A review of exercise pulmonary hypertension in systemic sclerosis

Abstract

In general, pulmonary vascular disease has important negative prognostic implications, regardless of the associated condition or underlying mechanism. In this regard, systemic sclerosis is of particular interest as it is the most common connective tissue disease associated with pulmonary hypertension, and a well-recognized at-risk population. In the setting of systemic sclerosis and unexplained dyspnea, the concept of using exercise to probe for underlying pulmonary vascular disease has acquired significant interest. In theory, a diagnosis of systemic sclerosis–associated exercise pulmonary hypertension may allow for earlier therapeutic intervention and a favorable alteration in the natural history of the pulmonary vascular disease. In the context of underlying systemic sclerosis, the purpose of this article is to provide a comprehensive review of the evolving definition of exercise pulmonary hypertension, the current role and methodologies for non-invasive and invasive exercise testing, and the importance of the right ventricle.

Introduction

Connective tissue disorders (CTDs) are often associated with diseases of the pulmonary vasculature. Systemic sclerosis (SSc), the CTD most commonly associated with pulmonary vascular involvement, is characterized by immune-mediated microvascular injury leading to progressive fibrosis of skin, lungs, and multiple other visceral organs. ¹ When this pattern of injury affects the pulmonary vascular bed, pulmonary arterial hypertension (PAH; Group 1 World Health Organization (WHO) classification) may develop. ² Of the autoimmune diseases classified by the American College of Rheumatology, SSc or the SSc-spectrum of disease has the strongest association with resting PAH (rPAH) and represent ~75% of a previously reported modern era CTD-PAH cohort. SSc-related pulmonary hypertension (SSc-PH) has a prevalence of ~25% ³ and may be as high as 69% in systematically screened and enriched SSc populations. ⁴ , ⁵ Importantly, left ventricular dysfunction (Group II), interstitial lung disease (ILD; Group III), and chronic thromboembolic PH (CTEPH; Group IV) are also potential mechanisms of resting PH (rPH) reported in SSc.^6–10

After the advent of angiotensin-converting enzyme inhibitors as treatment for SSc-associated renal crisis, SSc-related rPAH (SSc-rPAH) and/or ILD are now the leading causes of SSc-related death. ⁷ , ¹¹ , ¹² Survival of untreated SSc-rPAH is 50% at 1 year compared to 90% survival at 1 year for isolated SSc-ILD (without PH). ¹³ In the era of PH-specific therapy, the US-based multicenter, prospective REVEAL Registry reported a 3-year survival in incident cases of 51% in SSc-rPAH (n = 500) versus 76% in non-CTD-rPAH (n = 304; mostly idiopathic PAH (IPAH)). ¹⁴ This inferior survival in SSc-rPAH compared to IPAH is well documented by several investigators.^15–17 Given the poor prognosis of SSc-rPH (regardless of mechanism), it is important to optimize current screening efforts.

Exercise provocation is a powerful diagnostic tool that can be used to probe for pulmonary vascular disease (PVD) prior to the development of rPH, particularly in the assessment of a high-risk population such as SSc. The purpose of this comprehensive review is to examine the available literature with regard to the use of exercise in the evaluation of PVD in SSc. We will focus on the epidemiology and prognostic implications of exercise PH (ePH) as well as the dilemmas surrounding definitions and diagnosis of this clinically relevant phenotype of PH in SSc. For purposes of this article, PAH is classified as WHO Group I PH and is differentiated from the other PH WHO Groups (II–V). ²

Definitions

The definition of normal resting mean pulmonary artery pressure (mPAP) was originally established by Paul Wood in 1958 as <20 mmHg as assessed by right heart catheterization (RHC). ¹⁸ The definition of rPH is currently mPAP ⩾ 25 mmHg (at rest) and was the criteria used in the original National Institutes of Health (NIH) Registry. ¹⁹ In healthy individuals, the mPAP at rest is 14.0 ± 3.3 mm Hg with an upper limit of normal of approximately 20.6 mmHg. ²⁰ The upper limit of normal mPAP during exercise was first defined in 1987 as <30 mmHg, ²¹ without any established threshold for workload or cardiac output (CO). However, the concept of ePH has become a topic of extensive debate. In a systematic review of pulmonary artery pressure during exercise in >1100 normal subjects of varied age and gender, Kovacs et al. ²⁰ reported that 21% of non-trained normal subjects below 50 years of age at maximal exercise and half of normal subjects above 50 years of age at slight exercise demonstrated a mPAP ⩾ 30 mmHg. As a result of the heterogeneity in mPAP increases noted during exercise in normal subjects, the lack of an identified prognostic significance of ePH, and the lack of a standard for exercise testing (i.e. exercise type and modality), the entity of ePH was formally excluded from the diagnostic criteria of PH in 2008 at the 4th World Symposium on Pulmonary Hypertension. ²¹ , ²²

