If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Diastolic pulmonary gradient (DPG), the difference between pulmonary artery diastolic pressure and mean capillary wedge pressure, ≥ 7 mmHg is associated with pulmonary vascular remodelling.
•
Group 1 Pulmonary hypertension patients have broader distribution of DPG than group 2 pulmonary hypertension category.
•
Increasing DPG is associated with worse mortality in Group 1 pulmonary hypertension.
Abstract
Background
Diastolic pulmonary gradient (DPG), calculated as the difference between pulmonary artery diastolic pressure and mean pulmonary capillary wedge pressure ≥ 7 mmHg is associated with pulmonary vascular disease and portends poor prognosis in heart failure (HF). The prognostic relevance of DPG in group 1 pulmonary hypertension (PH) is uncertain.
Methods
Using the Pulmonary Hypertension Connection (PHC) risk equation for 225 patients in the NIH-PPH, the 5-year probability of death was calculated, which was then compared with DPG using a Cox proportional hazards model. Kaplan-Meier survival curves were determined for two cohorts using the median DPG of 30 mmHg as cutoff, and significance was tested using the log-rank test.
Results
The mean age was 38.1 ± 16.0 years old, 63% female, and 72% were “white”. The mean DPG was 31.6 mmHg ± 13.8 mm Hg and only 1.8% had a DPG <7 mm Hg. Increasing DPG was significantly associated with increased 5-year mortality even after adjustment for the PHC risk equation (HR 1.29 per 10 mm Hg increase). When DPG was dichotomized based on the median of 30 mm Hg, the HR for DPG >30 mm Hg with respect to 5-year mortality was 2.03. After adjustment for pulmonary artery systolic pressure (PASP), increasing DPG remained significantly associated with decreased 5 years survival (HR 1.99 for DPG > 30 mm Hg).
Conclusions
DPG is independently associated with survival in group 1 PH patients even after adjustment for the PHC risk equation or PASP. Patients with increased DPG had a 2-fold increased risk of mortality. The use of DPG for guiding treatment and prognosis in group 1 PH should be further investigated.
Combined pre and post capillary pulmonary hypertension
DPG
Diastolic pulmonary gradient
HFrEF
Heart failure with reduced ejection fraction
mPAP
Mean pulmonary artery pressure
NIH
National Institute of health
PAC
Pulmonary artery capacitance
PADP
Pulmonary artery diastolic pressure
PASP
Pulmonary artery systolic pressure
PCWP
Pulmonary capillary wedge pressure
PH
Pulmonary hypertension
PH-LHD
Pulmonary hypertension secondary to left heart disease
PPH
primary pulmonary hypertension
PPP
Pulmonary pulse pressure
PVR
Pulmonary vascular resistance
RSVWI
Right ventricular stroke work index
TPG
Transpulmonary gradient
1. Introduction
There has been growing interest in the hemodynamic marker, diastolic pulmonary gradient (DPG), which is calculated as diastolic pulmonary artery pressure (PADP) minus mean pulmonary capillary wedge pressure (PCWP) [
Prognostic value of the pre-transplant diastolic pulmonary artery pressure-to-pulmonary capillary wedge pressure gradient in cardiac transplant recipients with pulmonary hypertension.
]. DPG has been proposed to distinguish pulmonary vascular disease among patients with combined pre and post capillary pulmonary hypertension (CPCPH), previously referred to as ‘‘out-of-proportion’’ PH [mean pulmonary artery pressure (mPAP) ≥25 mmHg, PCWP >15 mmHg, and transpulmonary gradient (TPG) > 12 mmHg]. An elevated DPG (>7 mm Hg) may be associated with pulmonary vascular remodeling and has been shown to predict worse survival in individuals with PH secondary to left heart disease (PH-LHD) or WHO group 2 PH [
]. The range of absolute values of DPG in HF is relatively small with recent publications demonstrating that both low and high values of DPG in HF may be associated with adverse outcomes [
Characterization of pulmonary hypertension in heart failure using the diastolic pressure gradient: the conundrum of high and low diastolic pulmonary gradient.
