INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung volume reduction surgery involves removal of the most diseased lung tissue in patients with diffuse emphysema (1). Most patients respond by decreases in lung volume and dyspnea and improvement in exercise tolerance (1). It has been suggested that increases in lung elastic recoil (2), airway conductance and expiratory flow (4) are responsible for the reduction in lung volume. However, the precise mechanisms underlying the improvements in dyspnea and exercise tolerance remain largely unknown. Of note, the relationship between symptomatic improvement and change in pulmonary function appears to be rather weak (5).

Hyperinflation, a characteristic feature of emphysema, causes the respiratory muscles to operate at an unfavorable position on the length-tension curve (6), and also increases both the elastic and threshold mechanical load (7). Conceivably, improvement in respiratory muscle function may play a greater role in symptomatic improvements than do changes in pulmonary function. Indeed, studies in patients with chronic obstructive pulmonary disease (COPD) indicate that dyspnea is more closely related to respiratory muscle function than to airflow obstruction (8, 9).

Development of dyspnea with acute bronchoconstriction (10) and during whole body exercise (11) is associated with a considerable disparity between inspiratory effort and ventilatory output, so called neuromechanical uncoupling. Neuromechanical coupling takes into account the complex inter-relationships between respiratory drive, respiratory muscle performance, lung volume and lung mechanics (10). We hypothesized that lung volume reduction surgery produces improved diaphragmatic function with enhanced neuromechanical coupling of the diaphragm, and consequent reduction in dyspnea and improvement in exercise tolerance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Six men and one woman (age 62 ± 4 [SE] yr) with severe and fixed airflow limitation and roentgenographic evidence of emphysema were studied. The study was approved by the local human studies subcommittee and informed consent was obtained from each patient. Approximately 25% of each lung was resected via a median sternotomy (1). All patients were enrolled in a structured, supervised exercise rehabilitation program for a minimum of 6 wk before and 12 wk after surgery.

Evaluation of Pulmonary Function, Dyspnea and Exercise Performance

Esophageal (Pes) and gastric (Pga) pressures were separately measured with two thin-walled, latex balloon-tipped catheters (Erich Jaeger, Wurzberg, Germany) coupled to pressure transducers (MP-45; Validyne, Northridge, CA). A balloon containing 1.0 ml of air was positioned in the mid-esophagus (12); a gastric balloon containing 2.0 ml of air was advanced 70 cm from the nares. Transdiaphragmatic pressure (Pdi) was obtained by electronic subtraction of Pes from Pga. Airway pressure (Paw) was measured at the mouthpiece using a tap connected to a third transducer. Transpulmonary pressure was obtained by subtracting Pes from Paw.

Inspiratory flow () was measured with a heated Fleisch pneumotachograph (Model 3700; Hans Rudolph, Kansas City, MO) connected to a differential pressure transducer. Volumes were obtained by electronic integration of the signal. The duration of inspiration (TI) was calculated from the signal. Unless otherwise specified, the onset of inspiration was determined at the point at which inspiratory started, while the end of inspiration was determined as the point at which inspiratory ceased. Tidal volume (VT), mean inspiratory flow (VT/TI), and respiratory frequency were calculated as average values during one minute of recording. Intrinsic positive end-expiratory pressure (PEEPi) was measured during spontaneous breathing as the negative deflection in Pes between the onset of inspiratory effort (end- expiratory Pes) and the onset of inspiratory (13). Relaxation of the abdominal muscles at the onset of inspiration can contribute to the fall in Pes at the onset of inspiratory effort (14). Accordingly, any decrease in Pga at the onset of inspiratory effort was subtracted from the Pes signal (15). Dynamic lung compliance (Cdyn) was calculated as the ratio of change in volume (VT) over the change in transpulmonary pressure between instants of zero within the same breath (7), and the average value during one minute of recording was calculated. The pressure output of the respiratory muscles, quantified as pressure- time product (PTP), was calculated as the time integral of the difference between measured Pes and the estimated chest wall relaxation pressure (16). PTP per minute (PTP/min) was calculated as the product of PTP per breath and respiratory frequency (7).

