INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Maximum exercise tolerance is thought to be reduced in patients with severe chronic obstructive pulmonary disease (COPD) because ventilatory demand exceeds capacity (1). Many such patients, however, also develop lactic acidemia at a low metabolic rate, which contributes to the ventilatory requirement (2). Because respiratory muscles do not seem to be a significant source of lactic acid in such patients (3), it is possible that abnormal O₂ transport to, or utilization by, limb skeletal muscle is responsible.

Although the submaximal exercise cardiac output is appropriate for the metabolic rate in COPD (4, 5), peak exercise cardiac output is not (4), a finding similar to that in congestive heart failure (8). Whether this is a primary abnormality caused by pulmonary vascular disease and right heart dysfunction (9), or simply a reflection of the patient's terminating exercise because of exhaustion of ventilatory reserve, has not yet been resolved.

A growing body of literature also suggests limb skeletal muscle oxidative metabolism is abnormal in some patients with COPD (10). If such a peripheral abnormality is relevant to the depressed O₂max in COPD, it should be associated with reduced systemic O₂ extraction. We recently found abnormal extraction in a majority of patients with severe COPD that was correlated with a low lactate threshold (15). As with cardiac output, however, the normal widening of arterial-mixed venous O₂ content during incremental exercise (8, 16) might be truncated by exhaustion of breathing reserve. The purpose of the present investigation, therefore, was to determine if blunted cardiac output and systemic O₂ extraction in the exercising patients with COPD persist when the pulmonary mechanical limit is partially relieved.

In conditions complicated by increased turbulent airflow at rest (17), including COPD (18), work of breathing is reduced by breathing a gas less dense than room air such as a helium:oxygen gas mixture (He-O₂). During maximum exercise, normal older subjects increase ventilation while breathing He-O₂ (19) by similar mechanisms. In the current study, He-O₂ was used as a tool to lift the ventilatory mechanical "ceiling" in severe COPD during incremental exercise. It was hypothesized that if peak exercise cardiac output and/or systemic O₂ extraction fail to increase during ventilatory muscle unloading, they could represent primary exercise limits in COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Between 1995 and 1997, eight consecutive patients with severe COPD were recruited after initial evaluation for lung transplantation at Massachusetts General Hospital (MGH). The diagnosis of COPD was based on medical history and pulmonary function testing, which confirmed the presence of severe irreversible bronchial obstruction (postbronchodilator FEV₁ < 50% pred), hyperinflation (TLC > 100% pred), and reduced DL_CO (< 80% pred). All patients were ex-smokers; five carried a diagnosis of 1-antiprotease deficiency. None of the patients had a history of coexisting asthma or other lung, cardiac, peripheral vascular, or neuromuscular disease. There were two male and six female patients 36 to 56 yr of age. None of the patients had suffered an exacerbation or had been hospitalized in the preceding 3 mo, and all patients were considered to be in stable condition at the time of their exercise test. Five patients were receiving long-term oxygen supplementation; two were receiving systemic corticosteroids.

Patients with COPD were compared with 10 patients with normal exercise performance. This control group consisted of consecutive patients referred for clinically indicated cardiopulmonary exercise testing because of unexplained dyspnea or fatigue, but achieved a O₂max > 80% predicted (20) and had a breathing reserve (BR) at peak exercise of  11 L/min (21). This protocol was approved by the MGH Human Research Committee.

Physiologic Measurements

Height and weight were measured, and resting pulmonary function tests were performed on all subjects on the same day as the exercise test. Body mass index (BMI) was calculated as weight (kg)/height (m)². Pulmonary function testing included spirometry (P. K. Morgan Company, Chatham, UK), body plethysmography (W. E. Collins, Braintree, MA) and single-breath DL_CO (P. K. Morgan Company).

