Exercise testing provides additional information over resting variables and is a standard of care in the assessment of coronary artery disease with less subsequent cardiovascular events. In pulmonary hypertension (PH), right ventricular function is clearly an important determinant of survival. However, right ventricular impairment is currently assessed only during resting conditions. Exercise (stress) testing may provide further insight into the complex paradigm of right ventricular dysfunction, right ventricular–left ventricular interdependence and right ventricular–pulmonary artery coupling. Whether pulmonary artery pressure (Ppa)–flow relationships during exercise provides a window into earlier diagnosis of functionally significant PH or adds incrementally to our armamentarium of diagnostic tests and prognostic indicators in PH, is a topic of ongoing investigation. In this issue of the European Respiratory Journal (ERJ), the studies by Kovacs et al. [1] and Whyte et al. [2] discuss the potential clinical utility of identifying “normal” and “abnormal” pulmonary vascular response patterns to exercise in patients aged ≤50 yrs.
The clinical significance of a hypertensive Ppa response to exercise is currently uncertain. It may reflect a normal variant without clinical significance, or it may reflect a bona fide cardiopulmonary limitation and abnormal phenotype. In support of the latter, asymptomatic relatives of individuals with established idiopathic or familial PH have demonstrated an abnormal response to exercise [3].
Few studies have used right heart catheterisation (RHC) to address whether mean Ppa (P̄pa) >30 mmHg during exercise represents an aberrant pulmonary vascular system in “at-risk” populations [4–11]. Recently, Tolle et al. [9] evaluated a large cohort with exertional dyspnoea who underwent simultaneous cardiopulmonary exercise testing (CPET) and RHC. At maximal exercise, P̄pa and pulmonary vascular resistance (PVR) were highest, while resting pulmonary arterial hypertension was lowest in normal subjects and intermediate in patients with exercise-induced PH. These results suggest exercise-induced PH may be a clinically distinguishable phenotype in the continuum between normal and resting PH.
The 4th World Symposium on PH examined the literature involving healthy volunteers who underwent haemodynamic assessments at rest and during exercise [12]. A systematic review defined a normal resting P̄pa of 14±3.3 mmHg with an upper limit of normal of 20.6 mmHg in the supine position. Importantly, age did not affect resting P̄pa. However, during slight supine exercise P̄pa was significantly higher in subjects aged >50 yrs (29.4±8.4 mmHg) compared to those aged 30–50 yrs (20.0±4.7 mmHg) and <30 yrs (18.2±5.1 mmHg). The upper limit of normal for P̄pa in subjects aged <50 yrs was 29.0 mmHg versus 46.2 mmHg in subjects aged >50 yrs [12]. Consequently, the exercise component (P̄pa >30 mmHg) of the prior PH definition was abandoned [12]. This decision was based on the age-dependent nature of exercise P̄pa [13], limited data on normal subjects and the heterogeneity of the existing data, i.e. type, position and level of exercise [12].
In this issue of the ERJ, Kovacs et al. [1] expand on their prior systematic review and describe the changes in total pulmonary resistance (TPR) and PVR in healthy individuals during primarily supine exercise. This analysis is essential as it importantly accounts for cardiac output (CO) augmentation and the contribution of pulmonary capillary wedge pressure (Ppcw), variables complementary to single P̄pa measurements. Importantly, data were stratified by sex, age, type of exercise (i.e. cycle ergometry and treadmill exercise), body position (upright versus supine), and exercise level (slight, submaximal and maximal) as defined by physiological variables including, heart rate, work rate and oxygen uptake. The authors concluded that a moderate decrease in TPR and a mild decrease in PVR (0.95–1.45 dyn·s·cm−5·L−1·min−1) are seen during the exercise response in healthy normal subjects. The modest increment in transpulmonary gradient (TPG) relative to CO during exercise that results in a decrease in PVR is probably attributable to passive recruitment and distension of a normally compliant pulmonary circulation with active flow-mediated vasodilation [14, 15].