Despite these issues, ePH remains an area of active study and some experts hope to re-define the diagnostic criteria for this disease. Today, it is well accepted that an abnormal response to exercise is defined by the following criteria: (a) mPAP > 30 mmHg and (b) total pulmonary resistance (TPR = mPAP/CO) > 3 Wood Units (WU). In other words, the mPAP should not increase by more than 3 mmHg for each liter the CO increases during exercise.^23–26 However, this definition does not allow one to distinguish the etiology of the abnormal exercise response which may be the result of intrinsic PVD, left heart disease, parenchymal lung disease, increased intrathoracic pressure, advanced age, ²⁷ or a combination of the above. The assessment of the left heart filling pressure at rest and during exercise is assessed by the pulmonary artery wedge pressure (PAWP) and can further distinguish these phenotypes. Importantly, D’Alto et al. ²⁸ recently demonstrated the utility of a simple volume load during a resting RHC (rRHC) and the heterogeneity of the pulmonary hemodynamic response in SSc patients without rPH, particularly with regard to “unmasking” left heart disease. Unfortunately, there is no established threshold for the PAWP that distinguishes normal from abnormal during exercise. Studies have used a cut-off PAWP values ranging from 15 to 25 mmHg. Eisman et al. ²⁹ recently demonstrated that a ∆PAWP/∆CO slope > 2 mmHg/L/min during exercise distinguished abnormal (“heart failure with preserved ejection fraction (HFpEF)”) from normal controls and was associated with decreased exercise capacity and poor composite cardiac outcomes, although SSc was not represented in this cohort. This study begins to address what threshold value distinguishes normal from abnormal PAWP during exercise. A future, refined definition of ePH should include multiple variables to better characterize the pathophysiology underlying the abnormal exercise response for a given patient.

Epidemiology

The estimated prevalence of CTD-related rPAH is 2.3 to 10 cases per million ³⁰ , ³¹ and approximately 2.93 cases per million for SSc-related rPAH. ⁷ Among patients with CTD-PAH, the prevalence of SSc-PAH in two studies was as high as 74%–76%. ¹⁶ , ³⁰ In a meta-analysis of five European studies, rPAH was present in 5%–10% of SSc cases; ³² however, the composite literature reports that rPAH may be present in up to 27% of SSc patients.^3,13,33–35

Compared to rPAH in SSc, the epidemiology of ePH and ePAH in SSc is less clear. It has been reported that up to 50% of patients with CTD and normal resting mPAP had abnormal increases in mPAP during exercise echocardiography.^36–39 One study of 484 patients with CTD-PH (74% SSc) from five centers in the United Kingdom showed that 11% had ePAH (defined as mPAP ⩾ 30 mmHg and PAWP ⩽ 15 mmHg during exercise) measured by eRHC. ¹⁶ The prevalence of ePH in SSc measured by exercise Doppler echocardiography (eDE) has been reported to be even higher between 42% and 59%. ⁶ , ³⁶ , ³⁷ However, studies using eDE must be interpreted with caution as they may overestimate the prevalence of ePH. For example, using a threshold value for pulmonary artery systolic pressure (PASP) of >35 mmHg by eDE, Collins et al. ³⁶ found that 40%–44% of patients with SSc had ePH, whereas using a threshold of >55 mmHg decreased the prevalence to 10%–11%, which is more compatible to that reported in studies using RHC. ¹⁶

One of the first studies that evaluated ePAH using eRHC in SSc screened 54 patients by eDE; those who were positive for ePH (defined as an increase of 20 mmHg in PASP from baseline) were subsequently tested by eRHC. These authors found that 44% of the SSc cohort had ePH by eDE; however, only 62% of these patients eventually tested positive for ePAH by eRHC (defined as mPAP > 30 and PAWP < 18 at peak exercise), estimating an overall SSc-ePAH prevalence of 26%. ⁴⁰ In another study of an SSc cohort where ePH was defined as either an increase in eDE right ventricle systolic pressure (RVSP) by 40 mmHg from rest or a cardiopulmonary exercise testing (CPET)-derived peak VO₂ < 75% predicted, the authors report significant correlations at submaximal exercise between mPAP (above or below the median value of the cohort) and peak VO₂, 6-minute walk distance (6MWD), and pulmonary hemodynamics. ⁴¹ Thus, different criteria for defining ePH by eDE are likely needed in order to make this diagnostic tool more specific.

In addition to the technical challenges of defining ePH in this patient population, an abnormal exercise response in SSc may be confounded by the presence of HFpEF and/or ILD. ⁴ , ⁴² As ePH becomes a more widely accepted entity and our diagnostic testing and criteria are refined, we will gain a better understanding of disease prevalence and incidence.

Screening for SSc-PH

In general, screening for SSc-rPH by symptomatology is problematic as the usual, predominantly exertional complaints (dyspnea, fatigue, and chest pain), are mundane and may be attributable to concurrent ILD, HFpEF, generalized deconditioning/obesity, or the underlying SSc itself. Furthermore, more PH-specific symptoms (i.e. exertional syncope) may not be present until the onset of frank right heart failure ⁴³ when the diagnosis of rPH is likely already clinically established. In support of this concept, there is histopathologic evidence of pulmonary vasculopathy in up to 72% of SSc autopsies without clinical evidence of rPH. ⁴⁴ Therefore, there is a concern that our diagnostic modalities at rest may not be sensitive enough to detect early disease. In clinical practice, patients may have borderline PH (resting mPAP = 21–24 mmHg) or no PH by RHC and yet complain of exertional dyspnea without an apparent alternate explanation. In this setting, an abnormal increase in pressure (mPAP) relative to flow (CO) during exercise provocation may “unmask” underlying PVD. The natural history of an abnormal exercise response (i.e. ePH) in the setting of SSc has not been defined and is the subject of continued investigation.

