Cardiopulmonary Exercise Testing in the Functional and Prognostic Evaluation of Patients with Pulmonary DiseasesFerrazza A.M. · Martolini D. · Valli G. · Palange P.
Department of Clinical Medicine, Pulmonary Lung Function Unit, Sapienza University of Rome, Rome, Italy Corresponding Author
Exercise testing is increasingly utilized to evaluate the level of exercise intolerance in patients with lung and heart diseases. Cardiopulmonary exercise testing (CPET) is considered the gold standard to study a patient’s level of exercise limitation and its causes. The 2 CPET protocols most frequently used in the clinical setting are the maximal incremental and the constant work rate tests. The aim of this review is to focus on the main respiratory diseases for which exercise tolerance is indicated; for example, chronic obstructive pulmonary disease, interstitial lung disease, primary pulmonary hypertension and cystic fibrosis. This review also focuses on the variables/indices that are utilized in the functional and prognostic evaluation. The recognition of abnormal response patterns of ventilatory, cardiac and metabolic limitation to exercise may help in the diagnostic evaluation. In addition, CPET indexes can provide important functional and prognostic information regarding patients with pulmonary disease. Exercise indices, such as peak oxygen uptake (V’O2 peak), ventilatory equivalents for carbon dioxide production (V’E-/V’CO2) and arterial oxygen saturation (SpO2), have in fact proven to be better predictors of prognosis than lung function measurements obtained at rest. Moreover, useful information on the effects of therapeutic interventions may be obtained by CPET by studying the changes in endurance capacity during high-intensity constant work rate protocols.
Copyright © 2009 S. Karger AG, Basel
Exercise testing is being increasingly utilized in clinical practice to evaluate a patient’s level of intolerance to exercise and the possible underlying causes for this. Exercise testing is based on the principle that systems such as the respiratory or cardiovascular fail more easily and quickly while under stress. Exercise forces the systems to the limits of their tolerable ranges, and abnormal response patterns may be observed in patients with various diseases. Compared with normal subjects, those with pulmonary, cardiovascular or metabolic diseases have a reduced tolerance to exercise and show clear abnormalities in their physiologic adaptation to exercise in terms of the principal exercise variables (e.g. ventilation and heart rate). Exercise intolerance is defined as the inability to successfully complete a physical task that normal subjects would find tolerable. Importantly, in many chronic respiratory and cardiovascular diseases exercise intolerance cannot be adequately predicted from resting physiological measurements , such as forced expiratory volume in 1 s (FEV1), pulmonary diffusing capacity for carbon monoxide (DLCO) or ejection fraction.
Cardiopulmonary exercise testing (CPET) is considered the gold standard for exercise intolerance evaluation. During the test, patients are subjected to symptom-limited incremental exercise, breath-by-breath monitoring of cardiopulmonary variables [e.g. pulmonary O2 uptake (V′O2), pulmonary CO2 output (V′CO2), minute ventilation (V′E), heart rate (HR)], assessment of perceptual responses (e.g. dyspnea, leg discomfort) and measurements such as exercise-related arterial oxygen desaturation, dynamic hyperinflation and limb-muscle strength. Less expensive exercise tests are the 6-minute walking test and the shuttle walking test, which can be utilized to evaluate the level of exercise intolerance with measurement of distance, duration, heart rate, arterial O2 saturation; they give, however, less physiologic information about the determinants of exercise limitation.
CPET in Clinical Practice
The additional information that can be expected to arise from CPET in a particular patient can be summarized as follows.
(1) The level of exercise intolerance and identification of the mechanisms limiting exercise tolerance. As CPET should not be used for diagnostic purposes, in most instances the patient to be studied will have already presented with a primary diagnosis (e.g. chronic obstructive pulmonary disease, COPD). However, CPET can be useful in the differential diagnosis and should be focused on narrowing the range of suspected diseases. The pattern of cardiopulmonary and gas-exchange responses can distinguish between pulmonary and cardiac limitations to exercise and other pathophysiologic causes of exercise intolerance. Only in very specific conditions, such as exercise-induced bronchoconstriction and arterial O2 desaturation, can CPET be considered as a diagnostic tool.
(2) Functional and prognostic evaluation, by measuring exercise indexes that reflect poor exercise tolerance and system dysfunction.
(3) Evaluation of disease progression and response to interventions, by looking at changes in exercise variables (e.g. before and after treatment). The determination of the clinically important changes in a given outcome is complex and is based on considering multiple factors, including statistical variability, patient and physical assessment . The minimal clinically important difference is a useful tool to be utilized in clinical practice to define the minimal physiologic improvement (e.g. exercise duration) that reflects the positive effect of a given therapeutic intervention. Minimal clinically important difference values have been reported in literature for various outcomes in the different exercise protocols.
Two different types of exercise protocol have been demonstrated to be useful in the clinical setting.
(1) The maximal incremental test (i.e. CPET), in which the work rate is progressively increased until volitional fatigue is reached. The work rate increment should be selected carefully since the incremental exercise period should last 10–12 min. The measures obtained with this test are reproducible in the short term and are useful for diagnostic and prognostic purposes.
