2017 Volume 81 Issue 12 Pages 1871-1878
Background: Advanced left heart failure (HF) often accompanies post-capillary pulmonary hypertension related to RV afterload. Although pulmonary arterial capacitance (PAC), a measure of pulmonary artery compliance, reflects right ventricular (RV) afterload, the clinical utility of PAC obtained by echocardiography (echo-PAC) is not well established in advanced HF.
Methods and Results: We performed right heart catheterization in advanced HF patients (n=30), calculating echo-PAC as stroke volume/(tricuspid regurgitation pressure gradient-pulmonary regurgitation pressure gradient). The difference between the echo-PAC and catheter-measured PAC values was insignificant (0.21±0.17 mL/mmHg, P=0.23). Echo-PAC values predicted both pulmonary arterial wedge pressure (PAWP) ≥18 mmHg and pulmonary vascular resistance ≥3 Wood units (P=0.02, area under the curve: 0.88, cutoff value: 1.94 mL/mmHg). Next, we conducted an outcome study with advanced HF patients (n=72). Patients with echo-PAC <1.94 mL/mmHg had more advanced New York Heart Association functional class, higher B-type natriuretic peptide plasma levels, larger RV and lower RV fractional area change than those with echo-PAC ≥1.94 mL/mmHg. They also had a significantly higher rate of ventricular assist device implantation or death, even after adjustment for indices related to HF severity or RV function during a 1-year follow-up period (P<0.01).
Conclusions: Decreased PAC as measured by echocardiography, indicating elevated PAWP and RV dysfunction, predicted poorer outcomes in patients with advanced HF.
Left-sided heart failure (HF), especially at an advanced stage, frequently accompanies post-capillary pulmonary hypertension (PH)1–3 through elevated left ventricular (LV) filling pressures or increased pulmonary venous pressure. PH induces remodeling of the pulmonary arteries and increases right ventricular (RV) afterload.4,5 Persistent increased RV afterload causes dilatation of the thinner RV wall, leading to dysfunction6,7 and has been reported to be related to a poor prognosis in patients with left-sided HF.8
Pulmonary arterial capacitance (PAC), a measure of the compliance of the pulmonary artery and 1 of the indices of RV afterload, can be calculated at right heart catheterization (RHC) as the ratio of stroke volume (SV) to the pulmonary pulse pressure (PP).9–12 PAC calculated by RHC (PAC-RHC) shows a hyperbolic relationship with pulmonary vascular resistance (PVR), another indicator of RV afterload, and more closely correlates with the prognosis of HF patients than PVR.10 SV can be noninvasively estimated with Doppler echocardiographic indices. PAC can also be estimated echocardiographically, and has been reported to be a prognostic indicator in patients with pulmonary arterial hypertension.13 However, because the utility of noninvasive PAC measurements in HF has not been established, we investigated PAC estimated by echocardiography (echo-PAC) in advanced HF in 2 studies: a hemodynamics study and an outcome study.
This study was approved by the Ethics Committee of Osaka University Hospital.
For the hemodynamic study, patients with left-sided HF who had undergone both RHC and echocardiography on the same day in Osaka University Hospital between August 2010 and September 2014 were retrospectively assessed. RHC was performed for HF management or assessment for heart transplant eligibility. Inclusion criteria were (1) LV ejection fraction (LVEF) ≤35%; and (2) adequate tricuspid regurgitation (TR) jet and pulmonary regurgitation (PR) jet on Doppler echocardiography. A total of 30 patients were enrolled.
For the outcome study, patients who had been admitted with symptomatic left-sided HF to the same hospital between January 2010 and July 2013 were retrospectively analyzed. HF was defined based on current guidelines.14 Inclusion criteria were (1) LVEF ≤35%, (2) New York Heart Association (NYHA) functional class II–IV, and (3) age ≤65 years because in Japan patients over 65 years old are not eligible for LV assist device (LVAD). A total of 108 HF patients were reviewed. Patients with congenital heart disease (n=1), severe mitral regurgitation (MR) or aortic regurgitation caused by prolapse (n=2), previously implanted VAD (n=6), percutaneous cardiopulmonary support (n=1), low-quality images for PAC calculation, many without PR or TR (n=25), or loss to follow-up (n=1) were excluded. Patients with MR or TR caused by annular dilatation or tethering were included in this study. Finally, 4 patients in the hemodynamic study were also included in the outcome study.
