Abstract
Background
Echocardiographic metrics of right ventricular (RV) chamber size and function enhance prognostication, risk stratification, and measurement of therapeutic response in patients with pulmonary arterial hypertension (PAH), though the most effective metrics remain unclear.
Research Question
In a well-phenotyped cohort of patients with incident and prevalent PAH, can qualitative grades of RV echocardiographic function be established based on their association with functional outcomes, and do they demonstrate prognostic value beyond traditional risk scores?
Study Design and Methods
In the Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease Phenomics (PVDOMICS) program, 405 (prevalent, n = 336; incident, n = 69) participants were investigated. Multivariable linear regression examined associations with 6-minute walk distance and the Comparative Prospective Registry for Newly Initiated Therapies (COMPERA) and the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL) Lite 2.0 PAH risk scores. Penalized Cox regression was used to develop new models combining prior risk score variables with echo parameters. Cluster analysis combined with survival analysis adjusting for potential confounders was used to demonstrate prognostic significance.
Results
In both incident and prevalent PAH, reduced RV function was associated with increased N-terminal pro-B-type natriuretic peptide levels, reduced 6-minute walk distance, and increased COMPERA and REVEAL Lite 2.0 risk scores after adjusting for duration of PAH and relevant confounders. The addition of echocardiographic variables to models incorporating the COMPERA and REVEAL 2.0 scores yielded a 10% increase in the C-statistic. The severe RV dysfunction group was associated with increased all-cause mortality, with up to a threefold increase in mortality in multivariable models adjusted for relevant confounders, PAH duration, and invasive pulmonary vascular resistance.
Interpretation
Our results show that reduced RV function on echocardiography in PAH is associated with worsened outcomes in incident and prevalent PAH. Echocardiographic assessment of RV function provided additional value to existing PH risk prediction scores and invasive hemodynamics. Furthermore, defining severity of RV function through cluster analysis has important implications for risk prognostication, with potential application to monitor response to therapy.
Key Words: echocardiography, mortality, pulmonary hypertension, right heart
Take-Home Points.
Study Question: In a well-phenotyped cohort of patients with incident and prevalent pulmonary arterial hypertension, can qualitative grades of right ventricular (RV) echocardiographic function be established based on their association with functional outcomes, and do they demonstrate prognostic value beyond traditional risk scores?
Results: RV echocardiographic phenotyping provides added prognostic value in pulmonary hypertension (PH), with RV-pulmonary artery uncoupling, reduced speckle-tracking echocardiography-derived strain, and maladaptive remodeling linked to reduced 6-minute walk distance and increased PH risk.
Interpretations: In this study, cluster analysis and advanced modeling showed that RV dysfunction predicted mortality independently of traditional risk factors, highlighting the potential of echocardiographic phenotyping for enhanced risk stratification in PH.
The ability of the right ventricle to compensate for increased afterload is a key determinant of clinical outcomes in pulmonary hypertension (PH).1, 2, 3 The need for real-time, readily available, and noninvasive assessment of right ventricular (RV) function in PH has led to 2-dimensional echocardiography emerging as a valuable tool for PH screening and risk prognostication.4, 5, 6, 7, 8, 9, 10 Current guidelines,11,12 however, use binary cutoffs for RV chamber measurements and contractile indexes, which may not apply in high afterload settings such as in PH.1,3,13, 14, 15 More sensitive indexes of RV function, like global and free wall longitudinal strain assessed by speckle-tracking echocardiography (STE), have recently been systematically evaluated in large populations of PH,16 demonstrating strong predictive potential, but they have yet to be integrated into existing risk stratification frameworks. Detailed echocardiographic characterization of RV function is crucial for effective PH prognostication and therapeutic assessment. Furthermore, methodologically robust evidence supporting the use of echocardiographic right-sided heart metrics for risk stratification is a key research priority in pulmonary arterial hypertension (PAH).17,18
Despite evidence supporting the prognostic value of RV function in PH, current gold standard risk prediction scores minimally incorporate RV structural and functional echocardiographic parameters. For instance, risk scores derived from both the Comparative Prospective Registry for Newly Initiated Therapies (COMPERA) method8 and the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL) risk calculator7 fail to use echocardiographic RV functional indexes. More recently, the REVEAL-Echo score was derived in a subset of 2,400 patients with PAH in the REVEAL Registry for risk assessment, which showed prognostic significance of select echocardiographic indexes of RV function.19
To address these unmet needs and expand upon recent efforts in the REVEAL cohort, the National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI)-sponsored Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease (PVDOMICS)18 study conducted an in-depth analysis of participants and individuals with and at-risk for pulmonary vascular disease (PVD), incorporating detailed echocardiographic phenotyping. The current study had 3 primary objectives. First, we sought to identify the relationship between RV functional parameters from echocardiography and functional outcome measures in a large, well-phenotyped cohort of patients with incident and prevalent PAH. Second, we test the hypothesis that inclusion of echocardiographic indexes of RV function adds value to current PAH risk prediction models by using penalized Cox regression models. The third objective was to leverage cluster analysis as a clinically translatable tool for assessing RV dysfunction in Group 1 PAH risk stratification. Cluster analysis identifies distinct subgroups based on RV systolic function that may be overlooked by traditional risk scores, and it enables clinicians to tailor treatment strategies according to the degree of RV dysfunction, providing actionable insights even in the absence of external validation.
Study Design and Methods
Study Population
The PVDOMICS Network is a cross-sectional, observational study20 with incident and prevalent cases of PAH, those at risk of but without PVD (disease comparators), and healthy control participants recruited from 7 clinical sites across the United States with limited follow-up. Detailed inclusion and exclusion criteria for enrollment have been previously described and are provided in the supplemental methods.18 The PVDOMICS study obtained ethics approval from the institutional review boards at each enrollment site.
Risk Score Calculation
The COMPERA 2.0 score was calculated as previously described using World Health Organization functional class, N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels, and 6-minute walk distance (6MWD),8 according to the 4-strata model for patients with prevalent PAH. Risk was classified as follows: 1 point = low; 2 to 3 points = intermediate; and 4 points = high. The REVEAL 2.0 Lite score was calculated as previously described using NT-proBNP levels, 6MWD, World Health Organization functional class, systolic BP, heart rate, and estimated glomerular filtration rate.7 Risk was classified as follows: ≤ 5 points = low; 6 to 7 points = intermediate; and ≥ 8 points = high. The REVEAL 2.0 score and subsequently risk determination were calculated as described previously.6 All-cause hospitalization was not available and was excluded when calculating the REVEAL 2.0 score, and scores were calculated for patients with at least 7 of the features identified.
Echocardiographic Methods
Following calibration and standardization of ultrasound equipment, echocardiograms were obtained at each clinical site using a previously described standardized acquisition protocol21 and transferred to the PVDOMICS Echo Core Laboratory for masked analysis using a vendor-neutral analysis program (Syngo Dynamics, Siemens Medical). Detailed echocardiographic methods are provided in the supplement.
Statistical Approach
Baseline clinical and echocardiographic characteristics in study participants with prevalent and incident PAH are summarized as mean ± SD for normally distributed continuous variables, median (interquartile range) for non-normally distributed continuous variables, or percentages for categorical variables.
The cohorts were analyzed independently to identify differences in prognostic risk factors in incident vs prevalent PAH. Relationships between echocardiographic characteristics of RV function and NT-proBNP, 6MWD, COMPERA 2.0 Risk Score, the REVEAL and REVEAL Lite 2.0 Risk Scores, and mortality were determined by using multivariable linear regression, logistic regression, or Cox regression, respectively, adjusting for age, BMI, sex, race/ethnicity, and duration of PAH, where appropriate. For multivariable linear regression, the outcome variable was log transformed. Penalized (Elastic-Net) Cox regression models were generated to test the additive value of RV echocardiographic parameters to current gold standard risk scores and accuracy assessed by the C-statistic and time-varying area under the curve (AUC).
Cluster analysis (K-Means) was performed by using the following features: the ratio of basal RV to left ventricular diameter, echo-derived right atrial (RA) pressure, STE-derived 6-segment RV global longitudinal systolic strain (RVGLS), 3-segment RV free wall strain (RVFWS), and echocardiographic indices of RV to pulmonary artery (RV-PA) coupling (ie, tricuspid annular plane systolic excursion [TAPSE] to PA systolic pressure [PASP], fractional area change [FAC] to PASP, RVFWS/PASP, and RVGLS/PASP). The optimal number of clusters was determined by maximizing the silhouette score, which evaluates the ratio of intercluster to intracluster distances, and minimizing the Bayesian Information Criterion.
This analysis was conducted by using the sklearn Python package (version 3.8.12; Python Software Foundation) and identified 3 distinct groups as the most appropriate clustering solution. Group comparisons between clusters were performed with a Kruskal-Wallis or Fisher exact test, where appropriate. Multiple comparisons were accounted for using the Dunn test for post hoc analyses to adjust for pairwise differences while controlling for type I error.
Statistical analyses were performed by using Stata version 15.1 (StataCorp). Statistical significance was defined by a two-sided P value < .05. Detailed statistical methods are provided in the supplement.
