Chronic Thromboembolic Pulmonary Disease Phenotypes Using Upright Invasive Cardiopulmonary Exercise Testing

Authors: Pradhab Kirupaharan, Jehad Azar, Xiaofeng Wang, Yifan Wang, Alice Goyanes, Tereza Izakovich, Michael Z. Tong, Haytham Elgharably, Ihab Haddadin, Adriano R. Tonelli, Gustavo A. Heresi

Abstract

Patients with chronic thromboembolic pulmonary disease (CTEPD) without resting pulmonary hypertension (PH) frequently experience dyspnea which is not explained by stationary assessment and therefore requires cardiopulmonary exercise testing for further diagnostic insight. Ventilatory inefficiency in CTEPD is commonly identified; however, consensus definitions are lacking. Upright iCPET and careful definition of ventilatory inefficiency (i.e., lack of reduction in VD/VT with exercise) may allow optimal characterization of patients with chronic thromboembolic disease and specifically helps correctly identify patients with preload insufficiency who may otherwise be misclassified or undiagnosed.

Exercise intolerance is frequent in survivors of acute pulmonary embolism (PE) [1, 2]. It is well known that a minority of patients are ultimately diagnosed with chronic thromboembolic pulmonary hypertension (i.e. CTEPH) which is a disease hallmarked by resting pulmonary hypertension due to incomplete thrombotic resolution, aberrant vascular remodeling and small vessel arteriopathy [3]. Despite a more reassuring prognosis, patients with chronic thromboembolic pulmonary disease (CTEPD) without resting pulmonary hypertension can still experience significant symptom limitation and impaired quality of life [4]. These patients may demonstrate physiologic limitations directly attributable to vascular occlusion from chronic PE (most commonly, ventilatory inefficiency and/or exercise precapillary PH) which warrant distinct individualized treatments [5].

Recent guidelines recommend consideration of cardiopulmonary exercise testing (CPET) in the evaluation of patients with dyspnea or exercise limitation following acute pulmonary embolism [6]. Importantly, the incorporation of invasive hemodynamic measurements with right heart catheterization (RHC) and arterial catheterization during exercise is essential to confirm and accurately distinguish certain diagnoses like exercise precapillary PH, ventilatory inefficiency, preload insufficiency and occult diastolic heart failure [7, 8, 9, 10]. Previous authors have highlighted abnormal dead space ventilation, reduced stroke volume augmentation or both in patients with ongoing exertional dyspnea following pulmonary embolism [11]. The most precise definition of ventilatory inefficiency (or abnormal dead‐space ventilation) is debated without consensus amongst experts or contemporary guidelines. Moreover, the incidence of other pathophysiologic limitations, like preload insufficiency, in patients with CTEPD, is not well known and may be underrecognized. In this study, we report a cohort of patients with CTEPD without resting pulmonary hypertension evaluated with upright invasive cardiopulmonary exercise testing. We also propose defining ventilatory inefficiency using the lack of decrement of VD/VT with exercise using upright iCPET can help more accurately identify phenotypes and physiologic abnormalities in patients with CTEPD.

Study Design and Methods

This study was approved by our Institutional Review Board. We retrospectively identified 94 consecutive patients with chronic pulmonary embolism on chest imaging and persistent symptoms who underwent upright invasive cardiopulmonary exercise testing from March 2018 to December 2022. All patients underwent preceding evaluation with right heart catheterization in supine position for baseline hemodynamic determinations. All patients had chronic pulmonary embolism confirmed by thoracic imaging, after receiving more than 3 months of anticoagulation therapy. Patients without chronic PE on imaging (n = 8) or with resting supine PVR ≥ 3 Wood units (n = 25) were excluded (Figure 1).

