INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One of the pathophysiologic features of patients with chronic obstructive pulmonary disease (COPD) is the significant limitation of physical activity because of both increased ventilatory loads and reduced ventilatory reserves (1, 2). Patients with COPD usually develop a progressive dynamic hyperinflation because of expiratory airflow limitation during exercise (3). It has been increasingly recognized that dynamic hyperinflation is one of the most important factors leading to exercise limitation in these patients (4, 6).

Dynamic hyperinflation is characterized not only by a decreased capacity of the inspiratory muscles, which are forced to contract at a shorter operating length, but also by an increased demand on the inspiratory muscles, which have to overcome the intrinsic positive end-expiratory pressure (PEEPi) (7). Recently, in critically ill patients with COPD during resting breathing, total inspiratory pressure development has been partitioned into the portion required to overcome PEEPi and that to inflate the respiratory system (8). A similar analysis has never been performed in patients with COPD during exercise, nor has there been any study that describes how the diaphragm and RCMs contribute to overcome the PEEPi- imposed inspiratory threshold load and inflate the respiratory system in these patients. In the present study, we focused our attention on the inspiratory pressure contributed to the elastic load of the respiratory system, which significantly increased by PEEPi in patients with COPD during exercise. The aims of the present study were therefore twofold: in stable patients with COPD during exercise, (1) to partition the total inspiratory pressure to overcome the elastic load into the portion needed to overcome PEEPi and that available to inflate the respiratory system, and (2) to further assess the pressure contribution of RCMs relative to that of the diaphragm on the basis of this partition.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Sixteen patients (12 men and four women) participated in the study. All the patients gave informed consent to the experimental protocol, which was approved by the institutional research ethics committee. COPD was diagnosed in all the patients by a history of chronic bronchitis, clinical signs, and pulmonary function tests. The patient's pulmonary function data are given in Table 1. All the patients were in clinically stable condition without acute exacerbation of their respiratory symptoms for at least 1 mo prior to the experiment. Patients with evident cardiovascular disease and other systemic disease or conditions known to influence bicycle exercise were excluded.

Measurements

The experiments were conducted with the patients seated on an electronically braked bicycle ergometer (Ergo-metrics 800; Ergo-Line), breathing through a mouthpiece and wearing a noseclip. Respiratory flow was measured by a Fleisch No. 2 pneumotachograph (Fleisch, Lausanne, Switzerland). Two conventional balloon catheters were passed through the nose, with one positioned in the lower third of the esophagus and the other in the stomach to measure esophageal and gastric pressures (Pes and Pga). Transdiaphragmatic pressure (Pdi) was obtained as Pga  Pes. End-tidal carbon dioxide concentration was measured by a CO₂ analyzer (Ametek, Sunnyvale, CA), which sampled gas at the mouthpiece. During exercise, arterial oxygen saturation (Sa_O2) and the electrocardiogram were continuously monitored. Arterial blood pressure was measured at the end of each exercise work load.

Protocol

The patients were instructed to refrain from drinking coffee and using bronchodilators for at least 4 h prior to the exercise test. The patients breathed quietly for 5 min on the bicycle without pedaling to familiarize themselves with the equipment and the breathing circuit. The resting breathing data were recorded during the last 2 min of this period. The patients then commenced the exercise with a starting work load of 10 or 20 watts. The work load was progressively increased by 10 or 20 watts every minute until exhaustion. During resting breathing and at the end of each level of exercise, an inspiratory capacity (IC) maneuver was performed. The patients were instructed "to make a further maximal effort on top of a maximal inspiration" (4, 9), in order to measure the change in end-expiratory lung volume (EELV) during exercise caused by dynamic hyperinflation.

Data Analysis

The measured signals were preamplified, digitized, and recorded on a computer. The last five breaths at each work load were ensemble- averaged for further data analysis. VT, breathing frequency, minute ventilation (E), and respiratory timing were measured from the flow signal. During exercise, EELV was measured from the repeated IC maneuvers as the mirror image of IC (3, 10). Changes in end- inspiratory lung volume (EILV) were measured as the sum of EELV and VT.