Results
Population Characteristics
The final cohort consisted of 405 study participants, 336 of whom had prevalent PAH. Detailed clinical, echocardiographic, and invasive hemodynamic characteristics are provided in Table 1. The mean age was 53 ± 15 years, 97 (29%) were male, and 248 (74%) were White. Disease severity was reflected by > 30% of the cohort with intermediate to high risk as indexed by the COMPERA and REVEAL Lite 2.0 scores, a median NT-proBNP of 267 (111-924) pg/mL, and 31% of the cohort taking 3 or more PH medications. Results of pulmonary function testing were consistent with prevalent PVD, with reduced FEV1 and FVC of 75% ± 19% and 82% ± 18%, respectively, but a normal FEV1/FVC ratio. Invasive hemodynamics were consistent with PAH and PVD (mean PA pressure, 45 ± 14 mm Hg; pulmonary vascular resistance [PVR], 7 ± 4 Wood units) and RV dysfunction (RA pressure, 7 ± 5 mm Hg). Echocardiographic analyses also showed RV contractile dysfunction, as indicated by a mildly reduced TAPSE (19 ± 4 mm), impaired RV-PA coupling (TAPSE/PASP ratio 0.29 ± 0.13), and RA dilatation (RA indexed volume, 36 ± 19 mL/m2). Detailed clinical characteristics for the subgroup of patients with incident PAH as well as the combined cohort are provided in e-Table 1.
Table 1.
Baseline Clinical and Echocardiographic Characteristics of the Overall Cohort
| Characteristic | Incident + Prevalent PAH, n = 405 | Prevalent PAH, n = 336 | Incident PAH, n = 69 |
|---|---|---|---|
| Baseline clinical characteristics | |||
| Age, y | 54 ± 15 | 53 ± 15 | 62 ± 11 |
| PH disease duration, y | 3.9 (1.2-8.9) | 3.9 (1.2-8.9) | … |
| Male sex | 121 (30) | 97 (29) | 24 (36) |
| NYHA functional class | |||
| I | 37 (9) | 35 (10) | 2 (3) |
| II | 149 (37) | 133 (40) | 16 (24) |
| III | 193 (48) | 156 (46) | 37 (56) |
| IV | 19 (5) | 10 (4) | 9 (14) |
| Race | |||
| White | 297 (74) | 248 (74) | 49 (74) |
| Black | 49 (12) | 38 (11) | 11 (17) |
| Asian | 19 (5) | 18 (5) | 1 (2) |
| Othera | 37 (9) | 32 (10) | 5 (8) |
| BSA, m2 | 1.9 ± 0.3 | 1.9 ± 0.3 | 2.0 ± 0.3 |
| Systolic BP, mm Hg | 116 ± 18 | 113 (103-124) | 123 ± 20 |
| Diastolic BP, mm Hg | 69 ± 11 | 69 ± 10 | 73 ± 11 |
| Atrial fibrillation | 10 (2) | 9 (3) | 1 (2) |
| COMPERA score | |||
| 1 | 71 (18) | 67 (20) | 4 (6) |
| 2 | 161 (40) | 140 (42) | 21 (32) |
| 3 | 114 (28) | 94 (28) | 20 (15) |
| 4 | 55 (14) | 35 (10) | 20 (15) |
| REVEAL 2.0 score | |||
| Low (0-6) | 197 (49) | 175 (52) | 22 (33) |
| Intermediate (7-8) | 79 (20) | 64 (19) | 15 (24) |
| High (≥ 9) | 126 (31) | 97 (29) | 29 (44) |
| REVEAL Lite 2.0 score | |||
| Low (1-5) | 174 (43) | 153 (46) | 21 (32) |
| Intermediate (6-7) | 60 (15) | 52 (15) | 8 (12) |
| High (≥ 8) | 67 (17) | 54 (16) | 13 (20) |
| Laboratory values | |||
| Creatinine, mg/dL | 1.0 ± 0.5 | 0.9 (0.8, 1.1) | 1.0 ± 0.3 |
| NT-proBNP, pg/mL | 285 (111-1,050) | 267 (111-924) | 424 (115-2,165) |
| Pulmonary function testing | |||
| FEV1, % predicted | 75 ± 19 | 75 ± 19 | 75 ± 20 |
| FVC, % predicted | 81 ± 19 | 82 ± 18 | 80 ± 21 |
| FEV/FVC ratio, % predicted | 93 ± 11 | 93 (86, 99) | 94 ±1 3 |
| Dlco, % predicted | 54 ± 22 | 55 ± 22 | 45 ± 21 |
| 6MWD, feet | 380 ± 130 | 398 (313-467) | 322 ± 118 |
| PH medications | |||
| PDE5 inhibitors | 255 (63) | 243 (72) | 12 (18) |
| Endothelin receptor antagonists | 193 (48) | 188 (56) | 5 (8) |
| Prostacyclin | 152 (38) | 149 (44) | 3 (5) |
| Soluble guanylate cyclase | ... | 20 (6) | 0 (0) |
| PH calcium-channel blockers | 15 (4) | 13 (4) | 2 (3) |
| Non-PH calcium-channel blockers | 327 (81) | 296 (88) | 31 (47) |
| No. of PH medications | |||
| 1 | 82 (20) | 70 (24) | 7 (11) |
| 2 | 145 (36) | 139 (41) | 6 (9) |
| ≥ 3 | 80 (20) | 103 (31) | 1 (2) |
| Invasive hemodynamics | |||
| RA pressure, mm Hg | 7 ± 4 | 7 (4, 10) | 8 ± 6 |
| PASP, mm Hg | 71 ± 21 | 70 (54-86) | 68 ± 21 |
| Mean pulmonary artery pressure, mm Hg | 45 ± 14 | 44 (34-55) | 42 ± 13 |
| Pulmonary arterial wedge pressure, mm Hg | 11 ± 6 | 10 (7-14) | 10 ± 4 |
| Pulmonary vascular resistance, Wood units | 7 ± 5 | 6 (4-9) | 8 ± 5 |
| Cardiac output, L/min | 5.3 ± 1.7 | 5.1 (4.1-6.3) | 4.8 ± 1.6 |
| Cardiac index, L/min/m2 | 2.8 ± 0.9 | 2.7 (2.3-3.2) | 2.5 ± 0.8 |
| Pulmonary arterial saturation, % | 66 ± 10 | 68 (63-72) | 62 ± 10 |
| Echocardiographic parameters | |||
| Left ventricular ejection fraction, % | 61 ± 8 | 65 (65-60) | 58 ± 11 |
| RA major dimension, cm | 5.3 ± 0.8 | 5.2 (4.7-5.8) | 5.4 ± 0.8 |
| RA minor dimension, cm | 4.3 ± 0.9 | 4.2 (3.6-4.8) | 4.3 ± 0.9 |
| RA volume, mL | 67 ± 38 | 57 (41-82) | 69 ± 44 |
| RA volume index, mL/m2 | 36 ± 19 | 31 (22-45) | 36 ± 21 |
| FAC, % | 30 ± 10 | 31 (24-37) | 27 ± 10 |
| RV velocity time integral, cm | 14 ± 5 | 14 (12-17) | 12 ± 4 |
| TR grade | |||
| Trace | 100 (25) | 85 (25) | 15 (23) |
| Mild | 161 (40) | 132 (39) | 29 (44) |
| Moderate | 98 (24) | 85 (25) | 13 (20) |
| Severe | 20 (5) | 18 (5) | 2 (3) |
| TAPSE, mm | 19 ± 4 | 19 ± 4 | 17 ± 4 |
| RVGLS, % | –17 ± 5 | –17 ± 5 | –17 ± 10 |
| Echocardiographically derived PASP, mm Hg | 67 ± 22 | –19 ± 6 | 63 ± 18 |
| TAPSE/PASP ratio, mm/mm Hg | 0.26 (0.19, 0.37) | 68 ± 23 | 0.23 (0.18-0.34) |
| FAC/PASP ratio, %/mm Hg | 0.42 (0.28-0.65) | 0.26 (0.19-0.37) | 0.37 (0.24-0.64) |
| TDI Sʹ/PASP ratio, cm/s per mm Hg | 0.17 (0.12-0.23) | 0.42 (0.28-0.65) | 0.11 (0.08-0.16) |
| RVGLS/PASP ratio, %/mm Hg | 0.25 (0.17-0.37) | 0.18 (0.13-0.24) | 0.25 (0.18-0.37) |
Data are presented as mean ± SD or No. (%) for normally distributed continuous variables and categorical variables, respectively. Non-normally distributed or skewed distributions are presented as median (interquartile range). 6MWD = 6-minute walk distance; BSA = body surface area; COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; Dlco = diffusing capacity of the lung for carbon monoxide; FAC = fractional area change; NT-proBNP = N-terminal pro-B-type natriuretic peptide; NYHA = New York Heart Association; PAH = pulmonary arterial hypertension; PASP = pulmonary artery systolic pressure; PDE5 = phosphodiesterase type 5; PH = pulmonary hypertension; RA = right atrial; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RVGLS = right ventricular global longitudinal strain; TAPSE = tricuspid annular plane systolic excursion; TDI Sʹ = tissue Doppler imaging Sʹ velocity; TR = tricuspid regurgitation.
Other is non-White, non-Black, non-Asian on the Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease Phenomics coding sheet for race.