Rationale for exercise provocation in SSc: an evolving paradigm

There are compelling arguments supporting the notion and benefit of exercise provocation in SSc, particularly in patients with unexplained cardiopulmonary symptoms, and especially given the generally poor prognosis of SSc-rPAH compared to other types of Group I PAH.

It is likely that diagnosing PH earlier will improve outcome. For instance, Humbert et al. ⁴⁵ demonstrated that a rigorous screening algorithm using resting Doppler echocardiography (rDE) on all SSc patients (regardless of symptoms) led to the diagnosis of rPH with less functional limitation (lower New York Heart Association (NYHA) functional class) and significantly increased survival compared to a “routine practice SSc cohort.” It stands to reason that if an earlier diagnosis of rPH in SSc could in fact allow for more timely initiation of PH-specific therapy and less morbidity and mortality, perhaps establishing ePH (prior to the development of rPH) may offer additional value, and ideally alter the natural history of PVD. Second, several studies have identified “borderline PH (resting mPAP 21–24 mmHg by RHC)” as a clinically distinguishable intermediate between the “no PH” and “resting PH (resting mPAP ⩾ 25 mmHg by RHC)” phenotypes ⁴ , ⁵ , ⁴⁶ , ⁴⁷ with similar functional limitation and mechanisms for exercise intolerance compared to rPH, ³⁴ , ⁴⁸ and the potential for relative increased mortality. ⁴⁹ , ⁵⁰ A recent study suggests that ~40% of “borderline PH” related to SSc (pre-DETECT ⁵ algorithm) progress to rPAH after a mean of ~2.5 years. ⁵¹ However, the role of exercise provocation in an effort to distinguish those “no rPH” and “borderline PH” patients at risk for developing rPH over time remains unclear from these studies. Third, in a cohort of more than 400 patients with unexplained dyspnea (no SSc patients), Tolle et al. ⁵² demonstrated that ePH represents a clinically distinguishable phenotype between normal and rPH based on physiologic data obtained during exercise with simultaneous RHC and CPET. In this context, ePH has demonstrated prognostic relevance in several populations outside of SSc including asymptomatic valvular disease, ⁵³ , ⁵⁴ reduced ejection fraction cardiomyopathy, ⁵⁵ and in CTEPH after pulmonary thromboendarterectomy. ⁵⁶ Finally, sequential exercise pulmonary hemodynamics measured during exercise RHC (eRHC), unlike resting hemodynamic parameters, were independently associated with change in 6MWD in a rPAH cohort after a course of PH-specific therapy, ⁵⁷ highlighting a role for exercise provocation even in the setting of established rPAH. CPET, atrial pacing, eDE, other stress (i.e. dobutamine) echocardiography, cardiac magnetic resonance imaging (cMRI) during exercise, and eRHC are all modalities that have been employed for the diagnosis of ePH, and perhaps should be considered in a future diagnostic algorithm for PH. An exercise evaluation for patient-reported dyspnea on exertion, particularly in a heavily screened population such as SSc, may be an important tool for early detection of PH, particularly when there is no clear alternative explanation.

Physiology

In an effort to appreciate the clinical impact of ePH and the role of existing diagnostic modalities, it is important to understand the right ventricular (RV) and pulmonary vascular responses to exercise in normal patients as well as in those with established PVD at rest.

Pulmonary vascular physiology during exercise in normal subjects

In healthy subjects, as workload increases, peripheral tissues extract more oxygen at the microcirculation leading to a greater arterio-venous oxygen difference and increases in heart rate (HR) and stroke volume (SV) (i.e. CO) to meet the metabolic demand. As CO increases, mPAP also increases in a nonlinear fashion due to augmented flow across the pulmonary vascular bed, a relationship best represented by TPR (= mPAP/CO). With increasing levels of exercise in normal subjects, the TPR decreases, as the rise in CO is typically greater than that of the mPAP.^20,58–61

This decrease in pulmonary vascular resistance (PVR) is an adaptive response of the pulmonary circulation, resulting in the decreased RV afterload needed to efficiently maintain an augmented CO during exercise. This is postulated to occur by recruitment of under-perfused capillary beds normally available when transitioning from rest to exercise. ⁵⁹ However, PVR only describes the resistive component of afterload (Rp) on the RV, the other determinant being the elastic or pulsatile afterload defined as the pulmonary vascular compliance (Cp = SV/PP (pulmonary artery pulse pressure)). Cp decreases during exercise ⁶¹ as the rise in PP and mPAP limit the rise in SV and thereby restrict the rise in CO. It is possible that this reduction in Cp is a key limiting factor of exercise capacity in normal subjects, as well as patients with PH. ¹⁶ , ⁶¹ In contrast to the systemic circulation, the product of PVR (Rp) and pulmonary artery compliance (Cp) is a time constant in the pulmonary circulation that is characteristically not modified by clinical factors (i.e. presence/absence of PH-specific therapy, parenchymal lung disease) but is affected by PAWP which augments RV pulsatile load. ⁶²

