Abstract
Introduction Identification of elevated pulmonary artery pressures during exercise has important diagnostic, prognostic and therapeutic implications. Stress echocardiography is frequently used to estimate pulmonary artery pressures during exercise testing, but data supporting this practice are limited. This study examined the accuracy of Doppler echocardiography for the estimation of pulmonary artery pressures at rest and during exercise.
Methods Simultaneous cardiac catheterisation-echocardiographic studies were performed at rest and during exercise in 97 subjects with dyspnoea. Echocardiography-estimated pulmonary artery systolic pressure (ePASP) was calculated from the right ventricular (RV) to right atrial (RA) pressure gradient and estimated RA pressure (eRAP), and then compared with directly measured PASP and RAP.
Results Estimated PASP was obtainable in 57% of subjects at rest, but feasibility decreased to 15–16% during exercise, due mainly to an inability to obtain eRAP during stress. Estimated PASP correlated well with direct PASP at rest (r=0.76, p<0.0001; bias −1 mmHg) and during exercise (r=0.76, p=0.001; bias +3 mmHg). When assuming eRAP of 10 mmHg, ePASP correlated with direct PASP (r=0.70, p<0.0001), but substantially underestimated true values (bias +9 mmHg), with the greatest underestimation among patients with severe exercise-induced pulmonary hypertension (EIPH). Estimation of eRAP during exercise from resting eRAP improved discrimination of patients with or without EIPH (area under the curve 0.81), with minimal bias (5 mmHg), but wide limits of agreement (−14–25 mmHg).
Conclusions The RV–RA pressure gradient can be estimated with reasonable accuracy during exercise when measurable. However, RA hypertension frequently develops in patients with EIPH, and the inability to noninvasively account for this leads to substantial underestimation of exercise pulmonary artery pressures.
Abstract
This study shows that the right atrial component of the estimated pulmonary artery pressure equation is often overlooked, but quite important, and failure to account for this leads to substantial underestimation of the severity of exercise-induced PH http://bit.ly/32xgKTD
Introduction
Pulmonary hypertension is common in patients with a variety of cardiopulmonary diseases and is associated with increased morbidity and mortality [1–4]. Because the lungs display remarkable circulatory reserve, the presence of pulmonary hypertension at rest actually represents a rather advanced stage of pulmonary vascular disease, where changes in the left heart and pulmonary microcirculation have progressed dramatically [5, 6]. In patients with less advanced cardiopulmonary disease, pulmonary artery pressures increase only during physiological stresses such as exercise [7]. Like pulmonary hypertension at rest, the presence of exercise-induced pulmonary hypertension (EIPH) is associated with impaired aerobic capacity, worse haemodynamics, impaired ventricular-arterial coupling, and poor clinical outcomes [8–12]. Accurate identification of patients with EIPH may allow for more effective diagnosis and delivery of interventions to treat and prevent pulmonary hypertension-associated diseases [5, 13, 14].
Invasive haemodynamic exercise testing represents the gold standard to make this assessment, but technical complexity and cost may be barriers to widespread utilisation in practice [14–16]. Doppler echocardiography allows for noninvasive estimation of pulmonary artery pressures at rest based upon the velocity of tricuspid regurgitation and appearance and collapsibility of the inferior vena cava [17]. However, the assumptions upon which those estimations are based may be violated during exercise, particularly where right atrial pressures may increase dramatically. While commonly performed, data regarding the accuracy of estimated pulmonary artery pressure during exercise compared with invasive haemodynamic measurements are limited [18]. Accordingly, we performed a simultaneous cardiac catheterisation-echocardiographic study to determine the accuracy of Doppler echocardiography for the estimation of pulmonary artery pressures at rest and during exercise.