Clinical data, including patient characteristics, laboratory data, HF etiology, and medications, were collected at the time of echocardiography.
RHC was performed in a cardiac catheterization laboratory using a Swan-Ganz catheter. Pressure measurements were obtained at end-expiration while the patient was supine. The following hemodynamic parameters were recorded: pulmonary artery wedge pressure (PAWP), pulmonary artery systolic pressure (PASP), mean pulmonary artery pressure (mPAP), pulmonary artery diastolic pressure (PADP), mean right atrial pressure (mRAP), and cardiac output (CO). CO was measured using the assumed Fick equation. RV stroke work index (RVSWI), an index of RV systolic function, was determined as follows: (mPAP−mRAP)×SV index.15 The transpulmonary pressure gradient (TPG) was defined as the difference between mPAP and PAWP. The diastolic pulmonary vascular pressure gradient (DPG), which is reported to be an indicator of pulmonary vascular remodeling,16 was calculated as the difference between PADP and PAWP. PP was calculated as the difference between PASP and PADP. PVR was defined as TPG/CO. PAC calculated by RHC was defined as SV, which is obtained from CO/PP.9
Echocardiography was performed with the patient in the supine position using a Vivid 7 cardiovascular ultrasound system and a Vivid E9 cardiovascular ultrasound system (GE Healthcare, Milwaukee, WI, USA), operated by experienced sonographers who were blinded to patient data.17 Data analysis was performed using Echo-PAC Clinical Workstation Software (GE Healthcare, Milwaukee, Wisconsin). Echocardiographic measurements were obtained according to the guidelines of American Society of Echocardiography.18 Briefly, LV diastolic diameter (LVDd), LV systolic diameter (LVDs) and left atrial diameter (LAD) were measured on the parasternal long-axis view. LVEF was measured by the modified Simpson’s method in the apical 4- and 2-chamber views. Transmitral flow velocity curves were recorded to measure the peak early diastolic (E) and late diastolic (A) velocities. Tissue Doppler imaging at the mitral annulus level was obtained in the septal position in order to measure the early diastolic (E’) and late diastolic (A’) myocardial velocities, as previously described.19 MR and TR were graded on a 4-point scale based on color-flow Doppler in the apical 4-chamber view. RV end-diastolic diameter (RVDd) was measured at the midventricular level of the RV in end-diastole. Tricuspid annular plane systolic excursion (TAPSE) and RV fractional area change (RVFAC) were measured in a right ventricle-focused apical 4-chamber view. RVFAC was calculated as: (end-diastolic area−end-systolic area)/end-diastolic area. RV area was measured by tracing the RV endocardium. Echocardiographic SV was determined as the product of the LV outflow tract cross-sectional area and the LV outflow time velocity integral from pulsed wave Doppler. Echo-PAC was calculated as: SV/(TR pressure gradient [TRPG]−PR pressure gradient [PRPG]), as reported by Mahapatra et al.13 Intra- and interobserver variability of measurements for echo-PAC were analyzed in 10 randomly selected patients. In patients with a past history of mitral valve surgery, we excluded the tissue Doppler parameter and TAPSE for reliability reasons.20
The endpoint was death or the need for implantation of a LVAD. Patients were followed from the date of echocardiography for 1 year or until the endpoint, whichever was sooner. Follow-up data were collected in a blinded fashion via review of all available medical records.
