initial equilibrium pressure of arrested circulation
One obvious difference between the simple test system in Figure 3 and a living subject is that over a wide range of possible physiologic pressures in a given individual dynamic arterial compliance is not constant [14, 25]. Instead the compliance is “nonlinear” because the volume vs. pressure function for arteries is curved over the complete pathophysiologic range. Its slope, the dynamic compliance, varies as a function of pressure. However, a simple modification of the model in Figure 3 can be introduced to mimic the nonlinear characteristics of biological tissues.
as a function of pressure only. Here the reference pressure Pref is analogous to the reference tension T* in Fung’s original model. One can take Pref as the reference pressure at which the aortic compliance is normally determined, such as normal diastolic arterial pressure (80 mmHg) and convert linear compliances, Cref, into nonlinear ones, using Equation (16). In this way the numerical model of the mock circulation is easily modified for nonlinear dynamic compliance of the aorta by using expression (16) in place of Cref. As instantaneous distending pressure, P, becomes greater than Pref, the dynamic compliance becomes less than the nominal value, Cref. As instantaneous distending pressure, P, becomes less then Pref, the dynamic compliance becomes greater than the nominal value, Cref, mimicking the changes in slope of the nonlinear pressure-volume curve in Figure 1.
To make the model circulation go, a positive external, squeezing pressure having a half sinusoidal waveform is applied to the pump compartment. Instantaneous blood flows, volume changes, and pressure changes in all compartments of the model are computed during each discrete time step Δt = 0.00001 sec. These changes are numerically integrated, using the simple Euler method, to give time domain records of instantaneous volumes and pressures in all model compartments, as explained in detail in Appendix 2. The simulations begin with a cold start at the initial vascular volumes and a “mean circulatory pressure” of the arrested circulation of 10 mmHg in all compartments. The left ventricular pressure function is turned on, and at successive time steps, Δt, the following variables are computed to create a marching solution for flows, volumes, and pressures:
saved current flows (to compute derivatives)
integral of instantaneous left ventricular outflow (to compute true stroke volume).
Extraction of parameters from simulated pulse contours
To determine pulse wave velocity for the model circulation in Figure 3 the time interval between upstrokes of the aortic pulse waveforms in segments 2 and 8 was measured directly and divided into the center-to-center distance between the segments. The time difference was measured at a pressure corresponding to diastolic pressure plus 2 percent of the pulse pressure. These points marked the very beginning of the pulse upstroke and corresponded to aortic cross sections at the diastolic level. They are relatively insensitive to reflected wave augmentation in the arterial pressure waveform.
The representative cross sectional area of the aorta, as would be determined by ultrasound, was determined from the average of the diastolic radii of the mid aortic segments (which were quite similar). Blood density was taken as 1.03 g/cm3. Total aortic compliance was calculated using Equation (3b), with aortic volume computed as the product of mid-level cross section and the total length of the ten arterial segments. (That is, for the purposes of this simulation the effective length of the aorta was taken as the actual length.) The time interval th, representing the half time was computed from the contour of the function of aortic radius vs. time, working backwards from a time point just prior to the next diastolic point to avoid possible large oscillations after the peak of the pulse. Then stroke volume was computed using the half time formula of Equation (9). Integrated flow across the aortic valve (Equation (17)) was used as the reference, or actual stroke volume.
Estimation of aortic compliance
Pulse wave velocities calculated from pulse transit times for 1/3, ½, 1, and 3/2 normal aortic compliance were 1034, 857, 667, and 577 cm/sec respectively. Commercial systems give pulse wave velocities between about 8 and 10 m/sec for typical adult patients , who would be expected to have atherosclerosis and relatively stiff, less compliant, aortas. These are quite similar to those in the mock circulation for 1/3 and ½ normal aortic compliance.
Figure (4b) shows the ratio of stroke volume computed from pulse wave velocity and aortic cross section compared to actual stroke volume in the model over a range of aortic compliances from 1/3rd to 1.5 times normal. The halftime method of Equation (9) was used to estimate stroke volume. The agreement for the simple linear compliance model is satisfactory.
Dealing with nonlinear compliance
This form of expression (18a) shows that for small pulse pressures the correction factor approaches 1.00. For increasingly larger pulse pressures the correction factor becomes progressively less than 1.00.