(2) The constant work rate (CWR) test, in which the work rate applied is constant (usually 75–80% of peak work rate measured at maximal CPET) and patients are asked to exercise for as long as they can. The time to limitation (TLIM) is measured. TLIM is an important tool that is often used to highlight differences before and after a therapeutic intervention. Longer TLIM after an intervention is taken as evidence of improved exercise tolerance.
During an exercise test, the main symptoms that induce a patient to stop exercising are dyspnea and leg discomfort. Both are present in lung diseases. Dyspnea is an important reason for early exercise interruption in lung diseases. The mechanisms causing dyspnea are complex and are not fully understood. It is clear that dyspnea occurs when there is an imbalance between the central respiratory efferent drive and the response of the respiratory muscular pump. This imbalance may be consciously perceived as the distressing sensation of unsatisfied inspiration (neuromechanical uncoupling) . In lung diseases, the central respiratory efferent drive increases in response to the hypoxia, hypercapnia and acidosis that occur or increase during exercise. The carotid bodies are the chemoreceptors that drive ventilation in response to hypoxemia and acidosis . On the other hand, the muscular pump responds inadequately because of the deteriorating ventilatory mechanics, which result in an increased work and oxygen cost of breathing at any given ventilation and/or an increase in ventilatory requirement for a given level of exercise (reduction in breathing efficiency).
Leg discomfort is due to fatigue that occurs when muscle force output decreases for a given stimulus. The physiological mechanism responsible for the development of muscle fatigue is the failure of V′O2 to meet cellular O2 requirements. It has been postulated that pulmonary, cardiac, or vascular diseases impair O2 delivery to muscles and reduce the maximal O2 conductance capability of the O2 transport chain. Muscular dysfunction may also contribute to leg discomfort and has been well described in COPD .
In lung diseases, the mechanisms involved in oxygen transport are usually impaired. The wide range of physiopathologic mechanisms possibly involved in the reduction in O2 delivery and utilization at the peripheral level are likely to play an important prognostic role in patients with lung disease. The impairment in O2 transport is characterized by a low V′O2 measured at peak exercise (V′O2 peak). The causes of reduced VO2 peak in lung diseases are: (1) reduction in cardiac output, because of primary or secondary changes in vascular pulmonary resistances or secondary cardiac involvement; (2) reduction in arterial O2 content, because of reduction in total ventilation, alveolar-capillary diffusion and ventilation-perfusion mismatch, and (3) reduction in oxygen utilization in tissues, because of sedentarism or a poor nutritional state.
The ventilatory response to exercise in advanced pulmonary disorders is usually abnormal, whereas in early disease states it is less well characterized and appears to be more variable. Conventionally, the ratio of V′E at peak exercise to the estimated maximal voluntary ventilation (MVV) represents the assessment of the ventilatory limitation or of the prevailing ventilatory constraints [6,7]. Ventilatory limitation is judged to occur when V′E/MVV exceeds 85% . In lung diseases, the increase in V′E/MVV may reflect either the reduction in ventilatory capacity (reduction in MVV), the increase in ventilatory demand (increase in V′E), or both.
(1) MVV is usually calculated by multiplying the FEV1 by 40. This index is reduced in both obstructive and restrictive lung diseases.
(2) V′E response during exercise is affected by metabolic rate, the arterial carbon dioxide tension (PaCO2) and the physiological dead space fraction of the tidal volume (VD/VT).
The relationship among these variables is described by the following equation:
where 863 is a constant, PaCO2 is the arterial carbon dioxide partial pressure in Torr and VD/VT is the dead space (VD) expressed as a fraction of the tidal volume (VT). In lung diseases, for a given V′CO2 and PaCO2, V′E is usually increased because of a higher VD/VT. The adequacy of the ventilatory response to exercise is also expressed by the ratio V′E/V′CO2 that represents the liters of ventilation necessary to clear 1 l of CO2. In lung diseases, this ratio is usually increased, particularly in patients with pulmonary vascular disorders [9,10,11]. The slope of V′E-V′CO2 during exercise (from unloaded pedaling to the ventilatory compensation point) or the V′E/V′CO2 value at the anaerobic threshold (V′E/V′CO2 at AT) are often utilized in the functional assessment of patients with lung diseases.
Analysis of the flow-volume loops is also emerging as an important tool to assess the degree of airflow and ventilatory limitation during exercise in patients with obstructive lung diseases .
The efficiency of pulmonary gas exchange can be assessed by studying the magnitude of alveolar-arterial O2 partial pressure difference (P(A–a)O2), at rest and during exercise. Normally, arterial O2 partial pressure (PaO2) does not decrease during exercise and P(A–a)O2 at peak exercise usually remains below 20–30 Torr [13, 14]. In most patients with interstitial lung disease (ILD) and pulmonary vascular disease, lung gas exchange efficiency is impaired [15,16,17] as testified by an abnormal P(A–a)O2 (>30 mm Hg) at peak exercise accompanied by arterial O2 desaturation (pulse oxymetry arterial O2 saturation, SpO2, ≤88% and PaO2 ≤55 mm Hg are usually considered clinically significant). These changes reflect regional ventilation perfusion ratio (V′A/Q′) dispersion and alterations in pulmonary capillary transit time resulting from the recruited pulmonary-capillary volume becoming inadequate for the high levels of pulmonary blood flow.