Continuous data are expressed as mean±SD if symmetrically distributed or as the median (interquartile range) if asymmetrically distributed. Comparisons were performed with Student’s t-test or the Wilcoxon rank sum test. Categorical data are expressed as numbers and percentages, and compared with the chi-squared test. Natural log transformation (ln) was used to satisfy some model assumptions. The correlation between PAC obtained by echocardiography and by catheterization was assessed by Bland-Altman analysis. Linear regression analysis with a Spearman rank-order correlation coefficient was used to assess the correlations of echo-PAC with clinical characteristics, including RHC variables (PVR, PAWP, mPAP, RAP, DPG and RVSWI). Receiver-operating characteristic (ROC) curves were generated, and the area under the curve (AUC) was determined as a measure of the ability to detect elevated PVR (≥3 Wood units [WU]) and/or PAWP (≥18 mmHg), which are frequently used cutoff values in assessing HF severity, at any cutoff value. Event-free survival was estimated by the Kaplan-Meier method and group differences were compared using the log-rank test. Univariate analysis based on the Cox proportional hazards model was used to ascertain any effect on cardiac events (death or LVAD implantation) from each factor, and hazard ratios (HR) with 95% conﬁdence intervals (CI) were calculated. A multivariate model was used to adjust for the effect of variables associated with severity of HF (NYHA ≥3, ln [B-type natriuretic peptide (BNP)], LVEF), RV dysfunction (RVFAC, RVDd, total bilirubin) and PASP (TRPG). A 2-tailed P-value <0.05 was considered statistically significant. Statistical analyses were performed with JMP, version 11.0 (SAS Institute, Cary, NC, USA) and SPSS for Windows, version 20.0 (SPSS Inc., Chicago, IL, USA).
Table 1 shows the characteristics of the patients in the hemodynamic study. Reduced LVEF (mean 21±7%) with enlarged LVDd (72±10 mm), low systolic blood pressure (99±19 mmHg) and cardiac index (1.7±0.4 L/min/m2), and elevated plasma levels of BNP (median: 554 pg/mL, 25–75th percentiles: 231–1,009 pg/mL) were characteristic of advanced HF. Bland-Altman analysis showed that the difference between echo-PAC (median: 2.06 mL/mmHg, 25–75th percentiles: 1.16–3.38 mL/mmHg) and RHC-PAC (median: 2.04 mL/mmHg, 25–75th percentiles: 1.38–2.59 mL/mmHg) was not significant (0.21±0.17 mL/mmHg, P=0.23), and that echo-PAC at the higher end of the scale tended to exceed RHC-PAC (Figure 1). Echo-PAC had a hyperbolic association with PVR, and ln [echo-PAC] correlated in a linear fashion with ln [PVR] (Figure 2A, r=−0.45, P=0.01). Ln [echo-PAC] also correlated with ln [PAWP] (r=−0.74, P<0.001, Figure 2B), ln [mPAP] (r=−0.78, P<0.001, Figure 2C), and ln [RAP] (r=−0.65, P<0.001, Figure 2D), whereas it did not correlate with DPG (P=0.66, Figure 2E) or ln [RVSWI] (P=0.26, Figure 2F). Although many hemodynamic parameters correlated with each other in these subjects, there was no significant correlation between ln [PVR] and ln [PAWP] (P=0.61). In the multiple linear regression analysis, ln [echo-PAC] independently correlated with ln [PVR] and ln [PAWP] (ln [PVR]: ß=−0.38, P=0.001, ln [PAWP]: ß=−0.71, P<0.001). In the patients with high PVR (≥3 WU) and high PAWP (≥18 mmHg), echo-PAC and RHC-PAC were significantly lower than in those with low PVR (<3 WU) and low PAWP (<18 mmHg) (P<0.01, P<0.01, respectively, Figure 3). ROC analysis revealed that echo-PAC could predict the patients with both PVR ≥3 WU and PAWP ≥18 mmHg (P=0.02, AUC: 0.88, cutoff value: 1.94 mL/mmHg, sensitivity: 1.00, specificity: 0.73).
