|Home | About | Journals | Submit | Contact Us | Français|
Right ventricular (RV) failure from increased pulmonary vascular loading is a major cause of morbidity and mortality, yet its modulation by disease remains poorly understood. We tested the hypotheses that, unlike the systemic circulation, pulmonary vascular resistance (RPA) and compliance (CPA) are consistently and inversely related regardless of age, pulmonary hypertension (PH), or interstitial fibrosis, and that this relation may be changed by elevated pulmonary capillary wedge pressure (PCWP), augmenting RV pulsatile load.
Several large clinical databases with right heart/pulmonary catheterization data were analyzed to determine the RPA-CPA relationship with PH, pulmonary fibrosis, patient age, and varying PCWP. Patients with suspected or documented PH (n=1009) and normal PCWP displayed a consistent RPA-CPA hyperbolic (inverse) dependence; CPA=0.564/(0.047+RPA), with a near constant resistance-compliance product (RC) (0.48±0.17 seconds). In the same patients, the systemic RC was highly variable. Severe pulmonary fibrosis (n=89) did not change the RPA-CPA relation. Increasing patient age led to a very small though statistically significant change in the relation. However, elevation of the PCWP (n=8142) had a larger impact, significantly lowering CPA for any RPA, and negatively correlating RC (p<0.0001).
PH and pulmonary fibrosis do not significantly change the hyperbolic dependence between RPA and CPA, while patient age has only minimal affects. This fixed relationship helps explain the difficulty of reducing total RV afterload by therapies that have modest impact on mean RPA. Higher PCWP appears to enhance net RV afterload by elevating pulsatile, relative to resistive load, and may contribute to RV dysfunction.
When the left heart ejects blood into the systemic arteries, it must overcome both a mean resistive load, regulated by small peripheral vessels, and a pulsatile load related mostly to proximal aortic compliance. This geographic distribution of different vessels dominating resistive versus capacitive properties is important to understanding how left heart load varies with aging and disease. With aging, aortic stiffening and faster flow transmission to the periphery (enhancing wave reflection) results in systolic hypertension with less change in resistance.1–3 Increases in resistance with resulting elevation of mean pressure also reduces overall arterial compliance (e.g. essential hypertension in a young patient), although not as much as with primary stiffening of the thoracic aorta.
The pulmonary circulation is very different, though this disparity has only recently been highlighted, as pulmonary hypertension (PH) and right heart disease are attracting more attention. In a series of relatively small studies involving patients with and without PH, the Vonk-Noordegraaf laboratory showed that mean pulmonary vascular resistance (RPA) and pulmonary arterial compliance (CPA) are consistently inversely related.4–6 More recently, Bonderman et al reported similar findings in patients after pulmonary endartectomy.7 As a consequence, the product of resistance and compliance (RC time) is nearly constant and CPA can be predicted by knowing RPA. This means that the vessels responsible for pulmonary resistance and compliance are more or less the same (e.g. distal), and that this geographic distribution is unaltered by PH. The generalizability of these findings to other patient types, including patients with abnormal left ventricular (LV) function, different ages, or under conditions of acute hemodynamic stress is unknown. We performed a retrospective analysis of pulmonary (and systemic) arterial hemodynamics from several large patient populations to robustly test the RPA–CPA dependence, and to determine its sensitivity to pulmonary fibrosis, patient age, and pulmonary venous capillary wedge pressure (PCWP). Unlike the situation in systemic arteries, we found a hyperbolic RPA-CPA dependence for lung that varies minimally with the cause of PH (excluding World Health Organization (WHO) group II PH), with severe interstitial fibrosis, or with patient age. We also found that PCWP has a significant effect on this dependence resulting in a higher right ventricular (RV) pulsatile load.
For the Johns Hopkins cohort, the Institutional Review Board approved retrospective access to de-identified patient data under a HIPAA waiver. The Mayo Clinic and Columbia University Institutional Review Boards approved the prospective data collection for all individuals included in this analysis and signed consents were obtained. The hemodynamic data required for each patient analysis were mean cardiac output (thermodilution or Fick based), right atrial, mean, systolic, and diastolic pulmonary arterial and/or systemic arterial pressures, and heart rate. In addition, we obtained demographic information and primary cardiac or pulmonary disease diagnosis. Data were examined from four different databases:
To further test the reliability of the PCWP recordings from the clinical database, we randomly selected 50 patients from cohort C and 10 from cohort B, had the tracings reviewed blinded to patient by two cardiologists (RJT, OHC), their values averaged, and then compared to the recorded clinical entry, using Bland-Altman analysis (Supplemental Figure 1). The results show excellent agreement, (95% limits of agreement were −5.9 to 2.6 mmHg, with slight underestimation (−1.7 mmHg) in the clinical recorded database.
