INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic obstructive pulmonary disease (COPD) is characterized by a progressive clinical course that unfolds over years. As COPD worsens, the mechanisms that maintain normal ventilatory pump and circulatory function are stressed. The onset of either chronic hypercapnia or cor pulmonale generally heralds advanced COPD and is associated with poor survival. Indeed, the presence of resting hypercapnia has been recognized as a predictor of mortality and is a marker of advanced complicated disease (1). The exact prevalence of hypercapnia (Pa_CO2 > 45 mm Hg) in severe COPD remains unknown. However, some studies have reported an overall prevalence between 23 and 28% in patients with similar degrees of airway obstruction (2, 3).

Although respiratory muscle (RM) weakness or fatigue is frequently cited as a cause of hypercapnia (2, 4), the status of respiratory muscle function in hypercapnic patients with stable COPD is complex and controversial. Bégin and Grassino (2) reported that inspiratory muscle weakness (as expressed by maximal inspiratory pressure or PI_max) and the degree of airway obstruction (as expressed by FEV₁) were the most important determinants of chronic hypercapnia in clinically stable COPD. They postulated that chronic alveolar hypoventilation was most likely the result of a breathing strategy developed to avoid fatigue of weakened muscles having to overcome high inspiratory loads. In other words, hypercapnic patients behave as "wise fighters" who weigh their options and choose hypoventilation rather than risk respiratory muscle fatigue (2). On the other hand, several studies have documented that patients with COPD have a significant reduction in maximal inspiratory pressure (PI_max), which generally has been attributed to the effect of static and dynamic hyperinflation on inspiratory muscle function (5). Because many patients with COPD also exhibit a decrease in maximal expiratory pressure (PE_max), Rochester and Braun (7) have proposed that generalized muscle weakness also exists in many of these patients, and that the weakness is probably due to other factors such as malnutrition, hypoxemia, hypercapnia, weight loss, cor pulmonale, and electrolyte abnormalities.

Some studies in healthy subjects and in patients with COPD without resting hypoxemia suggest that experimental acute hypoxemia and hypercapnia per se may cause respiratory muscle weakness (8, 9). Hypoxemia has been associated with decreased endurance for exercise (8). Hypercapnia decreases the contractility and endurance time of the diaphragm in normal subjects (9). For all these reasons, it has been suggested that in severe COPD the presence of hypoxemia, hypercapnia, and the resulting acidosis may further affect RM function and thus worsen the hypercapnia (4).

We have shown that resting central drive and CO₂ response is similarly affected in eucapnic and hypercapnic patients when compared with that in normal subjects (10). Therefore, hypercapnia in stable COPD does not seem to result from a reduction in central drive. On the other hand, hypercapnia could result from overloading of ventilatory muscles working close to the fatigue threshold or from dysfunctional muscles affected by the hypercapnia itself. If this were the case, it would be expected that exercise capacity, and respiratory muscle function during exercise, should differ between eucapnic and hypercapnic patients. In this report, we extend our previous findings to include the metabolic, ventilatory, and RM function during exercise in the same 19 eucapnic and 13 chronic hypercapnic patients with severe stable COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

A total of 32 clinically stable patients with severe COPD participated in this study. The diagnosis of COPD was made according to the ATS criteria (11). Severe COPD was defined as a value of FEV₁ less than 50% of predicted. At the time of entry into the study the patients were clinically stable and receiving optimal medical therapy. The protocol was approved by the local Committee on Human Research and each patient gave informed consent to participate.

For comparison, and based on Pa_CO2, the patients with COPD were divided into two groups. (1) Eucapnic defined as a Pa_CO2 less than 44 mm Hg (19 patients), and (2) Hypercapnic with a Pa_CO2 greater than 45 mm Hg (13 patients). In a separate report, we have previously reported the results of resting central drive and response to CO₂ in these patients (10).

Pulmonary Function Testing and Blood Gas Analysis

Spirometry was performed with a volume displacement spirometer (Warren E. Collins, Braintree, MA), and FEV₁, FVC, and FEV₁/FVC were calculated according to the recommendations of the American Thoracic Society (12). FRC was measured in a body plethysmograph (Warren E. Collins) as described by Dubois and coworkers (13). Arterial blood samples were taken with the patients at rest and breathing room air. They were analyzed for oxygen tension (Pa_O2), carbon dioxide tension (Pa_CO2), and pH with appropriate electrodes (278 Blood Gas System; Ciba-Corning, Medfield, MA).