Lung volumes were measured by plethysmography and timed spirometry (17). Arterial blood was sampled from the radial artery while patients were breathing room air at rest. Blood gas values were measured using an automatic blood gas analyzer (Instrumentation Laboratory 1620, Lexington, MA). The magnitude of dyspnea was quantified using a visual analog scale, on which patients marked their response to the question, "How uncomfortable is your breathing?" (18). Six-minute walking distance was performed according to the standard procedure (19).

Evaluation of Respiratory Muscle Function

Compound diaphragmatic motor action potentials (CDAPs) were recorded bilaterally with surface electromyographic (EMG) electrodes placed at the seventh and eighth intercostal space and the anterior axillary line. In six patients, EMG recording of the right sternomastoid activity was obtained with hook-wire electrodes placed midway between the angle of the jaw and the clavicle (20). All EMG signals were amplified, band pass filtered (band width 10 Hz to 1 kHz; Gould Inc., Valley View, OH), and displayed on a storage oscilloscope (Gould Inc., Ilford, UK).

Bilateral phrenic nerve stimulation was performed using a magnetic stimulator (Magstim 200; Magstim Co. Ltd, Dyfed, Wales) with a 90-mm coil (P/N 9784-00). This device stimulates neuromuscular structures by inducing electrical currents in the tissue secondary to a time-varying magnetic field of brief duration (< 1 ms total pulse duration); at maximal output, the magnetic field is 2.0 Tesla. To achieve stimulation of the phrenic nerves, the patient's neck was flexed and the coil was placed over the cervical spine. While the patient relaxed at functional residual capacity (FRC), the site of optimal stimulation was determined by moving the coil between C₅ and C₇. Stimulus maximality was then assessed by progressively increasing the intensity of the stimulus until Pdi_tw displayed no further increases (12). This position was marked with a felt pen and all subsequent stimulations were performed at this point. Patients were studied without abdominal binding, waist belts removed and trousers unbuttoned; alterations in abdominal compliance during stimulations were minimized by instructing the patients to relax their diaphragm and abdominal wall muscles during each stimulation. Relaxation of the lower rib cage muscles, diaphragm and expiratory abdominal muscles was confirmed by the absence of EMG activity. Respiratory inductive plethysmography (Non-Invasive Monitoring Systems, Miami Beach, FL) was used to detect the direction in which the rib cage and abdomen moved and to identify end-expiratory lung volume (21).

The contribution of the diaphragm to tidal breathing was assessed by calculating the Pga/Pdi ratio (22). Changes in pressures were measured as the maximal deflection during a tidal inspiration compared with the Pga and Pdi values recorded at the onset of inspiratory effort. The average value of the Pga/Pdi ratio during one minute of recording was calculated.

Maximal transdiaphragmatic pressure (Pdi_max) was measured during Mueller maneuvers against an occluded airway at FRC (23). Diaphragmatic neuromechanical coupling was quantified using a modification of the approach of Lougheed and coworkers (10), viz., the quotient of VT (normalized by TLC) to tidal change in Pdi (normalized by Pdi_max). We employ the term neuromechanical coupling in accordance with current usage in the literature (10, 11), although a direct measurement of neuronal activity, such as motor neuron firing frequency, is not included in the calculation. Net diaphragmatic neuromechanical coupling was calculated using the maximal tidal change in Pdi from the onset of inspiratory , i.e., Pdi,_net (10, 11); Pdi,_net represents the Pdi available to produce inspiratory airflow. Total diaphragmatic neuromechanical coupling was calculated using the maximal change in Pdi measured from the onset of inspiratory effort, i.e., Pdi,_total. The average values of net diaphragmatic neuromechanical coupling, i.e., (VT/TLC)/(Pdi,_net/Pdi_max) ratio, and total diaphragmatic neuromechanical coupling, i.e., (VT/TLC)/(Pdi,_total/Pdi_max) ratio, during one minute of recording were calculated.