There is some debate about whether indirect measurements of SpO2 by pulse oximetry have a sufficiently high level of accuracy and reliability in measuring arterial oxygen desaturation. The 95% confidence limits of SpO2 values, relative to direct measurement of arterial oxygen saturation, have been shown to be approximately 4–5% .
A work rate of approximately 75–80% of peak V′O2 (or peak work rate) measured at incremental CPET has been utilized for CWR test. These tests are increasingly utilized in clinical practice because they better characterize exercise intolerance, improve prognostic power and better demonstrate the effects of therapeutic interventions [19, 20]. During these tests, 2 parameters are usually analyzed: the speed with which system responses attain (or attempt to attain) a new steady state (response kinetics), and the ‘endurance’ time to exercise limitation.
(1) During CWR tests, 3 phases of kinetic response of pulmonary gas exchange have been described. At exercise onset, the abrupt change in cardiac output accounts for the abrupt response in V′O2 that occurs within the first 20 s (phase I) . Phase II is a slower, exponential phase lasting 2–4 min. After phase II, the V′O2 reaches a steady state at work rates below AT, while at work rates above the AT (heavy work rates) a slow linear upward drift in V′O2 is typically observed (phase III). This upward drift is the consequence of the buffering of lactic acid that has been generated in the exercising muscles. Individuals with chronic lung disease usually show a slower time course of the V′O2 kinetics . In particular, the time constant (τ) that describes the rate of V′O2 development of phase II (quantified as the time taken to attain 63% of the steady state response) can be as much as 2- to 3-fold longer in patients with chronic lung diseases [23, 24]. The longer τ reflects mostly the reduction in O2 utilization at the muscular level, and in particular the impaired ability to increase aerobic metabolism. A slowing in the time constant of V′O2 (τV′O2) has also been described in sedentary and elderly subjects [23, 24] and can be reversed by endurance training [24, 25].
(2) During CWR testing a hyperbolic relationship between power output and endurance time, usually measured as TLIM, can be observed . The higher the work rate, the lower the TLIM. This power-duration relationship is described by the following formula:
TLIM = W′/(WR – CP) (2)
where CP is the critical power (the asymptotic work rate), and W′ is the curvature constant of the hyperbola and represents the amount of work that can be performed above the critical power (fig. 1a). The CP represents the highest work rate, somewhat above the AT, that can be sustained indefinitely because the blood lactate achieves a steady level .
|Fig. 1.a A schematic representation of work rate as function of time. The higher the maximal work rate (WRmax), the lower the tolerated duration. The power-duration relationship is hyperbolic and the critical power (CP) is the asymptotic work rate. b A schematic representation of change in the power-duration hyperbolic relationship after a therapeutic intervention. It may generate only a small upward shift in the hyperbola (from the solid to the dashed line), corresponding to a relatively small percentage increase in V′O2peak on an incremental test, but can generate an appreciable increase in tolerated duration (TLIM pre vs. TLIM post) on the CWR test at a given work rate (straight line).|
The importance of this relationship results also from the fact that therapeutic interventions (e.g. bronchodilators, oxygen, heliox and training) generate an upper shift of the hyperbola (fig. 1b) [19,28,29,30,31]. Thus, what may appear to be a relatively small percentage increase in V′O2 peak on an incremental test can correspond to an appreciable increase in TLIM on the CWR test. This accounts for the higher power of discrimination in therapy-induced changes and the greater fractional improvement in exercise tolerance for CWR than for incremental CPET. However, the magnitude of the TLIM improvement depends on the position of the pre-intervention work rate on the subject’s power-duration relationship line. In fact, with too low work rate (near or under the CP) the patients can continue exercise indefinitely, whereas if the work rate is too high, the patients tolerate exercise for only a short time, resulting in a reduction in fractional TLIM improvement.
Exercise Intolerance in COPD
In patients with COPD, the level of exertional dyspnea and exercise intolerance cannot be accurately predicted by resting lung function parameters (e.g. FEV1, DLCO, arterial blood gases). CPET is useful in such patients as it allows objective measurement of exercise tolerance, identification of the causes of the exercise intolerance and evaluation of the response to therapeutic intervention. Dyspnea is the most common exercise-limiting symptom in advanced COPD [32, 33], but leg discomfort is also frequently reported [34, 35]. Three major factors, alone or in combination, generate dyspnea, leg discomfort and early interruption of exercise: ventilatory abnormalities, pulmonary gas exchange abnormalities, and skeletal muscle dysfunction.