]. Although there has been growing enthusiasm around the utility of DPG as a hemodynamic maker, this interest has been primarily centered on the phenotype of PH-LHD. To our knowledge no previous studies have evaluated the association of DPG with clinical outcomes in patients with group 1 PH. To investigate the utility of DPG in predicting outcomes in group 1 PH, we queried the National Institute of health (NIH) primary pulmonary hypertension [(PPH), now group 1 PH) registry to assess the relationship between DPG and clinical outcomes.
2. Methods
The NIH established one of the first large registries of PPH patients (all idiopathic PAH, familial PAH, and anorexigen-associated) with the specific aim of characterizing demographic, clinical and laboratory findings of patients at the time of diagnosis, as well as determining the natural history of the disease with medical interventions. The publicly available registry dataset was accessed at https://biolincc.nhlbi.nih.gov/studies/pphreg. The methodology and enrollment have previously been published, including the risk equation for predicting 5 year survival probability [
]. Briefly, patients were enrolled into the registry from 32 medical centers throughout the US. PH was defined as a mean pulmonary arterial pressure of >25 mmHg at rest or 30 mmHg with exercise at catheterization. Patients were included in the study after the following secondary causes of PH were excluded: PH within the first year of life, congenital abnormalities of the heart, lungs, or diaphragm, pulmonary thromboembolic disease, sickle cell anemia, history of intravenous drug abuse, obstructive lung disease, interstitial lung disease, arterial hypoxemia, collagen vascular disease, parasitic disease affecting the lungs, pulmonary artery or valve stenosis, or pulmonary venous hypertension. This analysis was conducted with a de-identified public release of the NIH primary PH database (NCT00005357) from the NHLBI [
In order to evaluate the relationship between DPG and the clinical outcomes in the NIH-PH database, we analyzed patients who had the necessary invasive hemodynamic data to determine the Pulmonary Hypertension Connection (PHC) risk equation (mean right atrial pressure, mean pulmonary artery pressure, and cardiac index) and the DPG (pulmonary artery diastolic pressure, pulmonary artery wedge pressure). Patients were followed for 5 years, and time to death was reported within 5 years of enrollment. We calculated the DPG from the invasively measured right heart catheterization hemodynamic data. The 5-year probability of survival was calculated as a number between 0 and 1 using the PHC risk equation based on the mean pulmonary arterial pressure, mean right atrial pressure, and cardiac index. The calculated PHC 5-year survival probability, DPG, and pulmonary artery systolic pressure were then evaluated using a multivariable Cox proportional hazards model. Sensitivity and specificity were determined for DPG and survival, and the optimal cutoff for DPG was defined as that which maximized both sensitivity and specificity (i.e., where sensitivity and specificity were maximized and equal). Kaplan-Meier survival curves were determined for two cohorts using the median DPG of 30 mmHg as cutoff, and significance was tested using the log-rank test.
2.2 Statistics analysis
Statistical analysis was conducted using SAS 9.4 (SAS Institute, Cary, NC). Analyses of categorical variables were conducted using the χ2 test. The Fisher exact test was used for categorical variables with low frequencies. The Wilcoxon rank sum test was used to test for differences in continuous variables. Categorical variables are presented as frequencies with percentages, whereas continuous variables are described using medians and interquartile ranges (IQR). Survival analysis and Kaplan-Meier plots are used to show differences in adverse events between groups with stratification based on key predictor variables. The log-rank test was used to compare differences between groups. Multivariable Cox proportional hazards regression and multivariable logistic regression were used to model associations of multiple independent variables of interest with the endpoints of interest during follow-up. An alpha value of 0.05 was used for statistical significance.
3. Results
3.1 Baseline characteristics
Among 310 patients, 225 had the necessary invasive hemodynamic measurements for the determination of the PHC risk equation and the DPG, and these patients (N = 225) were used in the present analysis. There were no significant differences between included and excluded patients for variables of interest, including age, gender, BMI, DPG, PA systolic pressure, PADP, PCWP, cardiac index, and right atrial pressure (P > 0.10 for all).
