Relationship of Pulmonary Arterial Capacitance and Mortality in Idiopathic Pulmonary Arterial Hypertension


The purpose of this study was to determine if pulmonary vascular capacitance predicts survival in patients with idiopathic pulmonary arterial hypertension (IPAH).


The prognosis of patients with IPAH is difficult to predict, despite knowledge of clinical and hemodynamic parameters previously identified as predictors.


We proposed a capacitance index of stroke volume divided by pulmonary pulse pressure (SV/PP) and prospectively gathered data on IPAH patients who underwent a right heart catheterization. SV/PP was analyzed as a predictor of mortality after adjusting for other modifiers of risk.


During 4-year follow-up of 104 patients, 21 patients died. When compared with conventional markers, SV/PP was the strongest univariate predictor of mortality (hazard ratio 17.0 per ml·mm Hg−1decrease, 95% confidence interval 13.0 to 22.0; p < 0.0001). In successive bivariate analysis, SV/PP was the only predictor of mortality. In quartile analysis, the lowest SV/PP quartile had a 4-year mortality of 61%; the highest SV/PP had no deaths.


The capacitance index (SV/PP) is a strong independent predictor of mortality in patients with IPAH.


Abbreviations and Acronyms


confidence interval


one-second forced expiratory volume


hazard ratio


idiopathic pulmonary arterial hypertension


National Institutes of Health


pulmonary artery


pulse pressure


pulmonary vascular resistance


right atrium


right ventricle


stroke volume

Several clinical and hemodynamic parameters predict outcome in idiopathic pulmonary hypertension (IPAH), but no single marker has been shown to be highly predictive of mortality (1, 2, 3, 4, 5, 6, 7, 8). A multiparameter survival equation derived from logistic regression analysis of the National Institutes of Health (NIH) IPAH registry provides stronger predictive ability but is cumbersome and has not been prospectively tested in an era of vasodilator therapy (6).

Pulmonary arteriolar capacitance measures how much the aggregate pulmonary arteriolar tree will dilate with each contraction of the right ventricle (RV). It can be approximated by stroke volume divided by pulmonary pulse pressure (SV/PP) and should be inversely proportional to the workload on the right heart. Because low systemic capacitance is correlated with mortality in coronary artery disease and heart failure (8, 9, 10, 11), we hypothesized that pulmonary arteriolar capacitance is a strong predictor of outcome in IPAH.



Consecutive patients who underwent right heart catheterization for evaluation of pulmonary hypertension on echocardiogram from January 1, 1999, to December 31, 1999, were screened. Patients were included if complete invasive testing was done and confirmed a mean pulmonary artery (PA) pressure of >40 mm Hg, pulmonary capillary wedge pressure of <14 mm Hg, and pulmonary arterial resistance of >6 Woods Units. Patients were excluded if they were under 18 years of age, had pulmonary hypertension associated with an identified cause, or had an ejection fraction of <50%. Patients were also excluded if they did not give informed consent for research studies, as governed by Minnesota Statute 144.335. This study was approved by the Mayo Foundation Institutional Review Committee.

Pulmonary function testing

Pulmonary function testing was performed in most patients with the use of a constant-volume body plethysmograph. The single-breath technique with carbon monoxide was used to determine the diffusion capacity.


During right heart catheterization, a balloon-tipped thermodilution catheter was used to obtain right atrium (RA) pressures, PA diastolic, mean, and systolic pressure, and wedge pressure. Cardiac output was measured in triplicate by thermodilution technique. The Fick method was used if there was >15% variability of thermodilution output measurement or severe tricuspid regurgitation. Pulmonary vascular resistance (PVR), cardiac index, and stroke volume (SV) were calculated using standard formulas. Systemic blood pressure and blood gases were obtained via a radial artery catheter.