Protocol

Data were recorded in each patient 1-2 wk before and 3 mo after surgery. Patients were studied in the sitting position with the back supported at a 90-degree angle. After placement of all transducers, Pdi_max was measured as the best of four to six Mueller maneuvers. Oscilloscope recordings of Pdi provided visual feedback, and at least 1 min of rest was provided between each maneuver. Following 15 min of rest, , Pes, Pga, sternomastoid EMG and RIP signals were recorded during 1 min of resting breathing. During this period, patients were instructed to remain silent, breathe quietly and not to cough or sigh, so as to avoid the induction of twitch potentiation (23). Then, Pdi_tw was measured using 10 magnetic stimulations while the nose and mouth were closed. Dyspnea score while patients were resting comfortably in a chair, lung volumes, and 6-min walking distance were also recorded.

Data Analysis

Data were recorded and digitized at 2,000 Hz using a 12-bit analog-to-digital converter (CODAS; DATAQ Instruments, Inc., Akron, OH) connected to a computer (EMPAC Int. Corp., Fremont, CA). Individual twitch responses were rejected from analysis according to previously described criteria (23). Paired t tests were used to compare variables before and after surgery. Regression analysis was employed to calculate the correlation coefficient between different variables.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary Function, Dyspnea and Exercise Performance

Surgery produced an improvement in lung function, as reflected by an increase in forced vital capacity (FVC) and decreases in FRC, residual volume (RV) and Pa_CO2 (Table 1). Dyspnea at rest was reduced (p = 0.04) (Figure 1) and the distance covered during 6-min of walking increased (p = 0.002) (Figure 1). The change in 6-min walking distance and the decrease in dyspnea were correlated with each other (r = 0.84, p = 0.03). Neither the change in 6-min walking distance nor resting dyspnea were correlated with the change in lung volume. PEEPi decreased from 2.8 ± 0.7 to 1.3 ± 0.8 cm H₂O after surgery (p = 0.02). The decrease in PEEPi correlated with the increase in FVC (r = 0.83, p = 0.047), but not with the improvement in 6-min walking distance or with the decrease in dyspnea.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Six-minute walking distance (left) and dyspnea score, measured by the visual analog scale (right), before and 3 mo after surgery in seven patients with severe emphysema. The 6-min walking distance increased from 808 ± 115 (SE) to 1,198 ± 99 feet (p = 0.002). Dyspnea decreased from 57 ± 11 to 22 ± 6 mm (p = 0.04). Significant difference between pre- and postoperative values: *p < 0.05, and ***p < 0.01.

Diaphragmatic Contractility

Pdi_max increased from 80.3 ± 9.5 to 110.8 ± 9.3 cm H₂O after surgery (p = 0.034) (Figure 2). The change in Pdi_max correlated with the decrease in FRC (r = 0.90, p = 0.02), but not with the improvements in dyspnea (r = 0.25, p = 0.64) nor 6-min walking distance (r = 0.20, p = 0.71). Pdi_tw increased from 17.2 ± 2.4 to 25.9 ± 3.0 cm H₂O after surgery (p = 0.02) (Figure 2). The improvement in Pdi_tw tended to correlate with 6-min walking distance (r = 0.65, p = 0.11), but not with the improvements in dyspnea and lung function.

View larger version (11K): [in this window] [in a new window]   Figure 2.   Voluntary maximal transdiaphragmatic pressure (Pdimax) increased from 80.3 ± 9.5 (SE) before to 110.8 ± 9.3 cm H2O after surgery (p = 0.03) (left). Transdiaphragmatic twitch pressure (Pditw) response to phrenic nerve stimulation (right) increased from 17.2 ± 2.4 to 25.9 ± 3.0 cm H2O (p = 0.02). Significant difference between pre- and postoperative values: *p < 0.05 and **p < 0.025. (Pdimax data are based on six patients because one was unable to perform a Mueller maneuver.)

Respiratory Muscle Recruitment and Diaphragmatic Neuromechanical Coupling

The Pga/Pdi ratio increased from 0.45 ± 0.06 to 0.61 ± 0.04 after surgery (p = 0.01) (Figure 3). The change in Pga/Pdi ratio was correlated with the increase in Pdi_tw (r = 0.76, p = 0.049) and the improvement in FVC (r = 0.81, p = 0.03). The change in this ratio also tended to correlate with the improvement in 6-min walking distance (r = 0.64, p = 0.12) but not with the improvement in dyspnea score (r = 0.45, p = 0.31).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Transdiaphragmatic pressure (Pdi, continuous line) and gastric pressure (Pga, interrupted line) tracings during resting breathing before (left) and 3 mo after (right) surgery in a representative patient. The area enclosed between the Pdi and Pga signals is determined, at least in part, by the activity of the intercostal-accessory muscles relative to the diaphragm. After surgery the area between the Pdi and Pga signals decreased, indicating a decrease in the contribution of the intercostal-accessory muscles to tidal breathing and an increase in the diaphragmatic contribution.