To analyze the inspiratory pressure contributions, we first determined the Pes value at end-expiration during resting breathing. Because some stable patients with COPD had PEEPi caused by dynamic hyperinflation during resting breathing and the inspiratory flow was delayed relative to the initial inspiratory effort (11, 12), end-expiration throughout this study refers to the point at which Pes began to fall rather than to the start of inspiratory flow. The static pressure-volume curve of the chest wall (Pst,w) was drawn by passing it through the point of end-expiratory Pes during resting breathing. Assuming a normal slope of Pst,w curve for patients with COPD, the slope of Pst,w was obtained from the published data for healthy humans (13) taking the sex and age of the patient into consideration. To minimize the effect of expiratory muscle contraction on end-expiratory Pes (14), the value of the Pes at end-expiration during resting breathing was corrected by subtracting the expiratory rise in Pga whenever applicable, as suggested by Lessard and coworkers (15).

The measurement of the global inspiratory muscle pressure contribution (Pmus) is shown schematically in Figure 1. Point E is the end-expiratory Pes during resting breathing. The solid line passing point E is the predicted Pst,w curve. On the breathing loop, points A and B represent the Pes values at the beginning and end of inspiratory flow (zero flow point), respectively. The pressure differences given by AC and BD represent the Pmus for overcoming the PEEPi-imposed inspiratory threshold load and for the total inspiratory elastic load for a given EELV and EILV, respectively. The difference between BD and AC reflects the Pmus required to overcome the respiratory system elastance in order to inflate the respiratory system. The Pmus to overcome the inspiratory resistive load was measured as the peak inspiratory Pes relative to the value on a line connecting points A and B for a given volume, as given by FG. The pressure contribution of the diaphragm to overcome the PEEPi imposed inspiratory threshold load and to overcome the total inspiratory elastic load was measured as the Pdi value at the beginning and end of inspiratory flow (zero flow point), respectively, relative to Pdi baseline during resting breathing.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Schematic breathing loop during dynamic hyperinflation. Please see the METHODS (DATA ANALYSIS) for detailed explanation.

To quantify the pressure contribution of RCMs relative to that of the diaphragm, we calculated the Pdi/Pmus ratio. According to our current definition, Pmus, which reflects the inspiratory pressure contributed by both RCMs and diaphragm, was always negative, leading to a negative Pdi/Pmus ratio during breathing when Pdi > 0. Increasing Pdi/Pmus ratio (decreasing its negativity) requires Pdi to decrease for a given Pmus or Pmus to become more negative for a given Pdi, and therefore indicates an increase in the pressure contribution from RCMs relative to that from the diaphragm. When Pdi is zero, Pdi/Pmus ratio becomes maximal (zero). Under this condition, all the Pmus must be generated by RCMs.

The results were expressed as mean ± SE. The exercise data presented are from the first (E1), the median (E2), and the final (E3) work load. One-way repeated analysis of variance (ANOVA), paired t test, and Pearson's moment correlation were employed as appropriate to perform the statistical analyses. A p value less than 0.05 was considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

At exhaustion, the exercise work load accomplished was 36.2 ± 5.3% predicted. At peak exercise, compared with resting breathing, VT, f, and E were increased from 0.73 ± 0.04 to 1.04 ± 0.12 L, from 21.3 ± 1.6 to 29.8 ± 1.8 min¹, and from 15.17 ± 1.15 to 31.01 ± 2.56 L/min, respectively. The exercise reduced Sa_O2 from 93 ± 1 to 89 ± 2% and increased EELV by 0.27 ± 0.06 L or 9.87 ± 2.10% FVC.

The relationship between Pdi/Pmus ratio and Pga/ Pes ratio obtained during resting breathing for each patient is illustrated in Figure 2. The group mean Pdi/Pmus ratio was 0.88 ± 0.07. There was a significant linear correlation between these two ratios (r = 0.89, p < 0.0001).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Relationship between Pdi/Pmus and Pga/Pes during resting breathing. Each point represents the results from one patient.

It can be seen in Table 2 that to overcome the total elastic load, Pdi remained constant (p = 0.98) despite a progressive increase in the magnitude of Pmus (p < 0.001) with increasing exercise intensity. This resulted in a progressive decrease in negativity of the Pdi/Pmus ratio (p < 0.001) during exercise. The Pmus to overcome the inspiratory resistance also significantly increased during exercise (p < 0.001) (Table 2). As shown in Figure 3, the increase (decrease in negativity) in the total elastic Pdi/Pmus ratio at peak exercise were negatively correlated to the same ratio during resting breathing (r = 0.73, p < 0.001).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Correlation of the increase (decrease in negativity) in Pdi/Pmus at peak exercise and the same ratio during resting breathing. Each point represents the result from one patient. QB = quiet breathing; Ex = peak exercise.