Values are expressed as mean±SD, % (n), or median (interquartile range). ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin-receptor blocker; BP, blood pressure; BNP, B-type natriuretic peptide; CI, cardiac index; CO, cardiac output; DPG, diastolic pulmonary vascular pressure gradient; eGFR, estimated glomerular filtration rate; LAD, left atrial dimension; LVDd, left ventricular end-diastolic dimension; LVDs, left ventricular end-systolic dimension; LVEF, left ventricular ejection fraction; mPAP, pulmonary arterial mean pressure; MR, mitral regurgitation; NYHA, New York Heart Association; PAWP, pulmonary arterial wedge pressure; PVR, pulmonary vascular resistance; RVSWI, right ventricular stroke work index; TRPG, tricuspid regurgitation peak gradient.
Bland-Altman plot showing the relation between echo-PAC and RHC-PAC. Echo-PAC, pulmonary arterial capacitance obtained by echocardiography; RHC-PAC, pulmonary arterial capacitance obtained by right heart catheterization; SD, standard deviation.
Correlation of ln [echo-PAC] with (A) ln [PVR], (B) ln [PAWP], (C) ln [mPAP], (D) ln [RAP], (E) DPG, and (F) ln [RVSWI]. DPG, diastolic pulmonary vascular pressure gradient; echo-PAC, pulmonary arterial capacitance obtained by echocardiography; ln, natural log transformation; mPAP, mean pulmonary artery pressure; PAWP, pulmonary arterial wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; RVSWI, right ventricular stroke work index.
Distribution of echo-PAC (A) and RHC-PAC (B) in patients stratified by PAWP (≥18 mmHg or <18 mmHg) and PVR (≥3 Wood units [WU], <3 WU). *P<0.05 vs. patients with PVR <3 WU and PAWP <18 mmHg, #P<0.05 vs. patients with PVR ≥3 WU and PAWP <18 mmHg. Abbreviations as in Figures 1,2.
The mean difference between the measurements by the single observer (Y. Saito) was 0.09±0.21 mL/mmHg. Intraclass correlation coefficient (ICC) was 0.90 (P<0.001). The mean difference between the 2 independent observers (Y. Saito and T. Onishi) was 0.28±0.23 mL/mmHg. ICC was 0.87 (P<0.001).
Figure 4 shows the distribution of echo-PAC values in the outcome study. The median (interquartile range) value of echo-PAC was 1.35 mL/mmHg (0.93–2.29 mL/mmHg). Table 2 shows the clinical characteristics of the study subjects who were divided into 2 groups based on the cutoff values of echo-PAC for identifying both PVR ≥3 WU and PAWP ≥18 mmHg; the low echo-PAC group (≤1.94 mL/mmHg) and the high echo-PAC group (>1.94 mL/mmHg). The patients in the low echo-PAC group were in a more advanced NYHA functional class, and had a lower prevalence of atrial fibrillation and higher plasma levels of BNP and total bilirubin than those with high echo-PAC. In the linear regression analysis, ln [BNP] had a weak inverse correlation with echo-PAC (r=−0.25, P=0.03). The patients in the low echo-PAC group had higher mitral E/A ratios, PRPG, and TRPG, and tended to have larger RVDd than those with high echo-PAC. Although TAPSE did not differ between the 2 groups, the continuous value of echo-PAC weakly correlated with that of RVFAC (r=0.28, P=0.02) and RVDd (r=−0.28, P=0.01) in the linear regression analysis.
Cumulative frequency distribution of the pulmonary arterial capacitance values obtained by echocardiography (echo-PAC) in patients in the outcome study.
Values are expressed as mean±SD, % (n), or median (interquartile range). High echo-PAC, >1.94 mL/mmHg; Low echo-PAC, <1.94 mL/mmHg; CRT, cardiac resynchronization therapy; E/A, ratio of peak mitral E wave velocity to peak mitral A wave velocity; E/e’, ratio of peak mitral E wave velocity to peak early diastolic myocardial velocity at septal position recorded by tissue Doppler imaging; IVC, inferior vena cava; RPG, pulmonary regurgitation pressure gradient; RVDd, right ventricular diastolic diameter; RVFAC, right ventricular fractional area change; TAPSE, tricuspid annular plane systolic excursion. Other abbreviations as in Table 1.