Baseline Right Heart Catheterization and Hemodynamics

Baseline RHC was done in supine position in the outpatient setting. Patients could eat for up to 4 h before and drink water for up to 2 h before the test. Patients did not receive fluids before the test, even when the RAP was < 5 mmHg. We did not modify any pharmacological or non‐pharmacological interventions that patients were receiving before the test, i.e. we continued the prescribed anticoagulation and other treatments such as beta‐blocker or calcium‐channel blockers. RHC was performed under local anesthesia (1% lidocaine) through the right or left internal jugular vein. Pressure transducers were zeroed at the mid‐axillary line while supine. A 7.5 F pulmonary artery catheter (Biosensor International) was advanced through an 8.5 F introducer to the pulmonary arteries. Pulmonary artery catheter position was confirmed by pressure waveform analysis and fluoroscopic imaging. We then recorded right atrial pressure (RAP), systolic, diastolic and mean PAP (mPAP), and pulmonary artery wedge pressure (PAWP) using established protocols, waveform tracings and calipers. For RAP, we recorded mean values, averaged across the respiratory cycle. For PAP and PAWP, we recorded measurements both at end‐expiration and averaged across the respiratory cycle [12]. PAWP was recorded at mid “a” wave (without including the “v” wave). Cardiac output (CO) was measured by the thermodilution method, averaging values with less than 10% difference [13]. We calculated the cardiac index (CI = CO/body surface area) and PVR (using end‐expiratory determinations). All patients underwent radial artery catheterization for continuous arterial blood pressure monitoring and sampling of arterial blood at baseline and every stage of exercise.

Exercise Protocol and Measurements

After resting supine determinations, patients sat in an upright cycle ergometer. Pressure transducers were zeroed at the 4th intercostal level. Loaded exercise was completed in incremental stages (increasing 20 W every 2 min with a cadence of 60 rpm) under the supervision of an exercise physiologist. Exercise was terminated by exhaustion.

Patients were connected to a metabolic cart (Ultima CPX, MGC Diagnostics, Saint Paul, MN) using a 7450 V2 Series mask with MEDGRAPHICS Mask Adapter (Hans Rudolph Inc, Shawnee, KS) that was fitted for each patient to prevent air leaks, using a measuring tool provided by the manufacturer. The metabolic cart was calibrated before each procedure following manufacturer's recommendations [9]. Once the pulmonary catheter, arterial line, face mask and corresponding tubing were secured, we obtained resting sitting determinations. While resting in the sitting position, we obtained new hemodynamic and gas exchange determinations, and measured oxygen consumption (⩒O2). At baseline and every stage of exercise, we measured oxygen consumption (⩒O2); that together with hemoglobin concentration, arterial (radial artery) and mixed venous (distal port of the pulmonary artery catheter) O2 oxygenation (by co‐oximetry), were used to calculate cardiac output using Fick methodology where CO = ⩒O2 in milliliters per minute/(arteriovenous oxygen content difference × 10) [14]. Additional hemodynamic determinations were repeated at every stage of the exercise, and these include PAP, PAWP, RAP and CO all of which were averaged across the respiratory cycle. We calculated PVR at every stage. mPAP/CO and PAWP/CO slopes were reported using two point calculations incorporating mPAP, PAWP and CO determinations at baseline and peak exercise [15].

Dead space to tidal volume ratio (VD/VT) was calculated using the Bohr equation and direct measurement of mixed expired carbon dioxide (PECO2) and arterial carbon dioxide (PaCO2), where VD/VT = (PaCO2 – PECO2)/PaCO2). Maximal Voluntary Ventilation (MVV) was measured in each patient. Breathing reserve was calculated as the ratio between ((MVV‐VEmax)/MVV) x 100. Predicted values of ⩒O2 peak were calculated using the following equations proposed by Wasserman for adult males (⩒O2 peak = weight in kilograms x [50.72 ‐ (0.372 x age in years)]) and adult females ((weight + 43) x (22.78‐ 0.17 x age in years) [14].

Phenotype Criteria

We defined four distinct phenotypes presented in Table 1.

Statistical Analysis

Normally distributed continuous variables are presented as mean (SD), and groups were compared using one way ANOVA test. Continuous variables that are not normally distributed are presented as median (Q1, Q3), and groups were compared with Kruskal–Wallis H test. Categorical variables are presented as count (percentage) and groups compared with Fisher‐exact test. The level of statistical significance was set at p < 0.05 (two‐tailed). The statistical analyses were performed using the statistical package IBM SPSS version 22 (IBM).