In Figure 4, we partitioned the total elastic Pmus and Pdi into the fraction measured before the beginning of inspiratory flow (left panels), and the fraction measured between the beginning and end of inspiratory flow (right panels). Pmus before the beginning of inspiratory flow is equivalent to PEEPi measured dynamically without the influence from expiratory muscle activities (16). As shown in the left panels, during resting breathing, Pmus was detected before the beginning of inspiratory flow, suggesting the presence of PEEPi during resting breathing in our patients. During exercise, its amplitude progressively increased (p < 0.001), consistent with the increase in EELV, suggesting additional dynamic hyperinflation induced by exercise. During the same interval, Pdi also progressively increased (p < 0.001), suggesting a diaphragmatic pressure contribution to overcoming the PEEPi-imposed inspiratory threshold load.

View larger version (48K): [in this window] [in a new window]   Figure 4.   Pmus and Pdi before the beginning of inspiratory flow (left panels) and between the beginning and end of inspiratory flow (right panels). QB = quiet breathing; E1, E2, and E3 = the first, the median, and the final exercise work load. *Significantly different from QB.

As expected, Pmus between the beginning and end of inspiratory flow also progressively increased with increasing intensity of exercise (p < 0.001) (Figure 4, right upper panel). In contrast, Pdi significantly decreased at peak exercise (p = 0.026) (Figure 4, right lower panel), suggesting that the diaphragmatic pressure contribution to inflating the respiratory system decreased during exercise. In fact, in three patients, Pdi was less at the end than at the beginning of inspiratory flow, which made the Pdi available for inflating the respiratory system "negative" (Figure 5). Consequently, as shown in Figure 6, before the beginning of inspiratory flow, the negativity of Pdi/Pmus ratio progressively increased (p = 0.059) with increasing exercise work load (open bars), whereas between the beginning and end of inspiratory flow, the negativity of this ratio progressively decreased (p < 0.001) (hatched bars). At any given level of exercise as well as during resting breathing, the Pdi/Pmus ratio was always more negative before the beginning of inspiratory flow than that between the beginning and end of inspiratory flow (Figure 6).

View larger version (32K): [in this window] [in a new window]   Figure 5.   Experimental recordings from one patient who showed Pdi lower at the end of than at the beginning of inspiratory flow during exercise. The horizontal dotted line in the flow panels indicates zero flow. The two verticle dotted lines in each panel indicate the beginning and the end of inspiratory flow. Pes, Pga, and Pdi represent esophageal, gastric, and transdiaphragmatic pressures, respectively. Please note that at the work load of 30 and 50 watts, the Pdi was lower at the end of than at the beginning of inspiratory flow. This phenomenon did not exist at lower exercise work load and during resting breathing.

View larger version (24K): [in this window] [in a new window]   Figure 6.   Pdi/Pmus before the beginning of inspiratory flow (open bars) and between the beginning and the end of inspiratory flow (hatched bars). *Significantly different from QB; significantly different between open and hatched bars.

In Figure 7, Pmus and Pdi before the beginning of inspiratory flow were expressed as fractions of their respective values for the total elastic load. The fraction of the Pmus contributed to overcoming PEEPi was only slightly increased during exercise (p = 0.014) and never above 25%. In comparison, the same fraction for Pdi significantly increased during exercise (p < 0.001) and reached 60% at peak exercise. For a given work load, the fraction of the pressure contributed to overcoming PEEPi was always significantly greater for Pdi than for Pmus.

View larger version (17K): [in this window] [in a new window]   Figure 7.   Pdi (open circles) and Pmus (closed circles) before the beginning of inspiratory flow, each expressed as percent of the Pdi and Pmus developed for the total elastic load, respectively. *Significantly different from QB; significantly different between open and closed symbols.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In 16 stable patients with COPD during incremental exercise to exhaustion, we measured the inspiratory Pmus and Pdi for a given inspiratory effort, to reflect, respectively, the pressure contributions from the whole inspiratory muscles and the diaphragm alone. We further partitioned the total elastic inspiratory Pmus and Pdi a component used to overcome the PEEPi-imposed inspiratory threshold load and a component that accounts for the respiratory system inflation. We used the Pdi/Pmus ratio to estimate the relative pressure contributions of RCMs and the diaphragm for a given inspiratory task. It needs to be pointed out that the Pmus used to overcome the resistive load, which may account for an important part of the total Pmus generated in patients with COPD, also significantly increased during exercise (Table 2). However, our major interest was in the partition of the inspiratory pressure overcoming the elastic load, as this directly links to PEEPi. Hence, our following discussion will focus on the inspiratory pressures and their partition responsible for overcoming the elastic load of the respiratory system.

