The impact of right ventricular function on left ventricular mechanical cir culatory support

Introduction

It is estimated that approximately 14,000 patients are living with left ventricular assist devices, approximately half of which are not eligible for heart transplantation and are considered “destination LVADs” (1). As technology and medical management have improved the number of LVAD implantations have increased and there is no reason to think this trend will not continue as the incidence of heart failure is anticipated to increase due to multiple factors especially an aging population. In addition, the use of percutaneously placed left ventricular assist devices used for temporary hemodynamic support, for example for percutaneous coronary intervention with cardiogenic shock after ST elevation myocardial infarction, has also markedly increased. The website of Abiomed, the manufacturer of the Impella device, notes that over 400,000 devices have been sold, and the use of the Impella has increased by over 300% over the decade (2). It may be anticipated that many anesthesiologists and intensivists will be involved in the management of these patients.

The first use of a left ventricular assist device for support of a patient in cardiogenic shock after valvular heart surgery was reported by Liotta et al. in 1963 (3). Although only transiently successful, this proof of concept, and support from the National Institutes of Health, stimulated the development of multiple devices (4). These early devices generated pulsatile flow with pneumatic or electric pumps which displaced blood with one-way valves to control direction of flow. Although there were clinical successes, these early devices were limited to temporary use due to problems of hemolysis or thrombosis and were characterized by bulky equipment with multiple moving parts. The next major technological advance was the development of continuous flow devices which had fewer moving parts, were more durable, and were smaller (5). These devices created less hemolysis and were associated with less thrombosis. And recently pumps with magnetically levitated rotors have been developed, which further reduces hemolysis and improves durability.

Left ventricular assist devices

Currently there are two variants of continuous pumps, centrifugal, in which blood is “thrown” outwardly from a set of rotor blades (6) and axial, in which blood is pushed parallel to the shaft on which the rotors are mounted (7). The axial pump facilitated the development of the most popular temporary left ventricular assist device, the Impella (8). In this review we will focus on the Heartmate III, a surgically implanted centrifugal device and the only durable LVAD currently approved by the FDA (9) and the Impella, the most used temporary LVAD. There is an alternative device for temporary support, the TandemHeart. This system requires a transeptal puncture to place a drainage catheter in the left atrium, a centrifugal pump, and an inflow catheter typically placed in a femoral artery (10). We focus on the Impella in this review as it has been more widely deployed than the TandemHeart. Before we focus on right heart function in patients with left ventricular assist devices, we need to briefly consider the operation of these left ventricular assist devices.

The HeartMate III is surgically implanted, with resection of the apex of the left ventricle and creation of an opening for attachment of the pump. Blood drains from this opening in the ventricle to the pump and is then propelled via a graft into the side of the ascending aorta. (Figures 1,2). Power is delivered to the pump via a driveline that passes through the skin for connection to external power supplies. The system is managed by a system controller Figure 3). The clinician can set the speed (in rpms) of the device. The control console will display the power, which is required to generate the set pump speed, the flow that is generated by the set pump speed, and the pulsatility index. The latter is the difference between the maximum and minimum power (normalized to average power) over a 15 s interval. Changes in the PI may occur due to changes in maximum power or minimum power or both. The PI reflects the interplay of native heart function and the assist device. Device flow is a function of the difference between aortic pressure and left ventricular pressure, the pressure gradient against which the pump is working (referred to as the pressure head). Flow will increase as this gradient decreases and visa-versa. This pressure gradient is at its maximum during diastole and at its minimum during systole. Thus flow, and consequently power (a function of both flow and pressure) will fluctuate with the cardiac cycle. In clinical practice PI typically ranges from 1 to 10. If the patient is clinically stable there should not be significant variation in the PI. A PI change of more than 45% will trigger a PI event in which the LVAD controller automatically reduces pump speed. Interpretation of the PI should not be done in isolation and must consider other pump paramets as well as standard hemodynamic parameters. For example, an increasing PI is often interpreted as improving native heart function. But paradoxically it may also reflect hypovolemia or right heart failure (decreased left ventricular filling with decrease left ventricular pressure during diastole with increased variation in flow during the cardiac cycle and a resultant increased PI).

In contrast to the HeartMate III, the Impella is a left ventricular assist device intended only for temporary support of the circulation and can be placed percutaneously or surgically depending on the model. The device is a catheter-based system with catheter size ranging from 9 French to 19 French (Figures 4,5). The tip of the catheter is typically placed into the left ventricle via the femoral or axillary artery. The catheter contains a very small axial pump (Archimedes screw pump) embedded in the terminal section with an inlet port which sits in the left ventricle and an outlet port which sits just above the aortic valve. The axial pump draws blood from the left ventricle via the inlet port and ejects it into the aorta via the outlet port. The catheter is then connected to a control console.