In COPD patients, neuromechanical uncoupling is largely due to ventilatory abnormalities. These cause an increase in elastic and dynamic work that further results in an increase in the oxygen cost of breathing. At peak exercise, normal elderly subjects spend 13% of total V′O2 for breathing , whereas severe COPD patients may spend as much as 40% of their total V′O2 .
The pathophysiological core of ventilatory limitation in COPD is the expiratory flow limitation  that increases expiratory time. During exercise, VT increases and expiratory time decreases, resulting in end expiratory lung volume (EELV) increases above the baseline value [39,40,41,42,43,44,45,46,47,48,49]. The COPD patients have not enough time to breathe out all the inspired volume when the subsequent inspiration starts. This phenomenon, called dynamic hyperinflation (DH), may be considered a compensatory mechanism that permits the lung to reach a higher volume, greater pulmonary stretching, reduced airway collapse and, finally, a higher expiratory flow that reduces, as much as possible, the expiratory time. On the other hand, DH increases the work of breathing in 2 ways. First, it shifts the VT to the upper and relatively flat extreme of the pressure-volume relationship of the respiratory system (fig. 2) where there is lower compliance (the inspiratory muscles have to generate higher pressure to pump the same volume). Second, it generates an intrinsic positive end expiratory pressure due to higher EELV. At every inspiration, the inspiratory muscles have to work harder to reduce the intrinsic positive end expiratory pressure to the atmospheric pressure before the airway flow can start [50, 51]. The increases in VT are possible until the dynamic inspiratory reserve volume diminishes to <0.5 l [3, 41, 52], after which further increases in V′E require increases in the respiratory rate, which causes a greater reduction in the expiratory time and further DH.
|Fig. 2. A schematic representation of the volume-pressure relationship (Rhan diagram) for a COPD patient. During exercise the DH generates an upper-shift of the VT (filled triangles) to the relatively flat extreme of the pressure-volume relationship of the respiratory system where there is lower compliance and higher EELV. This generates an intrinsic positive end expiratory pressure (PEEPi).|
In order to evaluate exercise DH, serial measurements of inspiratory capacity are usually performed. Since total lung capacity is expected to remain unaltered, every reduction in inspiratory capacity is due to an increase in EELV [47, 49, 53]. Recently, in a multicenter study involving 463 patients with moderate-to-severe COPD, O’Donnell et al.  reported excellent reproducibility for exercise inspiratory capacity measurements (intraclass correlation coefficient = 0.85). The deleterious impact of DH on exercise tolerance has been highlighted by the finding of a strong correlation between inspiratory capacity and exercise tolerance .
Other ventilatory abnormalities commonly observed in COPD are the following. Ventilatory inefficiency, mostly due to an increase in wasted ventilation (high VD/VT). An increase in ventilation for a given level of exercise is typically observed in COPD patients [9, 10, 11]; this alteration worsens lung mechanics and DH. Exercise CO2 retention may also be observed in COPD patients. Within certain limits this can be considered an adaptive mechanism that, by increasing the PaCO2 and PACO2, induces a reduction in ventilatory requirement (relative to the PaCO2 level), but at the same time ameliorates ventilatory inefficiency. On the other hand, patients who retain CO2 have a similar, or even greater, fall in PaO2; however, during exercise PaCO2 and PaO2 usually remain unchanged (or PaO2 may increase). In fact, changes in blood gas during exercise cannot be predicted by resting physiologic measurements. The reason for the heterogeneity of exercise CO2 response remains obscure, but may include the following.
– Abnormalities of exercise flow-volume loop reflect significant ventilatory constraints even in patients with mild COPD who have an apparently normal ventilatory reserve at peak exercise, as ascertained by the peak V′E/MVV method .
– Ventilatory response to lactic acidosis is reduced due to mechanical restriction (e.g. DH). In patients with more severe COPD, ventilation-based methods for the detection of the AT or lactate threshold  are unreliable.
The mechanisms of hypoxemia in COPD patients during exercise have been studied extensively and reflect the combination of 2 alterations: the fall in mixed venous oxygen content and low ventilation/perfusion areas of the lung generating intrapulmonary shunting [58, 59]. In a large retrospective study, Hiraga et al.  evaluated the effect of exercise induced hypoxemia on mortality. The authors reported that the slope of the PaO2-V′O2 relationship (PaO2 slope) was an independent prognostic factor and that patients with a PaO2 slope ≤80 mm Hg·l–1·min–1 were associated with <20% survival at 5 years. Although arterial desaturation may occur with any type of intense leg exercise, several studies showed that the treadmill elicits more hypoxemia than the cycle ergometer [17,61,62,63].
Hypoxemia, which in some instances worsens during exercise, affects the availability of O2 to the muscles. In COPD, the effects of chronic hypoxemia are added to those of muscle changes from the disease itself and those of inactivity. These result in important leg discomfort during exercise. Reduction of daily activities, poor nutritional state and corticosteroid therapy are all implicated in peripheral muscle atrophy [64, 65], changes in muscle fibers  and deficit in muscular phosphocreatine and adenosintriphosphate . The limitation of oxidative metabolic processes has been also demonstrated by slower gas exchange kinetics [10, 11, 68, 69].