The mean age of patients was 38.1 ± 16.0 years old, 63% were female, and 72% had race reported as “white”. The mean DPG was 31.6 ± 13.8 mm Hg and the median DPG was 30 mm Hg (IQR 23–40 mm Hg). The distribution of DPG was broad (Fig. 1) compared with the typical distribution of DPG in HF characterized by a median DPG < 10 mm Hg and a cutoff value of 7 mm Hg. In fact, only 4/225 (1.8%) patients in this database had a DPG < 7 mm Hg. Other demographic and hemodynamic characteristics stratified by quartile of DPG are shown in Table 1.
Fig. 1Histogram of Diastolic Pressure Gradient Distribution. The histogram showing the distribution of the diastolic pressure gradient (DPG) is shown. Most of the patients with primary pulmonary hypertension in this study had DPG values in the range of 10–50 mm Hg.
Using logistic regression receiver operating characteristic analysis, we found that the median value of DPG (30 mm Hg) maximized sensitivity and specificity for mortality in this cohort of WHO group 1 pulmonary hypertension patients (Fig. 2). Sensitivity and specificity at this cutoff value were both slightly above 60%. Greater specificity (82%) could be achieved with a DPG cutoff of 40 mm Hg, (corresponding to the lower bound of the fourth quartile) although the associated sensitivity at this cutoff was 31%.
Fig. 2Sensitivity and Specificity of Diastolic Pressure Gradient for Survival. Sensitivity and specificity analyses for survival at 5 years are shown for different cutoffs of the diastolic pressure gradient (DPG). A cutoff of 30 mm Hg maximizes sensitivity and specificity at values greater than 60% for both.
3.3 Survival analysis and cox proportional hazards regression
In a multivariable Cox model adjusted for the PHC risk equation (Model 1), DPG as a continuous variable had a hazard ratio of 1.29/10 mm Hg (95% C.I. 1.13–1.48) for mortality (Table 2). In a separate Cox model adjusted for pulmonary artery systolic pressure (PASP) and the PHC risk equation (Model 2), DPG as a continuous variable remained significantly associated with decreased survival over 5 years with a hazard ratio of 1.33/10 mm Hg (95% C.I. 1.07–1.65) (Table 3).
With respect to discrete DPG subgroups, when DPG was grouped based on quartiles (≤23 mm Hg, 24–30 mm Hg, 31–40 mm Hg, and >40 mm Hg) (log-rank P = 0.003), increasing DPG was associated with increased mortality (Fig. 3). In a Cox proportional hazards model with DPG >40 mm Hg, diastolic pulmonary arterial pressure (Model 3), and the PHC equation, higher DPG > 40 mm Hg remained associated with increased mortality (Table 4). As shown in the table, a DPG >40 mm Hg confered a twofold increased HR for mortality even after adjustment for both the PHC equation and diastolic pulmonary pressure (p = 0.01). (Of note, the dichotomous rather than continuous form of DPG was used because diastolic pulmonary pressure is used for the calculation of DPG).
Fig. 3Kaplan Meier Survival Curves for Diastolic Pressure Gradient. Kaplan Meier curves are shown for patients with DPG divided based on quartiles into the following groups: ≤ 23 mm Hg, 24–30 mm Hg, 31–40 mm Hg, and >40 mm Hg.
The ROC area obtained with multivariable logistic regression for the probability of 5-year survival in the model containing DPG (P = 0.0008) and the PHC risk equation (P < 0.0001) was 0.70 (P < 0.0001) (Fig. 4A), which was significantly better than the ROC area for either the PHC risk equation alone (AUC = 0.64, P < 0.0001) (Fig. 4B) or DPG alone (AUC = 0.62, P = 0.01) (Fig. 4C).
Fig. 4ROC Analysis for Diastolic Pressure Gradient. Receiver Operating Characteristic (ROC) curves are shown for the multivariable logistic regression model for 5-year survival with the diastolic pressure gradient (DPG) and the PHC risk equation (A), the PHC risk equation alone (B), and DPG alone (C). Addition of the DPG the logistic regression model with the PHC risk equation increases the AUC from 0.64 to 0.70 (P < 0.0001).