The Impella CP can provide as much as 4 L/min of forward flow while the Impella 5.5 can provide up to 5 L/min. Placement of the catheter is critical and augmentation of cardiac output requires that the tip of the catheter with the inlet port sits in the left ventricle while the outlet port sits just above the aortic valve. The Impella 2.5 and CP are usually inserted percutaneously via the femoral artery while the Impella 5.5 is inserted via arteriotomy typically of the axillary artery.

The Impella control console displays flow generated by the Impella with a continuous display of aortic and left ventricular pressure and a continuous display of the power expended (Figure 6). These displays are important since if power output ceases to be pulsatile, improper placement needs to be considered. The control console also allows one to set the “P” level. This enigmatic term (“P” for performance) determines the rotor speed of the pump and is a discontinuous setting, ranging from 0 to 9. For the Impella CP P 9 indicates a pump speed of 51,000 rpm while P 1 indicates a pump speed of 25,000 rpm (11). Flow rates can be adjusted by changing the P setting. The Impella control console also notifies the provider of the status of the purge system. To minimize thrombosis and hemolysis, it is necessary to keep blood from entering the motor housing of the pump. The purge system uses a counter flow of a heparinized solution (or alternatively a bicarbonate-based solution) to create a pressure gradient that keeps blood out of the motor housing. The console will display both the pressure and flow of the purge fluid.

Anatomy and physiology of the right ventricle

The right ventricle is more complex than the left ventricle. In cross section it is crescent-shaped and in sagittal section, triangular. It can be roughly divided into three sections, inlet, apex, and outlet. While the right ventricle is larger than the left ventricle but has a thinner free wall, smaller mass, and small myocytes. The free wall of the right ventricle is also characterized by coarse trabeculations (12).

The myocytes of both the left and right ventricle are arranged in two layers, an outer circumferential layer, and an inner longitudinal, spiral layer (13). The right ventricle contracts in a peristalsis-type fashion from inlet to outlet and longitudinal contraction is the major contributor to overall function. It is noteworthy that left ventricular contraction is a significant contributor to right ventricle contraction, through septal contraction and that the contraction of the septum reflects twisting movement of the longitudinal spiral muscle layer.

Contraction of the right ventricle does not normally generate high pressure but rather streamlines the flow of venous return into the pulmonary artery and the right ventricle is best described as a low pressure, high compliance chamber. The high compliance of the right ventricle allows it to accommodate fluctuations in venous return without substantial changes in stroke volume. However, the right ventricle is significantly more afterload sensitive than the left ventricle.

The effect of left ventricular mechanical support on right ventricular function is complex and different for permanent surgically placed devices than for a percutaneously temporary device. It is important to emphasize that all the parameters used to clinically assess ventricular function are load-dependent and can vary significantly with changes in preload and afterload. The gold standard for assessing ventricular (left or right) systolic function is end systolic elastance, Ees, which is load-independent and may be defined by the simple equation.

Ees=Pes/(Ves−Vo)

where Pes is the end-systolic pressure, Ves is the end-systolic volume, and Vo is dead space term (14). This is the slope of end-systolic volume to end-systolic pressure ratios for pressure-volume loops generated by altering preload (typically by occluding the inferior vena cava) (Figure 7). Optimal cardiac performance occurs when Ees is matched to the load the ventricle is facing (15). This is best described by the arterial elastance, Ea, which is related to Pes (but not defined by) by the relationship.

Ea=Pes/SV

where SV is stroke volume (Figure 8). Given that SV is equal to Ved-Ved where Ved is end-diastolic volume, it is simple to demonstrate that

SV=(Ved-Vo)/(1+Ea/Ees)

This basic equation demonstrates that stroke volume will decrease as either Ea (afterload) increases or Ees (contractility) decreases if Ved stays constant. If the ratio Ea/Ees) increases stroke volume is preserved only if Ved increases. In other words, the consequences of decreased contractility are either decreased stroke volume or increased end diastolic volume. In simple terms there are two forms of heart failure, “forward heart failure” (decreased flow) or “backward heart failure” (congestion).

The key aspect of right ventricular function is a smaller value of the Pes implying lower contractility and increased afterload sensitivity.