After surgery, both net and total diaphragmatic neuromechanical coupling increasedfrom 0.49 ± 0.07 to 0.82 ± 0.11 (p = 0.03) and from 0.42 ± 0.07 to 0.78 ± 0.12 (p = 0.025), respectively (Figures 4 and 5). The improvement in net diaphragmatic neuromechanical coupling was correlated with the improvement in 6-min walking distance (r = 0.86; p = 0.03) and tended to correlate with the improvement in dyspnea (r = 0.76, p = 0.08). The improvement in total diaphragmatic neuromechanical coupling tended to correlate with the improvements in 6-min walking distance (r = 0.76, p = 0.08) and dyspnea (r = 0.64, p = 0.18). The improvement in both net and total diaphragmatic neuromechanical coupling correlated with the decrease in rib-cage muscle recruitment (r = 0.91, p = 0.01 and r = 0.82, p = 0.04, respectively).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Tidal volume, as a percentage of total lung capacity (VT/TLC, %) (upper panels) and transdiaphragmatic pressure, as a percentage of maximal transdiaphragmatic pressure (Pdi/Pdimax, %) (lower panels), during resting breathing before (left) and 3 mo after (right) surgery in a representative patient. The increase in Pdimax (from 64 to 135 cm H2O, data not shown) allowed the patient to maintain a constant VT/ TLC ratio using a much smaller fraction of diaphragmatic pressure generation (see text for details).

View larger version (12K): [in this window] [in a new window]   Figure 5.   Total diaphragmatic neuromechanical coupling, quantified as the quotient of tidal volume (normalized to total lung capacity) to tidal change in transdiaphragmatic pressure (normalized to maximal transdiaphragmatic pressure), i.e., (VT/TLC)/(Pdi,total/ Pdimax), improved from 0.42 ± 0.07 (SE) to 0.78 ± 0.12 after surgery (p = 0.025). Significant difference between pre- and post- operative values: *p < 0.05. (Data are based on six patients because one was unable to perform a Mueller maneuver).

Respiratory Drive and Pressure Output of the Respiratory Muscles

Cdyn, calculated as the mean of Cdyn value per breath recorded over one minute of resting breathing, showed a tendency to increase from 393 ± 14 ml/cm H₂O to 502 ± 17 ml/cm H₂O after surgery (p = 0.09). One patient developed a decrease in Cdyn after surgery; when this outlier was omitted, the increase in Cdyn in the remaining six patients reached statistical significance (p = 0.04) (Figure 6). Mean inspiratory flow (VT/TI) showed a tendency to decrease from 472 ± 56 to 372 ± 49 ml/s after surgery (Figure 6). One patient developed an increase in VT/TI after surgery; when this outlier was omitted, the decrease in VT/TI in the remaining six patients reached statistical significance (p = 0.01). Neither the change in Cdyn nor the change in VT/TI were correlated with the changes in dyspnea, lung function data, and 6-min walking distance.

View larger version (11K): [in this window] [in a new window]   Figure 6.   Dynamic compliance (Cdyn) (left) increased from 393 ± 14 (SE) before to 502 ± 17 ml/cm H2O 3 mo after surgery; the change was significant in six patients (p = 0.04). Mean inspiratory flow (VT/TI), (right) decreased from 472 ± 56 to 372 ± 49 ml/s 3 mo after surgery; the change was significant in six patients (p = 0.01).