Calculation of pulmonary arteriolar capacitance (SV/PP)

Pulmonary arteriolar capacitance reflects the ability of the pulmonary vessels to dilate during systole and recoil during diastole. By storing blood during systole, a high capacitance tree dampens the PA systolic pressure. By recoiling during diastole, a high capacitance tree increases the PA diastolic pressure. Thus, capacitance is inversely proportional to PP. The lower the SV, the lower the PP regardless of the vessel’s ability to dilate. Capacitance (ml·mm Hg−1) is therefore inversely proportional to PP and directly proportional to SV.()C=ΔvolumeΔpressure=Stroke;VolumePulmonary;Artery;Pulse;Pressure=SV [ml]PP [mm Hg]The SV/PP index has been validated in vivo in the systemic and splanchnic circulation with a 99% correlation with the lumped, two-parameter Windkessel model (12, 13, 14) (Appendix).

Calculation of NIH four-year predicted survival

As derived from a national registry (7), the probability that a patient with IPAH will survive four years = (0.47),A(x,y,z), in which A(x,y,z) = e(0.007325x+0.0526y−0.3275z)and where x = mean PA pressure, y = mean RA pressure, and z = cardiac index.

Follow-up evaluation

Clinical follow-up was conducted by clinical visits, mailed questionnaires, telephone contact, and research of the Social Security Death Index. In all cases, vital status was coded by personnel blinded to clinical and hemodynamic data.

Data analysis

The primary end point was all-cause mortality. Contingency tables were analyzed for association with the Pearson chi-squared test. Comparisons of the mean of a continuous variable between two groups were made with the Wilcoxon rank sum test. The follow-up duration was considered to represent the interval from the initial evaluation to the time of death or four years. A survival curve was estimated via the Kaplan-Meier method. Patients were stratified according to quartiles of pulmonary arteriolar capacitance for the analysis of survival; the log rank test was used to compare survival differences among these patient groups. Baseline clinical and hemodynamic variables for these patients were analyzed with the Cox proportional hazard model, and the hazard ratio (HR) with 95% confidence interval (CI) for each factor was given. Backward elimination techniques were used here to identify variables independently associated with mortality. Candidate predictors were pulmonary arteriolar capacitance, age, male gender, World Health Organization functional class, 6-min walk, heart rate, ejection fraction, NIH 4-year predicted survival, PVR, mean RA pressure, mean PA pressure, mean PA pressure multiplied by SV (RV stroke work), cardiac index, SV, 1-s forced expiratory volume, and total lung capacity. A p value of <0.05 was considered statistically significant.

Receiver-operator curves were constructed based on the sensitivity and specificity of a test at each level to predict survival.



One hundred seventeen patients underwent right heart catheterization for evaluation of suspected idiopathic pulmonary hypertension. Thirteen patients were excluded from further analysis because three were <18 years old, two had adult congenital heart disease, and eight had mean PA pressures of <40 or PVR of <480 dynes · s/cm3.

One hundred four patients (mean age 44 ± 11 years; 25 men) were included. Baseline demographics, clinical status, and pulmonary function tests are given in Table 1.