Peak Exercise Oxygen Uptake
The measurement of exercise tolerance (i.e. V′O2 peak) has been demonstrated to be a good predictor of mortality in COPD patients. Oga et al.  found that a V′O2peak <654 ml·min–1 was associated with 60% mortality after 5 years, and a V′O2peak >793 ml·min–1 was associated with 5% mortality after 5 years. A large retrospective study by Hiraga et al.  showed similar results, reporting a mortality of 62% for patients with V′O2peak <10 ml·min–1·kg–1.
The higher power of discrimination in therapy-induced changes and the greater fractional improvement in exercise tolerance (see CWR section) has been well established in patients with COPD (table 1). In a double-blind placebo-controlled trial (187 COPD patients), O’Donnell et al.  evaluated the effect of chronic tiotropium therapy on CWR exercise tolerance. The tiotropium group, showing pretreatment CWR exercise tolerance of approximately 8 min, had a post-treatment duration of CWR test 1.6 min longer than the placebo group.
|Table 1. Effects of different types of interventions on CWR tests in COPD|
The effect of acute administration of oxygen at elevated fractions on CWR tolerance in COPD patients has been evaluated in a single-blind randomized study . The patients did not have clinically significant hypoxemia. The authors showed an average exercise duration of 4.2 min on breathing air, which increased to 7.8 min on 30% oxygen and further increased to 10.3 min on 50% oxygen.
Palange et al.  investigated the effect of breathing heliox on exercise endurance time. The low gas density was associated with an increase in peak V′E and VT, a significant reduction in lung hyperinflation and dyspnea at isotime, and a significant 115% increase in TLIM.
In a double-blind study, Emtner et al.  determined the effect of oxygen inhalation and high intensity exercise training on CWR duration. All 29 COPD patients underwent training and were divided in 2 groups. During the training, 1 group received 3 l/min of supplemental oxygen by nasal cannula, and the other group received 3 l/min of supplemental air. Before training, CWR duration was 6.6 min in the group that breathed air and 5.2 min longer in the group that breathed oxygen. After training, CWR duration increased to 21.2 min while breathing air and to 25.8 min while breathing oxygen. These improvements were somewhat underestimated because several of the post-training tests were terminated by the laboratory staff at 30 min.
Schonhofer et al.  evaluated the effect of 3 months of nightly noninvasive ventilation in patients with chronic respiratory failure (i.e. 35 COPD patients and 24 patients with thoracorestriction). In the COPD group, the authors observed an improvement of exercise parameters after noninvasive ventilation (i.e. an increase in VT, SpO2, PaO2 and pH, and a decrease in PaCO2 and HCO3–), but did not observe changes in TLIM. Only in the non-COPD group did they observe an increase of TLIM after noninvasive ventilation.
A useful tool deriving from these studies is the minimal clinically important difference, which allows investigators to evaluate the clinically important change in a given outcome. However, as described above, in CWR tests the hyperbolic trend of the power-duration relationship causes variability in the magnitude of the TLIM improvement, depending on the position of the pre-intervention work rate and, consequently, the pre-treatment duration of the test. The studies [19,20,21,22,26,27,28,29,30,31] showed that, when the initial duration of the test is in the desirable range (4–7 min), the standard deviation of change is about 3 min and the minimal clinically important difference is about 1.5 min (one-half of standard deviation).
Exercise testing has also been studied as a possible predictor of mortality in COPD patients who are candidates for lung volume reduction surgery. The National Emphysema Treatment Trial  was a large multicenter clinical trial in which 1,218 patients with severe emphysema were randomly assigned to either lung volume reduction surgery or medical therapy. They reported that patients with predominantly upper-lobe emphysema and with a pre-operative peak work rate <25 for females and <40 for males had significant improvement in survival and functional outcomes at 3 years, compared with the medical treatment group. It has also been suggested that a preoperative V′O2 peak cutoff of <15 ml·min–1·kg–1 indicated a high risk of perioperative complication in major surgery .
Exercise Intolerance in ILD
ILD represents a broad and heterogeneous group of lung disorders that are characterized by a restrictive pattern and whose main exercise feature is the inability to expand VT appropriately to the increased metabolic demand. The same pattern may be observed in neuromuscular disorders, chest wall restriction and pulmonary resection, as is seen in lung parenchymal diseases.
As in other pulmonary conditions, in patients with ILD, intolerable dyspnea or leg fatigue on exertion seems to have a multifactorial origin. Lung mechanics and pulmonary gas exchange abnormalities represent the main causes of exercise intolerance ; however, cardiovascular abnormalities and peripheral muscle dysfunction may also contribute to reduced exercise capacity in ILD .