Pulmonary vascular physiology and gas exchange during exercise in rPH

The gold standard for objectively evaluating the ventilation–perfusion (V-Q) spectrum in a given patient at rest and during exercise is the multiple inert gas elimination technique (MIGET). Using MIGET, the typical PAH patient has a relatively preserved V-Q spectrum at rest (i.e. relatively matched ventilation and perfusion) with increased alveolar–arterial oxygen gradients (A-a gradient) best explained by blood flow to “low ventilation relative to perfusion” units (“low V/Q”) and low mixed venous oxygen saturation (the latter from reduced CO typical in PAH). ⁶³ During exercise, increased hypoxemia is generally driven by further reduction in mixed venous oxygen saturation, which subsequently influences the rate of diffusion equilibrium of oxygen in the lung. ⁶⁴ The V-Q spectrum at rest and exercise is further complicated by the presence of parenchymal lung disease (i.e. pulmonary fibrosis), ⁶⁵ which is certainly relevant in patients with SSc. There is a paucity of data regarding the TPR response in established PAH (without background PH-specific therapy) using eRHC; however, the value is likely between 6 and 12 mmHg/L.^66–68 Finally, CPET is another method capable of characterizing the exercise response in established PAH. Compared to controls, maximum oxygen consumption (VO₂max) occurs at a lower maximum workload. Inadequate CO leads to a disproportionate rise in minute ventilation in order to reduce the increased tissue/blood CO₂ levels that occur during exercise, ⁶⁹ and this characteristic physiology in PAH is termed “inefficient ventilation” and is represented by an elevated V_E/VCO₂ (either the slope based on multiple values or the actual value at the anaerobic threshold (AT)). ⁵² , ⁷⁰ Reduced peak HR, oxygen tension, and SV have also been reported. ⁵²

Diagnostics

There is no consensus regarding the optimal tests or protocols to evaluate exercise responses in patients with suspected ePH. In addition to eRHC, eDE, CPET, and cMRI have been utilized. Herein, we will review the current literature available regarding the above methodologies for the study of patients during exercise.

Diagnostic modalities

rDE

Transthoracic rDE is considered the screening test of choice for rPH. ⁷¹ If the rDE suggests rPH, the diagnosis is confirmed by a rRHC. As it stands today, the standard hemodynamic variables obtained by RHC (mPAP, PAWP, and CO) are also available by rDE.^72–76 In general, the reason rDE has not replaced RHC for a formal rPH diagnosis is that although the data appear to be accurate (close to the true value), it may lack precision (not reproducible) making it difficult to follow an individual patient, ⁷⁷ and there are other reports suggesting rDE data are both inaccurate and imprecise. ⁷⁸ This issue is likely related to the inherent variability of pulmonary hemodynamics in PAH as demonstrated by ambulatory hemodynamic monitoring. ⁷⁹ , ⁸⁰ There has also been considerable effort to standardize rDE parameters relevant to the structure and function of the right heart, but not all variables have defined thresholds.^81–83

Technical limitations related to rDE are even more problematic when performing eDE. Challenges include the timing of data acquisition which must ideally be done during steady-state exercise and not immediately after cessation of exercise, ⁸⁴ given that pulmonary hemodynamics can normalize soon after stopping exercise. ⁴⁰ Currently, eDE is considered “appropriate” in the consensus diagnostic algorithm for valvular heart disease in a limited number of clinical scenarios including moderate or asymptomatic severe aortic stenosis, aortic regurgitation, mitral stenosis, and mitral regurgitation.^85–88

In addition, the critical measurements and equations used to establish ePH using eDE are generally unreliable due to the varying hemodynamic responses that occur during exercise as compared to the resting state. ⁸⁰ , ⁸⁹ Changes in compliance of the right atrium and venous system during exercise make measurements of right atrial pressure less reliable; ⁹⁰ tricuspid valve regurgitation maximum velocity cannot be assessed in up to 26% of exercising patients ⁹¹ or is poorly assessed; and PVR and PAWP can only be indirectly measured by rDE and their accuracy and precision during exercise need further validation. ⁸⁶ , ⁹² Importantly, studies involving SSc patients have employed highly variable criteria in the definition of ePH using eDE including PASP >35 mmHg, ⁷ , ³⁶ , ³⁷  >50 mmHg, ⁶ and ∆PASP from rest to peak exercise of >18 mmHg ⁹³ or >20 mmHg. ⁴⁰ Table 1 provides a comprehensive list of prior studies which have used different forms of exercise, specifically to evaluate SSc patients. Interestingly, Codullo et al. ⁹³ found a ∆PASP >18 using eDE to have a sensitivity of 50% and specificity of 90% for predicting rPH in follow-up. In a CTD cohort, the cut-off of ∆mPAP/∆CO >3.3 by eDE best predicted future rPH. ¹⁰⁴ Nevertheless, eDE is widely available, less costly, and non-invasive compared to eRHC and therefore remains an attractive option for exercise evaluation. Once we standardize the methodology and data interpretation, the introduction of eDE may be an option in future screening algorithms for ePH and/or in the follow-up of established rPH.

Table 1.

Comprehensive list of studies evaluating ePH in the setting of SSc.