Results

Patient Characteristics

We included 59 consecutive patients in the study. Patient characteristics and baseline hemodynamic data are summarized in Tables 2 and 3. Pertinent circulatory, respiratory, metabolic and gas exchange data obtained by iCPET are shown in Table 4. The mean age at diagnosis was 51.7 years old Patients in the cohort were predominantly female (n = 44, 74.5%) and Caucasian (n = 49, 83%) with NYHA II/III functional class (n = 54, 91.5%) (Table 1). Patients with EPH were significantly older with median age of 70 (52–72, p = 0.003). BMI was greater than 30 kg/m^2^ in all groups. Comorbidities including diabetes mellitus, systemic hypertension and obstructive sleep apnea were also similar amongst all groups. Incidence of prior DVT, recurrent PE, and commonly recognized risk factors for chronic thromboembolic disease (i.e. history of cancer, thyroid replacement therapy, asplenia, antiphospholipid syndrome, connective tissue disease or inflammatory bowel disease) were similar amongst each group Quantitative clot burden scores by computed tomography (CT) were not statistically different between groups [16, 17].

Phenotyphe Classification

The most common phenotype was EPH (n = 16, 27%) followed by combined ventilatory inefficiency and EPH (n = 9, 15%), ventilatory inefficiency (n = 8, 13.5%), and preload insufficiency (n = 8, 13.5%). Eighteen patients (31%) did not fulfill any of these diagnostic criteria (Figure 1).

We compared phenotype classification according to three commonly used definitions of ventilatory efficiency (Figure 2). Use of specific VD/VT thresholds (i.e. VD/VT > 0.38 at rest and > 0.27 at anaerobic threshold) led to fewer patients being diagnosed with ventilatory inefficiency (n = 4, 7%), or ventilatory inefficiency with EPH (n = 6, 10%) and more patients fulfilling none of our diagnostic criteria (n = 18, 30%). Defining ventilatory inefficiency using VE/⩒CO2 criteria (i.e. VE/⩒CO2 > 34 at anaerobic threshold) led to a significant increase in patients being classified with ventilatory inefficiency with EPH (n = 14, 24%) or ventilatory inefficiency alone (n = 21, 36%). and reduced the number of patients being detected with preload insufficiency (n = 4, 7%).

Hemodynamic Determinations at Baseline

Baseline hemodynamic data are summarized in Table 3. During supine baseline hemodynamic assessment, patients with combined ventilatory inefficiency and EPH had the highest median mPAP 31 mmHg (10,24, p = 0.001) and PVR 2.27 WU (1.67,2.36, p = 0.003). Resting mRAP and PAWP were elevated and comparable between groups. Meanwhile, cardiac index was reduced (2.34 L/min/m2, 2.04, 2.84, p = 0.02) and significantly lower in patients with preload insufficiency.

Invasive Cardiopulmonary Exercise Testing

Circulatory, respiratory, metabolic and gas exchange data obtained by iCPET are shown in Table 4. Patients achieved a workload > 80 W without significant difference between groups. Patients achieved > 85% maximal predicted heart rate (MPHR) in all groups with no significant difference in % MPHR between groups. RER at peak exercise was above 1.1 and comparable amidst groups. There was a comparable limitation in peak ⩒O2 (% predicted and indexed to body weight) between groups.

Respiratory and Gas Exchange Measurements

Breathing reserves were lowest in patients with combined ventilatory inefficiency and EPH (11%, 5–35, p = 0.14). Breathing reserve was normal in the preload insufficiency and patients with none of the described phenotypes. Baseline spirometry (FEV1%, FVC%, FEV1/FVC ratio) and DLCO measurements were normal and comparable amongst groups. Baseline PaCO2 was lowest in patients with combined EPH and ventilatory inefficiency 33 mmHg (29,38, p = 0.006). VE/⩒CO2 at AT was highest in patients with combined ventilatory inefficiency and EPH 43 [39,43] compared to those with ventilatory inefficiency alone 39 [35,44], EPH 33 [29,36], preload insufficiency 35 [31,43] and none 32 [29,38], respectively p = 0.002. PET CO2 at AT was lowest in patients with combined ventilatory inefficiency and EPH 30 [29,31] vs 36 [31,39] in ventilatory inefficiency alone, 36 [34,36] in EPH alone, 33 [29,39] in preload insufficiency and 48 [38,48] in the none group (p = 0.008). Dead space to tidal volume ratio (VD/VT) was calculated using direct measurement of end tidal and arterial carbon dioxide in most cases (n = 46, 78%). Dead space to tidal volume ratio (VD/VT) was normal at rest and similar between groups. VD/VT at AT was highest in patients with combined ventilatory inefficiency and EPH 0.31 (0.3,0.32, p = 0.025). Pa‐ETCO2 gradient at peak exercise was high in patients with combined ventilatory inefficiency and EPH 4 [2,11], ventilatory inefficiency 3 [1,6], and EPH 3 [1,5], compared to those with preload insufficiency 1 (−5,4) and none ‐1 (‐6,‐2), p = 0.024. A‐a gradient at peak exercise was highest in patients with EPH with or without ventilatory inefficiency 28mHg [21,53] and 34mHg [25,41] respectively, p‐0.01).