From the above equation one could conclude that stroke volume could be increased by decreasing the ratio, Ea/Ees, for example, by enhancing contractility with inotropes or decreasing afterload. However, there are constraints on this approach to increasing stroke volume. Afterload reduction is limited by the dependence of coronary perfusion on blood pressure. Furthermore, ventricular efficiency (defined as the amount of work done per unit of myocardial oxygen consumption) also depends on the ratio Ea/Ees (or rather the inverse of this). Ideally this ratio is unity. As Ees/Ea decreases below one the ventricle becomes less efficient, and more energy is expended to maintain stroke volume (16). This carries the risk of subendocardial ischemia, especially when if blood pressure is marginal.

Determination of Ees and Ea has largely been a research activity given the difficulties of generating multiple pressure-volume loops. However, there has been a limited study in patients who have received a HeartMate III device using a ramp test where LVAD speed is altered (17). In this study they found heterogenous responses to the LVAD speed. In many patients the LVAD did not alter right ventricular contractility (Ees). In these patients RV diastolic compliance improved, and RV afterload decreased (Ea). However, there were other patients who also showed no change in contractility, decreased afterload but no change in diastolic compliance, possibly compromising stroke work and exacerbating right heart failure. These observations are intriguing and call out for further investigation given the small size of this study.

There are multiple mechanisms by which mechanical support of the left ventricle can impact right heart function. The unloading of the left ventricle reduces pulmonary artery pressure which then reduces right ventricular afterload. The unloading of the left ventricle also will reverse the shift of the interventricular septum to the right. This should improve the diastolic compliance of right ventricular function, a phenomenon noted in the study by Brener et al. (17). However, the increased cardiac output will increase flow to the right ventricle, and this increased preload may overload the right ventricle. The shift of the septum to the left may impair right ventricular contractility given that the septum is an important contributor to right ventricular contraction. Also, tricuspid regurgitation may increase due to leftward shift of the septum. Thus, paradoxically there are occasions where right ventricular function is improved by decreasing LVAD pump speed. In the case of permanent LVAD (HeartMate III) placement, there are other factors in play. The surgical approach involves pericardiotomy. This releases the constraint to overfilling of the right ventricle. It has also been demonstrated that pericardiotomy reduces the velocity of longitudinal myocardial contraction leading to increased transverse contraction which is more afterload sensitive (18). Also, surgical implantation disrupts the twisting contraction of the spiral layer of left ventricular myocytes and this also reduces the longitudinal contraction of the right ventricle.

Assessment of right ventricular function

While right heart failure can be manifested in peripheral edema or elevated jugular distention, these findings can be subjective and nonspecific. More specific identification of right heart failure requires echocardiography and/or invasive monitoring, especially in patients with mechanical support of the left ventricle.

In contrast to the more spherical left ventricle, the crescent-shaped renders rough “eyeball” assessment of right ventricular function is problematic. The American Society of Echocardiography recommends multiple parameters to assess right ventricular systolic function (19). These include fractional area change denoted FAC, tricuspid annular plane systolic excursion, denoted TAPSE, longitudinal velocity of contraction, denoted S, or right ventricular longitudinal strain, denoted LS). All these parameters can be measured in the four-chamber apical view. FAC is measured by tracing the endocardium in diastole and systole and is the ratio of the difference in diastolic and systolic areas divided by the diastolic area. The low reference value is 35%. FAC is considered a more reliable measure of function than RV ejection fraction, which requires assumptions about geometry for calculation. TAPSE is also frequently used to assess right ventricular systolic function, using M mode imaging to measure the peak longitudinal motion of tricuspid valve annulus during systole. The low reference value is 17 mm. The right ventricular systolic longitudinal velocity of excursion is measured using tissue Doppler imaging. A pulsed wave Doppler signal from a sample window in either the annulus or the RV free wall is used to determine the maximum velocity of excursion. The low reference value is 10 cm/sec. Right ventricular longitudinal strain is acquired using Doppler tissue imaging of the basal section of the right ventricular free wall to measure the fractional change in myocardial deformation. Although earlier thought to be less reliable than other measures due to technical challenges to acquisition, it has subsequently to be a reliable indicator of right heart dysfunction in patients with LVADs (20). The reported lower reference values range from −9.6% to −15.5%.

