Abstract
This study explores the prognostic utility of pulmonary arterial capacitance (PAC) in a diverse cohort of patients with pulmonary arterial hypertension (PAH) from a tertiary referral center and compares it with the prognostic utility of other hemodynamic parameters. PAC is a strong independent predictor of mortality in patients with PAH.
The importance of pulmonary arterial capacitance (PAC) in the pathogenesis and prognostication of pulmonary arterial hypertension (PAH) is increasingly recognized. 1 PAC is a strong predictor of mortality in patients with idiopathic PAH as well as in patients with PAH related to systemic sclerosis.2,3 However, the prognostic ability of PAC has not been demonstrated in other subtypes of PAH nor has it been compared with the prognostic ability of other hemodynamic parameters in patients with different subtypes of PAH. This report aims to address these objectives in a diverse, well-characterized cohort of patients with PAH.
METHODS
We identified all consecutive, adult patients with World Health Organization (WHO) group 1 PAH (both incident and prevalent disease) from the Tufts Medical Center Pulmonary Hypertension (PH) Registry who underwent a right heart catheterization (RHC) between July 1, 2001, and December 31, 2012. Patients who had WHO group 2, 3, 4, or 5 PAH were excluded from this analysis. Parameters measured during the RHC included heart rate (HR), mean right atrial pressure (RAP), mean pulmonary artery pressure (mPAP), pulmonary capillary wedge pressure (PCWP), and cardiac output (CO) by thermodilution method. These parameters were determined by manual review of the tracings. The pulmonary vascular resistance (PVR) was calculated as (mPAP – PCWP)/CO (Wood units), cardiac index (CI) as CO/body surface area (L/min/m2), pulmonary artery pulse pressure as PA systolic – PA diastolic, and PAC as stroke volume/pulmonary artery pulse pressure (mL/mmHg). Patients were followed up until death or October 1, 2015. None of the patients was lost to follow-up, and all patients who were still alive were censored at the end of follow-up.
Categorical variables were summarized as numbers and percentages, whereas continuous variables were summarized as mean value ± standard deviation (SD) or median (interquartile range [IQR]), as appropriate. Receiver-operating characteristic (ROC) curves were used to evaluate the ability of PAC, mRAP, mPAP, CI, and PVR to discriminate between patients who had died within the first year after RHC and those who were still alive. The area under the ROC curve (AUC) was compared using the DeLong method. Cox proportional hazards models were used to estimate hazard ratios for these hemodynamic parameters and mortality. These models were then adjusted for age, sex, and etiology of PAH. All analyses were conducted in R version 3.2.1.
RESULTS
A total of 105 patients were identified in the database and included in the analysis. Mean age (±SD) was 59 ± 13 years, 64% of patients were female, and 86% were non-Hispanic white. The most common etiology was idiopathic PAH (39%) followed by connective tissue disease (26%), portopulmonary hypertension (20%), and congenital heart disease (4%). Eighty patients (76%) had received a new diagnosis of PAH (i.e., had incident cases). At baseline, patients had significant pulmonary hypertension with median (IQR) RAP of 8 mmHg (5–13), mean mPAP of 46 ± 13 mmHg, mean CI of 2.4 ± 0.7 L/min/m2, median (IQR) PVR of 9 Wood units (5–12), and median (IQR) PAC of 1.4 mL/mmHg (1.0–2.3). The majority of the patients (91%) had moderate to severe impairment in exercise capacity (New York Heart Association functional class II and III) with a mean 6-minute walk distance of 346 ± 154 m. The median follow-up was 3.9 years. There were 60 deaths during the study period.
The AUCs for PAC, RAP, mPAP, CI, and PVR were not significantly different for discriminating 1-year mortality (0.61, 0.64, 0.60, 0.58, and 0.62, respectively). Combining PAC and PVR did not improve the predictive value for 1-year mortality as compared with either parameter alone (AUC for PAC + PVR combination was 0.61; P value was 0.96 and 0.54 when compared with AUC for PAC alone and PVR alone, respectively).
Although lower PAC and higher RAP, mPAP, and PVR were associated with increased risk of mortality, PAC was associated with the highest hazard (Table 1). A 1-mL/mmHg decrease in PAC was associated with a significantly increased hazard of death even after adjustment for age, sex, and etiology of PAH (hazard ratio [95% confidence interval]: 1.85 [1.31–2.59]). The hazard ratio for PAC was similar in the subset of patients with PAH other than idiopathic or connective tissue disease associated (data not shown).
Cox proportional hazard models for mortality
Note: CI: cardiac index; HR: hazard ratio; mPAP: mean pulmonary artery pressure; PAC: pulmonary arterial capacitance; PVR: pulmonary vascular resistance; mRAP: mean right arterial pressure.
Model adjusted for age, sex, and etiology of pulmonary arterial hypertension (portal hypertension vs. connective tissue disease vs. other).
DISCUSSION
In this study, we demonstrate that PAC is a superior prognostic predictor of mortality in patients with PAH of different etiologies compared with other hemodynamic parameters, including PVR, RAP, mPAP, and CI. Although the AUC for predicting 1-year mortality was low for all of the hemodynamic parameters, the hazard ratios, which are a more sensitive measure of incremental increase in risk, clearly demonstrate that a decrease in PAC translates to greater risk of death compared with increases in PVR, RAP, and mPAP or a decrease in CI.
The importance of PAC has been increasingly recognized in pulmonary vascular disease. At normal PCWP, the PAC and PVR have an inverse hyperbolic relationship; 4 therefore, PAC decreases significantly before an increase in PVR is detected. There is also evidence to suggest that, in the pulmonary vasculature, the distal small arterioles are the main contributors to both PVR and PAC. 1 These findings suggest that changes in PAC precede changes in PVR and could potentially serve as an early and more sensitive marker of disease progression in patients with PAH.
PAC is a major contributor to the right ventricular (RV) workload. 5 As PAC decreases, the RV workload increases, leading to RV hypertrophy, RV dilation, and eventually RV failure. 6 The RV function and its ability to adapt in the setting of pulmonary vascular disease play an important role in determining outcomes for patients with PAH, especially considering that most patients with PAH die of right heart failure. 7 Collectively, these factors help explain our finding that PAC has a better prognostic ability than PVR or CI in patients with PAH.
The prognostic ability of PAC has been previously demonstrated in patients with WHO group 2 PH associated with left heart disease,8,9 idiopathic and scleroderma-associated PAH,2,3 and pediatric PAH. 10 This is, to our knowledge, the first study to include a diverse, well-characterized population of patients with PAH to investigate the prognostic ability of PAC. We acknowledge the limitations inherent in a study that involves a small cohort and believe that future research should assess whether changes in PAC in response to pulmonary vasodilator therapies also have prognostic significance in patients with PAH.