As noted previously, 4/225 patients did have DPG <7 mm, which could be consistent with isolated post-capillary pulmonary hypertension. For this reason, we performed a sensitivity analysis for the major findings in this report with these 4 patients excluded and found minimal differences. For example, with these 4 patients excluded, in the Cox proportional hazards model with DPG and the PHC, the HR for mortality associated with DPG was 1.27 mm Hg/10 mm Hg (P = 0.0007). In the Cox proportional hazards model with DPG >40 mm Hg adjusted for PHC and PADP, the HR for mortality associated with DPG >40 mm Hg was 2.09 (P = 0.01). In the logistic regression model for mortality with DPG >40 mm Hg, the ROC area was 0.62 (P = 0.0003) and DPG OR for mortality of 2.80 (P = 0.0003). In the multivariable logistic model with PHC and DPG>40 mm Hg, the model ROC area was improved to 0.71 (P < 0.0001) with an OR for DPG > 40 of 3.08 (P = 0.0002), which improved upon the ROC area for PHC alone (0.65, P = 0.0002).
4. Discussion
The present analysis investigated the relationship between the hemodynamic variable DPG and the probability of 5-year survival in patients with WHO group 1 PH from the NIH-PPH database. We found that DPG had a broader distribution in patients with group 1 PH than is typically seen in patients with heart failure. Furthermore, increased DPG, when analyzed as a continuous variable, was associated with an approximately two-fold increased probability of death at 5 years. While a DPG value of greater than 7 mmHg in PH-LHD signals the presence of pulmonary vascular disease and increased adverse outcomes [
], a DPG threshold of greater than 30 mmHg in group 1 PH patients may provide a parallel means of risk stratification for these patients. This has important clinical implications in extending the application of DPG to patients with group 1 PH.
The burgeoning interest in DPG as a marker has been centered on ease of application and its ability to identify pulmonary vascular disease among patients with PH-LHD [
]. In such patients, the degree of elevation of pulmonary hypertension cannot be solely explained by the elevation in left sided filling pressures. DPG has been shown to be less sensitive to changes in loading conditions such as variations in stroke volume, and therefore a more appealing marker for pre-capillary pulmonary arterial remodeling in patients with group 2 PH [
]. To our knowledge the use of DPG in guiding risk stratification and prognosis in forms of PH other than group 2, PH-LHD, is yet to be reported.
DPG was first reported by Dexter et al. when they reported a series of invasively obtained hemodynamic parameters in both normal subjects and in selected disease states [
Studies of the pulmonary circulation in man at rest; normal variations and the interrelations between increased pulmonary blood flow, elevated pulmonary arterial pressure, and high pulmonary 'capillary” pressures.
]. In their series normal subjects had a mean PVR of 0.8 Wood units, with a mean DPG of 0 mmHg. Patients with atrial septal defects/patent ductus arteriosus and elevated PCWP, however, were excluded from this case series. Among those with PH (mean PAP 33.5 mmHg and mean PVR 0.75 Wood units) the mean DPG was 10 mmHg. Interestingly, for patients with Eisenmenger's complex (average mPAP of 87 mmHg, mean PVR 26.3 Wood unit), the mean DPG was 57 mmHg. As can be inferred from the aforementioned case series, elevation in DPG corresponded well to the degree of pulmonary vascular disease. Coincidentally, the PH patients in the Dexter's et al. case series (ASD/PDA and Eisenmenger complex) can be, under the current nomenclature, classified under WHO group 1 PH.