During the follow-up period (mean 236±151 days), 7 patients died and 28 patients underwent LVAD implantation. The patients in the lower echo-PAC group had a significantly higher rate of death or LVAD implantation than those in the high echo-PAC group (Figure 5, log-rank P<0.001). In the univariate analysis with Cox proportional hazards modeling (Table S1), lower echo-PAC values were also significantly associated with a higher incidence of death or LVAD implantation (HR: 0.90 for 0.1 mL/mmHg increase, 95% CI: 0.85–0.95, P<0.001). This association remained significant after adjusting for the parameters related to HF severity (NYHA functional class III–IV, ln [BNP], mitral E/A ratio, or LVEF) or RV function (RVDd, RVFAC, total bilirubin, TRPG, or TAPSE) (Table 3).
Kaplan-Meier plot of event-free (death or LVAD implantation) survival in patients in the high echo-PAC group (>1.94 mL/mmHg, blue line) and the low echo-PAC (≤1.94 mL/mmHg, red line) group. Echo-PAC, pulmonary arterial capacitance obtained by echocardiography; LVAD, left ventricular assist device.
Cardiac event: death or left ventricular assist device implantation. E/A (n=43), mitral peak E (early diastolic)/A (late diastolic); Echo-PAC, pulmonary arterial capacitance obtained by echocardiography; HR, hazard ratio; RVDd (n=69), right ventricular diastolic diameter; RVFAC (n=66), right ventricular fractional area change; TAPSE (n=51), tricuspid annular plane systolic excursion. Other abbreviations as in Table 1.
We had 2 major findings in this study. First, noninvasive echocardiographic estimation of PAC, an index of RV afterload, correlated with PAWP and PVR as well as with PAC invasively measured by RHC. In advanced HF, PVR and PAWP were independent variables for echo-PAC, and lower values of echo-PAC (<1.94 mL/mmHg) indicated an increased risk of a high-PAWP state accompanied by high PVR. Second, in advanced HF patients, those with lower values of echo-PAC (<1.94 mL/mmHg) had clinical parameters related to HF severity and echocardiographic indices related to RV dysfunction. Lower values of echo-PAC were associated with worse outcomes even after adjusting for indices of HF severity or RV function. To our knowledge, no study has evaluated the clinical significance of PAC noninvasively obtained by echocardiography in advanced HF. These results suggested that echocardiographic PAC measurement is a noninvasive index reflecting the integration of both PAWP and PVR, which affect RV function, and is helpful in identifying advanced HF patients at risk for poor outcomes.
PVR is commonly used as an index of RV afterload. It represents the arterial load under steady flow, whereas PAC reflects the arterial load under pulsatile flow, which plays an important role in the pulmonary circulation.21 Previous studies using catheterization revealed that elevated PAWP has an effect on the relationship between PAC and PVR, lowering PAC for a given PVR, and resulting in enhanced pulsatile RV afterload.11,12 The same correlation of the echocardiographic PAC measurement with PAWP and PVR was observed in this study. PAWP and PVR were independent determinants of the echo-PAC measurement. In patients with greatly elevated PVR, represented as reactive PH, PVR remains largely fixed, at least in the short term, even if PAWP is reduced by treatment.22,23 In patients with normal or mildly elevated PVR, PAC is reported to be a more important factor in RV afterload than PVR.24 In stratifying the patients in the hemodynamic study by differentiation of PH according to the guidelines issued by the European Society of Cardiology and the European Respiratory Society,25 echo-PAC values in those with passive PH (n=13, median: 1.19 mL/mmHg, 25–75th percentiles: 0.78–2.22 mL/mmHg) and those with reactive PH (n=3, median: 1.28 mL/mmHg, 25–75th percentiles: 0.79–1.94 mL/mmHg) were significantly lower than in those without PH (n=14, median: 3.31 mL/mmHg, 25–75th percentiles: 2.08–4.45 mL/mmHg), whereas PVR in those with reactive PH (6.39±2.39 WU) was significantly higher than in those with passive PH (2.58±1.08 WU) or without PH (2.31±0.77 WU). Echo-PAC values did not differ between passive and reactive PH. This result is consistent to a previous report on PAC obtained