Authors	Year	Study type	No. enrolled (SSc pts)	No. of patients with non-diagnostic study (%) a	No. studied (SSc)	Study population	Control group	Modality	Positioning	Definitions for positive exercise response	Definition of abnormal left-sided filling pressure (with exercise)	No. of ePH (%)	No. of ePH with PVH	Predictors of ePH
eDE
Alkotob et al. 37	2006	Prospective single center	65	–	65	SSc with r-sPAP ⩽ 35 mmHg	No	eDE	Bruce treadmill (exercise interrupted)	sPAP > 40 mmHg	–	30 (46)	–	Anti-centromere negative
Pignone et al. 94	2007	Prospective single center	27	–	27	lSSc with r-sPAP < 35 mmHg	No	eDE	Supine cycle ergometer	sPAP > 40 mmHg	–	18 (67)	–	None
Callejas-Rubio et al. 39	2008	Prospective single center	48	7 (15)	41	SSc with r-sPAP < 35 mmHg	No	eDE	Cycle ergometer	sPAP ⩾ 55 mmHg	–	11 (27)	–	↓DLCO (<72 %pred); ↑BNP (73 vs 42 pmol/mL)
Reichenberger et al. 95	2009	Prospective single center	33	–	33	SSc with normal r-sPAP	No	eDE/hypoxia b	Semi-supine cycle ergometer	sPAP > 40 mmHg	–	16 (48)	–	↑r-sPAP (26 vs 19 mmHg); ↓FEV1 (83 vs 97 %pred); ↓VC (84 vs 99 %pred); ↓TLC (97 vs 112 %pred)
D’Alto et al. 42	2010	Prospective multi-center	212	21 (10)	172	SSc with pTRV ⩽ 3 m/s	Yes	eDE	Supine cycle ergometer/exercise interrupted supine	sPAP ⩾ 48 mmHg	–	22 (13)	–	Presence of ILD
Gargani et al. 6	2012	Prospective multi-center	220	13 (6)	164	SSc with r-sPAP < 35 mmHg	Yes	eDE	Semi-supine cycle ergometer	sPAP ⩾ 50 mmHg + PVR ⩾ 3WU	–	69 (42)	–	↑r-sPAP (26 vs 19 mmHg); ↑p-sPAP (46 vs 43 mmHg)
Baptista et al. 96	2012	Prospective single center	38	8 (21)	23	lSSc with r-sPAP < 35 mmHg	No	eDE	Semi-supine cycle ergometer	sPAP ⩾ 50 mmHg	E/e′ ⩾ 13	11 (48)	2 (7)	–
Suzuki et al. 97	2014	Prospective single center	568	5 (<1)	494	SSc with r-sPAP < 50 mmHg	No	eDE	Master’s two-step (exercise interrupted)	sPAP ⩾ 50 mmHg	–	133 (27)	–	–
Voilliot et al. 98	2014	Prospective single center	63	15 (24)	45	SSc with r-sPAP < 35 mmHg	No	eDE	Semi-supine cycle ergometer	sPAP > 50 mmHg	–	21 (47)	–	↑r-sPAP (29 ± 6 vs 21 ± 5 mmHg); ↑r-mPAP (20 ± 4 vs 14 ± 3 mmHg); ↑mPAP/LVCO (5.8 vs 2.9 mmHg/L/min); ↑rPVR (2.6 vs 1.4)
eRHC
Saggar et al. 11	2010	Prospective single center	57	–	57	SSc with normal rRHC	No	eRHC	Supine cycle ergometer	mPAP > 30 mmHg + PAWP ⩽ 18 mmHg + TPG ⩾ 15 mmHg	mPAP > 30 mmHg + PAWP > 18 mmHg + TPG < 15 mmHg	21 (37)	12 (21)	↑rRHC PVR; ↓CO; ↑r-mPAP; ↑PAWP at rest; ↑TPG at rest; ↓FVC (<60 %pred)
Condliffe et al. 16	2008	Retrospective multi-center	42	–	42	SSc with abnormal eRHC c	No	eRHC	–	mPAP ⩾ 30 mmHg + PAWP < 15 mmHg	–	–	–	–
Keusch et al. 99	2014	Retrospective single center	37	–	37	SSc with normal rRHC	No	eRHC	Supine cycle ergometer	mPAP > 30 mmHg + PAWP < 15 mmHg	mPAP > 30 mmHg + PAWP > 15 mmHg	6 (16)	15 (41)	–
CPET
Dumitrescu et al. 100	2010	Prospective single center	32	4	28	SSc with dyspnea	No	CPET	Cycle ergometer	Peak VO2 < 75%; AT < 75%; VE/VCO2 nadir ⩾ 34; decreasing PETCO2	Peak VO2 < 75%; AT < 75%; VE/VCO2 nadir > 34; ∆PETCO2 neutral or increasing	11 (40)	5 (18)	–
Mixed modality
Chia et al. 101	2015	Prospective multi-center	25	2 (8)	23	SSc with r-sPAP < 36 mmHg	Yes	CPET/eDE	Bruce treadmill (exercise interrupted)	–	–	–	–	↓peak VO2; ↓HRR
Casella et al. 102	2017	Prospective single center	39	–	39	SSc + by DETECT algorithm	No	CPET/rRHC	Cycle ergometer	–	–	22 (56)	–
Dumitrescu et al. 48	2016	Prospective multi-center	173	–	173	SSc with dyspnea	No	CPET/rRHC	Cycle ergometer	–	–	15 (9)	–	↓Peak VO2
Nagel et al. 103	2015	Prospective single center	76	–	74 (eDE); 76 (eRHC)	SSc with dyspnea	No	eDE/eRHC	Semi-supine cycle ergometer	sPAP > 45 mmHg (eDE); mPAP > 30 mmHg (eRHC)	–	21 (29) d	–	↓6MWT, ↑NT-proBNP
Steen et al. 40	2008	Prospective single center	54	–	54 (eDE); 21 (eRHC) e	SSc with dyspnea	No	eDE/eRHC	Bruce treadmill (exercise interrupted-eDE)/supine dumbbell exercise	Change of > 20 mmHg (eDE); PAP > 30 mmHg + PAWP ⩽ 18 mmHg (eRHC)	–	24 (44) eDE/ 13 (24) f	–	↓DLCO (<50 %pred); ↑FVC/DLCO ratio (>1.6)
Kovacs et al. 41	2009	Prospective single center	29	–	29	SSc	No	CPET/eDE/eRHC	Semi-supine cycle ergometer	–	–	–	–	r-mPAP (> 7 mmHg), ↓rPVR