Cardiovascular and Hemodynamic Measurements

⩒O2/work slope, % predicted O2 pulse at peak exercise, % predicted cardiac output at peak exercise were comparable between groups. Heart rate at baseline 67 bpm (50,90, p = 0.008) and peak exercise 133 bpm (120,143, p = 0.016) were lowest in patients with combined ventilatory inefficiency and EPH.

Upright mean PAP at baseline 20 mmHg (19,22, p < 0.001) was abnormal and highest in patients with combined ventilatory inefficiency and EPH compared to other groups. mPAP at peak exercise was highest in patients with combined ventilatory inefficiency and EPH 44 mmHg [41,47] vs 43 mmHg [36,51] in EPH vs 24 [18, 19] in ventilatory inefficiency) vs 21 [18, 20]. in preload insufficiency vs 27 [21, 22] in none, p < 0.001. mPAP/CO at peak exercise was abnormally elevated and highest in EPH 4.47 WU (3.87, 5.3) and ventilatory inefficiency with EPH 3.82 WU (3.27, 4.16) compared to patients with ventilatory inefficiency 2.31 WU (1.47, 2.67), preload insufficiency 1.93 WU (1.51, 2.72) and none 2.10 WU (1.65, 2.41), p < 0.01. MPAP/CO slope reflected the same relationship and was highest in patients with EPH 5.34 WU (3.84, 7.08), and those with EPH and ventilatory inefficiency 3.41 (2.89, 4.73), p < 0.001. PVR at peak exercise was abnormally elevated and highest in patients with combined disease 2.21 WU (1.81, 3.12), and EPH 2.39 WU (2.05, 3.13), compared to ventilatory inefficiency 1.35 WU (0.91, 1.69), preload insufficiency 1.3 WU (1.00, 1.54) and none 1.05 WU (0.8,1.24), p < 0.001. Importantly, PAWP at peak and PAWP/CO slopes were within normal limits (< 25 mmHg and < 2 mmHg/L/min, respectively) in all groups.

Discussion

This is the first study to our knowledge which specifically analyzes a real‐world cohort of patients with confirmed chronic thromboembolic pulmonary disease without resting pulmonary hypertension using upright invasive cardiopulmonary exercise testing with radial artery catheterization. This study highlights advantages of this specific modality to help clarify mechanisms of exertional symptoms in the absence of resting pulmonary hypertension which may include exercise pulmonary hypertension with elevated PVR, ventilatory inefficiency, combined defects or preload insufficiency. We also highlight the use of VD/VT (specifically lack of decrement with exercise) measured by arterial blood gas to identify inefficient gas exchange due to pulmonary vascular disease as other definitions may either be too restrictive or less specific. Lastly, we show the specific value of upright positioning during iCPET to identify preload insufficiency which may be an underrecognized cause of dyspnea in CTEPD. Dyspnea and reduced functional capacity are prevalent in survivors of acute pulmonary embolism and can be seen in over 50% of patients [3, 18, 23, 24, 25, 26]. There is significant heterogeneity in post‐PE dyspnea and initial evaluation of residual symptoms despite adequate anticoagulation includes biochemical testing, echocardiography, ventilation perfusion scan and cross‐sectional imaging (i.e., CT pulmonary angiogram), to identify symptoms attributable to chronic PE. Right heart catheterization at rest is recommended to clarify the presence of pulmonary hypertension (i.e., chronic thromboembolic pulmonary hypertension or CTEPH).

