Right ventricular (RV) function predicts exercise capacity and clinical outcomes in subjects with pulmonary hypertension (PH).1–5 RV ejection fraction (RVEF), as measured by cardiac MRI, is considered the gold standard measurement of RV function. However, the use of MRI is limited by availability and local expertise. Hence, two-dimensional echocardiogram-based metrics remain of significant interest in clinical practice.5,6 As the predominant RV motion occurs in the longitudinal plane, tricuspid annular plane systolic excursion (TAPSE) is an easily attainable parameter that captures this motion with a normal value ≥17 mm and is predictive of clinical outcomes in different diseases.7–10However, the morphology and size of the RV change with worsening severity and among different phenotypes of pulmonary hypertension. 11 Hence, TAPSE’s inability to capture the transverse contribution of the RV free wall and septum would significantly affect the assessment of ventricular function in subjects with varying RV size and morphology.12,13 Notably, as the right ventricle dilates, the transverse motion of the RV plays an increasingly important role as opposed to the primary longitudinal motion recorded by TAPSE.1,5,14 In a canine model study, Yuchi et al. reported that TAPSE alone could be increased in a volume-overloaded state in severe RV dysfunction, but by normalizing TAPSE to RV size, it can better evaluate RV function by accounting for its load dependency.15In this study, we hypothesized that TAPSE divided by RV end-systolic area (TAPSE/RVA-S) and RV end-diastolic area (TAPSE/RVA-D) are better markers of right heart function in PH than TAPSE alone. We assess correlations of the TAPSE/RVA metrics with three different outcomes variables: 1) Pulmonary arterial compliance (PCa, pulmonary pulse pressure/stroke volume, normal value ≥3.0 mmHg.mL−1) as metric of pulsatile load impacting RV function and a unifying hemodynamic feature among different PH phenotypes, where a higher PCa protects the RV from pressure overload with exercise, while a lower PCa indicates vessel stiffness and is correlated with RV dysfunction,4,16–19 2) Exercise capacity (as peak oxygen consumption, peak VO2 as %predicted), 3) Adverse clinical outcomes (cardiovascular hospitalization and death) on one-year follow-up.
Methods:
The data that support the findings of this study are available from the corresponding author upon reasonable request.
In the present study, we conducted a retrospective review study of a cohort of subjects referred to the University of Wisconsin Pulmonary Hypertension clinic for undefined dyspnea. The subjects underwent exercise right heart catheterization with cardiopulmonary exercise testing (CPET) for exercise intolerance and suspected early-stage pulmonary hypertension. The subjects were then classified into four groups based on their rest and exercise hemodynamics based on studies from Lewis et al., Eisman et al. and the 2022 ESC/ERS Guidelines for pulmonary hypertension .20–22 From a database of 100 consecutive subjects who underwent exercise hemodynamic studies over a two-year period, 71 subjects were included in this study with full echocardiographic images (median, IQR time between echocardiogram and the exercise hemodynamic studies= 28, 80 days). Twenty-nine subjects were excluded due to inadequate echocardiographic images, left ventricular ejection fraction <50%, supplemental oxygen use, serum creatinine >2mg/dl or on dialysis, primary lung diseases with forced vital capacity < 60% predicted, and severe valvular heart disease. Patients with prior cardiothoracic surgery were excluded from our study. Of note, three subjects (among the 71 included subjects) had poor endocardial definition with systole, and their TAPSE/RVA-S and fractional area change were not reported. All subjects were in NYHA functional classes II-III. They underwent clinical evaluation, six-minute walk distance test (6MWD), echocardiograms, and pulmonary function testing.
Subjects were classified into the following four groups based on their rest and predominant exercise profile:
Control (n=12): All subjects had resting mean pulmonary artery pressure (mPAP) ≤ 20 mm Hg, pulmonary artery wedge pressure (PAWP) ≤ 15 mm Hg, and pulmonary vascular resistance (PVR) < 2 Wood units (WU) were classified as controls. With direct Fick cardiac output (CO) assessment during exercise, mPAP/CO slope was <3 mmHg/L.min−1 and PAWP/CO slope was <2 mmHg/L.min−1.