Methods
Subjects referred to the Mayo Clinic catheterisation laboratory for invasive haemodynamic exercise stress testing were prospectively enrolled between August 2011 and July 2013. Some participant data from this study have been published [19–24], but not as it relates to the evaluation of pulmonary hypertension. The study was approved by the Mayo Clinic institutional review board and the study was registered (NCT01418248). Written informed consent was provided by all patients prior to participation in study-related procedures. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
Study population
We prospectively enrolled 99 subjects who referred for a simultaneous echo-catheterisation testing with exercise in the evaluation of exertional dyspnoea and fatigue. One participant withdrew following consent and another developed complete heart block during right heart catheterisation, allowing for 97 subjects that completed the study. Subjects were divided into patients with and without EIPH, defined as a mean pulmonary artery pressure >30 mmHg during exercise with a total pulmonary resistance of >3 mmHg·min·L−1 [25, 26]. Post-capillary pulmonary hypertension was defined by EIPH with high left heart filling pressures (pulmonary capillary wedge pressure, PCWP) at rest (>15 mmHg) and/or with exercise (≥25 mmHg) [15, 22].
Study protocol
After providing consent, subjects underwent a history and physical examination, and comprehensive resting echocardiogram to familiarise the sonographer with optimal acoustic windows in the pre-procedure area. Cardiac catheterisation was then performed with simultaneous echocardiography and expired gas analysis at rest and during supine ergometer exercise. The first stage of exercise (20 W) was performed for 5 min to allow greater time for image acquisition, and was followed by graded 10-W increments in workload (3-min stages) to subject-reported exhaustion, which was held as long as possible to allow for imaging.
Catheterisation protocol
Patients were studied on their chronic medications in the fasted state after minimal sedation in the supine position as previously described [19–24]. Right heart catheterisation was performed through a 9-Fr sheath via the right internal jugular vein. Pressures in the right atrium (RA), right ventricle (RV), pulmonary artery, and PCWP were measured at end-expiration (mean of ≥3 beats) using 2-Fr high-fidelity micromanometer-tipped catheters (Millar Instruments, Houston, TX, USA) advanced through the lumen of a 7-Fr fluid-filled catheter (Balloon Wedge, Arrow; Teleflex, Wayne, PA, USA). Pressure tracings from the entire study were digitised and stored for offline analysis by one investigator with extensive experience in exercise haemodynamic assessment (BAB).
Mean RA and PCWP were taken at mid A-wave. The PCWP position was verified by typical waveforms, appearance on fluoroscopy, and direct oximetry (PCWP blood saturation ≥94%). Arterial blood pressure was measured through a 4–6-Fr radial arterial cannula. Oxygen consumption (V′O2) was measured from the expired gas analysis (MedGraphics, St Paul, MN, USA). Arterial–venous oxygen content difference (a–vO2diff) was measured directly as the difference between systemic arterial and pulmonary artery oxygen contents equal to the product of oxygen saturation×haemoglobin×1.34. Cardiac output or pulmonary blood flow (QP) was determined by the Fick method (V′O2/a–vO2diff) at baseline, 20 W and peak exercise.
Echocardiography
Two-dimensional, M-mode, Doppler and tissue Doppler echocardiography was performed according to current guidelines by experienced sonographers [17, 27, 28]. Echocardiographic data were obtained simultaneously with invasive assessment at rest and during all stages of exercise. All studies were interpreted offline and in a completely blinded fashion by a single investigator with extensive experience in resting and exercise echocardiographic assessment (GCK). Estimated RV–RA (eRV–RA) pressure gradient was measured from the velocity of the tricuspid regurgitation jet using the modified Bernoulli formula (=4×velocity2) [17, 29]. All patients had continuous-wave Doppler assessment through the RV outflow tract to exclude obstruction to flow and allow the equation of estimated right ventricular systolic pressure to estimated pulmonary artery systolic pressure (ePASP).
Estimated RA pressure (eRAP) was determined from the size and collapsibility of the inferior vena cava (IVC) and hepatic vein Doppler profile, coded as 5 mmHg (normal-sized IVC with >50% respiratory collapse and systolic forward predominant flow on hepatic vein Doppler), 10 mmHg (borderline/normal sized IVC with >50% respiratory collapse and equal degrees of systolic and diastolic forward flow on hepatic vein Doppler), 15 mmHg (enlarged IVC with >25% respiratory collapse and predominant diastolic forward flow on hepatic vein Doppler) or 20 mmHg (enlarged IVC with minimal or no collapse and solely diastolic forward flow or systolic flow reversal on hepatic vein Doppler), according to the Mayo Clinic protocol [29]. ePASP was then calculated as the sum of eRAP and eRV–RA gradient.