A limitation of the study by Kovacs et al. [1] was the paucity of individuals aged >50 yrs. Unlike younger patients, the overall decline in TPR and PVR is only modest in subjects aged 51–69 yrs and ≥70 yrs. Older individuals tend to have greater augmentation in Ppcw relative to blood flow during exercise but a lesser decline in TPG [13]. Furthermore, advancing age is associated with an increase in resting PVR secondary to normal age-related increases in P̄pa and decreases in CO, resulting in a doubling of resting PVR over five decades of life [13–18]. In the study by Kovacs et al. [1], subjects aged 51–69 yrs were limited to only 13 and four patients in supine and upright positions, respectively. Of these, calculated PVR was available for only eight (47%) subjects. Due to small patient numbers and the multiple variables being studied, we chose to focus only on interpretation of the data set limited to subjects aged <50 yrs and in the supine position.
An inherent weakness of interpreting PVR changes with exercise is the flow dependency of this variable. The PVR calculation assumes the P̄pa–CO relationship is constant and crosses the pressure axis at a value equal to left atrial pressure, where flow is theoretically zero. Accordingly, the acceptance of a single-point PVR for pressure–flow reslationships, as suggested by Kovacs et al. [1], assumes a homogenous response to exercise, irrespective of other factors such as age, relative hypoxia and unrecognised cardiac/respiratory disease, all of which may alter the distensibility of the resistive vessels [19].
Marked differences are seen in pressure–flow relationships in the pulmonary vasculature of healthy normal subjects relative to resting pulmonary arterial hypertension and systolic heart failure-associated PH (table 1). Ideally, assessment of exercise-induced increases in P̄pa should be interpreted relative to increases in blood flow (i.e. ΔPpa/ΔCO) and specific work rates, rather than relying on a single absolute Ppa threshold (i.e. 30 mmHg) or a peak exercise P̄pa. Reeves et al. [33] performed exercise measurements in 63 healthy young adults at rest and at least two levels of exercise and determined average ΔPpa/ΔCO ≈1 mmHg·min·L−1; alternatively, for each 1 L·min−1 increase in CO, there is an approximate 1 mmHg increase in P̄pa [33]. This pattern during exercise corresponds well to the cohort of Kovacs et al. [1] which reported ΔPpa/ΔCO ≈1.06 mmHg·min·L−1 at ≤50 yrs of age. However, both these studies also showed wide variability in individual responses of P̄pa and ΔPpa/ΔCO to exercise. This heterogeneity underscores the importance of defining “normality” specific for age, sex and cardiopulmonary limited populations.
The study by Whyte et al. [2] examined the haemodynamic response to mild-to-moderate exercise in patients aged <50 yrs with normal resting Ppa. These authors performed exercise haemodynamic assessments on 38 patients with symptoms of either exertional dyspnoea, abnormal echocardiogram or suspicion of pulmonary thromboembolic disease. Clearly, the patients were not healthy normal controls and presented with potential risk factors for PH, including history of connective tissue disease, anorexigen use or chronic thromboembolic disease. The inclusion of an “at-risk” population was further exemplified as individuals with an exercise P̄pa of <30 mmHg and an abnormal pressure–flow relationship (ΔPpa/ΔCO ≈1.63±0.70 mmHg·min·L−1) (table 2). During exercise haemodynamic assessment, individuals who increased P̄pa >30 mmHg (24 out of 38) tended to have a higher resting P̄pa and PVR. In particular, 88% (nine out of 11) of patients with a resting P̄pa of 21–24 mmHg developed P̄pa >30 mmHg on mild-to-moderate exercise. Despite a heterogeneous population, the baseline demographics between groups were well matched. Importantly, however, the patients who developed P̄pa >30 mmHg on exercise had a significantly lower diffusing capacity, and trended to have a lower 6-min walk test compared to the remaining 14 patients (478±99 versus 550±100 m; p=0.06). Several points arose from these conclusions. First, the authors appropriately acknowledge the limitation of the singular focus on P̄pa >30 mmHg during exercise. Secondly, the authors did not provide homogenous pressure–flow measurements since the number of data points varied among subjects. Thirdly, exercise was terminated either when subjects reached a workload threshold of 60 W or when P̄pa was >30 mmHg. Finally, despite reporting a causal relationship between a higher baseline resting P̄pa and achieving an exercise P̄pa of >30 mmHg, our independent analysis revealed that the ΔPpa/ΔCO of patients with exercise P̄pa >30 mmHg was significantly higher compared to those with an exercise P̄pa <30 mmHg (table 2). This post hoc analysis suggests that while a P̄pa cut-off of 30 mmHg may distinguish the degree of pulmonary vascular disease, it may not discern a normal from an abnormal exercise response. Alternatively, the ΔPpa/ΔCO ratio derived from the haemodynamic response to exercise may better delineate both the presence and severity of an underlying pulmonary vasculopathy.