Isolated post-capillary pulmonary hypertension (IpcPH, n=10): All subjects met rest hemodynamic criteria: mPAP > 20 mm Hg, PAWP > 15 mm Hg, and PVR < 2 WU.
Combined post-capillary pulmonary hypertension (CpcPH, n=24): All subjects met rest hemodynamic criteria: mPAP > 20 mm Hg, PAWP > 15 mm Hg, and PVR ≥ 2 WU.
Pulmonary arterial hypertension (PAH, n=25): All subjects met rest hemodynamic criteria: mPAP > 20 mm Hg, PAWP ≤ 15 mm Hg, and pulmonary vascular resistance ≥ 2 WU.
This study was approved by the University of Wisconsin institutional review board (IRB). The study complies with the guidelines of the Declaration of Helsinki.
All subjects underwent rest-to-exercise CPET with simultaneous right heart catheterization and expired gas analysis as previously described.4,23–25 Briefly, pulmonary hemodynamic measurements were obtained with a balloon-tipped, double-lumen, fluid-filled 7 Fr pulmonary artery (PA) catheter via an internal jugular vein approach. A 3 Fr right radial catheter was placed selectively in subjects with pre-test 6MWD test reporting systemic oxygen saturation <95% on room air. CO was measured via both direct Fick principle and thermodilution methods at rest, and via direct Fick with exercise. A nose clip and a mouthpiece with a saliva trap were placed to confirm an air leak-free system. The mouthpiece was connected to an umbilical adapter and further connected to Metabolic Ultima™ CardiO2® gas exchange analysis system (MedGraphics) for a breath-by-breath analysis. After recording rest hemodynamics, exercise was performed in a recumbent position on a stationary ergometer at 60 rpm on an incremental 10 W/min protocol. PA pressures and mean PAWP were recorded every 20–25 W until peak exercise to trend the pressures, which increased confidence in peak-exercise pressure readings (averaged over 3–5 respiratory cycles). We acquire meanPAP and direct Fick cardiac output at multiple points during exercise in an individual patient (every 20-25 watts). Symptom-limited exercise test with volitional fatigue was performed. Near peak exercise (while the subjects were still exercising), a mixed venous sample, oxygen consumption (VO2), and hemodynamic parameters (PA pressure, PAWP, and right atrial pressures) were recorded. The reported mPAP/CO slope is based on delta values between peak exercise and rest (delta exercise-rest mPAP/ exercise-rest CO).
All subjects underwent comprehensive, full transthoracic echocardiogram studies as a part of the work up for their exertional dyspnea and exercise intolerance. Standard echocardiographic and additional quantitative measurements were performed to determine RV end-diastolic and end-systolic areas.26 We then measured the RV end-diastolic and end-systolic areas; and calculated TAPSE/RVA-S by dividing TAPSE by end-systolic area and calculated TAPSE/RVA-D by dividing TAPSE by end-diastolic areas.
Baseline characteristics are presented as mean ± standard deviation for continuous variables and numbers and percentages for categorical variables, with percentage of total for binary categorical variables according to 4 groups as defined above. ANOVA was performed to analyze difference in means between the four subject groups, and post-hoc analysis was done using Tukey’s Honest Significance Difference test. Results with p-value < 0.05 were considered significant. The strength of association between quantitative variables was measured with Pearson correlation coefficient. Simple univariate linear regression was performed to predict pulmonary compliance at rest and peak VO2 (% predicted), rest and exercise PVR, VE/VCO2 slope, nadir ETCO2, mPAP/CO slope, and exercise cardiac output. Multivariate liner regression was performed for novel TAPSE/RVA metrics, adjusted for age, sex and BMI. Model assumptions were verified by examining model diagnostic plots. The pulmonary compliance and peak VO2 (%predicted) were categorized into two groups based on their median values. The median value of rest pulmonary compliance ≥3.0 mL/mmHg was consistent with prior literature, as a cut off for improved outcomes. 18 Similarly, median value of peak VO2 ≥50% predicted was also consistent with prior literature as a reliable cutoff for adverse outcomes.27 Receiver operating characteristic (ROC) curves with area under the curve (AUC) values were used to compare discriminative ability of the various parameters, and the optimal cut-points were defined by the Youden index. Survival from adverse outcomes (cardiovascular hospitalization and death) was measured from the date of exercise hemodynamic study over a one-year follow-up. The prognostic value of TAPSE/RVA metrics to predict adverse outcomes was determined using Kaplan–Meier survival with Log-Rank testing, and Hazard ratio reported with Cox proportional regression analyses adjusted for age and sex as biological variables. All statistical analyses and figures were conducted using R (version 4.0.3).