It should also be noted that echocardiography can be used to estimate parameters that are typically derived by invasive monitoring. CVP can be roughly estimated by observation of IVC diameter. Also, pulmonary artery systolic can be estimated from the velocity of a tricuspid regurgitation jet and pulmonary artery diastolic pressure can be estimated if there is a sufficient pulmonary artery regurgitant jet, using the Bernoulli principle. Cardiac output can be estimated from the velocity of flow out the aorta. However, while echocardiography can be used to estimate hemodynamic parameters, if one needs more precise, invasive monitoring may be more appropriate. Central venous pressure (> 16 mm Hg) is a criterion for right heart failure after LVAD implantation and a central venous catheter is helpful in identifying right heart failure as a cause of inadequate LVAD flow. However, a pulmonary artery catheter may be more useful as it not only provides central venous pressure but also pulmonary artery pressures and cardiac output. These parameters can be helpful in managing right heart failure.

There are several derived parameters that might be useful in evaluating right ventricular function. such as pulmonary vascular resistance (mean pulmonary artery pressure minus pulmonary artery occlusion pressure divided by cardiac output) may give insight into the use of pulmonary vasodilators. The pulsatility index given byWhere PAS is systolic pulmonary artery pressure and PAD is diastolic pulmonary artery pressure has been demonstrated to be predictive of right heart failure (21). Right ventricular stroke work index is a relatively load-independent measure of RV function (22). It is calculated asWhere SVI is the stroke volume index and PAP is pulmonary artery pressure (and the mean is utilized in the calculation). The ratio of CVP to PCWP (pulmonary capillary wedge pressure has also been suggested as a measure of right heart function (23). There is no straightforward evidence for which of these derived parameters is most useful, although the simplicity of the pulsatility index or the CVP to PCWP is clinically useful.

PI=(PAS-PAD)/CVP

RWSWI=(meanPAP/CVP)SVI0.0136

Treatment of right heart failure

The right ventricle is smaller than the left ventricle and with fewer and smaller myocytes. Hence it could be expected that the response to inotropes of the right ventricle might be less robust than the left ventricle. So it is vital in treating right heart failure to optimize preload and afterload. As noted, the right heart is very afterload sensitive so it is important to treat pulmonary hypertension. If the patient is intubated one efficient treatment is inhaled nitric oxide (24). These vasodilators can be largely be cleared by one pass through the lungs and hence create minimal systemic vasodilation. Inhaled nebulized prostaglandins can also be given to patients who are not intubated (25).

If inhaled agents are not indicated, systemic inodilators like dobutamine, and especially milrinone, may be given (26). Milrinone is often selected due to its systemic and pulmonary vasodilatory properties. Also, milrinone should cause less tachycardia. However, milrinone may compromise systemic blood pressure and hence coronary perfusion. In this case, vasopressors may be needed. The first line agent for treating hypotension is norepinephrine. It should be realized, however, that norepinephrine may increase pulmonary vascular resistance and right ventricular afterload. Vasopressin is a useful adjunct or even alternative to norepinephrine. The advantage is that low dose vasopressin does not increase pulmonary vascular resistance and may even decrease pulmonary vascular resistance via activation of V1 receptors (27).

Optimization of preload is a complex subject (28). In general, we assume that both left and right ventricles operate on an ascending Starling curve, but the right ventricular Starling curve is flatter than the left. Overdistention of the right heart can impair cardiac output via multiple mechanisms. Increasing size of the right ventricle increases wall tension and this may reduce myocardial flow. This can be particularly deleterious if there is also increased afterload since elevated pulmonary artery can lead to more diastolic and less systolic myocardial blood flow. Another mechanism for impaired function when preload is elevated is the creation or exacerbation of tricuspid regurgitation. Finally, overdistention of the right ventricle can cause leftward shift of the interventricular septum and this can impair left ventricular performance. Paradoxically for these reasons there are situations in which volume removal (via either diuretics or renal replacement therapy) can improve right ventricular function, in contrast to the assumption, based on consideration of the Starling curve, that the converse should occur. Optimization of right ventricle preload is often a challenging clinical problem.

In the context of left ventricular mechanical support, the first goal of right ventricular optimization is to facilitate acceptable LVAD function. If pharmacological therapy is not effective, consideration of right ventricular mechanical support is indicated. There are several right ventricular assist devices currently available for acute use. We will focus on the most commonly used devices, the ProtekDuo or the Impella RP. The ProtekDuo is a percutaneously placed (typically via the right internal jugular vein) dual lumen catheter (29). The device has an inflow port which should sit in the right atrium and drain blood via a 29–31 French lumen. The pump generates flow through a 16–18.5 French catheter with an outflow port that sits in the main pulmonary artery. Positioning is critical. If the catheter migrates too far forward, there can be shunting of flow or damage to the pulmonary artery while retrograde migration will overload the right ventricle exacerbating right heart failure. The commonly used pumps are the TandemHeart and the CentriMag systems. The TandemHeart pump is a centrifugal pump that can generate up to 5 L/minute of flow. If the patient also has impaired oxygenation, the TandemHeart can be coupled with an oxygenator. Note that this pump is also used as a temporary left ventricular assist drive with a drainage catheter placed via a transseptal puncture in the left atrium and an outflow catheter in a femoral artery. If used with an oxygenator it can also support extra-corporeal membrane oxygenation. The pump can generate up to 4 L/min flow. The CentriMag pump is a magnetically-levitated centrifugal pump that can be used as a left ventricular assist device when surgically implanted or as part of a larger ECMO system with oxygenator.