When the correction ratio (18a) is applied to the data in Figure 5(a) the results in Figure 5(b) are obtained. The systematic overestimation of stroke volume at larger pulse pressure is largely eliminated. The correlation coefficient (r2) for estimated vs. actual stroke volume in Figure 5(b) is 0.998.
Varying heart rate
A global emphasis on cost effectiveness in medicine has made it timely to revisit the issue of noninvasive measurement of cardiac output and stroke volume, which remains an open problem in biomedical engineering. Classical standard methods, inducing the Fick and indicator dilution techniques, require centrally placed right heart catheters to sample mixed venous blood or to gain access to the combined circulatory flow in the pulmonary artery. Pulse contour methods frequently require at least one invasively obtained calibration value, which should be repeated if mean aortic pressure changes, owing to the nonlinear compliance of the aorta. Bio-impedance methods [6, 27] are fully noninvasive, but somewhat difficult to calibrate, although novel front-to-back electrode placements can help to give a fully noninvasive estimate of cardiac ejection fraction, if not absolute stroke volume . The present paper proposes a new strategy, based on an old equation  that allows for fully noninvasive estimates of the absolute value of stroke volume, and in turn, cardiac output. The results show that it is at least theoretically possible to obtain accurate estimates of stroke volume in human adults over a wide range of aortic stiffnesses, heart rates, and ejection fractions. Moreover, the calculations can account gracefully for nonlinear aortic compliance. Unlike other non-invasive techniques for estimating cardiac output, which have been well reviewed [6, 29], the present method does not involve proprietary algorithms. The validity and assumptions of the approach are fully open to evaluation. Modifications and refinements can be made easily in response to future laboratory and clinical experience. Of course, pulse wave velocity based estimates of stroke volume would be precluded in patients with severe cardiac arrhythmias and in patients with intra-aortic balloon pumps. In such patients a pulmonary artery catheter may be justified.
Propagation of errors from multiple measurements
A potential weakness of the proposed technique is the precision with which pulse transit time and pulse wave velocity can be measured, since the computed compliance depends on the square of the pulse wave velocity. There is also potential error in the cross sectional area of the aorta near its midpoint, as measured by ultrasound, as well as in externally measured systolic and diastolic blood pressures and in the cardiac cycle length and halftime values.
And if ϵ = 0.05, then . If there are 5% errors each in the aortic cross section, pulse pressure, aortic length estimate, lumped timing factor, and pulse wave velocity, then the expected variation in computed cardiac output would be 14 percent. The 95% confidence limits, roughly two standard deviations, would be 28 percent of the mean value. This estimate is within the clinically acceptable limit of 30 percent variation, proposed by Critchley and Chritchley .
The potential lack of precision in noninvasively estimated cardiac output must be compared with the lack of precision in established “gold standard” techniques, which is also non-trivial . The Fick method, for example is an aggregate of independent measurements of oxygen uptake and arteriovenous oxygen difference. The indicator dilution method requires independent measurements of the amount of dye injected and the area under the dilution curve compared to baseline. The estimate using Equation (25) is similar to that of other alternative methods of measuring cardiac output as reviewed by Geerts , in which coefficients of variation ranged from 5 to 15 percent. Hence, then proposed noninvasive approach would appear to have potential precision similar to that of existing methods with fewer drawbacks and complications.
More precise measurements of the component variables, would lead to correspondingly better estimates of stroke volume. An important opportunity for increasing precision is the ability to sample and average data for 10 to perhaps 100 heartbeats using an automated data collection system. Random variations can be reduced substantially by such averaging. It remains as a challenge to experimentalists to create practical systems with sufficient precision for practical, clinical implementation.
In discussing the overall validity of the proposed method, however, it is important to distinguish between errors in accuracy introduced by constant biases and true random biological variation. Constant systematic biases may be introduced into the calculations, but are less troublesome than unpredictable random variations. For example, in determining the effective aortic length, L, based on superficial landmarks, it may well be that the actual aortic length is, on average, say, 10 percent more than the landmark based measurement, with individual patients varying somewhat about this value. In determining the precision of the method the constant average bias is not important. Clinicians can easily adapt to a constant 10 percent overestimate of stroke volume. Changes over time in the management of an individual patient remain easy to recognize and interpret. Further, with experience such constant biases can be reduced to acceptable levels.