Surgery caused PTP/min to decrease from 341 ± 35 to 259 ± 35 cm H₂O² · s/min (p = 0.034) (Figure 7). The change in PTP/min was correlated with the decreases in TLC (r = 0.90, p = 0.006) and tended to correlate with PEEPi (r = 0.67, p = 0.098), and with the decrease in dyspnea (r = 0.71, p = 0.07). No correlation was found between the change in PTP/min and the improvements in 6-min walking distance and diaphragmatic contractility. A crude estimate of the cost of CO₂ removal, calculated as the product of PTP/min and Pa_CO2 (7), decreased from 14.4 ± 1.6 × 10³ to 9.2 ± 1.1 × 10³ cm H₂O² · s · min¹ after surgery (p = 0.02) (Figure 7). The change in the PTP/min · Pa_CO2 product tended to correlate with the decrease in VT/TI (r = 0.72, p = 0.07).

View larger version (14K): [in this window] [in a new window]   Figure 7.   Pressure time product per minute (PTP/min) during spontaneous respiration (left) and the inefficiency of CO2 removal, calculated as the PTP/min · PaCO2 product (right), before and 3 mo after surgery. PTP/min decreased from 341 ± 35 (SE) to 259 ± 35 cm H2O/min (p = 0.034). Inefficiency of CO2 removal (PTP/min · PaCO2) decreased from 14.4 ± 1.6 × 103 to 9.2 ± 1.1 × 103 cm H2O2 · s · min1 after surgery (p = 0.02). Significant difference between pre- and postoperative values: *p < 0.05, and **p < 0.025.

Sternomastoid Inspiratory Recruitment

Of the six patients in whom sternomastoid EMGs were recorded, phasic inspiratory activity was recorded in two patients before surgery and in two patients after surgery; phasic activity was absent in the remaining four patients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung volume reduction surgery resulted in several beneficial effects with increases in diaphragmatic neuromechanical coupling, exercise performance and efficiency in CO₂ removal, and decreases in resting respiratory drive and dyspnea.

Diaphragmatic Contractility

A major adverse effect of hyperinflation is that it causes the inspiratory muscles to operate at an unfavorable position of the length-tension curve, with consequent reduction in the force generation capacity (6). By effecting a decrease in operating volume, lung volume reduction surgery should enhance their ability to generate pressure. Because of the linear relationship between maximal transdiaphragmatic pressure and lung volume (over the range of inspiratory capacity) (6, 24, 25), it is possible to predict the Pdi_max value that should occur as a consequence of the reduction in lung volume after surgery. In particular, Braun and coworkers (6) have shown that Pdi_max increases by 2.1 cm H₂O for each percent decrease in TLC. Six of our patients developed a decrease in FRC of 660 ml, or 8.6% of TLC. Thus, the postoperative Pdi_max was expected to be 98.1 cm H₂O or 25% higher than the preoperative value. Instead, the measured postoperative Pdi_max was 110.8 cm H₂O or 45% higher than the preoperative value.

Factors that might contribute to this disproportionate increase in Pdi_max include a learning effect, a decrease in dyspnea, an improvement in inspiratory muscle perfusion, and inspiratory muscle reconditioning. Coaching in the performance of a Mueller maneuver is necessary to obtain accurate measurements of Pdi_max. However, the involved learning effect appears to be retained for less than one week (26), making this factor an unlikely explanation for the observed increase in Pdi_max. A decrease in dyspnea due to a reduction in respiratory muscle load could cause a patient to cooperate better in the performance of the maneuver by favoring greater diaphragmatic recruitment. Similarly, if lung volume reduction surgery improves right ventricular function (2), the resulting improvement in inspiratory muscle perfusion may enhance muscle strength (27). The opportunity for enhanced respiratory muscle reconditioning arises through the ability to undertake more exercise due to improvements in pulmonary mechanics, respiratory muscle capacity and cardiovascular function (28).