SSc: systemic sclerosis; ePH: exercise pulmonary hypertension; PVH: pulmonary venous hypertension; eDE: exercise Doppler echocardiography; sPAP: systolic pulmonary artery pressure; r-sPAP: resting systolic pulmonary artery pressure; lSSc: limited SSc; BNP: brain- natriuretic peptide; ILD: interstitial lung disease; pTRV: peak tricuspid regurgitation velocity; p-sPAP: peak systolic pulmonary artery pressure; r-mPAP: resting mean pulmonary artery pressure; PVR: pulmonary vascular resistance; rPVR: resting pulmonary vascular resistance; mPAP: mean pulmonary artery pressure; LVCO: left ventricular cardiac output; PAWP: pulmonary artery wedge pressure; RHC: right heart catheterization; rRHC: resting RHC; eRHC: exercise RHC; CPET: cardiopulmonary exercise testing; DLCO: diffusion capacity of lungs for carbon dioxide; TLC: total lung capacity; TPG: transpulmonary gradient; CO: cardiac output; VO₂: oxygen consumption; VCO₂: carbon dioxide production; VE: expired ventilation; P_ETCO₂: end-tidal carbon dioxide; AT: anaerobic threshold; FEV₁: forced expiratory volume at 1 second; RA: right atrium; RV: right ventricle; 6MWT: 6-minute walk test; FVC: forced vital capacity; VC: vital capacity; HRR: heart rate reserve; ePAH: exercise pulmonary arterial hypertension; PH: pulmonary hypertension.

Only studies with N > 20 were included.

– Not available, not reported.

^aMay not represent all those excluded after enrollment.

^bHypoxia was induced via a face mask with delivery of 12.5% FiO₂ air mixture in supine position for 2 h.

^cThis study evaluated only patients with resting PH (315) or ePH (42). Therefore, comments on prevalence of ePH cannot be made.

^dUnable to determine absolute number of patients with +eDE.

^eOnly +eDE patients underwent eRHC in this study.

^feRHC proven prevalence documented as a percentage of the entire population (54 patients).

^gThis study used CPET to predict the presence of PH on resting RHC.

CPET

Gas exchange patterns during exercise using CPET are suggestive of resting PVD. Hypoperfusion of well-ventilated lung units (“high V/Q”) is the basic gas exchange abnormality seen in PVD, particularly during exercise. There is effectively a reduced reservoir of blood CO₂ available to the lungs for ventilation (VCO₂) at any given ventilatory effort (V_E) which can be expressed as either an increased V_E/VCO₂ ratio or a decreased end-tidal CO₂ (P_ETCO₂). ⁴⁸ In contrast to reduced peak VO₂ and AT which are typical CPET findings in patients with any cardiovascular limitation to exercise, parameters indicating inefficient ventilation (increased V_E/VCO₂ ratio and decreased P_ETCO₂) are more specific for PVD. Other findings supporting a PVD phenotype using CPET include decreased peak O₂ pulse (VO₂/HR) and a decreased ∆VO₂/Δwork rate. In addition, decreased P_ETCO₂ (at AT) and increased V_E/VCO₂ (at AT) have the added advantage of helping to distinguish between PVD and PH related to left heart disease (WHO Group II PH) in the setting of SSc.^47,100

Exercise not only decreases transit time across the alveolar-capillary bed but also increases the metabolic production of CO₂, thereby further exposing inefficient ventilation. In a cohort of 28 SSc patients without clinical concern for PVD (estimated PASP < 35 mmHg by rDE), 16 (~60%) patients had abnormal exercise gas exchange responses, 11 (40%) of whom had a CPET profile consistent with PVD. ¹⁰⁰ Importantly, half of the PVD-classified patients had significant spirometric restriction, while the other half had minimal to no restriction. These findings were further validated by the same group in a cohort of 173 SSc patients referred for possible PVD (n = 115 without PH; n = 48 with rPAH; n = 10 with WHO Group II PH, based on rRHC) and CPET evaluation. ⁴⁸ This group determined that peak P_ETCO₂ and nadir V_E/VCO₂ both distinguished PAH compared to the “no PH” and “WHO Group II PH” groups (AUC ⩾ 0.85 for both measures). Furthermore, a nadir V_E/VCO₂ threshold >45.5 showed a positive predicted value of 1.0 for PAH, and a peak VO₂ >18.7 mL/min excluded PAH with a negative predictive value of 1.0. Interestingly, Casella et al. ¹⁰² demonstrated that the V_E/VCO₂ slope (from early exercise to AT) was better at predicting SSc-PAH compared to V_E/VCO₂ at AT (AUC 0.96; p < 0.01) and discriminated PAH (higher values) from non-PAH patients with 100% sensitivity and 82% specificity at a threshold value of 38.5.