Relationship of RV Systolic Function and Functional Measures in Prevalent PAH
We evaluated the relationship between echocardiographic measures of RV function and RV-PA coupling and relevant functional outcomes (NT-proBNP and 6MWD) and risk scores (COMPERA, REVEAL Lite 2.0, and REVEAL 2.0). Regression analyses are presented in Table 2. A 1-unit decrease in most RV echocardiographic functional parameters was associated with a 30% to 84% increase in NT-proBNP or 1% to 6% reduction in 6MWD, with RVGLS exhibiting the strongest relationship (β = 110.5% [80.5, 145.4]). NT-proBNP and RVGLS/PASP showed the strongest relationship with 6MWD (β = 12.0% [5.8, 16.0]). Similarly, for both risk scores, a 1-unit decrease in any RV echocardiographic functional parameter yielded 1.4 to 2.1 odds for increased risk as indexed by the COMPERA score and 1.5 to 2.1 odds for increased risk by the REVEAL Lite 2.0 score. All associations remained significant after adjustment for age, sex, race, BMI, and PH duration, and if RV-PA coupling indexes were treated as continuous or dichotomized based on accepted criteria from prior invasive studies in a similar cohort of patients22, 23, 24, 25, 26 (e-Tables 2 and 3). Furthermore, associations remained significant after adjustment for RV basal diameter, suggesting that these associations are not simply a result of RV dilatation and remodeling (e-Table 4).
Table 2.
Relationship of RV Systolic Parameters With Clinical Outcome Measures in Study Participants With Prevalent PAH
| Parameter | Model 1 | P Value | Model 2 | P Value |
|---|---|---|---|---|
| NT-proBNP | ||||
| TAPSE, mm, n = 275 | –44.1 (–52.1 to –34.8) | <.001 | –43.9 (–51.7 to –34.7) | <.001 |
| RVFAC, %, n = 276 | –47.9 (–55.3 to –39.3) | <.001 | –53.4 (–59.4 to –46.4) | <.001 |
| TDI S', cm/s, n = 282 | –31.9 (–41.9 to –20.3) | <.001 | –29.6 (–39.9 to –17.4) | <.001 |
| Abs RVGLS, %, n = 235 | 83.2 (54.8 to 116.9) | <.001 | 110.5 (80.5 to 145.4) | <.001 |
| TAPSE/PASP, mm/mm Hg n = 274 | –49.9 (–56.9 to –41.8) | <.001 | –51.1 (–57.6 to –43.7) | <.001 |
| FAC/PASP, %/mm Hg n = 274 | –50.0 (–57.1 to –41.8) | <.001 | –54.3 (–60.2 to –47.6) | <.001 |
| TDI S'/PASP, cm/s × mm Hg n = 248 | –49.4 (–56.8 to –40.8) | <.001 | –50.1 (–57.4 to –41.6) | <.001 |
| Abs RVGLS/PASP, %/mm Hg n = 210 | –50.2 (–58.0 to –40.8) | <.001 | –54.1 (–60.8 to –46.3) | <.0001 |
| 6MWD | ||||
| TAPSE, mm, n = 260 | 3.5 (–1.5 to 8.8) | .17 | 7.5 (2.8 to 12.4) | .002 |
| RVFAC, %, n = 255 | 1.0 (–4.1 to 6.4) | .70 | 5.2 (0.4 to 10.2) | .03 |
| TDI S', cm/s, n = 264 | 1.8 (–3.1 to 7.0) | .47 | 3.5 (–1.1 to 8.3) | .14 |
| Abs RVGLS, %, n = 218 | –3.1 (–8.3 to 2.4) | .27 | –9.7 (–14.1 to –5.0) | <.001 |
| TAPSE/PASP, mm/mm Hg, n = 255 | 4.7 (–0.4 to 10.0) | .07 | 8.5 (3.8 to 1) | <.001 |
| FAC/PASP, %/mm Hg, n = 251 | 2.5 (–2.6 to 8.0) | .34 | 6.8 (2.0 to 11.9) | .005 |
| TDI S'/PASP, cm/s × mm Hg, n = 229 | 5.7 (0.2 to 11.4) | .04 | 9.1 (3.7 to 14.7) | .001 |
| Abs RVGLS/PASP, %/mm Hg, n = 193 | 5.5 (–0.9 to 12.4) | .09 | 12.0 (5.8 to 18.6) | <.001 |
| COMPERA 2.0 Risk Score | ||||
| TAPSE, mm, n = 288 | 0.60 (0.48 to 0.77) | <.001 | 0.50 (0.39 to 0.63) | <.001 |
| RVFAC, %, n = 289 | 0.61 (0.49 to 0.76) | <.001 | 0.48 (0.38 to 0.61) | <.001 |
| TDI S', cm/s, n = 295 | 0.69 (0.55 to 0.85) | .001 | 0.63 (0.50 to 0.80) | <.001 |
| Abs RVGLS, %, n = 243 | 1.74 (1.37 to 2.23) | <.001 | 2.55 (1.92 to 3.38) | <.001 |
| TAPSE/PASP, mm/mm Hg, n = 283 | 0.56 (0.44 to 0.71) | <.001 | 0.45 (0.35 to 0.58) | <.001 |
| FAC/PASP, %/mm Hg, n = 285 | 0.57 (0.46 to 0.72) | <.001 | 0.44 (0.34 to 0.57) | <.001 |
| TDI S'/PASP, cm/s × mm Hg, n = 258 | 0.48 (0.37 to 0.62) | <.001 | 0.43 (0.33 to 0.57) | <.001 |
| Abs RVGLS/PASP, %/mm Hg, n = 217 | 0.52 (0.40 to 0.69) | <.001 | 0.37 (0.27 to 0.52) | <.001 |
| REVEAL Lite 2.0 risk score | ||||
| TAPSE, mm, n = 230 | 0.56 (0.44 to 0.72) | <.001 | 0.45 (0.35 to 0.58) | <.001 |
| RVFAC, %, n = 230 | 0.66 (0.52 to 0.82) | <.001 | 0.47 (0.36 to 0.60) | <.001 |
| TDI S', cm/s, n = 234 | 0.66 (0.52 to 0.84) | .001 | 0.63 (0.49 to 0.81) | <.001 |
| Abs RVGLS, %, n = 200 | 1.58 (1.24 to 2.02) | <.001 | 2.68 (2.02 to 3.55) | <.001 |
| TAPSE/PASP, mm/mm Hg, n = 229 | 0.53 (0.41 to 0.67) | <.001 | 0.40 (0.31 to 0.52) | <.001 |
| FAC/PASP, %/mm Hg, n = 228 | 0.58 (0.46 to 0.74) | <.001 | 0.41 (0.32 to 0.53) | <.001 |
| TDI S'/PASP, cm/s × mm Hg, n = 206 | 0.48 (0.36 to 0.62) | <.001 | 0.41 (0.30 to 0.56) | <.001 |
| Abs RFGLS/PASP, %/mm Hg, n = 178 | 0.56 (0.42 to 0.74) | <.001 | 0.35 (0.25 to 0.49) | <.001 |
| REVEAL 2.0 risk score | ||||
| TAPSE, mm, n = 288 | 0.52 (0.42 to 0.64) | <.001 | 0.46 (0.36 to 0.57) | <.001 |
| RVFAC, %, n = 289 | 0.59 (0.48 to 0.72) | <.001 | 0.45 (0.36 to 0.57) | <.001 |
| TDI S', cm/s, n = 295 | 0.65 (0.53 to 0.79) | <.001 | 0.63 (0.51 to 0.78) | <.001 |
| Abs RVGLS, %, n = 243 | 1.85 (1.47 to 2.33) | <.001 | 2.64 (2.06 to 3.39) | <.001 |
| TAPSE/PASP, mm/mm Hg, n = 283 | 0.45 (0.39 to 0.61) | <.001 | 0.36 (0.28 to 0.46) | <.001 |
| FAC/PASP, %/mm Hg, n = 285 | 0.48 (0.39 to 0.61) | <.001 | 0.36 (0.28 to 0.46) | <.001 |
| TDI S'/PASP, cm/s × mm Hg, n = 258 | 0.44 (0.34 to 0.55) | <.001 | 0.40 (0.31 to 0.52) | <.001 |
| Abs RVGLS/PASP, %/mm Hg, n = 217 | 0.52 (0.41 to 0.67) | <.001 | 0.36 (0.27 to 0.49) | <.001 |
Model 1 is unadjusted, and model 2 is adjusted for age, BMI, sex, race/ethnicity, and duration of PAH. Coefficients shown are percentage for continuous outcomes and ORs for categorical outcomes. P values and 95% CIs are shown. COMPERA and REVEAL risk scores were categorized as low risk (COMPERA, 1-3; REVEAL ≤ 5), intermediate risk (COMPERA, 4-10; REVEAL, 6-7), or high risk (COMPERA, 11-12; REVEAL, ≥ 8). Abs RVGLS is the absolute value of RVGLS, as this is a negative number. Bolded font signifies statistical significance. 6MWD = 6-minute walk distance; COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; FAC = fractional area change; NT-proBNP = N-terminal pro-B-type natriuretic peptide; PAH = pulmonary arterial hypertension; PASP = pulmonary artery systolic pressure; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RVFAC = right ventricular fractional area change; RVGLS = right ventricular global longitudinal strain; TAPSE = tricuspid annular plane systolic excursion.
In addition, RV function was worse in patients with high risk according to COMPERA, REVEAL Lite 2.0, and REVEAL 2.0 (Table 3) scores. RVGLS was –14% ± 4% in patients with a COMPERA score > 10 (high risk) vs –20% ± 4% in patients with a COMPERA score < 4. A similar phenomenon was observed for the REVEAL Lite 2.0 score (low risk, –18% ± 4% vs –15% ± 4%; Dunn test, P = .0001). The same analyses were performed in incident PAH as well as the combined cohort and are presented in e-Tables 2-5 and demonstrated similar findings to those in prevalent PAH. Details are provided in the supplement.
Table 3.