by catheterization.26 Reactive PH is considered to be related to structural abnormality in pulmonary arteries caused by chronic elevation of pulmonary venous pressure,4 and PVR is a better marker than PAC for distinguishing patients with reactive PH from those with passive PH. On the other hand, PAC, obtained by echocardiography or catheterization, may be a possible early marker of functional changes in the pulmonary vasculature, compared with PVR, in advanced HF patients. Patients with pulmonary arterial hypertension have greatly elevated PVR, and PVR is therefore an important index in clinical practice. On the other hand, in HF patients, even in advanced HF, greatly elevated PVR is not common,27 suggesting that detecting patients with low PAC or modestly elevated PVR may be adequate for performing risk stratification. Therapy for advanced HF is changing in the current era because of improvements in LVAD therapy. As shown by the ROADMAP trial, which was a recent clinical trial evaluating extended eligibility of LVAD therapy in advanced HF patients with INTERMACS profile 4–628 and who did not have high PVR, the treatment strategy for patients who do not meet the conventional indication for LVAD therapy, but may have a poor prognosis, has received much attention. A clinical index for identifying those patients who could potentially benefit from LVAD therapy because of their poor prognosis is necessary. PAWP alone is not enough to identify the patients who would benefit from LVAD therapy and not enough to predict poor outcomes. On the other hand, lower PAC is considered to contain integrated information related to elevated PAWP and pulmonary vascular injury, which is related to RV dysfunction. In this study, noninvasively measured PAC was associated with levels of RV dysfunction and HF severity that were related to PAWP. These correlations may contribute to the predictive ability of PAC for poor outcomes in HF patients, as seen in previous reports on RHC-PAC.10,11,27 The present study expands those results and shows that PAC obtained noninvasively with Doppler echocardiography is also useful for identifying this subgroup of advanced HF patients at high risk for worse outcomes, who may be potential candidates for LVAD.
In this study, we assessed the clinical utility of echo-PAC measurements. Our results from comparing them with RHC-PAC values found that the echocardiographic values tended to be higher in the upper range when measured in patients without PH who had preserved values of RHC-PAC (4.5±2.1 mL/mmHg),27 whereas the difference in the lower range was small. There was no significant difference between SV by echocardiography and catheterization (mean difference: −1.7 mL, P=0.14). Although the estimated PP was also not significantly different from that estimated by RHC (mean difference: 0 mmHg, P=0.50), it was underestimated in some patients with lower PAC. These observations indicate that echo-PAC values may be suitable for detecting impaired, but not preserved, levels of PAC. Considering this trend, echocardiographic PAC measurements are likely to be useful to detect high-risk patients with advanced HF.
There are several limitations to note. First, SV calculated by LVOT velocity may not be equal to right-sided SV because of PR. However, measurement of RV outflow tract (RVOT) diameter is technically difficult, so measurement of the SV from the LVOT is considered to be more accurate than that from the RVOT. Second, this study was retrospective and conducted in a single institution; the sample size was small because we focused on patients with advanced HF. Although the echo-PAC measurements had sufficient predictive power for worse outcomes in this study, future multicenter studies will be needed.
Echocardiographic PAC values correlated with the severity of HF and RV systolic dysfunction, and might be a useful noninvasive marker of poor prognosis in patients with advanced HF.
We are grateful to Keiko Katsuki, JRDCS, and Yumiko Morimoto-Kobayashi, JRDCS, and Mariko Fujita for their technical assistance.
We have no conflicts of interest to disclose.
Supplementary File 1
Table S1. Univariate cox proportional hazard analysis for risk of cardiac event
Please find supplementary file(s);