The static recoil pressure of the lung at any given lung volume is increased in patients with ILD [75, 76]. Consequently, the pressure-volume relationship of the respiratory system looks relatively flat, meaning low compliance and a higher work rate for the inspiratory muscles at rest and during exercise. Moreover, during exercise, end-inspiratory lung volume increases to the upper extreme of the nonlinear part of the pressure–volume relationship with further rises in the elastic work rate of respiratory muscles until the VT cannot further increase (at approx. 50–60% of vital capacity). After this point, only abnormal rises in the respiratory rate can further increases the V′E. It is not uncommon to observe respiratory frequencies of >50 breaths/min–1 at peak exercise in ILD patients. Furthermore, tidal expiratory flow may exceed those of healthy individuals at a given lung volume, reflecting the reduction in inspiratory and expiratory time. The higher expiratory flow, together to expiratory flow limitation  is the basis of the failure to reduce EELV observed in these patients, and it may have negative consequences for exercise tolerance. However, during exercise, DH has not been described in ILD patients and it has been demonstrated that IC remains unaltered throughout exercise [78, 79].
In the early phases of ILD, arterial hypoxemia during exercise has been documented even before resting pulmonary function tests show overt impairment [80,81,82]. In advanced disease, the moderate or severe pulmonary fibrosis causes significant widening of the P(A–a)O2 at rest, with arterial oxygen desaturation during exercise [83,84,85,86]. The basis of the oxygen desaturation can be summarized as [83,84,85,86]:
– increased V′A/Q′ mismatch in the lungs;
– low mixed venous oxygen saturation that, in the setting of low ventilation/perfusion ratios, cannot be adequately saturated in the lung;
– decreased pulmonary capillary transit time.
Although arterial desaturation during exercise in ILD cannot be predicted on the basis of resting lung function, it has been found that there is a correlation between low resting DLCO and arterial hypoxemia during exercise .
Progressive parenchymal fibrosis, hypoxic vasoconstriction and reduction in lung volume are the main factors implicated in the reduced vascular bed and in increased pulmonary vascular resistance observed in ILD patients [87, 88]. The pathophysiologic consequence is the increase of arterial pulmonary pressure at rest and its further increase during exercise in order to maintain cardiac output [86, 89, 90]. In ILD, cardiac output is usually normal at rest and during low intensity bouts of exercise, but the rate of increase at higher work rates can be reduced in patients with more advanced disease [91, 92]. The advanced phase of the disease is characterized by right ventricular hypertrophy and cor pulmonale [86, 93]. Because of their relatively reduced stroke volume (SV), ILD patients have HR values higher than normal at submaximal levels of exercise [74, 94]. However, maximal HR is generally diminished because the patients stop exercise prematurely because of ventilatory limitation, and the cardiac reserve (difference between predicted and measured maximal HR) is adequate at the end of exercise. In patients with more severe cardiac impairment (e.g. sarcoidosis, scleroderma), the greater reduction in SV results in a reduced cardiac reserve [95, 96]. The failure to increase SV combined with O2 desaturation during exercise is well described by the trend of O2 pulse. This represents an important interpretational index that corresponds to the V′O2/HR ratio at a given level of exercise. Given the Fick Equation:
V′O2 = CO × (CaO2 – CvO2)
= HR × SV × (CaO2 – CvO2) (3)
the O2 pulse can be calculated as follows:
V′O2/HR = SV × (CaO2 – CvO2) (4)
Thus, the O2 pulse is the product of the SV and the arterovenous O2 content difference (CaO2 – CvO2). In ILD patients, the O2 pulse at peak exercise is lower and the rate of increase in the O2 pulse is reduced because of the reductions in SV and arterial O2 content.
In patients with ILD, CPET is a powerful tool that allows early detection of exercise-related ventilatory and gas exchange abnormalities. Moreover, several exertion parameters have been proven useful in predicting the prognosis of ILD patients. Miki et al.  retrospectively studied 41 patients with idiopathic pulmonary fibrosis and reported that the PaO2 slope (reduction of PaO2 in the time during exercise testing) was the most sensitive predictor of survival rate. They also found that V′O2peak, O2 pulse at peak exercise and V′E/V′CO2 at peak exercise were good prognostic indicators. The prognostic role of exercise-induced arterial hypoxemia in ILD patients has been previously shown by King et al. . In a large study involving patients with usual interstitial pneumonia, they predicted survival with a clinical/radiological/physiological scoring system and found that PaO2 at peak exercise accounted for as much as 10.5% of the maximum score.
Exercise Intolerance in Pulmonary Vascular Diseases
Pulmonary hypertension is a common feature that is reported in several lung and cardiac diseases. It is due to alterations in lung function and structure that result in higher pulmonary vascular resistance. More rarely, pulmonary hypertension can be a primary pathological condition of pulmonary vessels (i.e. primary pulmonary hypertension, PPH, and thromboembolic disease) in which the dynamic ventilatory mechanics are largely preserved during exercise. In this paragraph, we will describe only PPH patients who show a typical cardiopulmonary response pattern and have specific exercise indices for the assessment of disease severity, prognosis and response to interventions.
Pulmonary hypertension is associated with reduced exercise tolerance [99,100,101] secondary to ventilatory, gas exchange and cardiac abnormalities. Even during light physical activities, patients complain of intolerable dyspnea that is the main cause of early exercise termination.