Pdi_tw increased from 17.2 to 25.9 cm H₂O after surgery (p = 0.02). To the best of our knowledge, measurements of Pdi_tw following lung volume reduction surgery have not been previously published. As is the case with Pdi_max, Pdi_tw is also influenced by lung volume (12, 29). In an earlier study (12), we demonstrated that Pdi_tw changes inversely with lung volume: 7.6 cm H₂O for each liter (12). Since FRC decreased by 0.64 L after surgery, the postoperative value of Pdi_tw was expected to be 22.1 cm H₂O or 33% higher than the preoperative value. Instead, the measured postoperative Pdi_tw was 25.9 cm H₂O or 58% higher than the preoperative value. Unlike the situation with Pdi_max, Pdi_tw elicited by stimulation of the phrenic nerves (12, 23) affords a means of quantifying changes in diaphragmatic contractility that is independent of the influences of learning effect and patient cooperation. Thus, the observed increase in Pdi_tw after surgery provides firm evidence of improved diaphragmatic contractility. Like the Pdi_max data, the improvement in Pdi_tw was out of proportion to the observed changes in lung volume. The tendency towards a positive correlation between the postoperative improvement in Pdi_tw and 6-min walking distance suggests that diaphragmatic contractility plays an important role in the symptomatic improvement after surgery.

Respiratory Muscle Recruitment

Surgery produced an increase in the Pga/Pdi ratio (Figure 3), a qualitative index of diaphragmatic recruitment during tidal breathing (22). An increase in the Pga/Pdi ratio can be due to increased diaphragmatic contribution to tidal breathing or, when expiratory muscles contract, to a decreased recruitment of the expiratory muscles during tidal breathing (30). The latter possibility cannot have been a primary mechanism of the increased Pga/Pdi ratio after surgery, since an early inspiratory fall in Pga was never observed after surgery, and it was noted occasionally in only one patient preoperatively (a maximum fall in Pga of 0.5 cm H₂O before the onset of inspiratory flow). Moreover, the inspiratory outward movement of the abdomen in all patients both before and after surgery argues against the importance of this factor. A more likely mechanism for the increased Pga/Pdi ratio following surgery is a decrease in the activity of the intercostal-accessory muscles relative to that of the diaphragm (30). This alteration in the pattern of inspiratory muscle recruitment is probably due to increased diaphragmatic contractility (Figure 2), as suggested by the correlation between the increase in Pga/Pdi ratio and the improvement in Pdi_tw. The reduction in rib-cage muscle recruitment after surgery may be, at least in part, responsible for the observed decrease in dyspnea (Figure 1). Previous investigators have observed increased dyspnea when the contribution of the rib-cage muscles to tidal breathing increases (31, 32). It has been suggested that the rib-cage muscles bear a greater responsibility than the diaphragm in triggering the sensation of dyspnea (31), because the diaphragm possesses relatively few muscle spindles (33). Surgery could have decreased the activity of several rib-cage muscles, but the sternomastoid does not appear to be among them. The fact that there was no change in sternomastoid activity after surgery is not particularly surprising considering that most patients with COPD do not recruit this muscle during resting breathing (20). The rib cage muscles that are most likely to have undergone derecruitment after surgery are the intercostals, the scalenes, and the trapezii.

Respiratory Drive and Pressure Output of the Respiratory Muscles

An improvement in diaphragmatic contractility is expected to reduce respiratory drive (34). Mean inspiratory flow rate (VT/ TI) has been used as a measure of resting respiratory drive in healthy subjects and in patients with COPD (35). A significant decrease in VT/TI was observed in six of the patients after surgery (Figure 6). Two mechanisms could be responsible for this finding: worsening of pulmonary mechanics (36) or a decrease in respiratory motor output (35). The first possibility is excluded by the observed improvement in pulmonary mechanics after surgery. This suggests that the decrease in mean inspiratory flow rate represents a decrease in respiratory drive after surgery.

Mechanical work of breathing has been reported to decrease after lung volume reduction surgery (37). However, mechanical work can substantially underestimate O₂ consumption by the respiratory muscles, which is better reflected by pressure time product (PTP) (38), albeit not perfectly (39). We observed a 22% decrease in PTP/min after surgery (Figure 7). The decrease in PTP/min was correlated with the decreases in hyperinflation (TLC) and, to a lesser extent, the decreases in threshold load (PEEPi) and dyspnea. Decreases in respiratory drive (VT/TI) and operating lung volume (40) and increases in the efficiency of CO₂ clearance and airway conductance during inspiration may all have contributed to the postoperative decrease in PTP/min.