Table 1. Initial Characteristics of Patients

All Alive Dead p
 Total number 104 83 21 na
 Follow-up (days) 1,249 ± 515 1,501 ± 0 252 ± 243 na
 Age (yrs) 44 ± 11 45 ± 10 43 ± 11 0.6891
 Female/male 79/25 63/20 16/5 0.3033
 Mean RA pressure (mm Hg) 13 ± 6.8 12 ± 6.8 16 ± 6.1 0.0158
 Mean PA pressure (mm Hg) 52 ± 11 51 ± 11 54 ± 10 0.2586
 Pulmonary pulse pressure (mm Hg) 46 ± 13 44 ± 12 52 ± 14 0.0097
 Pulmonary capillary wedge pressure (mm Hg) 12.0 ± 1.8 11.9 ± 2.4 12.2 ± 1.8 0.6220
 Stroke volume (ml) 60 ± 22 65 ± 21 40 ± 14 0.0001
 Cardiac index (l/min/m2) 2.5 ± 0.8 2.7 ± 0.8 2.1 ± 0.7 0.0022
 Pulmonary capacitance (ml/mm Hg) 1.43 ± 0.73 1.60 ± 0.72 0.76 ± 0.18 0.0001
 PVR (dynes·s/cm−5) 1,250 ± 650 1,152 ± 582 1,639 ± 767 0.0018
 RV stroke work (mm Hg·cc) 3,017 ± 1,060 3,323 ± 998 2,201 ± 982 0.0010
 Heart rate (bpm) 82 ± 14 82 ± 15 86 ± 8 0.2418
 Systemic systolic blood pressure (mm Hg) 136 ± 26 137 ± 26 127 ± 24 0.1132
 Systemic diastolic blood pressure (mm Hg) 74 ± 12 75 ± 11 73 ± 16 0.5017
 Left ventricular ejection fraction (%) 61 ± 6 61 ± 6 62 ± 6 0.4966
 Right ventricular systolic pressure (mm Hg) 88 ± 20 87 ± 22 93 ± 12 0.2320
Pulmonary function (% expected)
 Total lung capacity 98 ± 3 99 ± 5 97 ± 4 0.0924
 1-s forced expiratory volume 82 ± 7 82 ± 9 80 ± 7 0.3458
 6-min walk test distance (m) 327 ± 89 334 ± 82 299 ± 111 0.1792
 WHO class
  I (%) 0 0 0 1.000
  II (%) 24 24 24 1.000
  III (%) 51 49 57 0.3213
  IV (%) 25 27 19 0.2393

PA = pulmonary artery; PVR = pulmonary vascular resistance; RA = right atrium; RV = right ventricle; WHO = World Health Organization.

Eighteen percent were on calcium-channel blockers before the study, 28% on beta-blockers, 9% on angiotensin-converting enzyme inhibitors, 25% on warfarin, and 5% on home oxygen before initial study. None were on prostanoids at baseline.

Hemodynamic results

Results of right heart catheterization are in Table 1. The average PA mean pressure and PP were 52 ± 11 mm Hg and 40 ± 13 mm Hg, respectively. There was modest correlation between these two pressures (r = 0.55; p = 0.01). The mean PVR was 1,250 ± 649 dynes·s·cm−5(15.6 ± 8.1 Woods Units).

The mean SV/PP at baseline was 1.43 ± 0.73 ml·mm Hg−1(range 0.40 to 3.77). The lowest quartile of 26 patients encompassed a capacitance index of 0.40 to 0.81 ml·mm Hg,−1, the second quartile 0.82 to 1.25 ml·mm Hg−1, the third quartile 1.26 to 2.00 ml·mm Hg−1, and the highest quartile 2.00 to 3.77 ml·mm Hg−1.

The NIH-predicted 4-year survival was 39 ± 18% (range 5% to 65%).


No patients were lost to follow-up and all patients were followed for four years or until death. During 356 person-years of follow up, 21 patients died and 4 underwent lung or heart-lung transplants. During the follow-up period, 52 patients were treated with epoprostenol, 48 with bosentan, 62 with calcium channel blockers, 11 with angiotensin-converting enzyme inhibitors, 21 with beta-blockers, and 32 with warfarin.

Patients who died had lower SV/PP, NIH-predicted survival, SV, cardiac index, and RV stroke work and higher RA pressure, PVR, and PP (Table 1).

In univariate analysis, SV/PP was the strongest predictor of mortality with an HR of 17.0 (95% CI 13.0 to 22.0; p < 0.0001) per ml·mm Hg−1decrease. The 4-year predicted survival from the NIH IAPH database was also a strong predictor of mortality, with an HR of 1.5 (95% CI 1.2 to 1.9; p = 0.0012) per 10% decrease in predicted survival, as was the cardiac index (HR 3.2 per l/min/m2decrease, 95% CI 1.5 to 6.8; p = 0.0015) (Fig. 1a).