Comparative Analysis of RV Functional Parameters Across PAH Risk Score Categories
| Parameter | Low Risk (1-3) | Low-Medium Risk (4-7) | Medium-High Risk (8-10) | High Risk (11-12) | P Value |
|---|---|---|---|---|---|
| COMPERA 2.0 | |||||
| TAPSE, mm | 21 ± 4 | 19 ± 4 | 18 ± 5 | 16 ± 4 | .0001 |
| RVFAC, % | 35 ± 8 | 30 ± 9 | 29 ± 11 | 25 ± 8 | .0001 |
| TDI S', cm/s | 13 ± 2 | 12 ± 3 | 11 ± 3 | 11 ± 3 | .0006 |
| RVGLS, % | –20 ± 4 | –17 ± 4 | –17 ± 5 | –14 ± 4 | .0001 |
| TAPSE/PASP, mm/mm Hg | 0.35 ± 0.14 | 0.30 ± 0.12 | 0.27 ± 0.14 | 0.21 ± 0.08 | .0001 |
| FAC/PASP, %/mm Hg | 0.61 ± 0.27 | 0.48 ± 0.23 | 0.33 ± 0.27 | 0.33 ± 0.18 | .0001 |
| TDI S'/PASP, cm/s × mm Hg | 0.25 ± 0.10 | 0.20 ± 0.08 | 0.17 ± 0.08 | 0.14 ± 0.05 | .0001 |
| Abs RVGLS/PASP, %/mm Hg | 0.39 ± 0.18 | 0.30 ± 0.16 | 0.26 ± 0.14 | 0.19 ± 0.07 | .0001 |
| REVEAL Lite 2.0 | Low Risk (≤ 5) | Intermediate Risk (6-7) | High Risk (≥ 8) | P Value |
|---|---|---|---|---|
| TAPSE, mm | 20 ± 4 | 19 ± 5 | 16 ± 4 | .0001 |
| RVFAC, % | 32 ± 9 | 30 ± 12 | 27 ± 10 | .005 |
| TDI S', cm/s | 12 ± 2 | 12 ± 3 | 11 ± 3 | .02 |
| RVGLS, % | –18 ± 4 | –17 ± 6 | –15 ± 4 | .002 |
| TAPSE/PASP, mm/mm Hg | 0.32 ± 0.13 | 0.29 ± 0.15 | 0.21 ± 0.08 | .0001 |
| FAC/PASP, %/mm Hg | 0.54 ± 0.26 | 0.46 ± 0.29 | 0.36 ± 0.19 | .0001 |
| TDI S'/PASP, cm/s × mm Hg | 0.22 ± 0.09 | 0.19 ± 0.09 | 0.14 ± 0.06 | .0001 |
| Abs RVGLS/PASP, %/mm Hg | 0.32 ± 0.16 | 0.28 ± 0.16 | 0.20 ± 0.09 | .0001 |
| REVEAL 2.0 | Low Risk (≤ 6) | Intermediate Risk (7-8) | High Risk (≥ 9) | P Value |
|---|---|---|---|---|
| TAPSE, mm | 20 ± 4 | 19 ± 4 | 16 ± 4 | .0001 |
| RVFAC, % | 32 ± 9 | 30 ± 10 | 27 ± 10 | .0009 |
| TDI S', cm/s | 12 ± 2 | 11 ± 3 | 11 ± 3 | .0008 |
| RVGLS, % | –18 ± 4 | –17 ± 5 | –15 ± 5 | .0001 |
| TAPSE/PASP, mm/mm Hg | 0.33 ± 0.13 | 0.27 ± 0.12 | 0.23 ± 0.11 | .0001 |
| FAC/PASP, %/mm Hg | 0.54 ± 0.26 | 0.44 ± 0.24 | 0.39 ± 0.23 | .0001 |
| TDI S'/PASP, cm/s × mm Hg | 0.22 ± 0.09 | 0.17 ± 0.08 | 0.16 ± 0.08 | .0001 |
| Abs RVGLS/PASP, %/mm Hg | 0.34 ± 0.20 | 0.27 ± 0.14 | 0.22 ± 0.11 | .001 |
Data are presented as mean ± SD or No. (%) for normally distributed continuous variables and categorical variables, respectively. Non-normally distributed or skewed distributions are presented as median (interquartile range). Ordinality was considered when conducting hypothesis tests. Overall model significance was assessed with a likelihood ratio test. Individual predictors were compared with Wald tests based on the asymptotic normality of maximum likelihood estimates. Bolded font signifies statistical significance. COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; FAC = fractional area change; PAH = pulmonary arterial hypertension; PASP = pulmonary artery systolic pressure; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RVFAC = right ventricular fractional area change; RVGLS = right ventricular global longitudinal strain; TAPSE = tricuspid annular plane systolic excursion.
Echocardiographic Predictors of Clinical Outcomes in Patients With Prevalent PAH
Kaplan-Meier analysis was next performed to identify echocardiographic predictors of all-cause mortality (Table 4). A total of 49 deaths were observed with a follow-up time of 51 months. Cox regression adjusted for age, BMI, race, sex, and PH duration revealed that all RV echocardiographic parameters, with the exception of tissue Doppler imaging Sʹ velocity (TDI Sʹ), was significantly associated with all-cause mortality. In all cases, indexes of RV-PA coupling had the strongest association (TAPSE/PASP hazard ratio [HR], 0.05 [0.01, 0.30]; FAC/PASP HR, 0.22 [0.08, 0.59]; TDI Sʹ velocity/PASP, HR 0.08 [0.02, 0.45]; and RVGLS/PASP HR, 0.15 [0.04, 0.55]).
Table 4.
Association of RV Echocardiographic Parameters and Mortality in Prevalent PAH
| Parameter | Model 1 |
Model 2 |
||
|---|---|---|---|---|
| Hazard Ratio | P Value | Hazard Ratio | P Value | |
| TAPSE, mm | 0.91 (0.84-0.98) | .02 | 0.91 (0.85-0.99) | .03 |
| RVFAC, % | 0.97 (0.94-0.99) | .03 | 0.95 (0.93-0.98) | .002 |
| TDI S', cm/s | 0.97 (0.86-1.08) | .56 | 0.98 (0.87-1.11) | .79 |
| RVGLS, % | 1.11 (1.03-1.19) | .004 | 1.15 (1.08-1.24) | <.001 |
| TAPSE/PASP, mm/mm Hg | 0.05 (0.01-0.30) | .001 | 0.02 (0.01-0.13) | <.001 |
| FAC/PASP, %/mm Hg | 0.22 (0.08-0.59) | .003 | 0.08 (0.03-0.24) | <.001 |
| TDI S'/PASP, cm/s per mm Hg | 0.08 (0.02-0.45) | .004 | 0.04 (0.01-0.26) | .001 |
| Abs RVGLS/PASP, %/mm Hg | 0.15 (0.04-0.55) | .004 | 0.04 (0.01-0.19) | <.001 |
Model 1 is unadjusted. Model 2 is adjusted for age, BMI, sex, race, and PH duration. Non-normally distributed variables were log transformed. Results were back transformed for reported hazard ratio. Bolded font signifies statistical significance. FAC = fractional area change; PAH = pulmonary arterial hypertension; PASP = pulmonary artery systolic pressure; RV = right ventricular; RVFAC = right ventricular fractional area change; RVGLS = right ventricular global longitudinal strain; TAPSE = tricuspid annular plane systolic excursion.
We next stratified survival curves based on accepted cutoffs for RV-PA uncoupling based on results of prior invasive studies.22, 23, 24, 25, 26, 27, 28 The cutoffs used were as follows: TAPSE/PASP < 0.31 mm/mm Hg; FAC/PASP < 0.71 mm Hg–1; TDI Sʹ/PASP < 0.18 cm/s per mm Hg; |RVFWS|/PASP < 0.17 mm Hg–1 (Fig 1). In all cases, uncoupled PAH patients had between 23% and 53% greater reduction in survival than coupled patients with PAH (log rank test, P < .0001 for all), which persisted after adjustment for age, sex, race, BMI, and PH duration. Moreover, patients with PAH and reduced RV-PA coupling exhibited a reduction in survival similar to that of patients with high-risk PAH as quantified by the COMPERA REVEAL 2.0 and REVEAL Lite 2.0 scores (e-Fig 1). Associations between RV-PA coupling ratios and outcomes persisted after adjustment for RV basal diameter, suggesting that unlike other RV contractile indices, the prognostic significance of echocardiographic coupling ratios is independent of changes in RV morphology (e-Table 6). As with our prior functional analysis, assessment of clinical outcomes was performed in the combined cohort and the subset of patients with incident PAH. As was true in patients with prevalent PAH, only RV-PA coupling ratios were associated with outcomes in patients with incident PAH after adjustment for age, sex, race, and RV basal diameter.
Figure 1.
RV-pulmonary artery uncoupling associates with worse survival in prevalent pulmonary hypertension. Echocardiographic indexes of RV-pulmonary artery coupling were dichotomized based on accepted cutoffs for uncoupling based on prior invasive studies. These are: TAPSE/PASP < 0.31 (A); FAC/PASP < 0.71 (B); TDI Sʹ/PASP < 0.18 (C); and RV FWS/PASP < 0.17 (D). The primary end point was all-cause mortality. P values are from Cox regression, adjusted for age, sex, race, BMI, and duration of pulmonary hypertension. PASP = pulmonary artery systolic pressure; RVFAC = right ventricular fractional area change; TAPSE = tricuspid annular plane systolic excursion; TDI Sʹ = tissue Doppler imaging Sʹ velocity.