The debate on the utility of DPG in PH-LHD is still evolving, as evidenced by conflicting studies and editorials addressing the prognostic implications of DPG in heart failure [
]. For example, Gerges et al. evaluated the DPG as a predictor of prognosis in CPCPH. They reported that patients with CPCPH and DPG ≥7 mmHg had significantly worse outcomes compared with those with DPG <7 mmHg. Furthermore, patients with a DPG ≥7 mmHg in their cohort of CPCPH had evidence of pulmonary arterial remodeling on pathological examination. However, subsequent publications, one by Tedford et al. and another by Tampakakis et al., failed to demonstrate the relationship of elevated DPG with outcomes in post-transplant and heart failure patients respectively [
Gerges et al. have recently analyzed the hemodynamic cutoffs that may further help in discriminating between idiopathic pulmonary arterial hypertension and post-capillary PH. They reported that the best hemodynamic discriminator between idiopathic pulmonary arterial hypertension and post-capillary PH was mean PCWP <12 mmHg and a DPG >7 mmHg. However, only patients with DPG >20 mmHg had significant hemodynamic improvement with treatment with prostaglandins. The authors suggested that DPG might be an important hemodynamic marker for guiding medical management [
]. Of note, the higher values and broader range of DPG in group 1 PH make it an even more robust parameter in this population.
4.1 Limitations
We acknowledge that this retrospective analysis from the NIH-PPH registry has some limitations. First, the mean age of patients was relatively younger compared to the other more contemporary PH registries [
]. Second, the prognosis and outlook of patients with PH in the current era may be different because of the availability of pulmonary vasodilator therapies [
]. In the past it was held that patients with a diagnosis of primary pulmonary hypertension (now group 1 PH) had a very poor prognosis because of the inexorably progressive nature of the disease and the fewer therapeutic options during the enrollment period of the NIH-PPH registry [
]. Third, there have been several iterations of PH guidelines, including several nomenclatural and classification changes, since the conclusion of the registry [
]. These limitations are offset by the strengths of the current study, including a focus on objective hemodynamic measurements, multicenter enrollment, and adjustment for the PHC risk equation, derived in the modern era of pulmonary vasodilator therapies [
In patients with group 1 PH, increasing DPG was strongly associated with increasing mortality during 5 years of follow-up, even after adjustment for the PHC risk equation. These findings support the use of the DPG as an important hemodynamic variable in the risk stratification of patients with WHO group 1 PH. Further studies are needed to evaluate the clinical applicability of this hemodynamic variable in the more contemporary era of pulmonary vasodilator therapies.
Author contributions
Dr. Mazimba is the guarantor of the entire manuscript.
Dr. Mazimba contributed with the study design, database, statistical analyses, writing and approving the manuscript.
Dr. Bilchick contributed to the, statistical analyses, figures, tables and writing the manuscript.
Dr. Mejia Lopez contributed with the tables, data interpretation and writing the manuscript.
Dr. Black contributed with drafting of the manuscript and literature search.
Dr. Bergin, Dr. Abuannadi and Dr. Kennedy contributed with study design, data interpretation and writing of the manuscript.
Dr’s. Mihalek and Tallaj contributed as external auditors and did critical read of the manuscript.
Financial/non-financial disclosures
Authors have no relevant financial or nonfinancial relationships to disclose.
Disclosures
Authors have no relevant financial or nonfinancial relationships to disclose.
References
Gerges C.
Gerges M.
Lang M.B.
Zhang Y.
Jakowitsch J.
Probst P.
et al.
Diastolic pulmonary vascular pressure gradient: a predictor of prognosis in “out-of-proportion” pulmonary hypertension.
Prognostic value of the pre-transplant diastolic pulmonary artery pressure-to-pulmonary capillary wedge pressure gradient in cardiac transplant recipients with pulmonary hypertension.
Characterization of pulmonary hypertension in heart failure using the diastolic pressure gradient: the conundrum of high and low diastolic pulmonary gradient.
Studies of the pulmonary circulation in man at rest; normal variations and the interrelations between increased pulmonary blood flow, elevated pulmonary arterial pressure, and high pulmonary 'capillary” pressures.
30219-0&id=gr1.jpg){kind=link}
30219-0&id=gr2.jpg){kind=link}
30219-0&id=gr3.jpg){kind=link}
30219-0&id=gr4.jpg){kind=link}