All exercise studies were performed in the MGH Cardiopulmonary Exercise Laboratory, using an upright cycle ergometer (CPE 2000; Medical Graphics Corporation [MGC], St. Paul, MN). After subjects fasted overnight, an arterial line was inserted percutaneously into a radial artery using a 20-gauge plastic catheter. A flow-directed, balloon-tipped, pulmonary artery catheter was inserted using the internal jugular vein approach, and directed into the pulmonary artery under fluoroscopic guidance. Systemic and pulmonary artery pressures were measured with HP 1290 A quartz pressure transducers (Hewlett-Packard Co., Andover, MA), which were calibrated before each study. The transducers were interfaced with an MT95K2 recorder (Astro-Med Inc., W. Warwick, RI), and mean end-expiratory values for right atrial pressure (PRA) and pulmonary artery pressure (Ppa) were obtained, in addition to systemic arterial pressure (BP). Three-milliliter samples of systemic and pulmonary artery blood were obtained at rest and during exercise and analyzed for PO₂, PCO₂, pH (Model 1620; Instrumentation Laboratories, Lexington, MA), and O₂ saturation, hemoglobin concentration [Hb], and O₂ content by co-oximetry (Model 482; Instrumentation Laboratories). One-milliliter samples of systemic artery blood were also analyzed for lactate concentration (Analox Instruments, London, UK).

Expired gases and minute ventilation (E) were measured breath-by-breath for patients with COPD using a commercially available metabolic cart (Model 2001; MGC) in which the methodology has been previously validated (22). The pneumatotachograph was calibrated using a 3-L syringe at five different flow rates, with errors of ± 2% accepted. The pneumatotachograph was recalibrated using He-O₂ immediately prior to all studies utilizing the gas. A zirconia cell O₂ analyzer and single beam infrared CO₂ analyzer were calibrated with room air and a 5% CO₂ /12% O₂ gas. In addition, using known gas mixtures of O₂ and CO₂ ranging from 21% O₂:0% CO₂  49% O₂:4% CO₂, the gas analyzers were recalibrated with balanced N₂ or He; mean differences of 0.34% for O₂ and 0.15% for CO₂ were detected. The phase delay between volume and expired gas fraction measurements was assessed with the wave form analyzer (MGC 2001) to ensure a correct product integral for experiments with and without He-O₂. For normal control subjects, expired gas flow was measured by a Model 47303 pneumatotachograph (Hewlett-Packard, Lexington, MA). Expired PO₂ and PCO₂ were measured continuously after passage through a 3-L mixing chamber by a mass spectrometer (Model 1100; Perkin-Elmer, St. Louis, MO) and averaged every 15 s. E, O₂, and CO₂ were derived from standard formulae by a Hewlett-Packard 9000 computer after analog-to-digital conversion. This system has been previously validated (23).

Exercise Protocol

Patients completed two trials of incremental exercise to a symptom-limited maximum, separated by a 1-h rest. The latter has been shown to be sufficient to allow a reproducible measurement of O₂max in COPD (24). The initial exercise test was performed during room air breathing for six patients. For two patients who were hypoxemic (Pa_O2 < 55 mm Hg) at rest, an FI_O2 = 0.30 was used at rest and during both exercise tests. The gas mixture in the first test is hereafter referred to as N₂-O₂. The second exercise test in patients with COPD was performed while breathing He-O₂ (21% O₂  79% He, n = 6; 30% O₂  70% He, n = 2). The N₂-O₂ test was always performed first because it was the clinically indicated study. Normal control subjects performed a single incremental exercise bout to exhaustion while breathing room air.

Five minutes of rest were followed by 2 min of unloaded cycling at 50 RPM and then work rate was continuously increased using a ramp protocol (25) at either 6.25 or 12.5 W/min in patients with COPD and 12.5 or 25 W/min in control subjects. Heart rate (HR), BP, PRA and Ppa were measured continuously while pulmonary artery occlusion pressure (Ppao) and a 12-lead EKG were obtained at rest and each minute of exercise. Two-milliliter blood samples were simultaneously drawn from the radial artery and distal port of the nonwedged pulmonary artery catheter during the last minute of the rest period and during the last 15 s of each minute during exercise.