The Impella RP uses the same technology used for left ventricular support and the only difference is the site of insertion usually a femoral vein. In parallel fashion to the Impella CP, it passes through the right heart into the pulmonary artery with a drainage port in the inferior vena cava and an outlet port in the main pulmonary artery (30). As with the left ventricular Impella, the axial pump is located in the tip of the catheter.

Discussion

The purpose of left ventricular assist devices is to provide systemic flow. If LVAD flow is not adequate, one should take a systematic approach to trouble shooting. Although the approach will vary depending on the device, there are problems common to any device.

The first step for diagnosing inadequate perfusion in any patient with an LVAD is to ensure that the device is actually functioning. This can often be evident on the device control console. For the HeartMate III one can usually confirm some flow by listening for the characteristic hum on auscultation. If the device is not functioning one should immediately make sure there is power to the device. If there is power, then check the driveline and all connections. Total failure of function with no obvious cause requires consultation with LVAD management experts. If the device is functioning, but with inadequate flow, one should consider problems at the device-patient interface such as inflow cannula obstruction, outflow cannula obstruction, or pump thrombosis. The latter are often manifested by increased power consumption. In the case of Impella devices, the trouble shooting of inadequate flow is similar. First, ensure that the device is actually functioning and not subject to kinking of the catheter or thrombosis (again an indirect clue to this is increased power expenditure). One additional consideration for Impella devices is catheter malposition, which can be suggested by nonpulsatile power output. Also, the latest generation of Impellas offer Smart Assist technology which analyzes, in real time, aortic pressure with an optical sensor as well as the pressure aortic-left ventricular pressure gradient as measured by the micro axial motor to confirm proper placement of the catheter.

Device-specific problems are rare (except for Impella catheter malpositioning) and much more common are problems which are specific to the patient. The primary problems to consider are hypovolemia, arrhythmias, tamponade, and right heart failure. These problems can be diagnosed by consideration of central venous pressure, pulmonary pressure, and most prominently echocardiography.

Both the HeartMate III and Impella are afterload-sensitive and preload-dependent. As noted earlier, as systolic blood pressure increases, the pressure head against which the pump is working, will increase and flow will decrease. Preload-dependence simply means that if the pump has no volume, it will not be able to generate flow.

Hypovolemia can result in suction events in which the ventricle collapses on itself due to negative pressure generated by continued flow. The HeartMate III automatically decreases speed in the event of a suction event. And in the event of a suction event, the Impella console will recommend a reduction in pump speed (the P level). But it is important to realize that suction events do not always reflect inadequate preload and may also reflect catheter malposition or thrombosis. The displays of aortic and left ventricular pressure on the Impella control console allow one to distinguish diastolic from continuous suction events. Diastolic suction events are characterized by negative diastolic left ventricular pressure that recovers to baseline by the onset of systole. In contrast continuous suction events have a negative diastolic left ventricular pressure that does not recover with the onset of systole. The distinction between diastolic and continuous suction events illustrates a key point. Suction events may be due to impaired filling of the left ventricle (as in the case of diastolic suction events) but there are other etiologies such as malposition of the catheter (continuous suction event). However, given that suction events may be triggered by causes other than device malposition it is prudent to utilize echocardiography in the differential diagnosis. The most common causes of impaired flow are hypovolemia vs. right failure. The right ventricle may simply not be able to fill the left ventricle, and this is best diagnosed by echocardiography. The treatment of hypovolemia is relatively simple but dealing with right heart failure is complex.

Right heart failure may be viewed as the Achilles heel of left ventricular mechanical support and is a problem for up to 50% of patients with mechanical support of the left ventricle. If right ventricular function is inadequate, we propose the following algorithm (Figure 9).

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Edward Bittner, Massachusetts General Hospital and Harvard Medical School, United States

Reviewed by: Maruti Haranal, U N Mehta Institute of Cardiology and Research, India

Jamshid Karimov, Cleveland Clinic, United States