Right Ventricular-Pulmonary Vascular Interactions

Supply and Demand: A Framework to Assess Right Ventricular-Pulmonary Vascular Interactions

Cardiopulmonary status is determined by the state of the right ventricle (RV), the pulmonary vasculature (PV), and their interactions, assuming adequate respiratory function. The RV-PV unit can be conceptualized as a series of pumps, which supply mechanical energy, and a network of large and small pipes, which simultaneously demand and dissipate this energy. Hemodynamically, RV energy supply is created by RV myocyte contraction, which depends on the preload (stretch), afterload (stress) imposed, and bioenergetics status (e.g., oxygen bioavailability, mitochondrial function, ATP levels, etc.). Moreover, RV pump function depends on left ventricular (LV) pump function; it has been estimated that 20–40% of RV energy supply (systolic pressure and volume outflow) is due to LV ejection in a healthy state (19, 30, 69). Hemodynamic energy demand in this framework is the cardiac output, largely determined by the LV and systemic vasculature, and transformed by the pulmonary vasculature. In particular, pulmonary vascular resistance and stiffness (inverse compliance) create a mean and pulse pressure demand from the cardiac output (flow demand); here, we denote the combination as PV demand (FIGURE 1).

Given this framework, methods to simultaneously assess RV supply and PV demand are essential to evaluation of RV-PV interactions. One such approach was proposed by Sagawa et al. for left ventricular-systemic vascular interactions based on pressure-volume (P-V) loops (FIGURE 2) (68). In P-V analysis, the ventricle is conceptualized as a sac that has a time-varying stiffness or elastance due to myocardial contraction and relaxation. Beginning at the start of isovolumic contraction, the elastance increases from an initial minimum to a peak value (E_max), and then returns to the minimum value again. The instantaneous elastance E(t) is defined mathematically as E(t) = P(t)/[V(t) − V₀], where P(t) is the instantaneous pressure, V(t) is the instantaneous volume, and V₀ is the volume intercept of the line connecting the P-V points at the same phase of the cardiac cycle at two (or more) different preload conditions (17, 48, 67, 68, 95) (FIGURE 2). Although V₀ can change with time (68), it is often assumed constant and equal to the value obtained at end systole (FIGURE 2).

Because it can be difficult to identify E_max from P-V curves, it is typically assumed equal to the end-systolic elastance (E_es) determined from a linear approximation to the end-systolic P-V relationship (ESPVR) (FIGURE 2). Although this approach is recommended by most experts (90), some accuracy is lost when ESPVR is assumed to be linear, since it is inherently curvilinear and typically convex to the pressure axis (11, 34). E_es is a commonly used load-independent measure of contractility that can be derived from P-V loops, but it is affected by changes in heart rate (52), reduction in coronary perfusion pressure (74), and ionotropic state (48).

When E_es is used to define hemodynamic energy supply, a complementary elastance of the vasculature is used to compute hemodynamic energy demand: the effective arterial elastance (E_a) is the slope of the line that connects the ventricular end-systolic point to the ventricular end-diastolic volume projected on the volume axis (FIGURE 2). E_a is a composite measure that is dependent on vascular resistance and compliance. It has the advantage (and disadvantages) of being a single metric of ventricular afterload. One disadvantage is that, like E_es, it is dependent on heart rate (73, 75).

The ratio of ventricular to vascular elastance (E_es/E_a), yields insight into the balance between hemodynamic energy supply and demand in the ventricular-vascular unit. It also provides a quantitative assessment of the adequacy of ventricular-vascular coupling (VVC). Experimental and modeling studies have demonstrated that, in the healthy beating heart, Ε_es/Ε_a = 1.5–2.0. Despite these healthy values of >1, E_es/E_a is often referred to as ventricular-vascular coupling efficiency. In a mechanical system, efficiency is defined as the ratio of output power to input power; transmission of power is maximized when the output impedance of the power-producing part and the input impedance of the power-receiving part are equal such that the efficiency of the system is equal to one. When energy dissipation occurs, efficiency is <1. Calculating an equivalent efficiency of the ventricular-vascular unit would require measurement of input power or energy (e.g., oxygen consumption), and output power or energy (e.g., stroke work), which is challenging and thus infrequently done (10, 41).