Data Analysis

Resting ventilatory and gas exchange data were obtained from the averaged final 30-s interval of the 5-min rest period. Exercise ventilatory and gas exchange data were averaged over contiguous 30-s intervals. VD/VT and AaPO₂ were calculated from standard formulae (26), and predicted peak exercise values for these variables were those of Wasserman and colleagues (26). Maximal voluntary ventilation (MVV) was calculated from a room air FEV₁ × 40 (21). Breathing reserve (BR) was calculated as MVV  E_max. A BR < 11 L/min was considered abnormal (21).

O₂max was defined as the highest 30-s averaged O₂ during the last minute of the symptom-limited exercise test. Predicted values for O₂max were those of Hansen and colleagues (21). The lactate threshold (LT) was defined for each exercise test as the O₂ at the intercept of the two linear regressions of a log-log plot of lactate concentration versus O₂ resulting in the least residual sum of squares (27). An LT was deemed absent if peak lactate concentration did not surpass two standard deviations above the resting mean for this laboratory (upper limit of normal = 2.0 mM).

Maximal predicted HR was estimated as 220  age in years (26). Cardiac output () was calculated from the Fick principle ( = O₂/ [Ca_O2  Cv_O2]), and stroke volume from  /HR. Predicted maximal cardiac output was calculated from predicted O₂max and an assumed maximal arterial-venous O₂ content difference of 15 ml /dl for healthy untrained subjects (16).

Oxygen delivery (DO₂) was calculated as  × Ca_O2. The measured peak exercise values for Ca_O2 and Cv_O2 in control subjects were used to determine percent predicted values in COPD. Pulmonary vascular resistance was calculated as (mean Ppa  Ppao)/. The systemic O₂ extraction ratio (O₂ER) difference was calculated as (Ca_O2  Cv_O2)/ Ca_O2.

Data are expressed as mean ± standard error of the mean (SEM). Discrete rest and exercise variables for patients with COPD with and without He-O₂ were compared by a two-tailed paired t test. Comparisons between patients with COPD with and control subjects were made by an unpaired t test. Continuous data were analyzed by simple linear regression. Computations were made with the Statview 4.0 statistical program (Abacus Concepts, Berkeley, CA). A p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Demographic and Resting Physiologic Data

Demographic data are shown in Table 1. Patients with COPD were younger than control subjects, and the COPD group included predominantly women, whereas the control group consisted mainly of men. Patients with COPD were shorter and had a reduced BMI. Hemoglobin concentration was not different. Pulmonary function test results for patients with COPD and control subjects are shown in Table 2. Patients with COPD were characterized by severe airflow obstruction, hyperinflation, and reduced DL_CO.

Incremental Exercise

Exercise tolerance and symptoms. Peak work load was severely depressed in patients with COPD, and failed to improve with He-O₂ (Table 3). While breathing N₂-O₂, dyspnea was reported as the exercise-limiting symptom by five patients with COPD, and both dyspnea and leg fatigue were reported by the other three. With He-O₂, dyspnea was limiting in four, and both dyspnea and leg fatigue contributed in four. All normal subjects reported leg fatigue as their sole limiting symptom at peak exercise.

Ventilation and gas exchange. All patients with COPD reached a pulmonary mechanical limit to exercise (breathing reserve < 11 L/min) while breathing N₂-O₂. He-O₂ was associated with a 32% increase in maximum E in patients with COPD (Table 3), which exceeded the calculated room air MVV (E_max /MVV = 1.14 ± 0.09 versus 0.86 ± 0.05, p = 0.001). None of the control subjects reached a pulmonary mechanical limit.

Peak exercise Pa_CO2 fell breathing He-O₂, though it remained higher than normal (Table 3). Patients with COPD had an abnormally high VD/VT and AaPO₂ with exercise while breathing N₂-O₂; VD/VT actually worsened with He-O₂.