Incorporation of CPET with exercise haemodynamic measurements enables assessment of the functional ramifications of abnormal pulmonary vascular response patterns by providing objective indicators of fitness (i.e. anaerobic threshold, peak oxygen uptake and oxygen uptake kinetics) that are closely related to quality of life and prognosis [34]. CPET also permits delineation of ventilatory efficiency (minute ventilation/CO2 uptake) and end-tidal carbon dioxide tension patterns have been shown to increase as the PH burden increases [35]. Simultaneous CPET and exercise haemodynamic assessment is best performed using a cycle ergometer to minimise upper body motion and a continuous ramp protocol to provide a smooth linear increment in work rate. Upright positioning most closely mimics normal physical activity but either upright or supine exercise can be utilised to derive Ppa–flow relationships. Arterial line placement during CPET is not essential but does allow precise systemic blood pressure measurement, as well as serial arterial blood gases, to assess important indicators of pulmonary vascular function, including dead space volume/tidal volume, alveolar–arterial gradient and Fick cardiac outputs when coupled with assessment of oxygen uptake and mixed venous saturations.
Coupling CPET with invasive haemodynamic measurements during exercise in future studies may permit more precise interpretation of haemodynamic measurements and their potential added value in the assessment of patients with suspected pulmonary vascular disease. While the feasibility of concurrent CPET with pulmonary haemodynamics is unlikely at most centres, the concept of performing both studies within a reasonable time-frame of each other may be a more practical consideration. Ultimately, exercise haemodynamics may ideally allow for an earlier diagnosis of PH and a more objective and reproducible clinical trial end-point. In fact, to date, small studies with phosphodiestersae-5 inhibitors in World Health Organization Group II PH have suggested that peak oxygen uptake improves proportionately to the degree of reduction in exercise PVR, and that Ppa response patterns correlate with peak oxygen uptake, whereas resting measurements do not [36].
The articles by Kovacs et al. [1] and Whyte et al. [2] provide invaluable insight into the complex methodology required to better understand the pulmonary vascular circuit using exercise haemodynamics. We emphasise that the use of a single haemodynamic variable, such as P̄pa or PVR, is probably inadequate to define a normal versus abnormal pulmonary vascular response to exercise.
Identifying the role of exercise pulmonary haemodynamics in the evaluation of pulmonary vascular disease and developing a standardised approach for when and how to perform exercise pulmonary haemodynamics, in the context of an evidence-based definition, is clearly needed. The articles by Kovacs et al. [1] and Whyte et al. [2] are important first steps in this endeavour. Finally, important areas remain open for investigation with regard to exercise and pulmonary vasculature, as follows. 1) Improved validation against invasive gold standards of noninvasive approaches (i.e. CPET or echocardiogram) for the study of pulmonary haemodynamics and right ventricular function. 2) Definition of the limits of age- and sex-related ″normal″ of averaged slopes of multi-point P̄pa–Q relationships at exercise in normoxic and in hypoxic conditions. 3) Characterisation of multi-point pulmonary vascular pressure–flow relationships in PH. 4) Definition of expected increase in pulmonary capillary pressure and TPG with increased Ppa at variable levels of exercise, and relationship to gas exchange. 5) Measurements of right ventricular function at exercise.
Acknowledgments
We would like to thank R. Vanderpool (Université Libre de Bruxelles, Brussels, Belgium) for providing statistical support.
Footnotes
Statement of Interest
A statement of interest for R. Naeije and R. Saggar can be found at www.erj.ersjournals.com/site/misc/statements/xhtml
- ©ERS 2012