The application of P-V loop analysis to the RV-PV unit was validated by Maughan and collaborators (48) with studies in isolated canine hearts. Clinical application of RV P-V loop analysis, often with significant approximations (as described below), has shown that lower values of E_es/E_a predict mortality in PAH (3, 8, 32, 44, 59). Higher values of E_es/E_a have been interpreted to mean that the RV supply is properly coupled to the demand of the pulmonary circulation (55) or that the RV-PV unit is operating with minimal energy cost (23, 66). However, some limitations in the application of this approach to RV-PV interactions exist. First, although E at end systole (E_es) is a good estimate of E_max in the LV, this is less true in the RV because of non-coincidence of end-ejection and end-systole (48). Second, RV P-V loops are frequently triangular in shape instead of square, which makes identification of the end-systolic point more difficult than in the LV (48, 62). Moreover, with increasing pulmonary vascular resistance due to PAH, the shape of RV P-V loops change: they become more square-shaped in mild PAH and then more trapezoidal in severe PAH (7, 53), which affects the calculation of E_max. Finally, V₀ varies more throughout the cardiac cycle in the RV than in the LV, such that the errors induced by assuming a constant V₀ are larger (48).

Approaches to examine RV-PV interactions other than P-V loop analysis do exist. One is pump function curves, which are traditionally built from measurements of the mean RV pressure that serves as a surrogate for PV demand, and stroke volume (SV) that serves as a surrogate for RV supply (86). There have been some limited pump function studies assessing RV function in PAH (58). However, the limitations of this approach, which include its sensitivity to changes in preload and neglect of pulsatile work components, have prevented it from being widely adopted in either clinical or preclinical studies.

A second alternative approach bases its estimation of PV demand on pulse-wave velocity and wave transit time in the heart, which are calculated through pulmonary vascular impedance (PVZ). Unlike pulmonary vascular resistance (PVR), which measures the opposition to steady flow, primarily determined by small-vessel resistance and a critical diagnostic criterion for PAH, PVZ measures the total opposition to flow through analysis of instantaneous arterial pressure and flow waves. PVZ provides the most comprehensive description of RV afterload, but its measurement is technically demanding (requiring simultaneously measured pressure and flow), and its computation is not standardized (to either frequency or time-domain analysis) (33, 80). Furthermore, PVZ does not enable straightforward assessment of PV demand coupling to RV supply.

Application to Clinical Studies

Two main factors limit the assessment of RV-PV interactions via P-V loop analysis in the clinical setting. First, simultaneous pressure and volume measurements in the RV are difficult to obtain. Recently, an FDA P-V catheter approved for human use has been employed in clinical studies on the left ventricle (21, 60), but the non-cylindrical shape of the RV limits accuracy, especially at small volumes. Second, although it is possible to vary preload through inferior vena cava occlusions in animals, such maneuvers are not feasible in humans due to the increased risks of additional procedures and hemodynamic compromise. To overcome this limitation, both Hsu et al. and Tedford et al. had subjects perform Valsalva maneuvers to increase intrathoracic pressure during invasive right heart catheterization (RHC) with a P-V catheter, which successfully decreased preload in patients with idiopathic PAH and systemic sclerosis-associated PAH (32, 81). Nevertheless, the majority of clinical studies use approximations of P-V analysis with the methods summarized below.

Acute Increases in PV Demand

An acute increase in RV supply in response to an acute increase in PV demand is a key component of adaptation to physiological stress, such as exercise, or pathological stress, such as pulmonary embolism. Preservation of RV-PV coupling in response to acute stressors has been investigated using acute hypoxia (20, 40, 63, 64) and acute thromboembolism (23, 36, 37, 47) (Table 1). In addition, the response of the RV to acute ischemia (50) and acute lung injury due to endotoxic shock have also been evaluated (45, 46). These studies provide key clinically relevant insights into RV function and adaptation in the setting of multiple models of increased demand, from mild to severe, that are summarized in Table 1.

Acute models of PAH also provide a time-efficient way to investigate the effect of drug therapies on RV-PV coupling (35, 37–40, 47, 50, 63, 64). As detailed in Table 2, various drugs have been investigated in large-animal models of acute pulmonary hypertension, including those that target the pulmonary vasculature (35, 63, 64), those that target the RV (37–39, 47, 50), and those that affect both (40). However, these studies typically assess only acute effects (10−120 min), thereby necessitating follow-up studies to determine whether any beneficial effects are sustained.

Insights Gained from Clinical Studies

Enabled by the novel technologies and the approximation methods described above, evaluation of VVC has provided important insights into the understanding of cardiopulmonary function, especially in PAH, in humans.