Cardiopulmonary exercise testing has also been recently recommended as a potential tool to highlight impairments related to chronic PE that are not evident with stationary testing (i.e., ventilatory inefficiency, exercise pulmonary hypertension) or to identify alternative causes of physiologic limitation [6]. Invasive cardiopulmonary exercise testing further combines conventional CPET with pulmonary arterial and radial arterial catheterization to provide comprehensive and incremental measurement of ventilation, gas exchange, hemodynamic response with progressive exercise and allows for direct cardiac output measurement through the Fick Method, and more precise gas exchange measurement [9, 15, 21, 27]. The value of arterial line catheterization, not only for optimal gas exchange analysis but to allow more precise cardiac output measurement was recently highlighted by Campedelli et al, wherein use of arterial line cardiac output calculation during iCPET was shown to more accurately classify exercise pulmonary hypertension particularly in those with Raynaud's, scleroderma or darker skin tones [21].

Patients may experience significant dyspnea due to increased alveolar dead space in areas of chronic vascular obstruction, leading to inefficient ventilation [27]. Despite increasing recognition of ventilatory inefficiency, a consensus definition has not been delineated in guidelines or expert statements. Numerous criteria at rest and during exercise have been proposed and may incorporate abnormal VE/⩒CO2 definitions (i.e., VE/⩒CO2 slope from onset to exercise, abnormal VE/⩒CO2 ratio at anaerobic threshold, VE/⩒CO2 nadir) or VD/VT thresholds at rest and anaerobic threshold (Hansen et al., 2015) [27]. Our study suggests that the lack of decrement of VD/VT between rest and anaerobic threshold, especially if determined via direct arterial blood gas measurement during iCPET, is a more specific sign of ventilatory inefficiency than VE/⩒CO2 at AT or specific VD/VT thresholds at rest or AT which have been previously used. For instance, defining disease using VE/⩒CO2 > 34 at anaerobic threshold, led to dramatic increases in classification of ventilatory inefficiency compared to VD/VT criteria which led to fewer patients being diagnosed with EPH, preload insufficiency or having none of our phenotypes (Figure 2).

Fernandes and colleagues describe only moderate correlation between VE/⩒CO2 and VD/VT at AT [11]. The fundamental advantage of VD/VT is demonstrated when examining the alveolar ventilation VE = 863 × ⩒CO2/(Paco 2 × [1 − VD/VT). Ventilation is therefore directly related to the VD/VT as well as CO2 production and partial pressure of arterial carbon dioxide. VE/⩒CO2 is therefore complex and dependent on a host of cardiopulmonary as well as peripheral factors (musculoskeletal, renal, and neurologic) which affect chemosensitivity and neural drivers of respiration (i.e. acid‐base balance, pulmonary baroreceptor input, mitochondrial and enzymatic activity, sympathetic activation, hypoxemia) and thus perturbations of VE/⩒CO2 may be unrelated to abnormal pulmonary perfusion or pulmonary vascular disease [20]. Therefore, while normal VE/⩒CO2 during exercise typically excludes significant pulmonary vascular disease, abnormal VE/⩒CO2 may overestimate ventilatory inefficiency due to excessive increases in minute ventilation stemming from unrelated conditions. For instance, hyperventilation syndromes and/or dysfunctional breathing unrelated to cardiopulmonary disease (i.e. due to underlying anxiety or apprehension during the procedure, or chronic pain and discomfort amongst other reasons) may increase VE/⩒CO2 due to predominant reductions in arterial carbon dioxide concentration rather than reductions in VD/VT due to alveolar dead space. Hyperventilation syndromes can also be quite prevalent in certain etiologies of preload insufficiency (i.e. POTS) which may explain why fewer patients characterized with preload insufficiency (4 vs 8 patients) when using VE/CO2 criteria in our cohort compared to either VD/VT criteria (Figure 2) [9]. Furthermore, these patients with preload insufficiency have inherent circulatory limitation due to low ventricular preload to exercise and may develop early anaerobic threshold with increased ⩒CO2 leading to excessive ventilation [27]. Therefore, VE/⩒CO2 may not be distinct enough to identify physiologic rationale to identify patients who could specifically benefit from reperfusion therapies in chronic thromboembolic disease without resting pulmonary hypertension. Healthy individuals decrease VD/VT with exercise due to tidal volume recruitment exceeding increases in dead space, and thus a lack of reduction in VD/VT with exercise should be considered inappropriate. The previously mentioned study of 40 patients with residual dyspnea following acute pulmonary embolism evaluated by (noninvasive) cardiopulmonary exercise testing found 35% of patients to have ventilatory inefficiency when defined by specific VD/VT thresholds (i.e., VD/VT > 0.38 at rest and > 0.27 at AT). However, we show that application of this VD/VT definition in our cohort led to fewer patients being diagnosed with ventilatory inefficiency (both with or without EPH) and more patients being characterized as fulfilling none of our phenotypes. Similar thresholds for VD/VT at peak exercise (i.e. > 0.28) have been previously shown to be poorly sensitive in detecting pulmonary vascular limitation in patients with pulmonary hypertension [28]. Therefore, restrictive VD/VT thresholds may overlook patients with subtle but inappropriate ventilatory responses. Increased arterial to end tidal CO2 (Pa‐ETCO2) gradient during exercise has also been shown to be a sensitive marker of inadequate gas exchange in patients with pulmonary vascular disease due to resting hypocapnia, exertional hypoxemia and resultant alterations to chemosensitivity [27]. Patients in our cohort with ventilatory inefficiency, EPH and combined impairments (which are all attributable to pulmonary vascular disease) thereby showed expected and significant increases in Pa‐ETCO2 gradients at peak exercise; 3 [1, 6] vs 3 [1, 5] vs 4 [2, 11], p = 0.024, respectively.