The main ventilatory alteration observed in patients with PPH is an elevated ventilatory response to exercise due to very high levels of wasted ventilation resulting from lung hypoperfusion (abnormally high V′A/Q′ ratio) [99,101,102,103]. These alterations may occur in the absence of spirometric abnormalities. The increases in the V′E-V′CO2 slope values (usually above 32) and in V′E/V′CO2 at AT (usually above 34) correlate with hemodynamic indices of pulmonary hypertension [101, 104] and with hemodynamic improvements after vasodilator therapy . Furthermore, V′E/V′O2 is higher at any given work rate, whereas V′E/MVV at peak exercise is usually normal. Finally, the breathing pattern is more rapid and shallow than in healthy subjects, due to tachypnea and reduction in VT expansion. Activation of pulmonary C fibers and/or vascular mechanosensor inputs may be responsible for these alterations .
Patients with PPH often present with arterial O2 desaturation on exercise, which results from a widening of the P(A–a)O2 [99, 101, 103]. During exercise, because of the high pulmonary blood flow in the reduced vascular bed, the red cell transit time is reduced below the critical value that permits full hemoglobin saturation [99, 101, 103]. In addition, the high levels of wasted ventilation result in a reduction in end-tidal carbon dioxide tension (PETCO2) that, at higher levels of exercise, reaches a lower value than in healthy subjects. PETCO2 values at AT <30 Torr, in the presence of unexplained dyspnea and exercise limitation, strongly suggest pulmonary hypertension . Moreover, Valli et al.  demonstrated that different exercise modalities (i.e. walking vs. cycling) influence ventilatory and gas exchange abnormality in PPH patients. During walking, V′E/V′CO2 at AT and V′E/V′CO2 values at peak exercise were significantly higher compared to cycling, and PETCO2 and SpO2 were lower (fig. 3). The authors suggest that the differences in body posture may have a significant effect on lung mechanics and lung perfusion and concluded that walking better reveals the degree of exercise gas exchange inefficiency.
|Fig. 3. Relationship between V′E/V′CO2 and PETCO2 values measured at AT during treadmill exercise (closed circles) and during cycling (open circles). In comparison with the results from cycling, treadmill exercise caused a rightward and downward shift in the relationship. As proposed by Yasunobu et al. , the likelihood of pulmonary hypertension (suspect, likely, or very likely) depends on reference values of PETCO2 versus V′E/V′CO2 at AT. Figure reproduced from Wasserman et al. .|
Cardiac responses are not specific for PPH [103, 107] and are similar to those described in patients with ILD . Pulmonary hypertension limits the normal increase in SV during exercise  with resulting increases in HR and low peak O2 pulse. A similar pattern of HR and O2 pulse has also been described in patients with cardiac diseases and severe skeletal muscle deconditioning .
In the past, exercise was considered to have unacceptably high risks in these patients and invasive techniques such as right heart catheterization were used to define the severity of disease, prognosis and the timing of transplantation. Since the late 1990s, several studies on exercise testing in pulmonary hypertension patients have been performed in order to assess its effectiveness in defining severity and prognosis. As for the CPET test, V′O2 peak is the most significant predictor of survival in PPH. Wensel et al.  prospectively studied 86 patients with PPH and reported that the strongest predictorsof impaired survival were low V′O2 peak (p < 0.0001) and lowsystolic blood pressure at peak exercise (p < 0.0001). Patients with V′O2 peak ≤10.4 ml·kg–1·min–1 were associated with a 50% risk of death at 1 year and 85% at 2 years, whereas in patients with V′O2 peak >10.4 ml·kg–1·min–1 the risk decreased to 10% for death at 1 year and 30% at 2 years. Furthermore, the group of patients with V′O2 peak ≤10.4 ml·kg–1·min–1 and a peak systolic blood pressure of <120 mm Hg had a higher 12-month mortality (77%) than those with none (3%) or with only 1 of the 2 risk factors (21%).
Exercise Intolerance in Cystic Fibrosis
Cystic fibrosis (CF) is a hereditary disease that affects the exocrine glands of the lungs, liver, pancreas and intestines, and which causes progressive disability. Lung disease results from injury to the airways caused by mucus build-up and the resulting inflammation and infection that lead to an obstructive syndrome.
Exercise tolerance is reduced in patients with CF [110,111,112,113,114,115,116,117,118,119], and it has been demonstrated that pulmonary function weakly correlates with peak exercise capacity [117,118,119]. This relationship should reflect the main role of ventilatory constraining to exercise limitation in CF. However, poor nutritional status [120,121,122] and peripheral muscle dysfunction [115, 116, 123] have also been demonstrated to reduce peak performance in CF. Thin et al.  showed that in CF patients with moderate-to-severe pulmonary disease, wasted ventilation depends on higher dead space and on the ventilatory pattern during exercise (e.g. high respiratory rate and low tidal volume). In particular, airway dead space is independent from VT, thus in cases of reduction in VT the patient has a greater proportion of dead space for each breath. However, several studies in CF patients with mild or moderate pulmonary disease reported increases in lactate levels [125,126,127,128] and early occurrence of the lactate threshold  during incremental exercise, indicating an increase in muscle anaerobic metabolism and suggesting that peak exercise is not limited by ventilation, but rather by nonpulmonary factors that lead to leg fatigue. A relatively low V′E/MVV at peak exercise and high HR reserve supported this interpretation. Ventilatory limitation prevails in patients with more severe pulmonary dysfunction.