There are alternative solutions to the problem at hand. The lack of evidence for the cost effectiveness of invasive hemodynamic monitoring using a pulmonary artery catheter  has prompted commercialization of systems to track changes in stroke volume by pulse contour analysis. These systems have been well reviewed [6, 13, 29, 33]. The PiCCO System (Pulsion Medical Systems, Munich, Germany) derives stroke volume estimates from the diastolic to systolic integral of the pulse contour, which is calibrated by invasive thermal dilution measurements, using a dedicated catheter typically placed in the femoral artery. The calibration must be repeated every time there is a significant hemodynamic change. The LiDCO plus system (LiDCO, Cambridge, UK) uses a proprietary pulse pressure algorithm (PulseCOTM) based on the change in power in the arterial tree and repeated lithium dilution for calibration. These systems allow continuous monitoring of changes in hemodynamics but still require placement of invasive catheters for calibration. The Flotrac system (Edwards Lifescience, Irvine, CA, USA) incorporates a specific transducer that may allow calibration using a non-central arterial line for accurate waveform acquisition with calibration based on nomograms using the age, weight, and sex of the patient to estimate compliance. The Flowtrac algorithm uses a multivariate scaling parameter reflecting the effects of vascular tone on pulse pressure that is computed from a polynomial equation. None of these systems is truly noninvasive.
A completely noninvasive approach has been described classically by Huntsman and coworkers , who used ultrasound to measure the volume of blood moving through the ascending aorta during systole. The blood velocity was calculated from Doppler measurements and the cross sectional area was measured in A mode or M mode. Stroke volume was calculated as V x ET x CSA, where V = the spatial average blood velocity in the aorta during systole, ET = ejection time and CSA = the cross-sectional area of the lumen. Of course the roughly 7 percent of stroke volume flowing into the coronary arteries would be missed by this method. The approach appears to be relatively dependent on the skill of the operator, requiring 10–15 minutes for the initial diameter assessment  by unhurried examination of the anatomy and repeated diameter determinations. In addition, the first velocity measurement may take as long as 5 minutes. Nonetheless, this classic paper shows the clinical feasibility of fully noninvasive measurements of stroke volume and cardiac output.
The present analytical and numerical modeling exercise suggests that a completely noninvasive pulse wave velocity based method for measuring stroke volume can in principle be accurate, despite pulse wave reflections within the aorta and nonlinear vascular compliance. That is, it can give the true value in the absence of noise. Follow-on clinical studies are needed to determine if the real-world precision of this approach is sufficient for monitoring the circulatory status of critically ill patients and if the method is practical and feasible in real-world clinical settings.
SBP as systolic arterial pressure
DBP as diastolic arterial pressure
PP as pulse pressure (systolic minus diastolic)
PH as halftime pressure (the average of systolic and diastolic pressure)
CVP as central venous pressure
tp as time to peak from the diastolic point
th as time from diastolic point to halftime point
T the cardiac period or cycle length
PH’ as PH minus CVP
For non-invasive estimates the terms tp and th can be determined from the time domain output of an external pulse pickup. Systolic and diastolic arterial pressures can be determined noninvasively as well by the usual cuff method and CVP by physical examination of the jugular veins.
Text Figure 3 illustrates a simple, 12-compartment numerical model of the circulatory system, for which numerical values for inertances, resistances, and compliances are determined as follows.
in units of mmHg/(ml/sec2), where instantaneous cross sections, A, are computed from segmental volume divided by the fixed segment length.
Systemic vascular resistance is estimated as normal mean arterial pressure divided by cardiac output or 95 mmHg/(5 L/min) = 95 mmHg/(83.3 ml/sec) = 1.14 mmHg/(ml/sec) in total. Each of the 9 parallel systemic vascular resistances in text Figure 3 was 10.3 mmHg/(ml/sec) or 9 times the total resistance.