Kuehne et al. found that VVC_sb from non-simultaneous MRI and RHC was reduced in PAH subjects compared with controls, despite increased RV supply (44). McCabe found the same result in patients with chronic thromboembolic pulmonary hypertension compared with controls using conductance catheter-based VVC_sb measurements (49). In a retrospective analysis of 134 patients who underwent non-simultaneous RHC and MRI as work-up for PAH, Sanz et al. found that PAH subjects had significant elevations in E_a and E_es,p compared with non-disease controls, as well as significant uncoupling as determined by decreased VVC_p (70).

Brewis et al. demonstrated a significant correlation between higher VVC_v and improved survival in patients with PAH. Furthermore, patients in this study who had stable or improved VVC_v in response to therapy demonstrated improved survival (8). However, as noted above, given the mathematical links between VVC_v and EF, it is unclear how VVC_v adds value to prognoses.

Multiple studies have evaluated the correlation between VVC_sb, VVC_p, and VVC_v. These invasive and non-invasive estimates of VVC have been found to be highly correlative with R values ranging from 0.79 to 0.93 (70, 83, 84). Two studies with a total of over 150 patients between them demonstrated that VVC_v correlates with survival, whereas VVC_p does not (8, 84). Additional studies are required to address the question of whether VVC, quantified without additional simplifications, is a better predictor of mortality than VVC_sb, VVC_p, or VVC_v (or RV EF). If time-resolution limitations of MRI can be addressed, MRI catheterization with preload reduction by Valsalva could be used to conduct these studies. Otherwise, comprehensive hemodynamic studies in large-animal models as described below will be required.

Physiological Insights from Large-Animal Studies

With appropriately chosen species, large-animal models have high fidelity to human cardiopulmonary physiology and pathology. Wauthy et al. evaluated the response to acute PAH caused by pulmonary arterial banding (PAB), hypoxia, or thromboembolization in dogs, goats, and minipigs (89). Significant interspecies differences were found, with mPAP, E_a, and E_es at baseline increasing from dogs to goats to minipigs. Under baseline conditions, VVC was high in all species, and all species demonstrated similar responses to acute hypoxia, with progressive increases in both E_a and E_es, and no significant change in their ratio. Consistent with these findings, both dog and goats were demonstrated to have similar responses to both proximal PAB and acute thromboembolism (89), which suggests that both baseline RV-PV interactions and the adaptive response of the RV-PV unit to acute stress is conserved across species.

Several large-animal models of chronic PAH have been described, including monocrotaline (MCT)-induced PAH in dogs (14, 15, 29, 51), progressive pulmonary artery banding in pigs (2), pulmonary venous banding to induce postcapillary pulmonary hypertension in pigs (61), and a combination of chronic thromboembolism and proximal PA coiling in pigs (2). These studies demonstrate the feasibility of creating large-animal models of chronic pulmonary hypertension with the expected RV hypertrophy and decreased RV EF. However, RV-PV interactions were not quantified in these studies. Guihaire created a swine model of chronic PAH that combined left PA ligation and sequential embolization of the right lower pulmonary lobe (Table 1) (27, 28, 57). Chronic RV pressure overload was demonstrated by significantly increased mPAP after 6 wk, which was accompanied by increased E_a and E_es. Despite the increase in E_es, it was not sufficient to compensate for increased E_a, and consequently VVC decreased until perfusion was surgically restored to the left lung (27, 28). This chronic large-animal model provides useful insights into the adaptation of cardiopulmonary function to chronic PAH. However, it is resource intensive, requiring multiple surgeries as well as serial fluoroscopic-guided right heart catheterization for directed embolization.

Continued development of large-animal chronic PAH models with high fidelity to human disease and evaluation of these models with P-V loop analysis should provide insight into disease pathophysiology as well as for a platform for testing the safety and efficacy of novel therapies.

Mechanistic Insights from Small-Animal Studies

Although the hemodynamic changes that result from acute and chronic PAH in small animals may be less similar to the clinical situation than those of large animals, small-animal, especially mouse, studies enable use of sophisticated genetic tools that allow for evaluation of the molecular drivers of RV-PV uncoupling. Critical to the study of PAH in small-animal models is the ability to measure RV supply and PV demand.

Our group was the first to describe the use of P-V loop analysis to evaluate RV-PV function in rodents in 2010, adopting technologies and methodology previously used to evaluate left ventricular function. This study demonstrated the feasibility and utility of using an admittance catheter to measure P-V loops in anesthetized, open-chested mice before and during vena cava occlusion to alter preload, permitting the determination of E_es, E_a, and VVC as well as measurement of a multitude of metrics of both diastolic and systolic RV function. This initial study was done in a mouse model of chronic hypoxic pulmonary hypertension, which has the limitation that the PAH is reversible and often mild (78) (FIGURE 4). Subsequently, we and others have used P-V loop analysis to characterize development of RV dysfunction in other rodent models of PAH (3, 16, 24, 25, 66, 87, 94).