As for gas exchange abnormalities, it has been demonstrated that oxygen desaturation during exercise is present only in the severe state [112, 129] while no substantial decreases in SpO2 are generally seen in CF subjects with mild and moderate lung involvement.
Furthermore, in CF subjects the main factor limiting exercise at peak is peripheral muscle effort. Moorcroft et al.  studied 104 adults with CF and reported that in 75% of cases, the score for muscle effort was higher than the score for dyspnea at peak exercise. They concluded that in patients with mild-to-moderate disease, nonpulmonary factors predominate in limiting exercise. Several factors contribute to reduce muscular function in CF patients. Poor nutritional status [120, 131, 132] and reduced habitual activity [133, 134] have been demonstrated to generate muscular atrophy in CF patients. However, recent studies have suggested that reduced muscle performance may be due to an intrinsic muscle defect [115, 116].
In CF patients, several rest indices (e.g. FEV1, PaO2, PaCO2, age, sex) have been used to classify the severity of their condition and to determine the timing for lung transplantation [135,136,137,138]. Poor nutritional status and recurrent respiratory infections (Pseudomonas species and Burkholderia cepacia) may also change the prognosis among patients with the same resting indices.
Previous studies showed the powerful prognostic meaning of V′O2 peak and other exercise indices in patients with CF. Nixon et al.  evaluated prospectively survival rates in 109 CF patients followed for 8 years after initial CPET. Patients with V′O2 peak ≥82% of that predicted had a higher survival rate (83%) than patients with V′O2 peak 59–81% and ≤58% of predicted (survival rates of 51 and 28%, respectively). In this study, age, sex, body mass index, FEV1, and PETCO2 at peak exercise were not independently correlated with mortality, whereas the authors calculated that patients with higher levels of aerobic fitness had a 3 times greater survival than patients with lower aerobic fitness levels. In the same years, Stanghelle et al.  obtained comparable results in a restricted group of 8- to 16-year-old boys with CF who were followed for 8 years. On the other hand, in a group of 92 adult patients with CF, Moorcroft et al.  found that only FEV1 was significantly correlated with mortality when independent factors were entered into a multivariate logistic regression. The best combination of specificity and sensitivity cut-off for FEV1 was 55% of the predicted value. However, V′O2 peak, work rate, V′E and V′E/V′CO2 at peak exercise showed significant differences between patients who survived and those who died.
More recently, Tantisira et al.  evaluated an index that combines the resting lung function (MVV) and ventilation at AT (an independent parameter of a patient’s ability to perform at maximal effort). They found that CF patients with a higher breathing reserve at AT (defined as V′E at AT/MVV) had an increased mortality rate.
Today, CPET protocols (i.e. the maximal incremental and the CWR tests) represent important tools for evaluating the degree of exercise intolerance in patients with pulmonary diseases (COPD, ILD, PPH and CF). Among the exercise tests, CPET represents the gold standard because it allows evaluation of the causes of exercise intolerance. Characteristic patterns of ventilatory, cardiac and gas exchange response may help in diagnostic evaluation, but CPET has not proven useful in diagnosing specific conditions.
Prognostic stratification is a major indication for exercise testing in patients with pulmonary disease. Several physiological indices obtained during exercise, such as V′O2 peak, distance during the 6-minute walking test, V′E-V′CO2 slope (or V′E/V′CO2 at AT) and arterial desaturation, have proven to better predict prognosis than resting lung and/or cardiac function (table 2).
|Table 2. Exercise-related prognostic indices in pulmonary diseases|
In clinical practice, information on disease progression and the effects of interventions may be obtained by changes in V′O2 peak and TLIM. They represent highly reproducible indices that allow investigators to register minimal changes in exercise tolerance. In COPD patients, the minimal clinically important difference for these variables has been demonstrated to be useful in the evaluation of the clinically important change in a given outcome. No data are available in the literature on the minimal clinically important difference in other lung diseases.
Finally, in major surgery, lung volume reduction surgery and lung transplantation, exercise testing is recommended for the preoperative evaluation of patients.
Prof. Paolo Palange
Department of Clinical Medicine, Pulmonary Lung Function Unit Sapienza University of Rome
Viale Università 37, IT–00185 Rome (Italy)
Published online: January 14, 2009
Number of Print Pages : 15
Number of Figures : 3, Number of Tables : 2, Number of References : 142
Respiration (International Journal of Thoracic Medicine)
Vol. 77, No. 1, Year 2009 (Cover Date: January 2009)
Journal Editor: Bolliger C.T. (Cape Town)
ISSN: 0025-7931 (Print), eISSN: 1423-0356 (Online)
For additional information: http://www.karger.com/RES