Segmental resistances to axial blood flow along the aorta, Ra0 and Ra2 through Ra8, are more than three orders of magnitude smaller than corresponding vascular resistances to drainage of blood from the aorta into the lumped venous compartment. These segmental resistances were initially estimated using Poiseuille’s law and then increased empirically to provide a physiologically realistic amount of damping of the arterial pulse waveform in the mock circulation. Final values for Ra0 and Ra2 through Ra8 were 0.005 mmHg/(ml/sec) and remained constant for all simulations. These aortic resistances are equal to each other (no tapering unless otherwise specified). In addition, the end resistances R1 and R9 are further increased to 0.1 mmHg/(ml/sec) to model the increased resistance of narrowing vessels in the carotid and femoral arterial trees. Similarly, the pump inflow and outflow resistances, Ri and Ro, are taken as small values = 0.01 mmHg/(ml/sec).
Using 280 msec as a normal ejection time , normal compliance of the aorta is estimated as 83.3 ml/40 mmHg multiplied by (1 – te/T) (text equation (5)) or 2.08 * (1–0.28/0.75) = 1.3 ml/mmHg. For ten equal aortic segments (no tapering) the compliance of each is 0.13 ml/mmHg normally. Total venous compliance is much larger than aortic compliance and is taken as 10 ml/mmHg. A nominal value for diastolic pump compliance is taken as 60 ml/5 mmHg = 12 ml/mmHg or stroke volume divided by the difference between end-systolic and end-diastolic pressure in the left ventricle.
For this initial model the length of each aortic segment is taken as 5 cm and the radius of the aorta is taken (uniformly, unless otherwise specified) as 1.5 cm. The initial volume of the venous reservoir is 2000 ml. The initial volume of the heart pump in an arrested circulation is modeled as 200 ml.
where Pmax is the peak applied pressure. To create a range of stroke volumes in the model Pmax was varied from 25 to 175 mmHg.
Here the max() functions simulate the effects of the inflow and outflow valves, permitting one-way flow only through the pump.
The cosine function in (38) represents the time derivative of Pext(t). Instantaneous pump volume and instantaneous pump pressure as functions of time are determined by numerical integration of expression (37) and expression (38).
Compartmental volume and pressure changes
Computational formulas for instantaneous aortic flows
Expressions of this form (45) are used as computational formulas for instantaneous aortic flows ia0, ia1, … through ia8.
As expected, for extremely small pulses the effect of nonlinearity is negligible. For increasingly larger pulses the correction factor becomes progressively less than unity.
For blood pressure of 120/80 Equation (54) gives 0.942 and Equation (55) gives 0.938. For blood pressure 180/40 Equation (54) gives 0.473 and Equation (55) gives 0.708. Hence, in general, and especially for higher pulse pressures, one must use the exact logarithmic expression of Equation (54).
The probable error method or delta method [30, 35] may be used to approximate variances of functions of random variables. If X is a random variable with mean μ and variance σ2, the variance σ2(f(X)) ≈ σ2(X) (f′(μ))2 where f′(X) is the first derivative of function f(X) with respect to X. To appreciate the approximation one can visualize the function f(X) as a graph with a tangent of slope f′(μ) at point (μ, f(μ)). By deduction from such a graph, it follows that the standard deviation of f(X) is approximately f′(μ) times the standard deviation of X, as long as f′(X) does not change greatly over the range of X. For the case of f(X) = ln(X), the delta method gives .