Incremental Exercise Test (Metabolic, Respiratory, and Cardiac Response)

Exercise test was performed on a cycle ergometer (Warren E. Collins) using a standard 1-min incremental cycle exercise protocol. Patients were started with a 2-min period of unloaded pedaling at 60 cycles/min, followed by 15-watt increments/min. The patients were strongly encouraged to cycle until discomfort or exhaustion was reported. Heart rate and rhythm were continuously monitored with a three-lead electrocardiogram, and blood pressure was measured with the cuff technique.

Minute ventilation and its components were measured using a pneumotachograph. The concentration of expired oxygen and carbon dioxide were analyzed breath by breath with a zirconium dioxide cell O₂ analyzer and an infrared CO₂ analyzer, respectively. These measures and flow signals were electronically integrated by a computerized system to yield 30-s averages of minute ventilation (E), respiratory duty cycle (TI/Ttot), tidal volume (VT), respiratory rate (RR), oxygen uptake (O₂), carbon dioxide output (CO₂) and, gas exchange ratio (R). The heart rate was also measured and used to obtain the heart rate reserve (HRR), O₂ pulse, and HR/VO₂. Maximal O₂ pulse (O₂-Pmax) was measured where O₂ pulse reached the plateau. Predicted maximal oxygen consumption (O₂max) was calculated using standard equations (14). The maximal voluntary ventilation (MVV) was measured in all patients by having the patients ventilate as deep and as fast as possible for 12 s, and this value was normalized for 1 min. The predicted maximal heart rate was obtained using the following formula: HRmax = 210  (Age × 0.65). Arterial blood gases were also obtained at peak exercise, as close as possible to the limitation of exercise. In no patient was the sample obtained later than 20 s after peak exercise. Arterial hemoglobin oxygen saturation (Sa_O2), was monitored noninvasively using a pulse oxymeter.

Respiratory Muscle Function Evaluation

We continuously monitored gastric (Pg) and pleural pressure (Ppl), using two thin-walled latex balloons (A and E Medical Corp., Farmingdale, NJ) passed transnasally, with one positioned in the stomach and the other in the midesophagus (15). Both balloons were secured at the nose. A separate transducer (Validyne Co., Northridge, CA) measured each pressure, and the calibrated output was continuously displayed on a strip chart recorder (Gould Inc., Rolling Meadows, IL). Mouth pressure was measured using a separate transducer, and the calibrated output was displayed on the same recorder. Transdiaphragmatic pressure (Pdi) was calculated as the difference between Pg and Ppl at end-inspiration (Pdi = Pgi  Ppli). Maximal Pdi (Pdi_max) was measured at FRC, and maximal inspiratory mouth pressure (PI_max), maximal Ppli (Ppli_max), and maximal Pgi (Pgi_max) pressures at RV by having the patient perform a maximal inspiratory effort against a partially occluded shutter. The patients were asked to maximally expand the chest and the abdomen and were coached in the performance of this maneuver until three reproducible results were obtained (16). Maximal expiratory mouth pressure (PE_max) was measured at TLC by having the patient forcefully exhale against a partially occluded airway. The highest value to three determinations was recorded as PE_max.

Continuous recording of Pg and Ppl were used to evaluate ventilatory recruitment (VMR) pattern (pg/Ppl). In real time, the calibrated signal from the pneumotacograph was displayed with the simultaneously obtained Pg and Ppl using a fast-response digital recorder (Hewlett-Packard, Waltham, MA). The data were recorded in a personal computer system using an analog-to-digital converter and a commercially available program. Over five consecutive tidal breaths, Pg-Ppl plots were constructed at rest and at peak exercise by averaging the end-inspiratory Ppl (Ppli) and Pg (Pgi) and end-expiratory Ppl (Pple) and Pg (Pge). The beginning and the end of inspiration were defined from the points of zero flow. Pg/Ppl was calculated using the following equation: Pg/Ppl = (Pgi  Pge)/(Ppli  Pple). In the analysis of Pg/Ppl, the more negative the value, the greater the contribution of the diaphragm in the generation of ventilatory pressures. As the inspiratory muscles of the rib cage and the expiratory muscles increase their participation in ventilation, the value of Pg/ Ppl becomes more positive (17). To further evaluate diaphragmatic function, we determined at rest, at peak, and at 5 mn postexercise, the ratios of mean Pdi/Pdi_max and the diaphragmatic tension time index (TTDI). Pdi during tidal volume (VT) was obtained by averaging five-consecutive tidal breaths, and TTDI was calculated as TTDI = mean Pdi/Pdi_max × TI/Ttot, where TI and Ttot represent inspiratory time and total respiratory cycle time measured from the flow signal, respectively (18, 19).