These data support the concept that there does exist a CPET signature raising the possibility of PVD not evident at rest. Nevertheless, there is no definitive CPET criteria that define resting or ePH. However, a combined modality using simultaneous eRHC and CPET may be a valuable tool and its application is well reported in PVD and unexplained dyspnea in non-SSc cohorts ⁵² but only in a single small SSc cohort (n = 29). ⁴¹ Finally, there are additional CPET parameters that may further characterize HFpEF physiology (compared to normal) during exercise, all related to the reserve (i.e. change from rest to peak exercise) responses in HR, SV, and AV oxygen difference, ¹⁰⁵ although this has not been studied to date in SSc.

In addition to its role in the evaluation for PAH, CPET has also been shown to help predict prognosis. For example, Grünig et al. ¹⁰⁶ demonstrated that peak VO₂ ⩽ 11.4 mL/kg/min was an independent predictor for survival in PAH. Hsu et al. postulated that the worse RV reserve response to exercise is likely responsible for the lower survival rates of SSc-PAH as compared to IPAH. In this study, inefficient ventilation based on CPET parameters had a strong correlation with poor right ventricular–pulmonary arterial coupling, a concept discussed further below. ¹⁰⁷

Right ventricle contractile reserve and ventricular–arterial coupling

A long-standing concept in the left heart failure literature known as “contractile reserve” (CR) has also been applied to the RV and refers to the change in ventricular systolic function in response to exercise or pharmacologic stress. CR is likely dependent on multiple factors affecting the RV in SSc, which include one or more of the following: intrinsic contractility, myocardial fibrosis, myocardial perfusion, neurohormonal activation, and afterload. Bossone et al. ⁷¹ evaluated SSc patients by eDE (without rPH) and used increased ∆PASP/∆confidence interval (CI) as a marker of poorer ventricular reserve. Grünig et al. ¹⁰⁶ evaluated both RHC-confirmed PAH and CTEPH cohorts with RV dysfunction (n = 124) using combined eDE and CPET. They determined that the group which increased PASP by greater than the median of the cohort (∆30 mmHg, i.e., greater RV contractile reserve) had increased mean 6MWD, mean peak VO₂ and lower HR, NT-proBNP, right atrial area, and better 1-, 3-, and 4-year survival compared to the group with <30 mmHg PASP change.

The RV CR likely reflects the capacity of the RV to function relative to the downstream afterload in the pulmonary vasculature, ¹⁰⁶ a concept termed ventricular–arterial (V-A) coupling. V-A coupling is the gold standard in understanding the relationship between RV contractility and pulmonary artery (PA) afterload and is determined by creating RV pressure–volume (PV) loops. Hsu et al. ¹⁰⁷ studied the PV loops of SSc-PAH and IPAH subgroups at rest and during CPET and found that though there were minimal differences at rest, the SSc-PAH group had evidence for V-A uncoupling compared to IPAH during exercise manifesting as decreased RV contractility, increased RV dilation, decreased total exercise time, and decreased RV ejection fraction. This same group evaluated endomyocardial biopsies from SSc-PAH, IPAH, SSc (dyspnea but no PH), and controls and found intrinsic differences of the SSc-PAH cardiac myocytes in terms of depressed sarcomere function and decreased maximum calcium-activated force which correlated with the previously reported reduction in RV contractile reserve at rest and RV dilation during exercise based on PV loop analyses. ¹⁰⁸ In aggregate, these data surrounding exercise-determined RV-CR suggest that the RV of the SSc-PAH patient may be primarily disadvantaged despite similar PA afterload and may explain the generally poorer survival when compared to IPAH.

cMRI during exercise

The recent advent of a “magnetic resonance imaging (MRI) compatible” exercise bicycle allows for the study of cardiopulmonary physiology during exercise using cMRI, although currently, this concept is generally only available in the research setting. In addition, improvement in imaging speed has led to the development of protocols for real-time assessment of cardiovascular function, as demonstrated by Le et al. ¹⁰⁹ in their use of an in-scan supine cycle ergometer with post-imaging resynchronization. Hemodynamic measurements such as cardiac index (CI) with cMRI have demonstrated inter-observer concordance, measure reproducibility, and accuracy in comparison with the direct Fick calculation of CO. ¹⁰⁹ , ¹¹⁰ In addition to pulmonary hemodynamics, cMRI can directly measure bi-ventricular end-systolic and end-diastolic volumes, right ventricular ejection fraction (RVEF), and SV. ¹⁰⁹ , ¹¹¹ , ¹¹² Furthermore, these cMRI parameters have been evaluated during high-intensity exercise ¹¹⁰ , ¹¹² and have proven to distinguish untrained healthy subjects from athletes ¹⁰⁹ and further characterize patients with rPAH (compared to healthy controls) and CTEPH post-endarterectomy with normal resting pulmonary hemodynamics. ⁵⁵ , ¹¹³

The current data suggest that cMRI is capable of providing all relevant pulmonary hemodynamics obtained by RHC, while providing additional volumetric data for both ventricles throughout the cardiac cycle. As with eDE, cMRI with exercise may be able to uncover ePH but with the added advantage of objectively characterizing the variable mechanisms of cardiac compensation during exercise. Furthermore, cMRI provides novel parameters such as pulmonary flow patterns obtained by phase-contrast MR angiogram that may be valuable in further characterizing V-A (un)coupling, Rp, and Cp. ¹¹⁴ Given the associated expense, issues of feasibility and the absence of studies using cMRI with exercise in SSc, it is currently limited to research protocols but may soon play a larger role in the diagnosis and prognosis of SSc-PH at rest and with exercise.

