Stroke Volume Monitoring: Novel Continuous Wave Doppler Parameters, Algorithms and Advanced Noninvasive Haemodynamic Concepts

Stroke volume (SV), cardiac output (CO) and blood pressure (BP) are fundamental and independent measures of cardiovascular function, and are essential for the accurate understanding of cardiovascular (CV) physiology, pathophysiology and the guidance of fluid, inotropes and vasoactive therapies [1]. However, in clinical practice, BP and BP-derived surrogates of SV are used interchangeably in place of direct and accurate flow measurements. This conflation of flow and pressure potentially compromises our understanding of CV physiology and limits the effectiveness of goal-directed therapy (GDT) and the clinical utility of haemodynamics [2, 3•, 4]. It may also result in inappropriate therapies, adverse outcomes and sub-optimal management [5‐7, 8•].

The arterial system is a network of branches of changing diameter, with changing BP, flow and pressure waveforms throughout its distribution [9, 10]. Further, regional vascular tone changes beat-to-beat under the control of the autonomic nervous system (ANS), diminishing the value and reliability of peripheral BP measurements as an analogue of SV. Understanding what is actually being measured, both where and when, is vital when evaluating the circulation and monitoring the effect of therapy.

Guyton observed that BP was the product of CO and systemic vascular resistance (SVR) or ‘total peripheral resistance’ [1]:However, as CO is the product of stroke volume and heart rate (HR), then:Thus SV is the fundamental unit of CV function as regards blood flow delivered from the heart to the circulation, and represents the combined influences of the heart, blood vessels, the ANS and even the physical properties of the blood itself. The SV and SVR are interdependent and compensatory variables acting in concert, conducted by the ANS, to preserve BP and perfusion. Importantly, they are treated with distinct therapies: stroke volume predominantly with fluid and inotropes, and SVR with vasoactive agents.

It is also clear that BP and SV are distinct measures and cannot be used interchangeably. Substitution of SV with BP-derived estimates of SV is a physiological example of mathematical coupling which undermines advanced haemodynamic monitoring. An independent and accurate method of measuring SV is essential to complement the measurement of BP in order to understand the function of the circulation, as Frank, Starling and later Guyton, established [11‐13].

Importantly, the role of fluid, inotropes and vasoactive interventions is directly targeted to change SV, so accurate and sensitive measurement of SV is central to diagnosis and therapy of the cardiovascular system.

While this circulatory overview is fundamental, the accurate measurement of SV has not previously been routinely available, making the effective implementation of basic haemodynamic monitoring impossible.

The accurate measurement of SV is central to understanding advanced haemodynamics and has become the Holy Grail of circulatory assessment. However, as SV, by definition, is the flow per beat through the ventriculo-arterial valves, a flow measurement technology must be used to measure flow across either the pulmonary or aortic valves (PV or AV). Historically, many technologies have been used to estimate SV, with most measuring flow surrogates at various points in the circulation from which an estimate of SV is made using a transfer function. The further the measures are made from the ventriculo-arterial valves, the greater the influence of the ANS and SVR variation, and the less agreement the method has with flow, and the less reliable the technology.

Simple oscillometric measures of brachial BP have been ubiquitous since the pioneering work of Riva-Rocci in 1896 [14]. Although the most common circulatory measurement, oscillometry has significant sources of error leading to questionable reliability [15‐20]. While measurement of the BP is routinely performed at the level of the brachial artery with an upper arm cuff, the choice of referencing systolic, diastolic or mean arterial pressure (MAP) further confuses applicability of the method. Regardless, oscillometric BP is a poor surrogate for SV as Guyton predicted and Wo and Shoemaker et al. demonstrated [1, 21].

Several methods measure CO, the minute output of the ventricle, including the Fick method and partial CO₂ rebreathing techniques, thermodilution with the pulmonary artery catheter (PAC) and other indicator dilution techniques. CO is then divided by HR to yield the average SV. From 1970 to 2010, the PAC was regarded as a reference standard, but recent evidence suggests that not only is its absolute accuracy questionable, it has limited sensitivity in terms of detecting changes in SV. Further, there is little evidence of outcomes benefit; it is associated with significant patient risk; it requires time to set up and perform, and has limited application to critically ill patients [22, 23•, 24‐28]. While the Fick [29] method is considered accurate and reliable, it is not generally used in critical care patients, let alone the haemodynamically unstable patient seen in other clinical environments.