Echocardiographic RV Function Parameters Improve Prognostication of Outcomes in Prevalent PH
We next tested whether inclusion of RV functional parameters from echocardiograms improve risk stratification from REVEAL 2.0, REVEAL Lite 2.0, and COMPERA. Time-varying AUC and the C-statistic data are shown in Table 5. The COMPERA, REVEAL Lite 2.0, and REVEAL 2.0 scores had a C-statistic of 0.78, 0.73, and 0.82, respectively, for predicting all-cause mortality. A total of 14 echocardiographic variables of RV function (e-Table 7) were incorporated into the risk prediction models that used all variables in the COMPERA or REVEAL risk scores as continuous variables. To prevent overfitting, a penalized Cox regression model was used. We first trained our models using the score as the only feature, which identified a C-statistic of 0.78, 0.73, and 0.82 for the COMPERA, REVEAL Lite 2.0, and REVEAL 2.0 risk scores for predicting all-cause mortality.
Table 5.
Effect of Echocardiographic Parameters of RV Function on the Prognostic Capabilities of the REVEAL and COMPERA Scores
| Parameter | C-Statistic | Average Time-Varying AUC |
|---|---|---|
| COMPERA Risk score | 0.78 | 0.82 |
| REVEAL 2.0 Lite Risk score | 0.73 | 0.81 |
| REVEAL 2.0 Risk score | 0.82 | 0.83 |
| COMPERA variables | 0.77 | 0.80 |
| REVEAL 2.0 Lite variables | 0.76 | 0.82 |
| REVEAL 2.0 variables | 0.81 | 0.85 |
| COMPERA variables + RV echo | 0.84 | 0.85 |
| REVEAL 2.0 Lite variables + RV echo | 0.88 | 0.89 |
| REVEAL 2.0 variables + RV echo | 0.83 | 0.88 |
Penalized Cox regression models (ElasticNet) were generated to avoid overfitting. C-statistic and average time-varying AUC are shown. A list of echocardiographic parameters of RV function is included in the Supplement. AUC = area under the curve; COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RV echo = right ventricular echocardiographic variables.
Scores alone are insufficient to statistically show prognostic improvement with incorporation of RV echocardiographic variables. Therefore, to test whether RV echocardiographic indexes enhance PAH risk prediction, each score variable was treated as an independent feature, and a penalized Cox regression model was trained to prevent overfitting. Penalized Cox regression takes all features as inputs as defined in e-Table 7 but penalizes large HRs, thereby forcing features to a regression coefficient of 0 (HR of 1). Features with non-zero coefficients for each of the models are shown in Figure 2. Incorporating RV functional variables improved the C-statistic by 7% and 12%, and average time-varying AUC improved by 5% and 7% when added to models generated by using the components of the COMPERA and REVEAL Lite 2.0 scores as independent features and continuous variables, respectively. When using features from the REVEAL 2.0 risk calculator, the C-statistic and average time-varying AUC improved by 5% and 3%. Incorporation of RV echocardiographic indices into the REVEAL 2.0 model resulted in minimal improvement in the C-statistic (0.83 vs 0.81) and modest improvement in the time-varying AUC (0.88 vs 0.85). Notably, the REVEAL 2.0 Lite + RV echocardiographic model performed similarly to the REVEAL 2.0 + RV echocardiographic model (Fig 3).
Figure 2.
Variable importance from penalized Cox regression models in prevalent pulmonary arterial hypertension. Penalized Cox regression models were generated based on a combination of variables in the COMPERA (A) and REVEAL Lite 2.0 (B) scores and RV Echo variables, and the REVEAL 2.0 scores and RV Echo variables (C). COMPERA variables include: 6MWD, NT-proBNP, and New York Heart Association functional class. REVEAL Lite 2.0 variables include: 6MWD, systolic BP, eGFR, heart rate, NT-proBNP, and eGFR functional class. Echocardiographic variables include RVFAC, echocardiogram-derived PASP (RVPS), RVOT VTI, echocardiogram-derived RAP, RV global longitudinal systolic strain, right ventricular free wall strain, TAPSE, tissue Doppler imaging Sʹ velocity, TAPSE/PASP ratio, FAC/PASP ratio, right ventricular free wall strain/PASP ratio, tissue Doppler imaging Sʹ velocity/PAPS ratio, RV basal diameter, and RAVi. Relative importance was calculated by mean and variance standardization of variables prior to fitting the penalized Cox regression model. 6MWD = 6-minute walk distance; COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; Dlco = diffusing capacity of the lung for carbon monoxide; eGFR = estimated glomerular filtration rate; NT-proBNP = N-terminal pro-B-type natriuretic peptide; PASP = pulmonary artery systolic pressure; RAP = right atrial pressure; RAVi = right atrial volume index; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RV Echo = right ventricular echocardiograph; RVFWS = right ventricular free wall strain; RVGLS = right ventricular global longitudinal strain; RVOT VTI = right ventricular outflow tract velocity time integral; RVSP = right ventricular systolic pressure; TAPSE = tricuspid annular plane systolic excursion.
Figure 3.
Cluster analysis identifies 3 groups of RVD. A, Normalized heatmap stratified according to agnostically identified clusters. Blue corresponds to a lower value and red a higher value. Compared with the mild RVD group, patients with severe RVD had reduced strain, reduced right ventricular-pulmonary artery coupling ratios, and right atrial dilatation. The moderate RVD group had preserved strain and normal RV geometry but also reduced right ventricular-pulmonary artery coupling. B, Kaplan-Meier analysis stratified according to clusters of right ventricular function. The overall P value is from a Cox regression adjusting for age, sex, race, BMI, and duration of pulmonary hypertension. C, Outcome measures stratified according to different clusters. ∗ refers to statistical significance between groups. 6MWD = 6-minute walk distance; FAC = fractional area change; NT-proBNP = N-terminal pro-B-type natriuretic peptide; PASP = pulmonary artery systolic pressure; RAVi = right atrial volume index; RVD = right ventricular dysfunction; RVFWS = right ventricular free wall strain; TAPSE = tricuspid annular plane systolic excursion.
These findings are clinically significant, as a 1-unit increase in the penalized Cox regression risk score reflects a 100% increase in the odds of the outcome compared to using COMPERA or REVEAL scores alone. The relative importance of all variables used in the model was assessed by standardizing and taking the absolute value of the regression coefficients (Table 2). Non-echocardiographic features that retained significance across models included 6MWD, NT-proBNP, age, and estimated glomerular filtration rate. Prognostically significant echocardiographic indices included echocardiographically derived pressures, RV-PA coupling ratios, and RA volume.
Cluster Analysis Identifies Mild, Moderate, and Severe Subgroups of RV Dysfunction
Recognizing the supervised nature of risk prediction models and the absence of external validation, we next applied unsupervised cluster analysis (Supplemental Methods) to stratify 336 patients in the prevalent PAH cohort into 3 agnostic groups: 92 (27%) with mild, 148 (44%) with moderate, and 96 (29%) with severe RV dysfunction. Table 6 summarizes baseline clinical and echocardiographic characteristics between these groups. Patients with RV dysfunction had the lowest percent predicted FEV1 (severe 70% ± 15% vs mild 79 ± 20%; P = .004) and FVC (severe 77% ± 16% vs mild 86% ±18%; P = .004). When combined with a normal FEV1/FVC ratio, these findings suggest severe PVD. Invasive hemodynamic studies further revealed that these patients had the highest filling pressures. Moreover, patients with severe RV dysfunction also had the highest mortality (log-rank test, P < .0001) (Fig 3), which remained significant following adjustment for age, sex, race, BMI, creatinine, PH duration, invasive PVR, and the COMPERA score (P = .03) but not the REVEAL score (P = .26). Leveraging this analysis, a novel calculator (Supplement) was derived, with applications in both clinical and research settings. Notably, the calculator as derived does not require measurement of all functional parameters.
Table 6.