- Mimoz O, Rauss A, Rekik N, Brun-Buisson C, Lemaire F, Brochard L: Pulmonary artery catheterization in critically ill patients: a prospective analysis of outcome changes associated with catheter-prompted changes in therapy. Crit Care Med 1994, 22: 573–579. 10.1097/00003246-199404000-00011View ArticleGoogle Scholar
- Kurita T, Morita K, Kato S, Kikura M, Horie M, Ikeda K: Comparison of the accuracy of the lithium dilution technique with the thermodilution technique for measurement of cardiac output. Br J Anaesth 1997, 79: 770–775. 10.1093/bja/79.6.770View ArticleGoogle Scholar
- Connors AF Jr, Speroff T, Dawson NV, Thomas C, Harrell FE Jr, Wagner D, Desbiens N, Goldman L, Wu AW, Califf RM, Fulkerson WJ, Viadillet H, Broste S, BGellamy P, Lynn J, Knaus WA: The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. JAMA 1996, 276: 889–897. 10.1001/jama.1996.03540110043030View ArticleGoogle Scholar
- Harvey S, Harrison DA, Singer M, Ashcroft J, Jones CM, Elbourne D, Brampton W, Williams D, Young D, Rowan K: Assessment of the clinical effectiveness of pulmonary artery catheters in management of patients in intensive care (PAC-Man): a randomised controlled trial. Lancet 2005, 366: 472–477. 10.1016/S0140-6736(05)67061-4View ArticleGoogle Scholar
- Harvey S, Stevens K, Harrison D, Young D, Brampton W, McCabe C, Singer M, Rowan K: An evaluation of the clinical and cost-effectiveness of pulmonary artery catheters in patient management in intensive care: a systematic review and a randomised controlled trial. Health Technol Assess 2006, 10: iii-iv. ix-xi, 1–133View ArticleGoogle Scholar
- Geerts BF, Aarts LP, Jansen JR: Methods in pharmacology: measurement of cardiac output. Br J Clin Pharmacol 2011, 71: 316–330. 10.1111/j.1365-2125.2010.03798.xView ArticleGoogle Scholar
- Rajzer MW, Wojciechowska W, Klocek M, Palka I, Brzozowska-Kiszka M, Kawecka-Jaszcz K: Comparison of aortic pulse wave velocity measured by three techniques: complior, SphygmoCor and arteriograph. J Hypertens 2008, 26: 2001–2007. 10.1097/HJH.0b013e32830a4a25View ArticleGoogle Scholar
- Bhatt S, Ghazale H, Dogra VS: Sonographic evaluation of the abdominal aorta. Ultrasound Clinics 2007, 2: 437–453. 10.1016/j.cult.2007.06.001View ArticleGoogle Scholar
- Remington JW, Noback CR, Hamilton WF, Gold JJ: Volume elasticity characteristics of the human aorta and prediction of the stroke volume from the pressure pulse. Am J Physiol 1948, 153: 298–308.Google Scholar
- Ferguson JJ, Julius S, Randall OS: Stroke volume–pulse pressure relationships in borderline hypertension: a possible indicator of decreased arterial compliance. J Hypertens Suppl 1984, 2: S397-S399. 10.1097/00004872-198408000-00012View ArticleGoogle Scholar
- Chemla D, Hebert JL, Coirault C, Zamani K, Suard I, Colin P, Lecarpentier Y: Total arterial compliance estimated by stroke volume-to-aortic pulse pressure ratio in humans. Am J Physiol 1998, 274: H500-H505.Google Scholar
- Jansen JR, Wesseling KH, Settels JJ, Schreuder JJ: Continuous cardiac output monitoring by pulse contour during cardiac surgery. Eur Heart J 1990,11(I):26–32.View ArticleGoogle Scholar
- de Simone G, Roman MJ, Koren MJ, Mensah GA, Ganau A, Devereux RB: Stroke volume/pulse pressure ratio and cardiovascular risk in arterial hypertension. Hypertension 1999, 33: 800–805. 10.1161/01.HYP.33.3.800View ArticleGoogle Scholar
- Posey J, Geddes L: Measurement of the modulus of elasticity of the arterial wall. Cardiovasc Res Cent Bull 1973, 11: 83–103.Google Scholar
- Bramwell J, Hill A: The velocity of the pulse wave in man. Proc R Soc Lond Ser B 1922, 93: 298–306. 10.1098/rspb.1922.0022View ArticleGoogle Scholar
- Huntsman LL, Stewart DK, Barnes SR, Franklin SB, Colocousis JS, Hessel EA: Noninvasive Doppler determination of cardiac output in man. Clinical validation. Circulation 1983, 67: 593–602. 10.1161/01.CIR.67.3.593View ArticleGoogle Scholar