The reproducibility of eRV–RA gradient and eRAP at rest, 20 W and peak exercise was assessed in 15 randomly selected patients. Intra- and inter-observer agreement was evaluated after the same observer and another experienced reader repeated the analysis using intraclass correlation coefficients.
Statistical analysis
Results are reported as mean±sd, median (interquartile range) or n (%). Between-group differences were compared by unpaired t-test, Wilcoxon rank sum test or Chi-squared test, as appropriate. Correlations between invasive haemodynamics and echocardiographic estimates were assessed using Pearson's correlation. The Bland–Altman method was used to assess agreement and bias between invasive and estimated, noninvasive haemodynamic measures.
Results
Subject characteristics
EIPH was present in 73 (75%) subjects. Of participants with EIPH, 49 (67%) had heart failure with preserved ejection fraction, eight (11%) had dilated cardiomyopathy, three (4%) had heart failure with reduced ejection fraction, five (7%) had primary valvular heart disease, seven (10%) had precapillary pulmonary hypertension (six had group I and one had group V pulmonary hypertension) and one (1%) had constrictive pericarditis. Compared to subjects without EIPH, subjects with EIPH were older, more obese, hypertensive and anaemic, and had higher N-terminal pro-brain natriuretic peptide levels (table 1). There were no differences in sex and other comorbidities. Subjects with EIPH were more likely to be treated with angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, β-blockers and loop diuretics. Jugular vein distention and peripheral oedema were more prevalent in EIPH compared to the non-EIPH group, while rales and gallop sounds were rare in both groups. Left ventricle size, mass and ejection fraction were similar between the groups, but the ratio of early diastolic mitral inflow to mitral annular tissue velocities (E/e′) and left atrial volume index were higher in subjects with EIPH (table 1). RV size and systolic function and tricuspid regurgitation severity were similar between the groups.
Baseline characteristics
Resting haemodynamics and echocardiographic-invasive relationships
Subjects with EIPH had higher right and left heart filling pressures, higher pulmonary artery pressures and lower QP by invasive assessment compared to those without EIPH (table 2). Heart rate and systolic blood pressure were similar between the groups.
Baseline invasive and noninvasive haemodynamics
Estimated RA pressure, RV–RA gradient and PASP were obtainable at rest in 97%, 57% and 57% of participants, respectively. The intra- and inter-observer agreement was consistently good at all stages (supplementary table S1). Resting eRAP and ePASP were higher in the EIPH group compared to the non-EIPH group (table 2). Significant correlations were observed between the simultaneous invasive and echocardiographic measurements for RA pressure (r=0.70, p<0.0001), RV–RA gradient (r=0.73, p<0.0001) and PASP (r=0.76, p<0.0001; figure 1). Bland–Altman analyses showed little bias for RA pressure (bias 0.2 mmHg, 95% limits of agreement −7–8 mmHg), RV–RA gradient (bias −1 mmHg, 95% limits of agreement −14–12 mmHg) and PASP (−1 mmHg, 95% limits of agreement −17–14 mmHg; figure 1).
Correlations and Bland–Altman plots between invasive and noninvasive haemodynamic parameters at rest. Modest correlations were observed between the simultaneous invasive and echocardiographic measurements for right atrial pressure (RAP; r=0.70, p<0.0001), right ventricular (RV)–right arterial (RA) gradient (r=0.73, p<0.0001) and pulmonary artery systolic pressure (PASP; r=0.76, p<0.0001). Bland–Altman analyses showed little bias for these estimates. e: estimated; LoA: limits of agreement.