Baseline Clinical and Echocardiographic Characteristics of Prevalent PAH by Groups of RVD
| Characteristic | Mild RVD (n = 92) | Moderate RVD (n = 148) | Severe RVD (n = 96) | P Value |
|---|---|---|---|---|
| Baseline clinical characteristics | ||||
| Age, y | 56 ± 14 | 51 ± 14 | 52 ± 15 | .05 |
| PH disease duration at enrollment, y | 3.4 (1.1-8.7) | 3.9 (1.7-9.1) | 3.9 (1.1-8.2) | .47 |
| Male sex | 67 (73) | 102 (69) | 70 (73) | .74 |
| NYHA functional class | .01 | |||
| I | 14 (15) | 14 (9) | 7 (7) | |
| II | 41 (45) | 56 (38) | 36 (38) | |
| III | 37 (40) | 74 (50) | 45 (47) | |
| IV | 0 (0) | 2 (1) | 8 (8) | |
| Race | .81 | |||
| White | 70 (76) | 104 (70) | 74 (77) | |
| Black | 8 (9) | 20 (14) | 10 (10) | |
| Asian | 5 (5) | 10 (7) | 3 (3) | |
| Othera | 9 (10) | 14 (9) | 9 (9) | |
| BSA, m2 | 1.9 ± 0.3 | 1.9 ± 0.3 | 1.9 ± 0.3 | .23 |
| Systolic BP, mm Hg | 116 ± 14 | 116 ± 18 | 111 ± 16 | .01 |
| Diastolic BP, mm Hg | 67 ± 11 | 70 ± 10 | 68 ± 10 | .32 |
| Atrial fibrillation | 1 (1) | 3 (2) | 5 (5) | .25 |
| COMPERA score | <.001 | |||
| 1 | 31 (34) | 33 (22) | 3 (3) | |
| 2 | 40 (43) | 70 (47) | 30 (3) | |
| 3 | 20 (22) | 36 (24) | 38 (40) | |
| 4 | 1 (1) | 9 (6) | 25 (26) | |
| REVEAL Lite 2.0 score, n (%) | <.001 | |||
| Low (1-5) | 50 (54) | 81 (55) | 3 (3) | |
| Intermediate (6-7) | 18 (20) | 13 (9) | 18 (19) | |
| High (≥ 8) | 3 (3) | 18 (12) | 33 (34) | |
| Laboratory values | ||||
| Creatinine, mg/dL | 0.9 ± 0.2 | 1.0 ± 0.4 | 1.1 ± 0.8 | .004 |
| NT-proBNP, pg/mL | 120 (58, 263) | 172 (96, 464) | 1104 (535, 2869) | .0001 |
| Pulmonary function testing | ||||
| FEV1, % predicted | 79 ± 20 | 76 ± 19 | 70 ± 15 | .004 |
| FVC, % predicted | 86 ± 18 | 82 ± 19 | 77 ± 16 | .004 |
| FEV/FVC ratio, % predicted | 92 ± 13 | 93 ± 10 | 90 ± 10 | .20 |
| Dlco, % predicted | 55 ± 20 | 57 ± 22 | 53 ± 22 | .40 |
| 6MWD, feet | 402 ± 141 | 402 ± 122 | 355 ± 127 | .02 |
| PH medications | ||||
| PDE5 inhibitors | 68 (74) | 109 (74) | 66 (69) | .57 |
| Endothelin receptor antagonists | 54 (59) | 79 (53) | 55 (57) | .73 |
| Prostacyclin | 31 (34) | 66 (38) | 62 (57) | <.001 |
| PH calcium-channel blockers | 3 (3) | 7 (4) | 3 (3) | .82 |
| Non-PH calcium-channel blockers | 81 (88) | 129 (87) | 86 (90) | .98 |
| No. of PH medications | .06 | |||
| 1 | 12 (13) | 17 (11) | 8 (8) | |
| 2 | 22 (24) | 34 (23) | 14 (15) | |
| ≥ 3 | 57 (62) | 95 (64) | 74 (77) | |
| Invasive hemodynamics | ||||
| RA pressure, mm Hg | 5 ± 4 | 7 ± 4 | 10 ± 5 | .0001 |
| PASP, mm Hg | 54 ± 13 | 74 ± 20 | 85 ± 18 | .0001 |
| Mean pulmonary artery pressure, mm Hg | 34 ± 7 | 47 ± 13 | 54 ± 12 | .0001 |
| Pulmonary arterial wedge pressure, mm Hg | 10 ± 5 | 12 ± 6 | 12 ± 6 | .01 |
| Pulmonary vascular resistance, Wood units | 4 ± 2 | 7 ± 4 | 10 ± 5 | .0001 |
| Cardiac output, L/min | 5.7 ± 1.5 | 5.5 ± 1.8 | 4.7 ± 1.8 | .0001 |
| Cardiac index, L/min/m2 | 3.0 ± 0.7 | 2.9 ± 0.9 | 2.5 ± 0.9 | .0001 |
| Echocardiographic parameters | ||||
| Left ventricular ejection fraction, % | 63 ± 5 | 61 ± 7 | 62 ± 10 | .10 |
| RA major dimension, cm | 5.0 ± 0.6 | 5.0 ± 0.7 | 5.8 ± 0.9 | .0001 |
| RA minor dimension, cm | 3.8 ± 0.7 | 3.9 ± 0.6 | 5.0 ± 0.8 | .0001 |
| RA volume, mL | 50 ± 26 | 52 ± 19 | 101 ± 42 | .0001 |
| RA volume index, mL/m2 | 27 ± 12 | 28 ± 9 | 55 ± 21 | .0001 |
| FAC, % | 40 ± 7 | 30 ± 6 | 22 ± 8 | .0001 |
| RV velocity time integral, cm | 17 ± 5 | 15 ± 5 | 13 ± 4 | .0001 |
| TR grade | <.001 | |||
| Trace | 39 (42) | 40 (27) | 6 (6) | |
| Mild | 44 (48) | 55 (37) | 33 (34) | |
| Moderate | 7 (8) | 35 (24) | 43 (45) | |
| Severe | 1 (1) | 3 (2) | 14 (15) | |
| TAPSE, mm | 22 ± 4 | 19 ± 4 | 15 ± 3 | .0001 |
| RVFWS, % | –25 ± 4 | –18 ± 4 | –13 ± 4 | .0001 |
| RVGLS, % | –22 ± 3 | –17 ± 3 | –13 ± 3 | .0001 |
| Echo-derived PASP, mm Hg | 52 ± 14 | 66 ± 22 | 83 ± 20 | .0001 |
| TAPSE/PASP ratio, mm/mm Hg | 0.44 ± 0.11 | 0.27 ± 0.08 | 0.19 ± 0.1 | .0001 |
Data are presented as mean ± SD or No. (%) for normally distributed continuous variables and categorical variables, respectively. Nonnormally distributed or skewed distributions are presented as median (interquartile range). Bolded font signifies statistical significance. 6MWD = 6-minute walk distance; BSA = body surface area; COMPERA = Comparative Prospective Registry for Newly Initiated Therapies; Dlco = diffusing capacity of the lung for carbon monoxide; FAC = fractional area change; NT-proBNP = N-terminal pro-B-type natriuretic peptide; NYHA = New York Heart Association; PAH = pulmonary arterial hypertension; PASP = pulmonary artery systolic pressure; PDE5 = phosphodiesterase type 5; PH = pulmonary hypertension; RA = right atrial; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; RV = right ventricular; RVD = right ventricular dysfunction; RVFWS = right ventricular free wall strain; RVGLS = right ventricular global longitudinal strain; TAPSE = tricuspid annular plane systolic excursion; TR = tricuspid regurgitation.
Other is non-White, non-Black, non-Asian on the Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease Phenomics coding sheet for race.
Discussion
In this large, multicenter cohort of PH patients, we demonstrate the prognostic utility of echocardiographic indexes of RV function beyond traditional risk prediction scores in PAH. We show that incorporating RV echocardiographic functional indexes into existing PH risk scores, such as COMPERA and REVEAL Lite 2.0, significantly enhanced their prognostic utility. In addition, agnostic unsupervised clustering based on RV function identified a subgroup of patients with severe RV dysfunction and adverse outcomes, even after adjusting for PH duration, current prognostic risk scores, and echocardiographically derived PVR. These findings underscore the critical role of RV contractile reserve and adaptive response to increased afterload in determining clinical outcomes in patients with PH. Furthermore, these results strongly support routine longitudinal echocardiographic monitoring of RV function and the rederivation of risk scores to integrate these indexes, offering a more comprehensive and clinically actionable framework for PH risk stratification.
The 2022 European Society of Cardiology/European Respiratory Society recommend echocardiographic evaluation of RV function in Group 1 PAH, emphasizing that RV afterload alone does not reflect disease progression.9,10 Despite established clinical significance, no studies have specifically delineated the relationship between a single RV functional echocardiographic parameter and functional outcomes, such as 6MWD, NT-proBNP, and PH risk. Although these associations are well characterized for the left ventricle,12,29 particularly in post-capillary PH, they remain poorly defined for the right ventricle, especially in the context of PAH, limiting the interpretation of impaired RV function in this population. The current study identified key relationships between echocardiographic indexes of RV function and clinically relevant outcomes measures in prevalent PAH. Notably, among all parameters of RV function, an increase (less negative) in STE-derived strain and a reduction in RV-PA coupling ratios emerged as the most sensitive predictor, consistent with previous reports.23,30, 31, 32, 33, 34 The current findings corroborate prior observations in both prevalent and incident PAH, reinforcing the broader application of echocardiographic phenotyping for both therapeutic monitoring and identifying patients at risk for PH. Furthermore, our use of advanced statistical modeling with penalized Cox regressions provides a practical methodology for integrating echocardiographic measures of RV function into traditional PH risk scores.
Prior studies have consistently shown that echocardiographic phenotyping alone is suboptimal for the diagnosis and assessment of therapeutic response in PH. However, noninvasive assessment of RV function significantly enhances PH risk prognostication when combined with hemodynamic and functional evaluations.16,27,31,32,35 The current findings align with these observations, supporting the expanded use of echocardiographic phenotyping to identify at-risk patients and monitor therapeutic response in both prevalent and incident PH. Moreover, our application of penalized Cox regression provides a practical approach for integrating echocardiographic measures of RV function into traditional PH risk scores. Although established risk scores such as COMPERA and REVEAL have proven prognostic value,7,8 adding echocardiographic indexes to existing risk scores is statistically inappropriate due to different training sets being used to develop these scores and identify relevant RV parameters.