RPA is calculated as (mean pulmonary artery (PA) pressure − PCWP)/cardiac output, expressed as mmHg•seconds•mL−1, CPA is estimated by stroke volume/PA pulse pressure, (mL•mmHg−1), as validated by several studies.1, 4, 9 RSA is (mean systemic arterial pressure − mean right atrial pressure)/cardiac output, and CSA = stroke volume/systemic arterial pulse pressure. The RC time (product of resistance and compliance) is therefore expressed as units of seconds.
Data are presented as the mean±SD. Curve fits (linear or non-linear) were generated and statistical analysis was performed using SigmaPlot version 11.0/Systat version 10.2. SigmaPlot uses Marquardt-Levenberg algorithm for curve fits. Comparison of various patient cohorts was performed with either unpaired Student t-test or Mann-Whitney Rank-Sum test as appropriate, or for multiple groups, by one-way ANOVA or ANOVA on ranks. Holm-Sidak or Dunn’s method was used for post-hoc multiple comparisons. Pearson chi-square test was performed for two-way cross-tabulation. An F-test was used to compare variances of pulmonary and systemic RC times. Other analysis, such as multiple linear regressions are indicated where appropriate. A p value <0.05 was considered statistically significant.
Demographic and hemodynamic data for all cohorts are summarized in Table 1. Figure 1A displays the scatter plot of RPA versus CPA from the 1009 patients in Cohort A. There was a consistent inverse dependence fit by the 2nd order hyperbolic decay: RPA =0.564/(0.047 + CPA); r2=0.74. By contrast, a similar plot of systemic arterial RSA versus CSA (Figure 1B, shown with curve fit to PA data) showed greater dispersion (r2 = 0.46). The inverse dependence between RPA and CPA was not dictated by their sharing of stroke volume (SV) in the numerator of one and denominator of the other. Removing SV from CPA (i.e. 1/pulmonary artery pulse pressure) and CO from RPA (PAmean−PCWP) yielded a similar relation (Supplemental Figure 2A). Reintroducing heart rate into the latter also did not alter the relation (not shown). To better quantify the disparity between circulations, RC time was plotted versus mean pressure for each respective vascular bed (Figure 1C). The RC time was narrowly constrained in the pulmonary system (mean 0.48±0.17 sec), but highly variable in the systemic arteries for any mean pressure (variance = 0.027 vs. 0.163; p<10−5). Lastly, we tested whether the RPA-CPA dependence was changed by lung interstitial stiffness in cohort (B) comprised of patients with severe pulmonary fibrosis. A near identical dependence was observed (Figure 1D) with the mean RC time (0.48±0.16 sec, Supplemental Figure 2B).
We next evenly divided the 1009 SPH/PH subjects into age tertiles to test the impact of patient age. Figure 2A highlights the very different effects of patient age on the pulmonary versus systemic vasculature. In the pulmonary plot, the distribution of patients in each age tertile was scattered throughout the hyperbola, whereas the data were clustered for the systemic circulation, with the oldest tertile dominating the lower right region, and youngest the upper left. Figure 2B depicts this in graphic form, showing the percentage of patients within each age tertile that lay within each compliance or resistance tertile for the two circulations. In the systemic circulation, there was a marked age-dependent shift from higher to lower compliance, and lower to higher resistance (both p<0.00001; p-values are for χ2 2-way cross-tabulation). The distribution was more even among age groups for the pulmonary data, and older patients were more prevalent in the high compliance tertile. As a result, the impact of patient age on the pulmonary RPA-CPA relation was small, though statistically significant. This was formally tested by log-transformation of each curve, and subsequent analysis of covariance, using age as a categorical (and continuous) variable (p<0.001; Supplemental Figure 3). For a median RPA of 3 Wood units, CPA fell from 2.64→2.15 mL/mmHg as age rose from 20→90, or a 19% decline over 70 years.