Technical Considerations

To measure Pmus during exercise in patients with COPD, Dodd and coworkers (3) used the Pst,w curve determined during voluntary relaxation as a reference. In the present study, we used a similar approach but did not ask our patients to perform relaxation maneuvers. It is usually very difficult for untrained subjects to truly and completely relax their respiratory muscles, so the Pst,w curve could not be reliably determined. The static chest wall compliance of patients with COPD has previously been shown to be comparable to that of healthy subjects (17). Hence, the slopes of the Pst,w curve we used were obtained from the published data of healthy subjects (13). This would induce random errors in our Pmus and therefore Pdi/Pmus ratio caused by the intersubject variation in Pst,w slopes. However, we were interested in changes in these parameters during exercise not their absolute values, and we showed that these changes were consistent not random. Hence, it is unlikely that such random errors, if present, would have significantly affected our major results and interpretations. We positioned the predicted Pst,w curve according to the end-expiratory Pes during resting breathing (Figure 1), a procedure used by many previous investigators (18, 20, 21). We assumed that with the presence of dynamic hyperinflation at rest, the fall in Pes at end-expiration indicated the beginning of inspiratory effort (11, 12). This is supported by a recent observation (15) in acutely ill patients with COPD during spontaneous breathing, which demonstrated that the rapid decrease in Pes and increase in Pdi at end-expiration were in phase and both were in phase with the start of the electromyogram of the diaphragm and accessary muscles. The value of the end-expiratory Pes during resting breathing was further carefully corrected for expiratory muscle activity when applicable using the method of Lessard and coworkers (15). We believe that with these procedures, contamination of Pmus by expiratory muscle activity was minimized.

Macklem and associates (22) proposed that the ratio between inspiratory transpulmonary pressure and abdominal pressure (PL /Pab) can be used to estimate the pattern of inspiratory muscle recruitment. As a modification of this method, the inspiratory Pga/Pes ratio has been frequently used to assess ventilatory muscle recruitment with a less negative value, indicating a proportionately greater contribution from RCMs (23- 27). The linear correlation between the Pdi/Pmus and this Pga/Pes (Figure 2) indicates that the change in the former reflects that in the latter during resting breathing. Two reasons prompted us to use the Pdi/Pmus ratio to partition the inspiratory pressures during exercise. First, Pmus measures the global inspiratory muscle pressure to overcome both lung and chest wall elastances for a given inspiratory effort at a given lung volume, so that it directly links to PEEPi. Second, inspiratory Pga and Pes were previously measured during exercise using either their own expiratory baselines (23) or expiratory baselines during resting breathing as reference (26). This may induce errors because, in patients with COPD, exercise causes elevation of EELV because of dynamic hyperinflation, and expiratory Pes and Pga baselines change because of both chest wall elastic recoil and expiratory muscle recruitment. Our measurement of Pmus takes the change in chest wall elastic recoil into consideration and minimizes the effect of expiratory muscle activity.

Differential Inspiratory Muscle Pressure Contributions to Breathing

We found a constant Pdi in spite of a progressive increase in the amplitude of Pmus to overcome the total elastic load during exercise in patients with COPD, leading to a progressive reduction in the negativity of the Pdi/Pmus ratio (Table 2). These results demonstrate that during exercise, patients with COPD, like healthy subjects, preferentially increase RCMs pressure contribution to meet ventilatory demands. This strategy is of particular importance in patients with COPD. In healthy subjects, EELV is usually reduced during exercise (28, 29). This improves the performance of the diaphragm as an inspiratory pressure generator. In patients with COPD, because of expiratory airflow limitation, EELV almost inevitably increases during exercise, as shown by previous (3, 10) and present results. Increasing volume seriously reduced Pdi amplitude (30), but it seemed to have less effect on RCMs pressure generation (31, 32). Thus, it may be expected that in patients with COPD, exercise ventilation may very much rely on RCMs function.