To mitigate this concern, we developed new models using the current cohort, incorporating echocardiographic indexes and continuous variables from the REVEAL Lite 2.0 or COMPERA scores. To prevent overfitting, penalized Cox regression was used, ensuring generalizability and improved performance over the individual REVEAL and COMPERA scores. Prognostically significant variables included RV-PA coupling ratios, echocardiogram-derived PVR, and echocardiographic indexes of adverse ventricular remodeling and congestion, each known to associate with clinical outcomes in PH.16,27,28,30,33,36, 37, 38 Moreover, penalized Cox regression showed that the REVEAL 2.0 and REVEAL 2.0 Lite models performed similarly, supporting prior observations that REVEAL 2.0 Lite accurately identifies risk groups comparable to those derived from the full REVEAL 2.0 model.7 The improvement observed with the addition of echocardiographic indices in both models, suggests some degree of over-parametrization in the REVEAL 2.0 score. These findings underscore the importance of using robust statistical techniques to prevent overfitting as statistical models become increasingly complex. We implemented such an approach here, in which the incorporation of echocardiographic indexes of RV function significantly enhances the prognostic capabilities of all PAH risk prediction models tested.
A significant limitation of using binary cutoffs for RV dysfunction for echocardiographic functional parameters have limited prognostic and therapeutic value in PH, particularly when used in isolation, as these parameters are load dependent. To address this limitation, we performed cluster analysis, identifying mild, moderate, and severe RV dysfunction using a combination of echocardiographic indexes. Our approach offers a more comprehensive and clinically useful method of phenotyping compared with the use of binary cutoffs. Societal guidelines recommend a similar holistic approach to RV phenotyping that integrates multiple RV functional parameters rather than relying on a single echocardiographic measure.11,12 In the specific context of PAH, using standard cutoffs for right heart assessment as outlined in guidelines would likely categorize the majority of patients as intermediate risk,39, 40, 41 which limits its clinical utility for precise risk prognostication.
The cluster analysis presented here mitigates several of these concerns by incorporating a combination of echocardiographic parameters rather than relying on single binary values for assessing RV function. Furthermore, by not relying on the acquisition of all indexes used for risk stratification, we are able to define a clear signature distinguishing intermediate from severe RV dysfunction based on echocardiographic parameters, enhancing the granularity of risk assessment. Specifically, we observe that RV-PA uncoupling is present in patients with intermediate dysfunction but is not noted in those with mild RV dysfunction. In contrast, patients with severe RV dysfunction were observed to have RV-PA uncoupling in addition to reduced STE-derived strain and RV dilatation. The majority of clustering algorithms have limited clinical value and are primarily used in research settings, as all parameters used in the model are required for effective group identification. However, we applied a clustering algorithm trained on the PVDOMICS data set to develop a calculator that estimates the probability of mild, moderate, or severe RV dysfunction, without requiring all parameters. Our innovative approach accommodates the variability in RV functional parameters across sites, providing flexibility for both research and clinical use. In addition, due to the unsupervised nature of this analyses, we offer a practical means of applying the results from supervised Cox regression analyses, enabling robust clinical implementation as more rigorously validated models are developed.
The current study has several limitations. First, as a cross-sectional examination of a large, observational cohort, we cannot establish causality or account for changes in RV function and therapeutic response over time. Although the majority of patients had prevalent PAH, limiting causal interpretation, associations remained significant after adjusting for PH duration. We additionally present ancillary analysis on a subgroup of incident PAH patients and identified significant associations between functional outcomes and echocardiographic RV-PA coupling. However, fewer associations were found in patients with incident PAH compared to the prevalent PAH subgroup, likely driven by reduced statistical power and smaller sample size. Both Cox regression and cluster analyses were developed using this cohort and not validated on an independent data set, which may limit their generalizability to other PH cohorts. Yet, these findings provide proof of concept that echocardiographic RV phenotyping significantly enhances PAH risk prediction. Future studies are needed to validate these results in a secondary cohort with comprehensive echocardiographic, functional, and clinical data.
Interpretation
To our knowledge, we present the largest study to date showing the added prognostic value of RV echocardiographic phenotyping in patients with well-characterized prevalent and incident PH. Our analysis showed that RV-PA uncoupling, reduced STE-derived strain, and indices of maladaptive chamber remodeling were associated with reduced 6MWD and increased PH risk. In incident PAH, FAC and FAC/PASP exhibited similar associations. Cluster analysis revealed that RV dysfunction was associated with mortality independent of PH duration, PVR, and other known risk factors. Advanced statistical modeling, as a proof of concept, further found that RV echocardiographic phenotyping enhanced risk prognostication in PH beyond traditional risk scores. Further studies are required to confirm these findings and to test the statistical risk prediction models and cluster analysis in a well-phenotyped and validated PH cohort.
Funding/Support
The PVDOMICS Program received NIH/NHLBI grants from U01HL125218, U01HL125205, U01HL125212, U01HL125208, U01HL125175, U01HL125215, and U01HL125177, as well as funding from the Pulmonary Hypertension Association. Funding for this work was also supported by the National Scleroderma Foundation (M. M.), Department of Defense PR191839 (S. C. M.), Department of Defense PR231648 (M. M.), NIH/NHLBI R01HL162851 (M. M.), NIH/NHLBI R01HL159055 (J. K.), NIH/NHLBI R01HL114910 (P. M. H.), NIH/NHLBI R01HL170090, NIH/NHLBI R01HL152724 (P. J. L.), and NIH/NHLBI R01HL163960 (A. R. H.).
Financial/Nonfinancial Disclosures
The authors have reported to CHEST the following: M. M. reports financial support from the NHLBI; and a relationship with the NHLBI, US Department of Defense, and the National Scleroderma Foundation that includes funding grants. Unrelated to this work, M. M. serves on the Data Safety Monitoring Board for Advarra, Inc. A. R. H. serves as a consultant to Merck, Janssen, Gossamer Bio, United Therapeutics, and Tenax Therapeutics; and is a stockholder in Tenax Therapeutics. None declared (V. P. J., R. O., H. M., A. A., G. B., S. E., R. P. F., P. M. H., N. S. H., E. M. H., J. K., D. K., A. B. L., P. J. L., J. A. L., S. C. M., R. M., M. M. P., E. B. R., W. H. W. T., C. L. J., F. P. R., R. B.).
Acknowledgments
Author contributions: Each author has made substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data; or have drafted the work or substantively revised it and have approved the submitted version and to have agreed both to be personally accountable for the author's contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature. All authors contributed to the revision and final approval of the manuscript.
Other contributions: The authors thank the other investigators, the coordinating center, and the PVDOMICS participants for their valuable contributions.
Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.
Additional information: The e-Figure and e-Tables are available online under “Supplementary Data.”
Footnotes
M. M. and V. P. J. contributed equally to this study.
Supplementary Data
References
- 1.Vonk Noordegraaf A., Westerhof B.E., Westerhof N. The relationship between the right ventricle and its load in pulmonary hypertension. J Am Coll Cardiol. 2017;69(2):236–243. doi: 10.1016/j.jacc.2016.10.047. [DOI] [PubMed] [Google Scholar]
- 2.Sanz J., Sánchez-Quintana D., Bossone E., Bogaard H.J., Naeije R. Anatomy, function, and dysfunction of the right ventricle: JACC State-of-the-Art Review. J Am Coll Cardiol. 2019;73(12):1463–1482. doi: 10.1016/j.jacc.2018.12.076. [DOI] [PubMed] [Google Scholar]
- 3.Vonk Noordegraaf A., Chin K.M., Haddad F., et al. Pathophysiology of the right ventricle and of the pulmonary circulation in pulmonary hypertension: an update. Eur Respir J. 2019;53(1) doi: 10.1183/13993003.01900-2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dardi F., Boucly A., Benza R., et al. Risk stratification and treatment goals in pulmonary arterial hypertension. Eur Respir J. 2024;64(4) doi: 10.1183/13993003.01323-2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ghio S., Mercurio V., Fortuni F., et al. A comprehensive echocardiographic method for risk stratification in pulmonary arterial hypertension. Eur Respir J. 2020;56(3) doi: 10.1183/13993003.00513-2020. [DOI] [PubMed] [Google Scholar]
- 6.Benza R.L., Gomberg-Maitland M., Elliott C.G., et al. Predicting survival in patients with pulmonary arterial hypertension: the REVEAL Risk Score Calculator 2.0 and comparison with ESC/ERS-based risk assessment strategies. Chest. 2019;156(2):323–337. doi: 10.1016/j.chest.2019.02.004. [DOI] [PubMed] [Google Scholar]