During exercise in healthy individuals, vascular dilation and recruitment of a compliant pulmonary circulation are necessary to maintain or reduce pulmonary vascular resistance (< 1.5 WU) and maintain mean pulmonary artery pressures less than 30 mmHg at a cardiac output < 10 L per minute [19, 29]. Exercise pulmonary hypertension can be defined by elevated mPAP/CO (> 3 mmHg/L/min) between rest and peak exercise using multipoint slope data, though other authors have highlighted that using mPAP/CO > 3 mmHg/L/min at peak exercise maintains excellent sensitivity and specificity [19]. Though it is essential to identify a precapillary component of disease when implicating chronic thromboembolism, and thus this definition should be further clarified by elevated PVR during exercise [30]. One extensive study of patients with unexplained dyspnea by Huang et al., further enrich the definition of exercise pulmonary hypertension using a maximal exercise PVR cut off of 1.5 WU (Huang et al., 2017). Exercise PH with elevated PVR was the most common phenotype in our cohort (n = 18, 30%). Interestingly, this study also informs us that patients with exercise PH with elevated PVR may be symptomatic despite normal predicted (%) O2 pulse at peak exercise and normal predicted (%) CO at peak exercise.

Preload insufficiency is increasingly appreciated as a cause of unexplained dyspnea that requires iCPET in the upright position for diagnosis [15]. This condition is hallmarked by inadequate ventricular preload leading to reduced stroke volume augmentation and resultant cardiac output during exercise which can be caused by a host of conditions that may impair intravascular volume, venous flow, or tone [9]. Germane to patients with chronic thromboembolic disease, venous return may be impaired in patients with thrombosis or stenosis of the inferior vena cava, pelvic or lower extremity veins. For example, two of the eight patients diagnosed with preload insufficiency in our cohort had contributory calcification and stenosis of the inferior vena cava. By identifying 13.5% of patients with preload insufficiency we highlight a specific group of patients who may have been mischaracterized as having low stroke volume reserve, circulatory limitation or elevated VE/⩒CO2 due to chronic PE rather than underlying preload insufficiency.

Nearly one third of our patients with persistent dyspnea had no impairment attributable directly to underlying thromboembolic disease (i.e. EPH, ventilatory inefficiency, or both). This is consistent with work by previous authors who have shown that symptom limitation is not always related to persistent vascular obstruction and may be related other comorbidities or muscular deconditioning [31].