Figure 3A displays the RPA-CPA relationship from the 8,142 patients in Cohort C. The curve fit based on the PH/SPH population is reproduced for comparison. Unlike Cohorts A and B, in whom PCWP was in the normal range, the broader patient group had many subjects with reduced CPA for a given RPA. This change depended on PCWP, as those patients with a PCWP ≤10 mmHg (black) lay on the previously derived curve, whereas those with PCWP≥20 mmHg had a disproportionate decline in CPA (red). We again confirmed that this change was not driven by SV (Supplemental Figure 4A). Converting the data to a log (RPA)-log (CPA) plot (Figure 3B) showed the impact of PCWP was continuous, with a downward-leftward shift in the relation with higher PCWP. The magnitude of the shift was much greater than that observed by age. For example, at a given RPA of 3 Wood unit, CPA is be lowered from 3.34 to 1.65 to 0.82 as PCWP increases from 0 to 25 to 50 mmHg respectively (CPA range of 2.52 ml/mmHg). Increased PCWP therefore also resulted in a lower RC time, due principally to lower compliance (Figure 3C, RC = −0.0063•PCWP + 0.46, r2=0.98, p<0.002). This is also displayed in RC versus mean pressure plots (Supplemental Figure 4B). The magnitude of RC decline was much greater than that observed with patient age.
To further test the impact of PCWP on the RPA-CPA dependence, we examined an subgroup of Cohort C (n=207) who had a diagnosis of heart failure and two RHC’s at different time points; one when PCWP was ≤ 10 mmHg and another when PCWP ≥ 20 mmHg. We also examined 24 patients (Cohort D) with early-stage HFpEF, meaning a PCWP ≤15 mmHg at rest, but in whom PCWP rose ≥ 25mmHg during supine exercise. As shown in Figure 4A and B, an elevated PCWP in the same individual at different time points or with exercise shifted the RPA-CPA relation downward to the left, reducing the RC time constant (0.43±0.15 to 0.28±0.12 seconds; 0.43±0.17 to 0.26±0.10 seconds); both p<0.001). The shift in both curves was statistically significant (Supplemental Figure 5A, B). Thus, unlike mechanisms involved with PH, both acute and chronic elevation of PCWP may enhance the pulsatile relative to resistive load on the right heart.
After nearly half a century, there has been broad acceptance that the systemic arterial circulation combines a resistive and capacitive load, and that these can vary at least partially independent of one another. The pulmonary vasculature is very different, coupling resistance and compliance in a very constrained manner so long as pulmonary venous pressure is low. Increasing PCWP appears to alter this behavior to augment right heart pulsatile load. The current study establishes key properties for right heart-pulmonary vascular coupling and illustrates a previously unappreciated deleterious impact of left heart pressures on pulsatile RV load. This is important, as right ventricular dysfunction is a major independent predictor of death from cardiac and/or pulmonary vascular disease.10–13
It has previously been shown that the RPA-CPA relation does not change with treatment of PH.5 The consistency and shape of the RPA-CPA relation, which our study confirms in large patient groups with normal-range PCWP, have important clinical implications. The relationship’s overall predictability means that a simple set of RHC data defines a given patient’s position on a shared continuum curve, allowing one to possibly predict a therapeutic target. Based on this relation, clinicians may be able to estimate how much RPA must be lowered to have any meaningful change in CPA, and thus, pulsatile (and net) afterload. It also indicates that unlike the proximal aorta in the systemic circulation, the main pulmonary artery adds relatively little to overall pulmonary vascular compliance, since if the PA stiffened independent of resistance with PH, the RC time would decline. This is further supported by work of Saouti et al6, who determined that proximal pulmonary arteries contribute only 19% to overall compliance and that, unlike systemic arteries, pulmonary vascular compliance is distributed evenly throughout the peripheral lung in conjunction with resistance. Small age-dependent changes in CPA also support this notion and are consistent with little rise in pulse wave velocity from the main PA to peripheral lung with aging.14
While some change in main PA distensibility can occur with disease or age, this is small and has less impact on RV load than what is determined by the peripheral vessels. The flatness of the curve at elevated RPA means that resistance must decline substantially to meaningfully impact net RV loading, since pulmonary compliance would still be quite low. This was first suggested by Lankhaar et al4, and may underscore why hemodynamic measurements including RPA have been generally unreliable endpoints for clinical PH management.15 One can appreciate this problem by plotting the average pre- and post- treatment RPA from three large therapeutic PH trials involving sildenafil, treprostinil, or prostacyclin (Figure 5).16–18 A high baseline RPA (0.62,0.77, or 0.96 mmHg*S*mL−1 respectively) and modest decline with treatment (0.50, 0.71, 0.63 mmHg*S*mL−1, respectively) would mean little to no change in estimated compliance, thus maintaining a high RV pulsatile load. While these therapies are clinically used, one would anticipate more effective treatment would need to reduce RPA further or differentially enhance CPA to also impact pulsatile load.