A subsequent study from our group by Wang et al. examined the development of RV dysfunction and RV-PV uncoupling over time in a progressive model of hypoxia and SU5416 treatment (HySu exposure), a model of severe PAH that recapitulates the plexiform pulmonary vascular lesions seen in human PAH in rodents (1) (FIGURE 4); however, SU5416 may directly impact RV function since decreased angiogenesis and capillary density are associated with RV failure (6, 77). Both RV systolic pressure and E_a increased throughout the 28-day HySu exposure, whereas E_es only increased with HySu exposure up to 21 days, with no further increases seen between 21 and 28 days (with cardiac output increasing up to 14 days and then declining). Corresponding to these findings, VVC was initially preserved and even increased at 14 days of HySu exposure and then decreased consistent with uncoupling by 28 days. Through the rigorous investigation of RV function at multiple time points, this study demonstrates initial adaptation of the RV and subsequent maladaptation and failure after prolonged exposure to increased afterload (87).

Alaa et al. utilized P-V loop analysis to differentiate between PAH with preserved cardiac index (CI) and PAH with reduced CI in a MCT-induced model of PAH in rats (3). MCT has the advantages of causing severe PAH that recapitulates some of the pathological hallmarks of the disease in humans, such as distal smooth muscle hypertrophy, obliteration of small pulmonary arteriole lumens, and plexiform lesions (22, 91): however, it has the disadvantage of causing cardiac inflammation that may impair RV supply independent of increased RV demand (92). As expected, MCT-induced PAH led to increased RV systolic pressure, increased E_a, and increased E_es with decreased VVC. Interestingly, compared with PAH with preserved CI, PAH with reduced CI was associated with significantly higher E_a and E_es, but no significant difference in VVC. Moreover, although all MCT-treated animals demonstrated RV dilation, there was no difference in RV dilation between the preserved and reduced CI groups. This study, in which the inability to preserve CI was ultimately related to diastolic dysfunction (3), demonstrates that even the relatively comprehensive assessment of the cardiopulmonary unit afforded by P-V analysis can miss critical pathophysiological changes if only the standard metrics E_a, E_es, and VVC are evaluated.

As with large-animal studies, P-V analysis has been used to evaluate therapeutic interventions in rodent studies. Rungatscher et al. demonstrated S-nitroso human serum albumin (S-NO-HSA) treatment resulted in improvement in E_a, E_es, and VVC in a rat model of RV volume overload (66). Zeineh et al. showed a possible positive inotropic effect of iloprost, a prostacyclin, which resulted in increased VVC in a MCT rat model of PAH (94). De Man et al. demonstrated that β-blockade with bisoprolol delayed progression to RV failure and improved VVC in a MCT rat model of PAH (16). Despite promising results in small-animal studies, few of these findings are translatable into human therapies due the current limitations in large-animal studies for PAH and the fact that often therapies effective for small animals are not effective in large-animal models or human trials.

With the establishment of P-V loop analysis as a robust tool evaluating small-animal models of PAH, the impact of genetic modifications on RV-PV coupling has been investigated. Golob et al. demonstrated a benefit of genetically impaired collagen turnover on RV-PV coupling in a HySu model of PAH (25) and no effect from mitochondrial DNA mutation on RV-PV interactions (24) using P-V analysis. These studies provide important insights; however, more investigation is needed to understand the underlying molecular mechanisms that regulate RV function and RV-PV interactions.

P-V loop analysis in small-animal models provides a unique opportunity to study and understand the effects of molecular and cellular events leading to RV-PV uncoupling and RV failure. Future studies should help determine how modifications to cellular and extracellular structure impact RV-PV interactions, while providing a better understanding of the signaling pathways responsible for the transition from maintained RV function in the presence of increased afterload to RV failure.

Conclusions

The optimum management of PAH remains elusive. Evaluating the adequacy of hemodynamic interactions between the PV and RV through ventricular-vascular coupling is a powerful tool that should help in understanding disease development and progression as well as in the development, translation, and monitoring of novel therapies. Recommended areas of future work include:

• 1)

  Long-term studies on prognostic relevance of RV-PV interactions during both rest and exercise for PAH patients

• 2)

  Development of large-animal models that recapitulate pulmonary vascular disease phenotypes and enable validation of pressure-only and volume-only approximations of VVC in healthy and diseased states

• 3)

  Consideration of RV-PV interactions in mechanistic studies to account for drivers of both RV supply and PV demand.