Maximum O₂ uptake and blood lactate. A modest increase in O₂max occurred with He-O₂ (Table 3), but it remained markedly abnormal. A lactate threshold occurred for six (75%) patients with COPD, both with and without He-O₂, and for all normal subjects. The LT was abnormal in patients with COPD breathing N₂-O₂ (16 ± 2% pred O₂max) and He-O₂ (19 ± 2% pred O₂max) versus normal subjects (46 ± 3% pred O₂max, p < 0.0001 for each). The LT did not change with He-O₂.

Systemic O₂ delivery. Peak exercise was markedly reduced in COPD (Table 4), and it accounted for approximately two thirds of the reduction in O₂max, according to the Fick equation (Figure 1). Peak failed to improve with He-O₂. The submaximal exercise cardiac output response ( versus O₂ slope and y-intercept) was normal, however, during both N₂-O₂ and He-O₂ breathing (Figure 2). A depressed peak HR was responsible for some of the reduction in max in COPD, and neither it nor the stroke volume changed with ventilatory muscle unloading (Table 4). Peak exercise PRA, Ppa, and Ppao did not differ between groups. PVR, however, was markedly elevated in COPD and did not change with He-O₂.

View larger version (32K): [in this window] [in a new window]   Figure 1.   Change from normal in Fick principle variables at peak exercise in severe COPD. *p < 0.05, patients with COPD, He-O2 versus N2-O2 test.

View larger version (16K): [in this window] [in a new window]   Figure 2.      O2 relationships shown for patients with COPD breathing N2-O2 and He-O2 during exercise, and normal control subjects. A    O2 slope and intercept were derived for each subject based on resting and peak exercise data; illustrated lines represent group mean slopes and intercepts. There were no differences in slopes (6.3 ± 0.6, 5.6 ± 0.8, 6.2 ± 0.3 ml/ml; patients with COPD breathing N2-O2 and He-O2, and normal control subjects, respectively, p > 0.05 for all comparisons) or intercepts (2.3 ± 0.4, 2.6 ± 0.5, 2.4 ± 0.6 L/min; patients with COPD breathing N2-O2 and He-O2, and normal control subjects, respectively, p > 0.05 for all comparisons) between groups.

Peak exercise Ca_O2 was abnormal in patients with COPD breathing N₂-O₂, and it increased modestly with He-O₂ (Table 4 and Figure 1). The mechanism of increase was a leftward shift in the oxyhemoglobin dissociation curve because peak exercise Sa_O2 increased with He-O₂ without a change in Pa_O2. A trend towards increased peak exercise [Hb] with He-O₂ (15.4 ± 0.9 versus 15.0 ± 0.8 g/dl, He-O₂ versus N₂-O₂, p = 0.11) also contributed to the increase in peak exercise Ca_O2. Peak exercise DO₂ did not change with He-O₂.

Systemic O₂ extraction. Peak exercise Cv_O2 was abnormally high in patients with COPD breathing N₂-O₂ (Table 4), accounting for approximately one third of the reduction in O₂max (Figure 1), and did not decrease with He-O₂. The relationship between blood lactate and systemic O₂ extraction ratio, with mean values at rest, LT, and peak exercise is illustrated in Figure 3. O₂ER was abnormally low at the LT and at peak exercise and failed to significantly increase with He-O₂. When patients were subdivided on the basis of O₂ desaturation  85% with exercise (n = 4), no difference in peak exercise O₂ER was noted.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Relationship between blood lactate and oxygen extraction ratio (O2ER) using mean values at rest, the lactate threshold (LT), and peak exercise. *p < 0.05 versus normal subjects; p = 0.06 versus normal subjects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, He-O₂ was used to mechanically unload the ventilatory muscles and determine if O₂ transport or utilization limits incremental exercise in patients with severe COPD. Increased E and reduced hypercapnia suggest respiratory muscle unloading did in fact occur with He-O₂ at peak exercise. These changes are similar to those seen in the older normal human breathing He-O₂ during intense exercise (19). Despite raising the ventilatory ceiling with He-O₂, abnormal peak exercise O₂, cardiac output, and systemic O₂ extraction failed to improve proportionally. These data suggest that abnormalities of O₂ transport and utilization may contribute to the depressed O₂max in severe COPD.