The use of pressure transducers to measure intra-arterial BP waveforms, with accompanying transfer functions to estimate SV, have also proven unreliable. This is particularly true in hyperdynamic circulations where these methods have limited value for guidance of interventions [30‐34]. These methods, which include LiDCO (LiDCO, London, UK), PiCCO (Pulsion Medical Systems, Munich, Germany), Deltex CardioQ-ODM+ (Deltex, Chichester, UK) and Vigileo/Flotrac (Edwards LifeSciences, Irvine, CA, USA), are predicated on the assumption that the ANS is constant between and within individuals and over time, resulting in SVR being mathematically predictable. This flawed assumption results in insensitivity to acute SV change, despite most devices recommending intermittent calibration to a alternate monitoring technology such as thermodilution or oesophageal Doppler. Additionally, the algorithms used by different manufacturers are proprietary and not directly comparable, resulting in absolute output values of poor agreement and interchangability, with none as accurate as Doppler ultrasound [30‐36]. Further, any predictability of the ANS functions, vascular tone and fluid dynamics breaks down in many common clinical situations. These include, but are not limited to, high, low and changing outputs as in sepsis and trauma, during dyssynchronous breathing, in dysrhythmias, with vasoactive interventions, in right ventricular overload states, peripheral vascular disease and in children, precisely the patients in whom the measurement of SV is critical [30]. This results from the mathematical coupling of BP and SV and is a fundamental source of error in haemodynamics. This limits the applicability of BP-derived haemodynamics and underlies the failure of many GDT protocols; guiding therapy by BP and a BP surrogate of SV can only be as effective as guiding it by BP alone [3•, 4, 5, 34‐37].

Other devices use plethysmographic or pulse oximetric measurements in the digits (Masimo, Irvine, California, USA; Nihon Kohden, Tokyo, Japan; Edwards LifeSciences, Irvine, California, USA), and then estimate the central circulation. However, as the measure reflects blood flow characteristics in the distal digital circulation, the region most markedly affected by circulatory redistribution, it is unlikely to provide any more than the circulatory status of the digit and provide little reliable information on the central circulation.

Similarly, non-invasive methods based on thoracic electrical impedance or reactance such as NiCOM (Cheetah Medical, Wilmington, Delaware, USA) have failed to demonstrate sufficient accuracy and sensitivity in critical care practice [37‐45].

Unfortunately, with the exception of techniques based on the Fick principle, the reliability of all of these methods for measurement of SV has been found to be generally inadequate for monitoring and therapy in the critical care environment where haemodynamic variability is common and can be critical [27, 34‐40].

Doppler ultrasound as used in echocardiography is a reliable method for measuring flow volumes, providing validated direct transvalvular measurement of SV and CO [36, 46, 47, 48••, 49‐54]. The ultrasonic cardiac output monitor (USCOM 1A, Uscom Ltd., Sydney, Australia) is a specialised Doppler SV monitor derived from echocardiography, which has similarly been validated for CO ranging from 0.12 to 18.3 L/min [55‐56]. These non-invasive flow measurements can be integrated with BP measures to generate integrated flow-pressure parameters and monitor changes predicted by pressure-volume loops. This may lead to improved understanding of cardiovascular physiology and pathophysiology, and ultimately to physiologically guided therapy with improved outcomes in adults and children [2, 58, 59].

The Doppler equation defines the relationship between frequency and velocity, and from velocity, the flow volume can be calculated by multiplying by the cross-sectional area (CSA) of the vessel. For a valve of radius r with blood flowing at a velocity V, and with a circular CSA, then at steady-state, the quantity of fluid flowing per unit time (Qt) or flow volume is:As the blood flow is pulsatile, we need to know the mean velocity of the blood flowing across the aortic or pulmonary valve. The ejection waveform at the aortic valve approximates a triangle, with a normal velocity range of zero at the base to around 1.4 m/s at the peak as shown in Fig. 2. The duration of systole is approximately 350 ms, and diastole around 450 ms, giving a total cycle time of around 800 ms (HR = 75 bpm) denoted by t. The mean velocity of ejection can be calculated from the area under the ejection curve by integrating the velocity with respect to time t. This is the velocity-time integral or vti, and is known as the stroke distance (sd) as it is the average distance red blood cells travel per heart beat, normally around 25 cm.

SV is the product of the stroke distance, vti, and the cross-sectional area (CSA):where CSA = πr². The ejection Doppler flow profile can be traced to calculate vti, but the radius of the ventriculo-arterial valve is required to calculate CSA and thereby, SV. There are several possible methods to obtain the CSA.

The diameter of the aortic or pulmonary valve, or the left or right ventricular outflow tract (LVOT/RVOT) can be measured using echocardiography [62]. However, the value for velocity that is used to calculate vti must be measured at exactly the same point as the CSA of the flow to comply with the continuity equation. For example, where a vessel narrows from a maximum diameter and cross-sectional area CSA1 to a minimum cross-sectional area at CSA3 and then widens again to its original size (Fig. 3), then the continuity equation states that the product of the velocity and CSA (the flow) at any one point equals the product of the velocity and CSA at any other point. Thus, as the vessel narrows, the CSA diminishes but the velocity increases.

The LVOT varies in diameter across its course (Fig. 4) and the continuity equation determines that the flow volume across any of these points, d1, d2, d3 and d4, must be equal, and so the velocity must vary as in Fig. 3. Coupling the exact position for measuring both Doppler flow and CSA is essential to prevent violation of the continuity equation, but this is technically challenging in practice. Fortunately, the physics of continuous wave (CW) Doppler simplifies the method.