In the pre-procedure area, where patient position could be modified more than in the catheterisation laboratory, eRAP, RV–RA gradient and PASP were obtainable at rest in 97%, 81% and 80% of participants, respectively. Correlations between these non-simultaneous invasive and echocardiographic measurements remained robust for RA pressure (r=0.72, p<0.0001), RV–RA gradient (r=0.70, p<0.0001) and PASP (r=0.74, p<0.0001).
The criteria for eRAP as suggested by the American Society of Echocardiography (ASE) [17] which differs to that used at the Mayo Clinic, also correlated with invasive RA pressure, but less well (r=0.45, p<0.0001). A Bland–Altman plot showed little bias between the ASE-based eRAP and invasively measured RA pressure, but wider limits of agreement (bias 0.6 mmHg, 95% limits of agreement −11–12 mmHg; supplementary figure S1) than those between eRAP estimated by the criteria used at the Mayo Clinic and true RA pressure (figure 1). The superiority of the Mayo criteria as compared to the ASE criteria was consistently observed across different categories of eRAP (supplementary table S2). The lateral tricuspid valve annulus E/e′ ratio correlated only weakly with RAP (r=0.34, p=0.003), with a lateral tricuspid annular E/e′ ratio >6 having a sensitivity of 73% and specificity of 56% for a RA pressure ≥15 mmHg.
Echocardiographic-invasive relationships during exercise
Peak exercise workload (37±14 versus 71±26 W, p<0.0001) and peak V′O2 (8.2±2.7 versus 14.1±4.3 mL·min−1·kg−1, p<0.0001) were both markedly impaired in subjects with EIPH as compared to those without EIPH. During submaximal (20 W) and peak exercise, subjects with EIPH displayed higher left and right heart filling pressures, with lower QP compared to non-EIPH subjects (table 3).
Submaximal and peak exercise invasive and noninvasive haemodynamics among patients with obtainable estimated pulmonary artery systolic pressure with an assumed exercise right atrial (RA) pressure of 10 mmHg (ePASP10)
The number of subjects with obtainable ePASP during 20 W and peak exercise decreased substantially to 16 (16%) and 15 (15%), respectively. This was largely caused by an inability to determine eRAP during exercise (68% both during 20 W and peak exercise). In a subset of patients with obtainable eRAP (n=31 and n=30 during 20 W and peak exercise, respectively), eRAP correlated with invasively obtained RA pressure during exercise (r=0.73, p<0.001), with an underestimation of true RA pressure by 5 mmHg (supplementary figure S2). In this smaller cohort with both eRAP and eRV–RA assessment (n=16 and n=15, respectively), estimated PASP was correlated with invasive measurements, but underestimated true PASP by 3–6 mmHg on average, with wide limits of agreement (supplementary figure S3).
In contrast, the eRV–RA gradient was obtainable in 54% and 45% of subjects at 20 W and peak exercise (table 3). Correlations between invasive and echocardiographic measurements for RV–RA gradient during 20 W and peak exercise were moderately strong (r=0.73, p<0.0001 and r=0.72, p<0.0001, respectively), with little bias (bias −0 mmHg, 95% limits of agreement −16–16 mmHg and bias −1 mmHg, 95% limits of agreement −19–16 mmHg, respectively; figure 2).
Correlations and Bland–Altman plots between the simultaneous invasive and echocardiographic measurements for right ventricular (RV)–right arterial (RA) gradient. Correlation between RV–RA gradient remained significant during a) 20 W and b) peak exercise (r=0.73, p<0.0001 and r=0.72, p<0.0001, respectively), with little bias. e: estimated; LoA: limits of agreement.
Given the inability to assess eRAP during exercise in the majority of subjects, a common assumed RA pressure value of 10 mmHg was next used to provide an estimate of exercise PASP (ePASP10). While ePASP10 correlated well with invasively measured PASP during exercise (r=0.68, p<0.0001 during 20 W and r=0.70, p<0.0001 during peak exercise), it systematically underestimated invasive PASP, with wide 95% limits of agreement (bias 7 mmHg, 95% limits of agreement −15–30 mmHg; bias 9 mmHg, 95% limits of agreement −13–30 mmHg, respectively; figure 3).