Results:
A total of 71 subjects were divided into four groups. Their baseline characteristics and echocardiography results are shown in Table 1. The control group subjects were younger (56.4 ± 14.5 years) and had a lower BMI (28.3 ± 4.8 kg/m2). Compared to other cohorts, control subjects also had fewer comorbidities. At baseline, control subjects performed better at 6MWD (437 ± 68 meters) and had a lower BNP value (93 ± 86 pg/mL). Compared to other cohorts, PAH subjects were more likely to have connective tissue disease (44%).
The baseline hemodynamic measurements at rest and peak exercise were included in Table 2. Per definition, control subjects had mPAP at rest ≤ 20 mmHg, which increased to 32 ± 7 with peak exercise; both were significantly lower than the other three cohorts. IpcPH cohort had a lower baseline mPAP than CpcPH and PAH groups and a lower mPAP at peak exercise. The control group and PAH subjects had similar PAWP at rest and peak exercise, which were lower than those of the CpcPH and IpcPH groups. The control group had a significantly lower mPAP/CO compared to CpcPH and PAH and a significantly lower PAWP/CO compared to CpcPH and IpcPH. The control cohort achieved a higher peak heart rate and exercised longer. The cardiac output and cardiac index were significantly higher in the control group, with a two-fold augmentation at peak exercise. Both CO and CI were lower in CpcPH, IpcPH, and PAH groups, with a lower augmentation at peak exercise. Pulmonary compliance (PCa), calculated as stroke volume divided by pulse pressure, was significantly higher in the control group at rest (compared to CpcPH and PAH) and peak exercise (compared to all other groups).
Baseline transthoracic echocardiogram results were summarized in Table 2. TAPSE of the four cohorts were not statistically different (p=0.38). As previous studies have suggested, TAPSE/PASP (PA systolic pressure, per echocardiogram) was higher in the control group compared to CpcPH and PAH (p<0.01). RV FAC was significantly higher in the control group compared to CpcPH (p=0.01). Lateral E/e’ was not significantly different across all four groups (p=0.17).
Both TAPSE/RVA-D and TAPSE/RVA-S were significantly higher in the control group than the PAH and CpcPH groups but were not statistically significantly higher than IpcPH. The box and whisker plots of four subject groups are shown in Figure 1.
Pulmonary compliance is the main component of pulsatile RV afterload. On Pearson correlation (Figure 2) and univariate linear regression analyses (Supplementary Table S1), there was a significant correlation of PCa with TAPSE, TAPSE/RVA-S, and TAPSE/RVA-D. On multivariate analyses (including age, sex, and BMI), both TAPSE/RVA-S and TAPSE/RVA-D (p<0.001) remained significant (Supplementary Table S1). Moreover, on ROC analyses (Table 3), both TAPSE/RVA metrics performed better compared to TAPSE and RV FAC in predicting poor pulmonary compliance at rest (PCa <3 mL/mmHg). Based on the ROC analysis, an optimal threshold was 1.5 for TAPSE/RVA-S (sensitivity=0.75, specificity=0.81) and 1.1 for TAPSE/RVA-D (sensitivity=0.72, specificity=0.78). Both TAPSE/RVA metrics outperformed TAPSE alone, and either RV end-diastolic or end-systolic area in the regression analysis for PCa (Supplemental Table S1).
Peak VO2 is a metric of exercise capacity and can risk-stratify subjects with pulmonary hypertension4. On Pearson correlation (Figure 2), peak VO2 (% predicted) had a significant correlation with TAPSE/RVA-D, and approached significance with TAPSE/RVA-S (p=0.08). Univariate and multivariate linear regression analysis continue to reveal a significant correlation of peak VO2 with TAPSE/RVA-D (p=0.008) and TAPSE/RVA-S (p=0.04) (Supplementary Table S1). Similarly, on the ROC analysis to identify peak VO2 <50% (Table 3), the AUC was significant for both TAPSE/RVA-S (optimal threshold=1.5, p=0.04) and TAPSE/RVA-D (optimal threshold=1.1, p=0.02). Both TAPSE/RVA measurements had stronger correlations with peak VO2 <50% than TAPSE and RV FAC. Again, TAPSE/RVA was found to have a stronger correlation compared to TAPSE and RV areas (Supplemental Table S1).