Statistical Analysis

Data are presented as mean ± SD. The differences in static lung function, blood gases, resting and peak exercise metabolic, and RM function parameters, between eucapnic and hypercapnic patients with COPD were evaluated using the t test for independent samples. The analysis was also repeated using Wilcoxon's rank sum/Mann-Whitney U test for variables with large variances. The results were similar, irrespective of the method used. For all tests, a p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary Function and Gas Exchange

The mean values for the physical characteristics, pulmonary function test, and blood gas parameters from the eucapnic and hypercapnic patients are shown in Table 1. Patients in both groups were similar in age, weight, and height (p > 0.05). FVC and airway obstruction (FEV₁) were different in the two groups (p < 0.001). There was no significant difference in the results of TLC, FRC, and RV between eucapnic and hypercapnic patients (p > 0.05). In contrast, hypercapnic patients had lower MVV and resting Pa_O2 (p < 0.05).

Exercise Response

Resting and peak exercise metabolic measurements, and blood gases from eucapnic and hypercapnic patients are shown in Table 2. As shown in Figure 1, patients in both groups had similar exercise capacity as expressed by peak O₂ uptake (O₂max, O₂/kg). The maximal work capacity was 71 ± 51 Watts in eucapnic and 48 ± 49 Watts in hypercapnic patients, (p > 0.05). Hypercapnic patients had lower E_max. Because they also had a lower MVV, the breathing reserve value (VE/MVV%) was similar in both groups. There was no significant difference in the results of cardiovascular (Max HR%) and O₂ delivery indices (O₂ Pulse max, HR/VO₂) between eucapnic and hypercapnic patients. Peak exercise Pa_O2 was lower and Pa_CO2 was higher in the hypercapnic group (p < 0.05). There was no difference in VD/VT between the two groups at rest or during exercise.

View larger version (7K): [in this window] [in a new window]   Figure 1.   The maximal oxygen uptake at peak exercise was similar in eucapnic and hypercapnic patients with COPD.

The respiratory rate at rest (19 ± 6 versus 19 ± 3) and at peak exercise (33 ± 13 versus 32 ± 7) was similar in both groups. The TI/Ttot was slightly lower in hypercapnic (0.36 ± 0.13) versus eucapnic (0.39 ± 0.12) patients, but this difference was not statistically significant. The VT at peak exercise was significantly lower in hypercapnic (0.74 ± 0.32 L) than in eucapnic (1.05 ± 0.39 L) patients, p = 0.027. However, because the TI was shorter in the hypercapnic (0.66 ± 0.05 s) versus eucapnic (0.71 ± 0.03 s) patients, there was no difference in the VT/TI. The value at peak exercise was 1.45 ± 0.4 L/s in the eucapnic group and 1.1 ± 0.5 L/s in the hypercapnic groups, p = 0.27.

Respiratory Muscle Function

As shown in Figure 2, marked reductions in maximal ventilatory pressures (PI_max, PE_max, and Pdi_max) were noted in the eucapnic and hypercapnic patients with COPD (20). However, no significant differences were found between the two groups.

View larger version (22K): [in this window] [in a new window]   Figure 2.   The maximal inspiratory (PImax), expiratory (PEmax), and transdiaphragmatic (Pdimax) pressures were similar in eucapnic and hypercapnic patients with COPD.

The mean values of VMR data at rest and at peak exercise for the two groups of patients are shown in Table 3. Resting Ppli was slightly but significantly more negative in the hypercapnic patients (p < 0.05). There was no significant difference in the results of resting and peak Pgi, Pple, Pge, and Pdi. The Pg/Ppli at rest and during exercise were similar in eucapnic and hypercapnic patients, as shown in Figure 3 (p > 0.05).

View larger version (8K): [in this window] [in a new window]   Figure 3.   Plots of gastric (Pg) and pleural (Ppl) pressures in cm H2O, measured at end-inspiration (I) and end-expiration (E). The slopes were similar at rest (left panel ) and during exercise (right panel ) in eucapnic (squares) and hypercapnic (circles) patients with COPD.

Diaphragmatic load measures (mean Pdi/Pdi_max and TTDI) were similarly increased in the two groups of patients (Figure 4).