Correlations and Bland–Altman plots between the simultaneous invasive and echocardiographic measurements for pulmonary artery systolic pressure (PASP) during exercise. While estimated PASP with an assumed exercise right atrial pressure of 10 mmHg (ePASP10) correlated well with direct PASP during a) 20 W and b) peak exercise (r=0.68, p<0.0001 and r=0.70, p<0.0001, respectively), it underestimated true PASP, with wide 95% limits of agreement (LoA). RV: right ventricular; RA: right arterial; e: estimated.
We further investigated whether estimated RA pressure at rest could predict eRA pressure during peak exercise. Echocardiographic RA pressure at rest was correlated with eRA pressure during peak exercise (r=0.66, p<0.0001), allowing for prediction of peak eRA pressure using rest eRA pressure (peak eRAPpredict=0.83×baseline eRAP+5.4 mmHg, r2=0.43, p<0.001). Peak ePASPpredict (calculated as eRV–RA gradient during peak+peak eRAPpredict) was correlated with direct PASP (r=0.75, p<0.0001; n=45) with relatively little bias (5.3 mmHg), but wide limits of agreement (−14–25 mmHg; supplementary figure S4).
Contributions of right atrial hypertension to PASP during exercise
PASP is equal to the sum of the RV–pulmonary artery pressure gradient and RA pressure. The individual contributions of changes in RA pressure and the RV–RA gradient to total change in PASP during exercise are shown in figure 4. On average, the change in RV–RA gradient accounted for 59% of the change in PASP, while the change in RAP during exercise accounted for 41% of the increase. The underestimation of PASP using assumed RAP values was more pronounced in EIPH subjects and in those with higher RAP during exercise (figure 4 and supplementary figure S5).
Changes in direct and echocardiographic pulmonary artery systolic pressure (PASP), right ventricular (RV)–right arterial (RA) gradient and right atrial pressure (RAP) during exercise. a) Echocardiography could reasonably estimate direct change in RV–RA during exercise. However, it underestimated direct change in RAP either using measured peak RAP or assumed RAP. This led to substantial underestimation of direct change in PASP by echocardiography; b) the underestimation of true PASP using assumed RAP values was more pronounced in exercise-induced pulmonary hypertension (EIPH) subjects and in those with higher RAP during exercise. ePASP10: PASP with an assumed exercise RAP of 10 mmHg; e: estimated.
Diagnostic implications
Finally, we examined the diagnostic ability of exercise stress echocardiography for identification of EIPH. When restricted to the subset of patients with obtainable ePASP, resting ePASP as a continuous variable demonstrated a modest diagnostic accuracy for EIPH (area under the curve (AUC) 0.70) with a sensitivity of 62% and a specificity of 70% at the optimal cut-off value of 37 mmHg (table 4). Within the subgroup of patients with measurable data during exercise (54% at 20 W and 46% at peak exercise), ePASP10 >40 mmHg increased sensitivity to detect the development of EIPH, but reduced specificity, with no improvement in discrimination (AUC 0.72 and 0.67, respectively; table 4). Additionally, applying ePASP10 as a continuous variable displayed moderate discrimination of the groups, with AUC of 0.70 and 0.76 at 20 W and peak exercise, respectively (table 4). The diagnostic ability of peak PASPpredict (prediction of peak RA pressure using eRAP at rest) was slightly improved (AUC 0.81; table 4) Results were similar when excluding EIPH patients with pulmonary hypertension at rest (supplementary table S3).