The optimal management for CTEPD without resting PH remains unclear and should be individualized based on the underlying exercise impairment and severity of symptom limitation. In contrast to CTEPH, there is no consensus recommendation for pulmonary thromboendarterectomy for patients with CTEPD without resting pulmonary hypertension [22] CTEPD patients without resting PH do not have impaired survival from cardiopulmonary pathology. However, symptomatic patients with CTEPD without resting PH can derive substantial symptomatic benefit from carefully selected mechanical treatment to alleviate persistent perfusion defects at expert PH centers [4, 22]. In one observational study of patients with CTEPD who underwent pulmonary thromboendarterectomy (PTE), all nine patients reported improved functional classification (from WHO FC II/III. These authors concluded this was likely due to improvement in post‐operative RV‐SV response and normalization of ventilatory efficiency [32]. Taboada and colleagues also demonstrated significant improvements in 6‐min walk distance, functional classification (WHO FC II/III to WHO FC I/II) and quality of life scores in selected CTEPD patients treated with PTE [33]. Wiedenroth and colleagues also demonstrated feasibility and improvement in functional classification in a group of ten inoperable CTEPD patients who underwent balloon pulmonary angioplasty (BPA) [34]. Seventeen patients in our cohort underwent mechanical therapy with PTE or BPA, 14/17 patients had consistent iCPET data supporting pulmonary vascular limitation (i.e. ventilatory inefficiency, EPH, or combined disease). Importantly, a substantial proportion of patients did not have abnormal physiology on iCPET to warrant mechanical intervention for CTPED, underscoring the importance and usefulness of careful exercise testing with pulmonary and radial artery catheterization to rule out pathology in patients with significant symptom burden and exercise limitation. We acknowledge that there were rare exceptions (2 patients classified as none, and 1 classified as having preload insufficiency) with significant chronic thromboembolic disease who were offered mechanical therapy due to refractory symptoms (i.e., chest pain) despite extensive comorbidity optimization.

Our study does have limitations. The sample size of this retrospective cohort was small, albeit comparable to recent physiologic or hemodynamic studies in assessing post‐PE dyspnea. Exercise testing continues to evolve and has become more sophisticated at our institution even since the time of patient inclusion for this study (2018–2022). For instance, we recognize and value the use of multipoint slope data for classification of exercise pulmonary hypertension and exercise induced diastolic dysfunction, however due to missing data in earlier patients we elected to use our composite definition which has also been supported as being clinically relevant in separate literature [35]. We did not incorporate PAWP/CO thresholds to exclude patients in our definition of exercise pulmonary hypertension. This EPH cohort of patients was older with more prevalent diabetes mellitus, hypertension and OSA and therefore included a few patients with PAWP/CO > 2 however the median PVR at peak exercise was significantly elevated highlighting that some patients may have combined post and precapillary disease which should not preclude individualized treatment of chronic thromboembolism. We also acknowledge that numerous PVR thresholds at peak exercise have also been described in large cohorts, however we elected to use our cited definition based on large retrospective experience from a highly experienced center [19]. Future studies of patients with CTEPD should aim to enroll patients prospectively with a larger cohort and matched control with incorporation of post‐intervention physiologic data to better understand the therapeutic implications of diagnostic phenotyping.

Author Contributions

Pradhab Kirupaharan: Participated in data collection, analysis, literature review and interpretation of results, drafting and critical revision of the final manuscript. Jehad Azar: Participated in the design of the study, data collection, interpretation of the results, drafting and writing of the manuscript. Xiaofeng Wang: Participated in the data analysis, interpretation of the results and critical revision of the manuscript. Yifan Wang: Participated in the statistical analysis, interpretation of the results and critical revision of the manuscript. Alice Goyanes: Participated in interpretation of the results, writing and critical revision of the manuscript. Tereza Izakovich: Participated in literature review, interpretation of the results and critical revision of the manuscript. Haytham Elgharably: Participated in interpretation of the results, writing and critical revision of the manuscript. Michael Z. Tong: Participated in interpretation of the results, writing and critical revision of the manuscript. Adriano R. Tonelli: Participated in statistical analysis, interpretation of the results, writing and critical revision of the manuscript. Gustavo A. Heresi: Participated in the design of the study, data collection, statistical analysis, interpretation of the results, writing and critical revision of the manuscript for important intellectual content and final approval of the manuscript submitted.

Ethics Statement

This study has obtained ethics approval from the Cleveland Clinic Institutional review board with approved use of patient data for the present study (IRB #16‐872, IRB #8097). Informed consent was obtained from all patients.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.