Although this study was not aimed at evaluating the effects of He-O₂ on pulmonary gas exchange during exercise, we were surprised in finding higher peak exercise Sa_O2 with He-O₂, despite no change in Pa_O2 and with an elevated VD/VT. The increase in Sa_O2 (and Ca_O2) with He-O₂ must have occurred because of a leftward shift in the oxyhemoglobin dissociation curve, in turn caused by relative hypocapnia. We speculate that CO₂ excretion was less efficient with He-O₂ in the current study because it preferentially favored ventilation of the more diseased airways dominated by turbulent flow.

Cardiac Output

We found, as did others, a blunted peak (4), but normal submaximal exercise cardiac output response (4, 5) in COPD. The inference that O₂max is limited in COPD by pulmonary mechanics alone, and not by cardiac output, should be made with caution, because patients with advanced congestive heart failure also maintain the normal versus O₂ slope of 5 to 6 ml/ml (8). In the present investigation, neither peak exercise heart rate nor stroke volume improved with respiratory muscle unloading, despite being given the "opportunity." This suggests the possibility of a primary, rate-limiting abnormality of cardiac function during exercise in severe COPD.

Cardiac output was decreased at peak exercise in part by a reduced peak heart rate. Chronotropic incompetence during exercise has been reported in both right (28) and left (29) ventricular failure which may be due in part to afferent signals that arise from the exercising muscles (29). Alternatively, relative bradycardia might have occurred on the basis of reduced lean muscle mass (30).

Reduced peak exercise stroke volume in COPD has been explained by reduced left ventricular preload because of right heart afterload (9, 31); however, left ventricular underfilling can also occur as a consequence of ventricular interdependence in the right ventricular volume overload state. Because peak exercise PVR was markedly elevated and Ppao was not different versus normal in the current study, the former mechanism likely predominated. In addition to a diminished and poorly recruitable pulmonary circulation in the resting emphysematous patient, dynamic hyperinflation may have mitigated the fall in PVR during exercise.

Systemic O₂ Extraction

Another major finding of the present study is that peak exercise systemic O₂ extraction is abnormal in COPD and, as with cardiac output, fails to improve with ventilatory muscle unloading. The normal /O₂ relationship during exercise in COPD has been used in the past to exonerate systemic O₂ extraction (32). Although it is true that patients with pure oxidative myopathies typically have an increased /O₂ slope during exercise (33), an apparently normal relation may result from coexisting abnormalities of cardiac output and extraction. Two prior studies measured mixed venous blood O₂ content during maximum incremental exercise in COPD and found subsets of patients with blunted O₂ extraction (6, 7). This does not provide conclusive evidence that abnormal systemic O₂ extraction limits O₂max in COPD because (as with cardiac output) a primary pulmonary mechanical limit could have truncated otherwise normal extraction (8, 16). This hypothesis was refuted in the current study, however, because systemic O₂ extraction did not improve when the ventilatory mechanical load was lightened by He-O₂.

Consideration of the normal relationship between lactate threshold and systemic O₂ extraction ratio suggests that the latter is abnormal in COPD during the submaximal domain of exercise, when ventilatory mechanics are not thought to be limiting. Weber and Janicki (8) have shown in normal subjects and in patients with congestive heart failure of varying degrees of severity that the lactate threshold occurs relatively reproducibly when the O₂ER exceeds 0.60. Our control subjects' LT occurred at an O₂ER of 0.57, but in the patients with COPD the O₂ER at LT was less than 0.50. These data, similar to those from our prior noninvasive study (15), suggest that abnormal extraction and early lactic acidemia in COPD may be causally related.

Abnormal systemic O₂ extraction in COPD could occur because of a mismatch in perfusion and skeletal muscle metabolism, an inherent defect of muscle oxidative capacity, or both. Functional abnormalities of systemic vasoregulation have been described in current smokers (34), and a systemic microangiopathy has been reported in some patients with COPD (35), but the question of abnormalities in macrocirculatory or microcirculatory structure and function in COPD remains unanswered.