Diagnostic accuracy for identification of exercise-induced pulmonary hypertension
Discussion
In this simultaneous echocardiographic–cardiac catheterisation study we observed that Doppler-estimated ePASP at rest was well correlated with invasively measured PASP, with little bias, but ePASP was not obtainable in >40% of subjects due to inability to image the tricuspid regurgitation Doppler envelope. During exercise, there was a progressive increase in the number of patients with incomplete echocardiographic data to determine PASP (∼85%), mainly due to the inability to determine eRAP. Using an assumed RAP of 10 mmHg during exercise increased the feasibility of obtaining estimated PASP, but resulted in a significant underestimation of true PASP, especially among patients with more severe EIPH, with wide limits of agreement, due to the fact that marked RA hypertension develops in many patients with EIPH during exercise. Among the 46–54% of subjects where tricuspid regurgitation Doppler could be performed during exercise, exertional ePASP10 improved sensitivity to identify the development of EIPH, but decreased specificity, with no improvement in overall discrimination compared to resting data, but estimation of eRAP during exercise from resting eRAP improved discrimination. These data confirm that the RV–RA gradient can be accurately estimated in many patients during exercise, but failure to account for the severity of RA hypertension during stress, which is often dramatic, leads to marked underestimation of EIPH severity in many patients.
Accuracy of noninvasive estimates of pulmonary artery pressures at rest
Pulmonary hyertension is associated with worsening exercise capacity, increased mortality and greater morbidity in patients with a variety of cardiopulmonary diseases [1–4]. Doppler-derived PASP is often used in the noninvasive evaluation of pulmonary hypertension. Several studies have reported modest correlations between ePASP and directly measured PASP at rest [30, 31], but simultaneous echocardiographic–catheterisation assessments are more limited [18, 32–35]. Except for one study examining group I pulmonary hypertension [33], ePASP has been found to be well correlated with invasively measured PASP at rest (r=0.72–0.97) with acceptable limits of agreement [18, 32, 34, 35]. The current data confirm and extend upon the utility of ePASP as a noninvasive tool to estimate PASP at rest. However, even with highly trained research sonographers, a significant proportion of subjects (>40%) did not have a reliable measurement of ePASP at rest, due to inability to obtain tricuspid regurgitation velocity. The proportion of patients with missing data in this study is very similar to a previous meta-analysis [31]. Even when tricuspid regurgitation spectra are obtainable, it has been reported that accuracy of ePASP depends vitally on the quality of the tricuspid regurgitation Doppler envelope [18]. In the current study, a greater proportion of participants had evaluable data when evaluated in the pre-procedure area, where greater freedom is available to change positions, and the inability to change from the supine position in the invasive laboratory probably contributed to the lower number of participants with adequate data.
Estimating the RV–pulmonary artery pressure gradient during exercise
The importance of identifying exercise-induced pulmonary hypertension has been increasingly recognised in view of its diagnostic, prognostic and potentially therapeutic utilities [8–12]. Like resting pulmonary hypertension [1–4], exercise-induced elevation in pulmonary artery pressures due to pre- and post-capillary aetiologies is associated with increased morbidity and mortality [36–38], and has recently been adopted as an important therapeutic target for both drug and device therapies [39–42]. Invasive exercise haemodynamic testing provides a direct measurement of pulmonary artery pressures with exertion and thus serves as the gold standard, but is difficult to apply as broadly given its invasive nature, technical complexity and cost.
There is increasing enthusiasm for broader utilisation of exercise stress echocardiography studies for clinical purposes, including evaluation for exertional dyspnoea [13]. However, few studies have examined the reliability of PASP estimates during exercise [34, 43]. Only one study has reported assessments using simultaneous catheterisation-echocardiographic evaluation at rest and during exercise [18]. In this study, van Riel et al. [18] observed that despite wide limits of agreement (−30–34 mmHg), noninvasive PASP was modestly well correlated with invasive PASP (r=0.57) with small bias (1.9 mmHg). However, in this study the authors compared invasive and noninvasive RV–RA gradient, not absolute PASP, which is the clinical variable of interest that determines RV afterload. The current data demonstrate how even with robust correlations between invasive and noninvasive RV–RA gradients, there may be substantial error in the estimation of PASP owing to the difficulty with noninvasively assessing RA pressure.