View larger version (36K): [in this window] [in a new window]   Figure 4.   Diaphragmatic load (mean Pdi/Pdimax) and tension time index during exercise was similar in eucapnic and hypercapnic patients with COPD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There are two important findings in this study. First, exercise capacity is similarly reduced in eucapnic and hypercapnic patients with severe stable COPD, but hypercapnic patients achieve lower ventilation with exercise. Second, this is not due to changes in RM function as respiratory muscle strength, ventilatory muscle recruitment at rest and during exercise, and diaphragmatic load are equally affected in both groups of patients. Therefore, hypercapnia in patients with stable COPD at rest and during exercise is most likely due to an incapacity of the lungs to increase ventilation as a function of the increased respiratory muscle work, and not to an adaptive strategy adopted by the pump as a whole, or to respiratory muscle weakness or dysfunction.

The past years have witnessed an increasing interest in the factors that determine exercise performance in patients with COPD (21). Although decreased ventilatory capacity has been proposed by several studies as the main determinant of exercise tolerance (21), the role of other factors such as deconditioning, hypoxemia, dynamic hyperinflation, peripheral muscle weakness, and hemodynamic compromise continue to be a subject of controversy. The relationship between hypercapnia, exercise performance, and respiratory muscle function in COPD remains unknown. Although several studies have evaluated simultaneously many potential factors associated with exercise capacity (21), most of them have assessed patients with a wide and variable range of airway obstruction, and none has compared exercise parameters and RM function during exercise between eucapnic and chronic hypercapnic patients with severe stable COPD.

The first important finding in the present study is that exercise capacity (O₂max and watts), cardiovascular response, and O₂ delivery indices are equally reduced in eucapnic and hypercapnic patients. This suggests that peripheral muscle performance is similarly affected in all patients with stable COPD independent of the degree of hypercapnia, otherwise there would have been differences in oxygen utilization during exercise between the two groups (21). Despite similar oxygen uptake, hypercapnic patients achieved a lower ventilation and Pa_O2 and manifested higher Pa_CO2 at peak exercise. These findings are very likely the result of decreased alveolar ventilation in the hypercapnic patients, which could have occurred via several mechanisms.

The first is the development of an adaptive strategy by the ventilatory pump to delay or prevent respiratory muscle fatigue. This has been frequently cited as a cause of hypercapnia in COPD. Begin and Grassino (2) reported that hypercapnic patients differ from eucapnic patients in that they have a higher inspiratory muscle load and a lower PI_max. They postulated that chronic alveolar hypoventilation was most likely the result of a breathing strategy developed to avoid overloading of the inspiratory muscle, leading to fatigue and possible irreversible failure. This interesting hypothesis was based on indirect evidence since no measurements of central drive or exercise performance was performed in those patients. In that study, the investigators reported that inspiratory muscle loading among hypercapnic patients was variable, thereby recognizing that hypercapnia could not be entirely explained by a combination of respiratory load and inspiratory muscle weakness. Nevertheless, the concept of a "wise" breathing strategy adopted by the respiratory muscles to prevent further damage and fatigue has remained a logical and attractive one.

In a separate report on this same group of patients, we showed that baseline central drive and mouth and pleural occlusion pressure response to CO₂ were similar between hypercapnic and eucapnic patients (10). Our findings and those of Scano and coworkers (30) indicate that central drive is relatively well preserved in hypercapnic patients with COPD, and that the response of the respiratory centers is, in a way, unimodal, increasing with increasing severity of airflow obstruction. These findings agree with those of DeTroyer and coworkers (32) who, using diaphragmatic electromyography, also showed increased neuronal diaphragmatic activation in patients with COPD independent of the level of CO₂. We now extend our analysis to include the evaluation of central drive during exercise. The similar value of VT/TI observed in our study provide further evidence that central drive is similarly affected in eucapnic and hypercapnic patients even during exercise.

If the pump could adopt a strategy to prevent respiratory muscle fatigue, it most likely would result in differences in intrathoracic pressures at rest that should become more prominent during exercise. This was not the case in our study, as shown in Table 3 and Figure 2. Indeed, the ventilatory muscle recruitment pattern and the magnitude of the ventilatory pressures at rest and during and after exercise was the same in both groups of patients. This indicates that hypercapnia is not solely the consequence of an "adaptation" of the central controller aimed at preventing respiratory muscle failure or fatigue.