Lactic acidemia and poor peripheral O₂ extraction are hallmarks of the oxidative myopathies (33), another possibility in severe COPD. Using biopsies, Jakobsson and coworkers (13) found enzymatic evidence for augmented glycolytic (increased phosphofructokinase activity) and decreased aerobic capacity (citrate synthase) in the resting quadriceps femoris of 18 patients with advanced COPD, which failed to reverse after long-term oxygen therapy. Maltais and colleagues (14) found biopsy evidence for an interrelationship between reduced limb muscle oxidative enzyme activity and an excessive increase in blood lactate during exercise in patients with COPD. Using ³¹P-magnetic resonance spectroscopy (MRS) of exercising calf muscle, Payen and colleagues (10) found a higher inorganic phosphate to phosphocreatine (Pi/PCr) ratio, decreased intracellular pH (pHi), and slower PCr resynthesis during recovery in seven patients with COPD compared with control subjects, suggesting impaired oxidative phosphorylation. Kutsuzawa and colleagues (11) also found abnormally low forearm PCr/(PCr + Pi) values for the normalized work rate in COPD. On the other hand, Thompson and coworkers (12), found ³¹P-MRS indices of mitochondrial function were normal during recovery and concluded that (unmeasured) skeletal muscle O₂ delivery was responsible for decreased oxidative metabolism during exercise in COPD. Mannix and coworkers (36) calculated the relative contributions of anaerobic sources and oxidative phosphorylation to exercising skeletal muscle ATP production in COPD. They also concluded that ATP production from oxidative phosphorylation is decreased in COPD, but that it can be reversed by supplemental inspired O₂. Thus, how much of the abnormal peripheral skeletal muscle oxidative metabolism in COPD is related to inadequate O₂ delivery is not yet known.

Deconditioning might appear to be the explanation for impaired extraction; its prevalence is high in COPD (2, 32), it is associated with reduced limb muscle mass (37) and oxidative enzyme activity (38), some of which can be increased by endurance training (39). Conversely, systemic O₂ extraction is relatively preserved (16), however, and noninvasive markers of the lactate threshold are only marginally abnormal (21) in simple detraining. The relative contribution of deconditioning to abnormal systemic O₂ extraction and O₂max in severe COPD remains to be determined.

Study Limitations

The possibility of a type II error for cardiac output and systemic O₂ extraction changes is recognized. If a larger sample size resulted in a (small) statistically significant increase in those Fick variables, the major finding of the study would be unchanged. This is because the magnitude of increase in and extraction were inappropriate for a 30% increase in peak exercise E. The study was also potentially limited by a lack of treatment randomization. Because maximum incremental exercise testing in COPD, repeated 1 h apart, elicits a reproducible O₂max (24), it is unlikely that significant bias was introduced. Another source of bias stems from the use of a nonmatched control group; patients with COPD were younger and predominantly female compared with control subjects. To circumvent this form of potential bias, values were presented as a percent of predicted, in addition to absolute values, whenever possible. Although there are no standard predicted values for systemic O₂ extraction ratio during exercise, healthy young subjects and patients with congestive heart failure all approach a value of approximately 0.75 at maximum exercise (8). Therefore, it is unlikely that maximal systemic O₂ extraction during exercise is significantly altered by age. To our knowledge, there are no data comparing maximal systemic O₂ extraction during exercise between men and women, however. Finally, it must be recognized that since the current study involved a relatively young group of patients with severe COPD and a majority suffering from alpha-1-antiprotease deficiency, the results may not extend to the general COPD population.

Conclusions

This study suggests that some patients with severe COPD have blunted cardiac output and systemic O₂ extraction responses to incremental exercise that cannot be fully accounted for by limiting pulmonary mechanics. Abnormalities of O₂ transport and utilization may contribute to the depressed O₂max in these patients.