- Ciaccio EJ, Drzewiecki GM: Tonometric arterial pulse sensor with noise cancellation. IEEE Trans Biomed Eng 2008, 55: 2388–2396.View ArticleGoogle Scholar
- Kemmotsu O, Ueda M, Otsuka H, Yamamura T, Okamura A, Ishikawa T, Winter DC, Eckerle JS: Blood pressure measurement by arterial tonometry in controlled hypotension. Anesth Analg 1991, 73: 54–58.View ArticleGoogle Scholar
- Sato T, Nishinaga M, Kawamoto A, Ozawa T, Takatsuji H: Accuracy of a continuous blood pressure monitor based on arterial tonometry. Hypertension 1993, 21: 866–874. 10.1161/01.HYP.21.6.866View ArticleGoogle Scholar
- Voges I, Jerosch-Herold M, Hedderich J, Pardun E, Hart C, Gabbert DD, Hansen JH, Petko C, Kramer HH, Rickers C: Normal values of aortic dimensions, distensibility, and pulse wave velocity in children and young adults: a cross-sectional study. J Cardiov Magn Reson 2012, 14: 77. 10.1186/1532-429X-14-77View ArticleGoogle Scholar
- Blacher J, Asmar R, Djane S, London GM, Safar ME: Aortic pulse wave velocity as a marker of cardiovascular risk in hypertensive patients. Hypertension 1999, 33: 1111–1117. 10.1161/01.HYP.33.5.1111View ArticleGoogle Scholar
- Asmar R, Benetos A, Topouchian J, Laurent P, Pannier B, Brisac AM, Target R, Levy BI: Assessment of arterial distensibility by automatic pulse wave velocity measurement. Validation and clinical application studies. Hypertension 1995, 26: 485–490. 10.1161/01.HYP.26.3.485View ArticleGoogle Scholar
- Jezewski J, Roj D, Wrobel J, Horoba K: A novel technique for fetal heart rate estimation from Doppler ultrasound signal. Biomed Eng Online 2011, 10: 92. 10.1186/1475-925X-10-92View ArticleGoogle Scholar
- Babbs CF, Weaver JC, Ralston SH, Geddes LA: Cardiac, thoracic, and abdominal pump mechanisms in cardiopulmonary resuscitation: studies in an electrical model of the circulation. Am J Emerg Med 1984, 2: 299–308. 10.1016/0735-6757(84)90124-4View ArticleGoogle Scholar
- Drzewiecki G, Field S, Moubarak I, Li JK: Vessel growth and collapsible pressure-area relationship. Am J Physiol 1997, 273: H2030-H2043.Google Scholar
- Fung YC: Biomechanics: Mechanical Properties of Living Tissues. New York: Springer-Verlag; 1981.View ArticleGoogle Scholar
- Geddes LA, Voelz MH, Babbs CF, Bourland JD, Tacker WA: Pulse transit time as an indicator of arterial blood pressure. Psychophysiology 1981, 18: 71–74. 10.1111/j.1469-8986.1981.tb01545.xView ArticleGoogle Scholar
- Babbs CF: Anterior-posterior impedance cardiography: a new approach to accurate, non-invasive monitoring of cardiac function. Cardiovasc Eng 2010, 10: 52–65. 10.1007/s10558-010-9094-zView ArticleGoogle Scholar
- Alhashemi JA, Cecconi M, Hofer CK: Cardiac output monitoring: an integrative perspective. Crit Care 2011, 15: 214. 10.1186/cc9996View ArticleGoogle Scholar
- Crumpler TB, Yoe JH: Chemical Computations and Errors. New York: John Wiley & Sons, Inc.; 1940.Google Scholar
- Critchley LA, Critchley JA: A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput 1999, 15: 85–91. 10.1023/A:1009982611386View ArticleGoogle Scholar
- Stevens K, McCabe C, Jones C, Ashcroft J, Harvey S, Rowan K: The incremental cost effectiveness of withdrawing pulmonary artery catheters from routine use in critical care. Applied Health Econ Health Pol 2005, 4: 257–264. 10.2165/00148365-200504040-00008View ArticleGoogle Scholar
- Marik PE: Noninvasive cardiac output monitors: a state-of the-art review. J Cardiothorac Vasc Anesth 2013, 27: 121–134. 10.1053/j.jvca.2012.03.022View ArticleGoogle Scholar
- Lance VQ, Spodick DH: Heart rate–left ventricular ejection time relations. Variations during postural change and cardiovascular challenges. Br Heart J 1976, 38: 1332–1338. 10.1136/hrt.38.12.1332View ArticleGoogle Scholar
- Oehlert GW: A note on the delta method. Am Stat 1992, 46: 27–29.MathSciNetGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.