The second possibility is that in spite of similar driving stimulus, the respiratory muscles of hypercapnic patients are weaker, more dysfunctional, or fatigued. Our findings argue otherwise. In our previous report on the same patients, we had shown that overall respiratory muscle force and lung volume, the prime determinants of respiratory muscle length, are similar in both groups. In the present report we expanded the analysis to include diaphragmatic load and ventilatory muscle recruitment pattern during exercise. The only difference between the groups was the development of a more negative pleural inspiratory pressure in the hypercapnic group at rest that was not observed during exercise. This small difference indicates that the respiratory muscles of hypercapnic patients were capable of generating similar if not greater pressures than eucapnic patients. This small difference in Ppli did not alter the Pg/Ppl slope, indicating similar RM recruitment. Likewise, the similar end-expiratory pleural pressure in both groups indicate no difference in end-expiratory lung volume (dynamic hyperinflation) or intrinsic positive end-expiratory pressure. Given that both group of patients had similar respiratory muscle strength, ventilatory muscle recruitment pattern, and diaphragmatic load during and after exercise, neither weakness nor dysfunction or fatigue of the respiratory muscles are the most important reasons for hypercapnia in stable COPD.

One other theoretical explanation for respiratory muscle dysfunction in hypercapnic COPD is the effect of hypercapnia per se on respiratory muscle function. Data from animal studies indicate that high levels of carbon dioxide impair diaphragmatic contractility (32, 33). Fitzgerald and coworkers (33) demonstrated that acidosis caused by acute hypercapnia, or by a reduction in the bicarbonate level, decreases the force of contraction of the rat diaphragm in vitro. Likewise, in a study using anesthetized dogs, Schader and coworkers (34) observed a reduction in diaphragmatic pressure while having the animals breathe CO₂ during spontaneous inspiratory efforts or during bilateral electrophrenic stimulation at constant length and geometry. Later, the same group studied the effect of acute increases in CO₂ on diaphragmatic contractility and performance in four normal subjects (9). They showed that hypercapnia reduces the capacity of the diaphragm to generate force during voluntary contraction. Moreover, when breathing was performed against a small resistance, hypercapnia produced electromyographic changes consistent with diaphragmatic fatigue. The exact mechanisms by which CO₂ affects muscle contractility remain unknown and continue to be a subject of controversy. However, it has been postulated that its effect is mediated by an alteration of muscle pH (35). Low pH has a detrimental effect on the contractile function of the muscle through several possible mechanisms: decreasing the affinity of the troponin for calcium, increasing the binding of calcium by the sarcoplasmic reticulum, and reducing the rate of glycolysis and thus ATP resynthesis. The increased partial pressure of blood CO₂ may affect any of these mechanisms, rendering the diaphragm less efficient as a force generator. If the hypercapnic patients in our study had decreased ventilation during exercise, it could have been argued that it was due to the effect of hypercapnia on respiratory muscle function. Our results negate this hypothesis, as respiratory muscle strength, load, and recruitment pattern were similar in both groups. Indeed, our study shows that the effect of compensated hypercapnia in stable COPD on respiratory muscle function is minimal, if any, and was certainly not the cause of diminished ventilation in our hypercapnic patients. The mean Pa_CO2 with exercise increased minimally in the eucapnic patients, but this had no effect in respiratory muscle function, which did not differ from that of the hypercapnic patients.

This study does not pretend to minimize the possible role of muscle dysfunction and fatigue in the development of progressive hypercapnia resulting from acute decompensation of COPD or increased mechanical load. Neither does it deny the possible role of some protective mechanism that leads patients to stop exercise at levels of respiratory muscle load that is close but still below the fatigue threshold. On the other hand, the findings argue against the hypothesis of muscle weakness, dysfunction, or fatigue and adaptive compensation of the pump, as the most important causes of chronic hypercapnia in severe stable COPD. It is more likely that the decrease in ventilation is the result of the mechanical changes brought about by progressive airflow obstruction. Indeed, our data indicate that the increased CO₂ results from failure of the lungs to wash out the produced CO₂ at an increased level of physiologic ventilatory drive. This makes neuroventilatory coupling failure a more attractive explanation for CO₂ retention in these patients.

Our findings have important clinical implications in that those therapeutic maneuvers aimed at increasing drive to the muscles or increasing respiratory muscle contractility in hypercapnic patients with COPD are unlikely to be effective. On the other hand, therapeutic maneuvers may need to be directed at assisting the pump when it fails. This is probably heralded by sudden increases in CO₂ with uncompensated reductions in the pH.

In summary, we have shown that exercise capacity and ventilatory muscle function, at rest and during exercise, are similarly impaired in eucapnic and hypercapnic severe COPD. These findings demonstrate that compensated elevation of CO₂ is not an important cause of respiratory muscle weakness in stable patients with COPD. In addition, the findings argue against the hypothesis of adaptive compensation or respiratory muscle dysfunction as the most important mechanism in the genesis of chronic stable hypercapnia. Our data indicate that hypercapnia in stable COPD results from decreased ventilation despite similar ventilatory effort in these patients.