The importance of right atrial hypertension
Consistent with van Riel et al. [18], we found a reasonable correlation between invasive and echocardiographic measurements for the RV–RA gradient during exercise (r=0.72–0.73), with little bias. However, we observed that it was very difficult to obtain diagnostic quality imaging of both the tricuspid regurgitation spectrum and IVC to allow for estimation of PASP, obtainable in only 15% of participants during exercise.
An important observation from this study was that people with EIPH frequently develop substantial RA hypertension during exercise, and this contributes substantially to the total PASP, often exceeding the pressure elevation attributable to the RV–RA gradient (supplementary figure S6). Exertional RA hypertension frequently develops from abnormalities in RV–pulmonary artery coupling, common in patients with pulmonary vascular disease, as well as in patients with increased pericardial restraint, such as patients with obesity [11, 44]. In addition, there may be greater redistribution of blood from the capacitance veins to the right heart during exercise that contributes to worsening RA hypertension [45].
To explore whether use of an assumed RA pressure might produce acceptable results, we next estimated ePASP using a fixed eRAP of 10 mmHg (ePASP10), and observed a modest correlation with directly measured PASP during exercise. However, this estimation systematically underestimated true PASP, especially among patients with EIPH (figure 4). Thus, an important clinical implication from this study is that Doppler-echo assessments of PASP during exercise may substantially underestimate the severity of pulmonary hypertension, and this underestimation is greatest in the population that is most afflicted by pulmonary hypertension and abnormalities in RV–pulmonary artery coupling during exercise. Using assumed values of RAP during exercise was associated with robust sensitivity to detect EIPH (table 4), but at the cost of specificity (43–57%), which would increase the risk of diagnosing normal patients with EIPH.
An estimation of PASP by the addition of the Doppler-derived RV–RA gradient and exercise eRAP predicted by resting eRAP correlated with the true measured exercise PASP, and provided an improved metric to discriminate EIPH from no pulmonary hypertension, underscoring the potential clinical value for an accurate estimate of exercise RA pressure in the assessment of pulmonary vascular haemodynamics, although limits of agreement were wide. Other modalities to estimate RA pressure during exercise, such as imaging of the internal jugular vein or superior vena cava warrant future study to address this problem [46].
Limitations
This is a single-centre study from a tertiary referral centre and as such has inherent flaws relating to selection and referral bias. The sample size is moderate, but this is the largest study for a comparison of simultaneous catheterisation and echocardiography during exercise, and the only study to also estimate simultaneous right atrial pressure to calculate PASP. Although echocardiography was performed by rigorously trained, dedicated research sonographers, the requirement for imaging in the supine, draped patient in the catheterisation laboratory may have limited the availability of Doppler data and IVC assessment.
Conclusions
In patients with obtainable echocardiographic images, the RV–RA pressure gradient can be estimated with reasonable accuracy and precision during exercise. However, RA hypertension develops frequently in patients being evaluated for pulmonary hypertension during exercise, and the inability to noninvasively account for high RAP leads to substantial underestimation of exercise PASP, which may lead to underappreciation of the burden of pulmonary vascular disease. Further study is necessary to identify methods to accurately estimate RAP during exercise to improve assessment of EIPH.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-01617-2019.SUPPLEMENT
Supplementary figure S1. A Bland-Altman plot between the ASE-guided RAP and invasively measured RAP at rest. A Bland-Altman plot between the ASE-guided RAP and invasively measured RAP at rest showed a little bias, but relatively wider limits of agreement (bias 0.6 mmHg, 95% limits of agreement -11 mmHg to 12 mmHg) (n=85). ASE: American Society of Echocardiography; RAP: right atrial pressure. ERJ-01617-2019.FIGURE_S1