Direct 2D ultrasonic measurement of the valve radius introduces potential errors secondary to oblique insonation and also from the spatial resolution of 2D ultrasound [60], both of which contribute significantly to errors in CSA calculation and thereby of SV. Mark et al. reported ‘2D measurement of aortic diameter was time consuming and poorly correlated with invasive measurements (r=0.31)’ [63]. Figure 6 illustrates the error resulting from measuring the flow diameter with an oblique ultrasound beam. The apparent CSA becomes oval with an increased measured diameter of the vessel. While the linear error of r may be small, once calculated as CSA = πr², any error is squared, leading to a significant error in the calculated CSA. An error in radius measurement of 3 mm results in over-measurement of the CSA from 346 to 572 mm², an error of 65%, which results in a flow volume measurement error of the same magnitude.

To ensure accuracy of the flow radius measurement, The American Society of Echocardiography (ASE) recommend averaging at least three measures of the flow diameter, and in dysrhythmia, 5 or more may be required [62]. While this improves the accuracy of echocardiography measurements, it increases the examination time considerably.

Measurement of the valve area using 2D ultrasound is further limited by the spatial resolution of the pulsed ultrasound beam. The frequency of the transducer f is inversely proportional to the wavelength λ, hence resolution increases directly with frequency:where C = velocity of ultrasound, f = frequency of the ultrasound beam and λ = wavelength of the pulse [61]. If each 2D ultrasound pulse is between 2 and 4λ, the spatial pulse length (SPL) of a transducer where f = 5 MHz is 1.2 mm, and the spatial resolution is 0.6 mm, as 2D axial spatial resolution = SPL/2.

Errors in measurement of the CSA, whether from oblique insonation or poor 2D spatial resolution, are the biggest single source of inaccuracy in echocardiographic measurement of SV. If changes in SV of 15% or less are to be accurately measured during clinical therapy, then this measurement requires meticulous technique and repeated measures to achieve acceptable sensitivity.

Once a reliable measure of SV has been acquired, combining this with other common physiologic measurements allows the calculation of a number of important parameters of cardiovascular and pulmonary performance.

Once the SV and CO are known, then combining these with arterial saturation SaO₂ and haemoglobin concentration [Hb] allows calculation of the volume of oxygen delivered per minute, DO₂. Dividing this by body surface area (BSA) yields oxygen delivery index (DO₂I):

The Fick equation [29] states:where VO₂ = minute consumption of oxygen, CaO₂ = arterial oxygen content and CvO₂ = venous oxygen content. If the venous and arterial oxygen contents are known, then it follows that VO₂ is simply:Again, this can be indexed to BSA as VO₂I [70]. The oxygen extraction ratio is then simply the ratio of VO₂ to DO₂.

Flow and pressure are the fundamental physical parameters of the cardiovascular system. While much of our understanding of disease and treatment has derived solely from the measurement of BP, combined flow-pressure measurements allow the calculation of pressure-volume coupling and combined flow-pressure parameters [71‐75].

With accurate measures of SV, direct measures of ventriculo-arterial coupling can be derived with input of BP values. For example MAP = CO x SVR [1]. However, if we assume that blood flow requires kinetic energy (KE) and BP represents potential energy (PE), then the product of the two energy components is the total stroke work (SW) of the heart in milli-Joules (mJ). Power is the rate of performance of work, so dividing SW by the time of one heart beat (60/HR) provides the work per unit time, while SW × HR is cardiac power output (CPO) for 1 min [76‐80].

Arterial elastance (Ea) is the ratio of pressure change to volume change in the pressure-volume loop, or dP/dV, and can be noninvasively calculated as pulse pressure (PP) divided by SVwhere PP = BP_systolic − BP_diastolic, and SV = left ventricular end diastolic volume − left ventricular end systolic volume (LVEDV − LVESV) [81‐84].

Arterial compliance (Ca) is the reciprocal of elastance or dV/dP and can be noninvasively calculated as SV/PP [82].

These combined flow-pressure parameters show increased predictive value for outcomes in heart failure, hypertension and cardiogenic shock [76‐80], provide an improved understanding of the Frank-Starling mechanism [79] and improve guidance of fluid, inotropes and vasoactive interventions.

Currently, there are three Doppler-based methods in clinical practice.

Doppler ultrasound has been used since 1980 to measure cardiovascular blood flow, and its applications and protocols for use are now well established. Doppler ultrasound is a sophisticated technology and its application to advanced haemodynamics depends on the training and skill of the operator, and an understanding of the unique physical properties of Doppler ultrasound [91•]. New specialised Doppler ultrasound devices deliver real-time SV measures simply and rapidly, and, combined with central and brachial BP monitoring, provide multi-parametric indices which can provide non-invasive flow-pressure parameters which improve on simple BP monitoring and which may provide novel insights into cardiovascular physiology, pathophysiology and therapy.