Natural history of ePH and the subsequent development of rPH

The mean time period from the time of SSc diagnosis to the development of PAH is reported to be 10–15 years. ¹¹⁵ Chang et al. ¹¹⁶ report that in 361 SSc patients with normal baseline pulmonary hemodynamics by DE, 26% and 14% progressed to have mild-to-moderate and severe PH, respectively. Furthermore, Condliffe et al. ¹⁶ reported in a registry approach that more than 50% with WHO functional class I/II PAH progressed to functional class III/IV, had a 20% decline in 6MWD, or demonstrated worsened rRHC parameters at a mean of 2.3 years from the time of PAH diagnosis.

The natural course of ePH in SSc, however, is still not entirely clear. Some reports suggest that ePAH may be a stable variant that does not progress to resting PAH, ⁴³ , ¹¹⁷ but a majority of the evidence suggest that ePAH is part of the PAH spectrum, functioning as an intermediate physiology between normal and resting PAH. ⁷ , ⁵¹ Saggar et al. ¹¹⁸ and Condliffe et al. ¹⁶ found that of those SSc patients fulfilling criteria for ePH by eRHC, 25% and 19% subsequently developed resting PAH, respectively, with a mean time to development of ~2.3 years. ¹⁶ In addition, 17% of SSc-ePAH required advanced pulmonary vasodilator therapy within 3 years of an ePAH diagnosis. ¹⁶ Kovacs et al. ¹¹⁹ reported a progressively abnormal exercise response by eDE and CPET over time (99 SSc patients followed for ~4 years); however, only a fraction of these patients went on to develop rPAH (0.75 cases per 100 patient-years). Table 2 demonstrates the available literature with regard to SSc-ePH progression to rPH and mortality.

Table 2.

Mortality and disease progression of SSc patient with documented ePH.

Authors	Year	No. of patients with ePH	Diagnostic modality	Mean follow-up (weeks)	Treatment	Mortality rate %	No. of patients who progress to resting PH (%)
Reichenberger et al. 95	2009	16	eDE	36	None	–	1 (6.3)
Condliffe et al. 16	2008	42	eRHC	60	Mixed a	1-, 2-, 3-, 4-, and 5-year mortality rate: 14, 24, 45, 86, and 91	8 (19)
Saggar et al. 118	2012	11	eRHC	24	Ambrisentan	None	3 (25)
Kovacs et al. 119	2017	8 b	eRHC	48	Mixed a	4	3 (3) c

SSc: systemic sclerosis; ePH: exercise pulmonary hypertension; PH: pulmonary hypertension; eDE: exercise Doppler echocardiography; eRHC: exercise right heart catheter; TPVR: transcatheter pulmonary valve replacement.

– Not reported.

^aMixed = some received treatment, while others remained untreated.

^bThis study evaluated 99 SSc patients with eDE and those with high risk features for PH were evaluated by eRHC (28), 8 of whom met criteria for ePH (mPAP > 30 mmHg, TPVR > 3WU) at baseline and 11 of whom met criteria at follow-up.

^cOf the 99 patients, 3 progressed to resting PH at follow-up, unclear from the data how many of those patients had ePH at baseline.

Treatment

Current consensus guidelines do not recommend PH-specific therapy for ePH. Although PH-specific therapy for SSc-related rPAH has been shown to improve exercise capacity, resting hemodynamics, quality of life, and even survival^95–98 little data exist for ePH.

Castelain et al. ⁶⁷ studied the effect of intravenous epoprostenol over a 6-week follow-up period, and they found an improvement in 6MWD as well as exercise PVR by eRHC. The effects of ambrisentan on 12 patients with SSc-related ePAH over 24 weeks of treatment were evaluated by Saggar et al. ¹¹⁸ with baseline and 24-week eRHC parameters. A significant decrease in mPAP (42–37; p = 0.02) mmHg, PVR (247–161; p = 0.003) dynes·s/cm⁵, and transcatheter pulmonary valve replacement (TPVR; 406–313; p < 0.01 (dynes·s/cm⁵)) in addition to increases in the CO, SV, and SV index are reported. In addition, 72% of patients had a >10% increase in 6MWD with a mean improvement of 44 m. Similar findings were demonstrated with bosentan in a SSc cohort (n = 10) with “borderline PH” which demonstrated progression of PVD prior to treatment. ¹²⁰ Segrera et al. ¹²¹ reported similar hemodynamic improvements with ambrisentan in an ePH (non-SSc) cohort with concurrent CPET parameters showing a trend toward increased peak VO₂. These studies suggest that treatment of ePH complicating SSc may have benefit. Clearly, however, randomized controlled trials are needed to best understand whether PH-specific therapy has a role in this setting.

Conclusion

PH in patients with SSc is common and portends a poor prognosis. Currently, there is no formal consensus definition for ePH and no recommendations for PH-specific therapy. However, the presence of ePH in SSc appears to be a clinically discernable phenotype from normal which may progress to rPH. The use of exercise testing has emerged as an important consideration in the evaluation of patients with SSc and respiratory symptoms. Advances in imaging techniques continue to improve the non-invasive assessment of the cardiovascular response to exercise; however, invasive RHC continues to be the gold standard. Whether or not ePH should be treated in a similar fashion to rPH is still not known, although early reports suggest some benefit of PH-specific therapy in this group. Further study is needed to better understand the pathophysiology, natural history, and optimal treatment approach in these patients.