Supplementary figure S2. Correlations and Bland-Altman plots between the simultaneous invasive and echocardiographic measurements for RAP during exercise. In a subset of patients with obtainable estimated right atrial pressure (eRAP) (n=31 and n=30 during 20 W and peak exercise, respectively), eRAP correlated with invasively obtained RA pressure during exercise (r=0.73, p<0.001), with an underestimation of true RA pressure by 5 mmHg. Abbreviations as in supplementary figure S1. ERJ-01617-2019.FIGURE_S2
Supplementary figure S3. Correlations and Bland-Altman plots between the simultaneous invasive and echocardiographic measurements for PASP during exercise. In the smaller cohort with both eRAP and eRV-RA assessment (n=16 and n=15), estimated pulmonary artery systolic pressure (ePASP) was correlated with invasive measurements, but underestimated true PASP by 3-6 mmHg on average, with wide limits of agreement. Abbreviations as in supplementary figure S1. ERJ-01617-2019.FIGURE_S3
Supplementary figure S4. Correlations and Bland-Altman plots between the predicted peak PASP using eRAP at rest and simultaneous invasively measured PASP. Echocardiographic RAP at rest was highly correlated with eRAP during peak exercise (r=0.66, p<0.0001), allowing for a prediction of peak eRAP using eRAP at rest (peak eRAPpredict = 0.83*Baseline eRAP + 5.4 mmHg). Peak ePASPpredict (eRV-RA gradient during peak + peak eRAPpredict) was correlated with direct PASP (r=0.75, p<0.0001, n=45) with little bias though limits of agreement were relatively wide (5.3 mm Hg, LOA -14 mmHg to 25 mm Hg). Abbreviations as in supplementary figures S1 and S3. ERJ-01617-2019.FIGURE_S4
Supplementary figure S5. Baseline, submaximal (20 W), and peak exercise for simultaneous invasive and echocardiographic measurements for RAP, RV-RA gradient, and PASP. The underestimation of PASP using assumed RAP values (ePASP10) was more pronounced in exercise-induced pulmonary hypertension (EIPH) subjects. Data are mean±SE. Abbreviations as in supplementary figures S1 and S3. ERJ-01617-2019.FIGURE_S5
Supplementary figure S6. Representative case for the underestimation of true PASP by stress echocardiography. Exercise hemodynamics and echocardiography in a patient presenting exertional dyspnea and normal EF. Pulmonary artery (red) and RA (blue) pressures were moderately elevated at rest. Echocardiography could accurately estimate both RV-RA gradient and RAP at rest. During exercise RAP increased dramatically to 38 mmHg, without an increase in RV-RA gradient (direct PASP 67 mmHg). While exercise RV-RA gradient could be estimated by echocardiography, inferior vena cava image was unobtainable during exercise. Estimated PASP (39 mmHg) using assumed RAP of 10 mmHg substantially underestimated direct PASP by 28 mmHg. This underestimation could even misclassify this patient as having normal pulmonary hemodynamics response with exercise. Abbreviations as in supplementary figures S1 and S3. ERJ-01617-2019.FIGURE_S6
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01617-2019.Shareable
Acknowledgements
The authors thank the staff of the Earl Wood Catheterisation Laboratory and patients who agreed to participate in this study.
Footnotes
This article has an editorial commentary: https://doi.org/10.1183/13993003.02385-2019
This article has supplementary material available from erj.ersjournals.com
This study is registered at clinicaltrials.gov with identifier NCT01418248.
Conflict of interest: M. Obokata has nothing to disclose.
Conflict of interest: G.C. Kane has nothing to disclose.
Conflict of interest: H. Sorimachi has nothing to disclose.
Conflict of interest: Y.N.V. Reddy reports grants from NIH (T32 HL007111), outside the submitted work.
Conflict of interest: T.P. Olson has nothing to disclose.
Conflict of interest: A.C. Egbe has nothing to disclose.
Conflict of interest: V. Melenovsky reports grants from the Czech Healthcare Research Grant Agency (17-28784A), outside the submitted work.
Conflict of interest: B.A. Borlaug reports grants from NIH (R01 HL128526, R01 HL 126638, U01 HL125205, U10 HL110262), outside the submitted work.
Support statement: This study was supported by an award from the Mayo Department of Cardiovascular Diseases. B.A. Borlaug is supported by R01 HL128526, R01 HL 126638, U01 HL125205 and U10 HL110262.
- Received August 14, 2019.
- Accepted November 2, 2019.
- Copyright ©ERS 2020
References