Additionally, we performed univariate and multivariate regression analyses for TAPSE/RVA and its correlation with rest and exercise PVR (Supplemental Table S2), VE/VCO2 slope and nadir ETCO2 (Supplemental Table S3), mPAP/CO slope and exercise cardiac output (Supplemental Table S4). In comparison to TAPSE and RV areas, TAPSE/RVA was found to be consistently and strongly correlated with all those parameters.
To evaluate the prognostic values of TAPSE/RVA metrics, adverse cardiovascular events were recorded within 12 months of the exercise hemodynamic study. The adverse outcomes included cardiovascular hospitalization and death. Kaplan-Meier time-to-event analysis was performed, and the results are shown in Figure 3. TAPSE alone at the cutoff of 17 mm was not associated with adverse events. On the contrary, both TAPSE/RVA-D (with a cutoff of <1.1) and TAPSE/RVA-S (with a cutoff of <1.5) were significantly associated with increased adverse events (log rank p<0.01). Among TAPSE/RVA-D <1.1 (vs. >1.1), out of 37 patient, 3 died and 20 had cardiovascular hospitalization (vs. out of 34 patients, 1 died and 6 had cardiovascular hospitalization). Among TAPSE/RVA-S <1.5 (vs. >1.5), out of 34 patient, 3 died and 17 had cardiovascular hospitalization (vs. out of 33 patients, 1 died and 7 had cardiovascular hospitalization). With Cox Proportional hazard analysis adjusted for age and sex, the hazard ratio for adverse outcomes were: TAPSE = 0.99 (CI: 0.88-1.02, p=0.74), TAPSE/RVA-D = 0.38 (CI: 0.15-0.98, p=0.04), and TAPSE/RVA-S = 0.44 (CI: 0.25-0.80, p<0.01).
In 31 subjects among the overall cohort (n=71), cardiac MRI-based RVEF data was available within 3 months of the echocardiogram (median days=54, IQR 49). The correlation of RVEF were stronger with TAPSE/RVA metrics (TAPSE/RVA-D r=0.378, p=0.03, and TAPSE/RVA-S r=0.328, p=0.07) than TAPSE (r=0.082, p=0.68) and RV FAC (r=0.012, p=0.94).
Among subjects dichotomized (normal vs abnormal) based on TAPSE/RVA-D >1.5 vs <1.5 (n=34 vs 37), RVEF was 54%±14 vs 44%±13 (p=0.03) and incidence of adverse clinical outcomes were 20% vs 62% (p<0.01). Similarly, based on TAPSE/RVA-S=1.1 (n=33 vs 34), RVEF was 56±13 vs 44±14 (p<0.01), and incidence of adverse clinical outcomes were 24% vs 59% (p=0.02). Lastly, based on TAPSE=17mm (n=43 vs 28), RVEF was 50±15 vs 47±14 (p=0.16), and incidence of adverse clinical outcomes were 39% vs 46% (p=0.49).
To validate these findings, we prospectively identified a separate cohort of 20 subjects with invasive CPET data and completed echocardiographic data (< 3 months between the two studies). These subjects were dichotomized based on the traditional TAPSE cutoff of 17mm, TAPSE/RVA-D cutoff of 1.1 mm/cm2, and TAPSE/RVA-S cutoff of 1.5 mm/cm2 (results summarized in Supplementary Table S5). Peak VO2 was statistically different using either TAPSE/RVA as a cutoff (p=0.02 for TAPSE/RVA-D, p=0.03 for TAPSE/RVA-S), but not with TAPSE (p=0.20). Similarly, both TAPSE/RVA were better at discriminating rest PCa and mPAP/CO slope than TAPSE. With a follow-up limited to only 6 months, there was only one incidence of adverse outcome out of twenty subjects. Hence, the clinical outcome data was not tested for statistical significance.