The RPA-CPA relation’s sensitivity to pulmonary venous pressure introduces a new way of considering the hemodynamic consequences of elevated left-sided filling pressures. We initially considered that a high PCWP might impact parenchymal stiffness, with the lung acting as more of a wet than dry sponge. Distensibility of small vessels in the lung tissue is enhanced when surrounded by compressible air, but this would be diminished if the parenchyma stiffened. One counter to this theory is the data showing severe pulmonary fibrosis does not generate the same effect. An alternative is that PCWP is the downstream pressure that amplifies a peripheral pulse reflection, thereby augmenting systolic pulmonary arterial pressure (PAP) and leading to a decline in total compliance.
The impact of PCWP on pulmonary arterial and thus right heart loading is likely relevant to clinical symptoms in heart failure patients. Such individuals become dyspneic during exercise when PCWP frequently rises. Lewis et al19 recently showed that symptoms and clinical outcomes of patients with LV dysfunction correlate better with augmentation in PAP than the change in PCWP. At first glance, this may seem in contrast to the findings in this study. However, we would suggest this observation could be explained by the effect of PCWP on the RPA-CPA relationship. Elevations in PCWP lowers CPA for a given RPA, resulting in enhanced pulmonary arterial wave reflections and augmentation of the systolic PAP, and thus, mean PAP. Therefore, the higher proportional rise in mean PAP compared to PCWP could be explained by the indirect effect of rising PCWP lowering CPA. Our findings also indicate that with the rise in PCWP, there will be enhanced pulsatile RV load, which would further limit RV ejection, and in turn LV filling. Indeed, in the patients in whom mean PAP augmentation plateaued during exercise (indicating RV dysfunction), prognosis was worse.
Our study has several limitations. We relied on the recorded clinical indication for RHC for identifying patients with suspected PH or known PH, and while the actual diagnosis was more mixed and indeed some did not have PH, the consistency of the RPA-CPA relation despite this further supports the idea of its constancy so long as PCWP is in the normal range. Similarly for all cohorts, we relied on the original operator’s recordings and interpretation of hemodynamic data. It is possible that interpretations of tracings may have been different among individual operators. However, the primary data were all interpreted by heart failure cardiologists or pulmonologists, and our blinded review found minimal inter-observer error on a random subset of studies. Lastly, we relied on an indirect estimate of total pulmonary vascular compliance. Alternative approaches using pulsatile pressure-flow analysis are difficult and impractical for large population studies, but the estimation method has favorably compared to such alternative approaches in smaller controlled studies.1–3,4
In conclusion, the pulmonary circulation and the afterload it imposes on the RV are very different from the systemic circulation. PH, interstitial fibrosis, and patient age do not appear to have much effect on the inverse, hyperbolic RPA-CPA relationship. Because of the consistency of this relationship, one can estimate pulmonary vascular compliance from resistance using a simple equation, identify where a given patient lies relative to normal, and possibly anticipate what therapy would need to achieve for a robust clinical benefit. The findings that a higher PCWP may impact pulmonary arterial and thus RV pulsatile loading offers new insight into the symptomatology of HFpEF and other left heart failure disorders.
Right ventricular (RV) dysfunction is a major independent predictor of death in patients with elevated afterload from pulmonary hypertension (PH). We analyzed data from >1000 right heart catheterizations to show that unlike the systemic circulation, pulmonary vascular resistance (RPA) and compliance (CPA) display a highly predictable, inverse correlation, and that this relationship is not significantly altered by PH or severe pulmonary fibrosis, and only minimally by aging. This means that unlike the proximal aorta in the systemic circulation, the main pulmonary artery contributes relatively little to overall CPA, and is little impacted by age. The consistent RPA-CPA relation allows the prediction of RPA decline required by a given treatment to adequately lower CPA and thus reduce pulsatile and net RV afterload. Current mono-therapies for PH are not reducing pulsatile afterload in most patients – identifying an area of clinical need. We also found that RPA-CPA relationship is sensitive to changes in pulmonary venous pressure (mean pulmonary capillary wedge pressure (PCWP)), as elevation of this pressure lowers CPA for any given RPA, augmenting RV pulsatile afterload. In patients with exertional dyspnea and a preserved ejection fraction who exhibit a marked rise in PCWP during exercise, we found a disproportionate decline in CPA and thus increased pulsatile load. This identifies a novel mechanism whereby left side diastolic dysfunction contributes to RV load.
Funding Sources: Funding was provided by NIH-NHLBI Grants 5P50HL084946, K23 HL086714, T32-HL-07227, KL2-RR024156, Fondation Leducq and Peter Belfer Laboratory, The Robert Wood Johnson Physician Faculty Scholars Program, and the Herbert and Florence Irving Scholar Award.
Conflict of Interest Disclosures: Dr. Lederer is a consultant for Gilead and on the clinical trial steering committee for Intermune.