A postoperative increase in elastic recoil (2, 4) and expiratory flow (2), probably account for the greater reduction in RV than in TLC (13% versus 4%, respectively). A decrease in FRC reduces the operating lung volume (40), which not only increases the capacity of the inspiratory muscles to generate pressure, but should also reduce the chest wall elastic load more than it increases lung elastic recoil (40). Furthermore, a decrease in the operating lung volume allows patients to breathe at a lower, more desirable portion of the lung pressure-volume curve (6). Indeed, all but one of our patients exhibited an increase in Cdyn (Figure 6).

Surgery produced a decrease in Pa_CO2 from 42 to 36 mm Hg (p = 0.02). The simultaneous reductions in Pa_CO2 and PTP/ min after surgery resulted in a significant decrease in the PTP/ min · Pa_CO2 product (7) (Figure 5), a crude estimate of the inefficiency of the ventilatory pump in clearing CO₂ (7). That is, surgery resulted in the more efficient clearance of CO₂ (Figure 5). This improvement may have resulted from decreased dead space ventilation, improved pulmonary blood flow (2) and/or improved pulmonary mechanics. Not surprisingly, the improvement in CO₂ clearance tended to correlate with the decrease in respiratory drive (VT/TI) (r = 0.72, p = 0.07).

Surgery produced an increase in the diaphragmatic contribution to tidal breathing (Pga/Pdi) (Figure 3) and enhanced diaphragmatic contractility (Pdi_max, Pdi_tw) (Figure 2). Consequently, the diaphragm operated at a smaller fraction of its maximal force-generating capacity (Pdi/Pdi_max). Moreover, the load imposed upon the diaphragm (PEEPi, Cdyn) was decreased. Therefore, the ventilatory output (VT, as a fraction of TLC) for a given degree of diaphragmatic contraction (Pdi, as a fraction of Pdi_max) increased after surgery (Figure 4). Improved neuromechanical coupling of the diaphragm can decrease dyspnea both at rest and during exercise through alterations in respiratory center output and/or afferent signals from peripheral sensory receptors (8, 41).

Dyspnea is presumed to result from the radiation of neuronal impulses from the brain-stem respiratory centers to the cortex, which results in conscious awareness of the respiratory motor command ("corollary discharge") (8, 11, 41). Afferent signals from peripheral receptors, providing proprioceptive information about muscle and chest wall displacement, tension development, and possibly changes in respired volume and flow, play an additional role (11, 41). The corollary discharges in combination with the proprioceptive signals probably provide information about the appropriateness of the ventilatory response for a given effort. Not surprisingly, improvements in dyspnea at rest and the 6-min walking distance were correlated with the improvement in net neuromechanical coupling of the diaphragmbased on changes in the Pdi available for inspiratory (10, 11)and, to a lesser extent, with total diaphragmatic neuromechanical coupling. This improvement in diaphragmatic neuromechanical coupling and, to a certain extent, the decreased rib-cage muscle contribution to tidal breathing may be key mechanisms explaining the symptomatic improvement achieved by lung volume reduction surgery. These results are in agreement with Lougheed and colleagues (10) and O'Donnell and colleagues (11), who have shown that net neuromechanical coupling of the respiratory muscles determines the degree of perceived dyspnea in patients with asthma (10) and COPD (11), and with Celli and colleagues (32), who have shown that dyspnea triggered by rib cage muscle recruitment may contribute to decreased exercise endurance.

In summary, lung volume reduction surgery effected an increase in diaphragmatic contractility that could not be fully explained by a decrease in operating lung volume; we suspect that muscle reconditioning was responsible for much of the remaining improvement. The enhancement in diaphragmatic contractility probably accounted for the de-recruitment of the rib cage muscles and the decrease in respiratory drive. Surgery also produced a decrease in the pressure output of the respiratory muscles, presumably due to decreased mechanical load and increased efficiency of CO₂ clearance. These alterations in diaphragmatic contractility, mechanical load and lung volume improved the neuromechanical coupling of the diaphragm, which, combined with de-recruitment of rib-cage muscles, may account for the decrease in dyspnea at rest and the increase in 6-min walking distance after surgery. In conclusion, symptomatic improvement following lung volume reduction surgery is due to improvement in respiratory muscle function, greater than can be accounted for by a decrease in operating lung volume, and enhanced neuromechanical coupling of the diaphragm.