Although there is an exercise-induced increase in the pressure contribution to elastic load from RCMs relative to that from the diaphragm in both healthy subjects and patients with COPD, differences exist between these two populations. On the basis of our previous observations (33, 34) in healthy seated humans breathing quietly, inspiratory Pdi and Pes are about 10 and 4 cm H₂O, respectively. Assuming an equal lung and chest wall compliance, Pmus should be about 8 cm H₂O. This leads to a Pdi/Pmus ratio of about 1.25 in normal subjects, compared with the same ratio of 0.88 in our patients with COPD. These results, consistent with the observation of Martinez and coworkers (25), suggest that patients with COPD already have increased RCMs pressure contribution at rest to compensate for the diaphragmatic weakness resulting mainly from chronic hyperinflation (35). Therefore, the ventilatory reserves of these muscles are presumably low and their ability to further increase their contribution during exercise may be reduced. This is supported by the negative correlation between the magnitude of increase (decrease in negativity) in Pdi/Pmus ratio during exercise and this ratio during resting breathing in our patients (Figure 5). A greater increase in RCMs pressure contribution (more increase in Pdi/Pmus ratio) was seen during exercise in those patients with less RCMs recruitment (more negative Pdi/Pmus ratio) at rest.

Partitioning of Inspiratory Pressure during Dynamic Hyperinflation

Another important difference, which may alter the differential inspiratory muscle pressure contributions, is the dynamic hyperinflation exaggerated during exercise in patients with COPD. As a result, additional inspiratory pressure is needed to overcome the PEEPi-imposed inspiratory threshold load. In previous attempts (25, 27) to investigate ventilatory muscle recruitment patterns in patients with COPD, inspiratory pressures were measured between the beginning and the end of inspiratory flow, leaving the initial inspiratory pressure before the beginning of inspiratory flow for counterbalancing PEEPi (11, 12) completely ignored. In addition, there has been no attempt to partition the total elastic inspiratory pressure into a portion for overcoming PEEPi and a portion for inflating the respiratory system, nor has there been any attempt to examine the relative inspiratory pressure contribution from RCMs and the diaphragm on the basis of this partition in patients with COPD during exercise.

We found that the magnitude of Pmus before the beginning of inspiratory flow increased from 1.94 during resting breathing to 4.45 cm H₂O during peak exercise (Figure 4, left upper panel). This is in agreement with the previous notions that PEEPi may exist in stable patients with COPD during resting breathing as well (11, 12) and that the exercise- induced dynamic hyperinflation in patients with COPD may be modest (4). Although this Pmus only accounted for as much as 25% of the total elastic Pmus (Figure 7), it is a substantial load on breathing during exercise, especially when considering that this PEEPi-induced load is applied throughout inspiration. To overcome this load, Pdi also progressively increased during exercise representing diaphragmatic contribution (Figure 4, lower left panel). Consequently, the corresponding Pdi/Pmus ratio tended to decrease (increase in negativity) during exercise, suggesting a proportional or even a predominant diaphragmatic pressure contribution to the PEEPi-imposed inspiratory threshold load. As the diaphragm may be preferentially impaired as a pressure generator by increasing EELV, this pattern implies that the drive to the diaphragm must have disproportionately increased over that to RCMs.

However, our patients failed to maintain the same pattern of inspiratory pressure contribution throughout the entire inspiration. The Pdi between the beginning and end of inspiratory flow significantly reduced during exercise despite an increased Pmus (Figure 4, right panels). Under such conditions, increases in the RCMs pressure contribution are required to compensate not only for increased VT excursions but also for the reduced share by the diaphragm during flow. Furthermore, in three patients, the Pdi value was even lower at the end of than at the beginning of inspiratory flow during exercise (Figure 5). This requires the RCMs not only to completely take over the work of inflating the respiratory system, but also to generate, during the period of inflation, additional pressure in order to maintain the constant Pmus needed to overcome the PEEPi throughout the whole inspiratory effort. These results strongly suggest the RCMs as the predominant pressure generator once flow starts during exercise in patients with COPD. This conclusion is further supported by the fact that at any work load as well as during resting breathing, the Pdi/Pmus ratio was significantly higher (less negative) during flow than before the beginning of inspiratory flow (Figure 6).

In summary, based on our present results, we conclude that in patients with COPD during exercise to exhaustion, (1) there is a progressive increase in total inspiratory pressure contribution to elastic load from RCMs relative to that of the diaphragm, and the magnitude of this increase appears to depend on the RCMs reserves during resting breathing; (2) most of the diaphragmatic pressure contribution occurs before the beginning of inspiratory flow, to overcome the PEEPi-imposed inspiratory threshold load; (3) RCMs pressure contribution predominates during the period of inspiratory flow, not only for overcoming the elastic load caused by increased VT, but also for compensating for the diaphragmatic contribution that is gradually lost with increasing work load. As a result, the RCMs are heavily loaded. This will amplify the exertional breathlessness, as respiratory sensation is usually related specifically to RCM activity (36). The inspiratory muscle training programs for patients with COPD should therefore be directed towards also emphasizing the training of RCMs.