Figure 1. Univariate analysis of potential parameters of all-cause mortality. FEV-1 = 1-s forced expiratory volume; NIH = National Institutes of Health; PA = pulmonary artery; PVR = pulmonary vascular resistance; RA = right atrium; SV/PP = capacitance (stroke volume/pulmonary pulse pressure); WHO = World Health Organization.

In quartile analysis, patients with the lowest capacitance index had a mortality of 62%, whereas those with the highest capacitance index had no deaths (Figure 2, Figure 3).The four transplant patients were in the two lowest SV/PP quartiles.

Figure 2. Kaplan-Meier survival curves by pulmonary vascular capacitance quartiles.

Figure 3. Mortality for each pulmonary vascular capacitance quartile.

Because there were only 21 deaths, multivariate analysis was not done. In successive bivariate analysis SV/PP was the sole independent predictor of mortality. Other markers added no prognostic information to SV/PP (Table 2).

Table 2. Bivariate Predictors of All-Cause Mortality Against SV/PP

Variable Hazard Ratio of Variable With SV/PP p of Variable With SV/PP Hazard Ratio of SV/PP With Variable p of SV/PP With Variable
NIH 4-year 0.93 (0.67–1.30) 0.71 16.4 (12.6–21.3) <0.0001
Cardiac index 0.97 (0.42–2.09) 0.95 16.6 (12.6–21.8) <0.0001
RA pressure 1.03 (0.96–1.10) 0.37 16.4 (12.7–21.1) <0.0001
PVR 0.96 (0.86–1.08) 0.09 17.0 (10.4–27.7) <0.0001
SV 0.97 (0.92–1.02) 0.28 16.5 (14.1–19.2) 0.0002
PP 0.97 (0.93–1.01) 0.15 16.9 (13.1–21.8) 0.0002
Mean PA · SV 0.96 (0.91–1.01) 0.14 15.4 (12.3–20.8) 0.0001

NIH 4-year = National Institutes of Health 4-year predicted survival; SV = stroke volume; other abbreviations as in Table 1.

Receiver-operator curves showed SV/PP to be the most discriminating marker (Fig. 4).

Figure 4. Receiver-operator curves for potential predictors. Abbreviations as in Figure 1.


In this study, high pulmonary vascular capacitance (SV/PP) strongly predicts survival in IPAH. An SV/PP of <0.81 ml·mm Hg−1predicted a <40% probability of survival at 4 years, and an SV/PP of >2.00 ml·mm Hg−1predicted a 100% survival. Other clinical and hemodynamic variables measured in this study did not add additional prognostic information.

The total energy the ventricle must expend to propel a given volume is inversely proportional to capacitance (12). When the ventricle expels its stroke volume, the bolus can either proceed immediately to resistance vessels or be stored in capacitance vessels before advancing downstream. Because not all blood is sent downstream at once, the load on the heart is lower. In addition a lower capacitance is associated with an earlier reflected wave which has been suggested to further increase the load on the heart (15, 16, 17). The relationship between vessel capacitance and outcome has been shown for the systemic circulation. This is the first study which demonstrates the utility of measuring the capacitance and its effect on the right side of the heart.

Study limitations

Our method of estimating pulmonary arteriolar capacitance has been validated in mathematical models but not in human lungs. The SV/PP at initial catheterization does predict mortality in patients with IPAH, and mortality is due to right heart failure; however, it is unknown from this data whether capacitance is fixed or it changes over time. Other predictive parameters, such as O2consumption and brain natriuretic peptide (3, 18), were not studied. There is a modest separation in the confidence intervals surrounding the hazard ratios of a lower SV/PP. Nonetheless, the area under SV/PP’s receiver-operator curve of 0.9 suggests that SV/PP has an excellent ability to distinguish between survivors and nonsurvivors in this series.


In this initial study SV/PP is a strong predictor of survival in IPAH. Other markers added no prognostic value.