Discussion:
In this study, we report novel metrics of TAPSE divided by RV areas among different hemodynamic phenotypes of pulmonary hypertension. Compared to TAPSE alone and RV FAC, TAPSE/RVA metrics comprehensively capture right heart function across variable severity and different phenotypes of PH, as RV morphology and size change.28 Moreover, we found that TAPSE/RVA metrics were strongly correlated with pulsatile RV afterload (pulmonary compliance) and exercise capacity (peak VO2) in pulmonary hypertension. Optimal cutoffs of 1.1 for TAPSE/RVA-D and 1.5 for TAPSE/RVA-S were superior to TAPSE alone in predicting cardiovascular hospitalizations and death (Figure 4).
TAPSE and RV fractional area change are used as markers for right ventricular function. 7–10 The RV wall is comprised of an inner layer of longitudinally arranged myocardial fibers and an outer layer of circumferentially arranged myocardial fibers.5,9,28 In pulmonary hypertension and subsequent right heart failure, RV dilates, and the circumferential contraction of the outer layer becomes more important in RV function.1,5,14,29 Hence, TAPSE would be inadequate in measuring the RV function as it does not capture the transverse component of the ventricular contraction.5 While RV fractional area change can capture RV’s longitudinal and transverse contraction, its measurement using echocardiography is often limited by inconsistent windows, poor endocardial definition of RV walls, and trabeculations on the endocardial surface.1,6,30 It also fails to account for the RV outflow tract and results in the underestimation of RV volumes.31 The current study combines both of these metrics (TAPSE and RV areas) to reduce the limitations and improve utility of 2D echocardiography in pulmonary hypertension.
The current study is focused particularly on pulmonary compliance, as this is a clinically available metric of pulsatile RV afterload and predictive of clinical outcomes in a wide spectrum of PH phenotypes, especially left heart disease.16–19,32–35 While the PCa loses its prognostic ability as a singular metric in multivariate analyses in most studies18,19, it remains an important physiological metric and signifies the role of pulsatile pulmonary load, especially in post-capillary PH. As a metric of pulsatile load, PCa is possibly an oversimplified metric, while bioengineering metrics of pulmonary vascular impedance are more sophisticated and have been shown to correlate with RV dysfunction with better precision 23,36. Further research is warranted into bioengineering metrics of pulmonary impedance. As a second outcomes variable, peak VO2 is widely accepted as a metric of exercise capacity and proven as a strong prognostic marker of clinical outcomes in different cardiopulmonary diseases. 6,31 This study reports that pulmonary compliance and peak VO2 are better correlated with TAPSE/RVA metrics, in comparison to traditional echocardiogram-based metrics (TAPSE and FAC). This is further supported by a stronger ability of TAPSE/RVA-S <1.5 and TAPSE/RVA-D <1.1 to identify adverse clinical outcomes at one-year follow-up (Figure 3).
Limitations
It is a single-center, retrospective study with a moderate sample size with 29 subjects excluded from final data analyses due to missing data (Supplemental Table S6). We further divided the entire cohort based on hemodynamic phenotype leading to a dilution of sample size within each group has a small sample size, which may contribute to type II error. Using an echocardiogram to calculate the right ventricular area is technically challenging given its limited windows, poor endocardial definition of RV walls, and trabeculations on the endocardial surface, 1,6,30 however, we only included subjects with good windows and excluded subjects with poor endocardial definition of RV free wall. Lastly, the current study excluded patients with prior cardiothoracic surgery to reduce bias from post-cardiotomy patients 37. While we tested validation in a small prospective cohort internally, future larger studies are warranted to validate TAPSE/RVA metrics in specific phenotypes of PH and heart failure.
Conclusions:
In summary, TAPSE/RVA metrics account of varying RV morphology in PH by normalizing TAPSE to the right ventricular area and can serve as comprehensive echocardiographic measure of right heart function. These novel metrics correlate with pulsatile RV afterload (pulmonary arterial compliance) and exercise capacity (peak VO2). TAPSE/RVA-D <1.1 mm/cm2 and TAPSE/RVA-S <1.5 mm/cm2 have prognostic value in identifying cardiovascular hospitalization and death among subjects with pulmonary hypertension.
AE
Dr. Ahmed El shaer
Internal Medicine · General Internal MedicineUniversity of Wisconsin–Madisonahmede