- 7.Benza R.L., Kanwar M.K., Raina A., et al. Development and validation of an abridged version of the REVEAL 2.0 Risk Score Calculator, REVEAL Lite 2, for use in patients with pulmonary arterial hypertension. Chest. 2021;159(1):337–346. doi: 10.1016/j.chest.2020.08.2069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hoeper M.M., Pausch C., Olsson K.M., et al. COMPERA 2.0: a refined four-stratum risk assessment model for pulmonary arterial hypertension. Eur Respir J. 2022;60(1) doi: 10.1183/13993003.02311-2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Humbert M., Kovacs G., Hoeper M.M., et al. 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension: developed by the task force for the diagnosis and treatment of pulmonary hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS). Endorsed by the International Society for Heart and Lung Transplantation (ISHLT) and the European Reference Network on rare respiratory diseases (ERN-LUNG) Eur Heart J. 2022;43(38):3618–3731. [Google Scholar]
- 10.Kovacs G., Bartolome S., Denton C.P., et al. Definition, classification and diagnosis of pulmonary hypertension. Eur Respir J. 2024;64(4) doi: 10.1183/13993003.01324-2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mukherjee M., Rudski L.G., Addetia K., et al. Guidelines for the echocardiographic assessment of the right heart in adults and special considerations in pulmonary hypertension: recommendations from the American Society of Echocardiography. J Am Soc Echocardiogr. 2025;38(3):141–186. doi: 10.1016/j.echo.2025.01.006. [DOI] [PubMed] [Google Scholar]
- 12.Lang R.M., Badano L.P., Mor-Avi V., et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2015;16(3):233–271. doi: 10.1093/ehjci/jev014. [DOI] [PubMed] [Google Scholar]
- 13.Lankhaar J.-W., Westerhof N., Faes T.J., et al. Quantification of right ventricular afterload in patients with and without pulmonary hypertension. Am J Physiol Heart Circulatory Physiol. 2006;291(4):H1731–H1737. doi: 10.1152/ajpheart.00336.2006. [DOI] [PubMed] [Google Scholar]
- 14.Focardi M., Cameli M., Carbone S.F., et al. Traditional and innovative echocardiographic parameters for the analysis of right ventricular performance in comparison with cardiac magnetic resonance. Eur Heart J Cardiovasc Imaging. 2015;16(1):47–52. doi: 10.1093/ehjci/jeu156. [DOI] [PubMed] [Google Scholar]
- 15.Vonk Noordegraaf A., Haddad F., Bogaard H.J., Hassoun P.M. Noninvasive imaging in the assessment of the cardiopulmonary vascular unit. Circulation. 2015;131(10):899–913. doi: 10.1161/CIRCULATIONAHA.114.006972. [DOI] [PubMed] [Google Scholar]
- 16.Mukherjee M., Mathai S.C., Jellis C., et al. Defining echocardiographic degrees of right heart size and function in pulmonary vascular disease from the PVDOMICS study. Circulation Cardiovasc Imaging. 2024;17(10) doi: 10.1161/circimaging.124.017074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Elinoff J.M., Agarwal R., Barnett C.F., et al. Challenges in pulmonary hypertension: controversies in treating the tip of the iceberg. A Joint National Institutes of Health Clinical Center and Pulmonary Hypertension Association Symposium Report. Am J Respir Crit Care Med. 2018;198(2):166–174. doi: 10.1164/rccm.201710-2093PP. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hemnes A.R., Beck G.J., Newman J.H., et al. PVDOMICS: a multi-center study to improve understanding of pulmonary vascular disease through phenomics. Circ Res. 2017;121(10):1136–1139. doi: 10.1161/CIRCRESAHA.117.311737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.El-Kersh K., Zhao C., Elliott G., et al. Derivation of a risk score (REVEAL-ECHO) based on echocardiographic parameters of patients with pulmonary arterial hypertension. Chest. 2023;163(5):1232–1244. doi: 10.1016/j.chest.2022.12.045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pulmonary Vascular Disease Phenomics Program PVDOMICS (PVDOMICS). ClinicalTrials.gov identifier: NCT02980887. Updated January 27, 2025. https://clinicaltrials.gov/study/NCT02980887
- 21.Jellis C.L., Park M.M., Abidov A., et al. Comprehensive echocardiographic evaluation of the right heart in patients with pulmonary vascular diseases: the PVDOMICS experience. Eur Heart J Cardiovasc Imaging. 2022;23(7):958–969. doi: 10.1093/ehjci/jeab065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tello K., Wan J., Dalmer A., et al. Validation of the tricuspid annular plane systolic excursion/systolic pulmonary artery pressure ratio for the assessment of right ventricular-arterial coupling in severe pulmonary hypertension. Circ Cardiovasc Imaging. 2019;12(9) doi: 10.1161/CIRCIMAGING.119.009047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ünlü S., Bézy S., Cvijic M., Duchenne J., Delcroix M., Voigt J.-U. Right ventricular strain related to pulmonary artery pressure predicts clinical outcome in patients with pulmonary arterial hypertension. Eur Heart J Cardiovasc Imaging. 2023;24(5):635–642. doi: 10.1093/ehjci/jeac136. [DOI] [PubMed] [Google Scholar]
- 24.Guazzi M., Bandera F., Pelissero G., et al. Tricuspid annular plane systolic excursion and pulmonary arterial systolic pressure relationship in heart failure: an index of right ventricular contractile function and prognosis. Am J Physiol Heart Circulatory Physiol. 2013;305(9):H1373–H1381. doi: 10.1152/ajpheart.00157.2013. [DOI] [PubMed] [Google Scholar]
- 25.Guazzi M. Use of TAPSE/PASP ratio in pulmonary arterial hypertension: an easy shortcut in a congested road. Int J Cardiol. 2018;266:242–244. doi: 10.1016/j.ijcard.2018.04.053. [DOI] [PubMed] [Google Scholar]
- 26.Gorter T.M., van Veldhuisen D.J., Voors A.A., et al. Right ventricular-vascular coupling in heart failure with preserved ejection fraction and pre-vs. post-capillary pulmonary hypertension. Eur Heart J Cardiovasc Imaging. 2018;19(4):425–432. doi: 10.1093/ehjci/jex133. [DOI] [PubMed] [Google Scholar]
- 27.Gami A., Jani V.P., Mombeini H., et al. Prognostic value of echocardiographic coupling metrics in systemic sclerosis-associated pulmonary vascular disease. J Am Soc Echocardiogr. 2025;38(2):115–126. doi: 10.1016/j.echo.2024.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Jani V.P., Strom J.B., Gami A., et al. Optimal method for assessing right ventricular to pulmonary arterial coupling and subclinical right ventricular dysfunction in older aged healthy adults: the Multi-Ethnic Study of Atherosclerosis. Am J Cardiol. 2024;222:11–19. doi: 10.1016/j.amjcard.2024.03.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Young J.B., Dunlap M.E., Pfeffer M.A., et al. Mortality and morbidity reduction with candesartan in patients with chronic heart failure and left ventricular systolic dysfunction: results of the CHARM low-left ventricular ejection fraction trials. Circulation. 2004;110(17):2618–2626. doi: 10.1161/01.CIR.0000146819.43235.A9. [DOI] [PubMed] [Google Scholar]
- 30.Tunthong R., Salama A.A., Lane C.M., et al. Right ventricular systolic strain in patients with pulmonary hypertension: clinical feasibility, reproducibility, and correlation with ejection fraction. J Echocardiography. 2023;21(3):105–112. doi: 10.1007/s12574-022-00593-6. [DOI] [PubMed] [Google Scholar]
- 31.Mercurio V., Mukherjee M., Tedford R.J., et al. Improvement in right ventricular strain with ambrisentan and tadalafil upfront therapy in scleroderma-associated pulmonary arterial hypertension. Am J Respir Crit Care Med. 2018;197(3):388–391. doi: 10.1164/rccm.201704-0789LE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Mukherjee M., Mercurio V., Tedford R.J., et al. Right ventricular longitudinal strain is diminished in systemic sclerosis compared with idiopathic pulmonary arterial hypertension. Eur Respir J. 2017;50(5) doi: 10.1183/13993003.01436-2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Fine N.M., Chen L., Bastiansen P.M., et al. Reference values for right ventricular strain in patients without cardiopulmonary disease: a prospective evaluation and meta-analysis. Echocardiography. 2015;32(5):787–796. doi: 10.1111/echo.12806. [DOI] [PubMed] [Google Scholar]
- 34.Sachdev A., Villarraga H.R., Frantz R.P., et al. Right ventricular strain for prediction of survival in patients with pulmonary arterial hypertension. Chest. 2011;139(6):1299–1309. doi: 10.1378/chest.10-2015. [DOI] [PubMed] [Google Scholar]
- 35.van de Veerdonk M.C., Marcus J.T., Westerhof N., et al. Signs of right ventricular deterioration in clinically stable patients with pulmonary arterial hypertension. Chest. 2015;147(4):1063–1071. doi: 10.1378/chest.14-0701. [DOI] [PubMed] [Google Scholar]
- 36.Mukherjee M., Mercurio V., Balasubramanian A., et al. Defining minimal detectable difference in echocardiographic measures of right ventricular function in systemic sclerosis. Arthritis Res Therapy. 2022;24(1):146. doi: 10.1186/s13075-022-02835-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Unlu S., Farsalinos K., Ameloot K., et al. Apical traction: a novel visual echocardiographic parameter to predict survival in patients with pulmonary hypertension. Eur Heart J Cardiovasc Imaging. 2016;17(2):177–183. doi: 10.1093/ehjci/jev131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Mukherjee M., Ogunmoroti O., Jani V., et al. Characteristics of right ventricular to pulmonary arterial coupling and association with functional status among older aged adults from the Multi-Ethnic Study of Atherosclerosis. Am J Cardiol. 2023;196:41–51. doi: 10.1016/j.amjcard.2023.03.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Hjalmarsson C., Kjellström B., Jansson K., et al. Early risk prediction in idiopathic versus connective tissue disease-associated pulmonary arterial hypertension: call for a refined assessment. ERJ Open Res. 2021;7(3):00854–2020. doi: 10.1183/23120541.00854-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kylhammar D., Kjellström B., Hjalmarsson C., et al. A comprehensive risk stratification at early follow-up determines prognosis in pulmonary arterial hypertension. Eur Heart J. 2018;39(47):4175–4181. doi: 10.1093/eurheartj/ehx257. [DOI] [PubMed] [Google Scholar]
- 41.Hoeper M.M., Kramer T., Pan Z., et al. Mortality in pulmonary arterial hypertension: prediction by the 2015 European pulmonary hypertension guidelines risk stratification model. Eur Respir J. 2017;50(2) doi: 10.1183/13993003.00740-2017. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.