INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung volume reduction surgery (LVRS), a surgical procedure that involves resection of 20 to 30% of the volume of emphysematous lungs, has attracted a great deal of attention since the original report by Cooper and colleagues in 1995 (1). In that report, 20 patients who underwent LVRS had significant improvements in lung function, gas exchange, and exercise capacity after surgery. There was no early or late mortality reported. Since then, several recent studies have shown similar but variable improvements in lung function (2), gas exchange (6), and exercise capacity (7, 8) after bilateral LVRS with stapling resection. Moreover, several longer term follow-up studies have now reported that the improvements in lung function, exercise capacity, and quality of life (QOL) are maintained for up to 2 to 3 yr after LVRS (9), whereas others have reported a small, gradual decline in lung function (10). Despite widespread enthusiasm about the benefits of LVRS in patients with end-stage emphysema, caution has been raised about the merits of LVRS in the absence of carefully designed prospective controlled clinical trials examining its effects over standard therapy (11). This healthy skepticism is fostered in part by the lack of unified inclusion or exclusion criteria, lack of complete testing in all treated patients, and the inability to independently determine the short-term effects of LVRS in contrast to intensive medical therapy and pulmonary rehabilitation.

To address these latter concerns, we conducted a prospective, randomized, controlled trial examining the effects of intensive medical therapy and pulmonary rehabilitation versus intensive medical therapy, pulmonary rehabilitation, and LVRS on functional status, gas exchange, symptom-limited maximal exercise performance, and 6-min walk distance. The goal of this study was to objectively evaluate the independent short-term physiologic impact of LVRS over maximal medical therapy, including intensive pulmonary rehabilitation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Selection

Patients with severe diffuse emphysema meeting criteria shown in Table 1 were enrolled into the study. Patients had stopped smoking for at least 6 mo and remained seriously symptomatic despite optimized medical therapy (bronchodilators, systemic corticosteroids, home oxygen therapy). All patients signed informed consent approved by our institution's human research committee.

Two hundred patients with severe nonbullous emphysema were screened for enrollment into the study. Thirty-seven patients met study criteria (18.5%) and were enrolled into the trial. Patients excluded were referred for lung transplantation (n = 59), had severe underlying psychological or emotional issues (n = 32), extreme deconditioning (n = 19), severe pulmonary artery hypertension or coronary artery disease (n = 22), and presence of pulmonary nodules on radiologic screening (n = 31). The trial was prospective, randomized, and controlled, but allowed crossover of patients from the medical to the surgical arm after they had completed evaluation after 3 additional months of medical therapy and pulmonary rehabilitation. Physiologic measurements of lung function and exercise capacity were performed at baseline, after 8 wk of outpatient pulmonary rehabilitation, and 3 mo after additional pulmonary rehabilitation or LVRS. Patients crossing over to LVRS from the medical arm were studied a fourth time, 3 mo after LVRS (Figure 1).

View larger version (24K): [in this window] [in a new window]   Figure 1.   Algorithm showing patient evaluation and treatment plans via the medical and surgical arms of the trial.

Eighteen patients were enrolled into the medical arm; 15 of the 18 patients completed 3 mo of additional pulmonary rehabilitation. A total of 32 patients underwent LVRS. These consisted of 19 patients who were originally randomized to undergo LVRS and 13 crossover patients from the medical arm.

Physiologic Measurements

Pulmonary function testing was performed (System 6200 Autobox DL Plethysmography; SensorMedics Corp., Yorba Linda, CA) following American Thoracic Society guidelines (12). Spirometry and lung volume determination (by both helium gas dilution and plethysmography) were performed. Trapped gas is reported as the difference between FRC measurements by plethysmography and helium dilution methods. Diffusion capacity was measured by single-breath technique (13).

Maximal voluntary ventilation was determined by instructing patients to breathe rapidly and deeply for 12 s while measuring expired volume. Arterial blood gases were performed after patients inspired the prescribed level of supplemental oxygen at rest for at least 20 min. Oxygenation is reported as a ratio of arterial oxygen tension to fraction of inspired oxygen (Pa_O2/FI_O2). All pulmonary function data are presented in absolute numbers and as percent of reference normal values.

Exercise Testing

All patients underwent incremental maximal treadmill exercise to symptom-limited maximum (Precor 9.4sp; Precor Inc., Bothell, WA). During the test, oxygen uptake (O₂), oxygen pulse, minute ventilation (E), tidal volume (VT), and respiratory frequency (f) were continuously measured by a metabolic cart (SensorMedics 2900). Oxygen saturation by pulse oximetry (N-200; Nellcor, Chula Vista, CA) and multiple lead electrocardiography (ECG Horizon; SensorMedics) were continuously recorded. Supplemental oxygen was individualized for each patient to prevent oxygen desaturation during exercise. The same level of prescribed oxygen was used in all subsequent exercise tests. An identical exercise protocol was used in all patients on serial testing to allow comparison of different metabolic parameters at similar workloads. Six-minute walk distance was measured in the conventional manner as previously described (14).

Surgical Technique

LVRS was performed via median sternotomy and bilateral stapling resection with the goal of removing 20 to 40% of the volume of each lung. High-resolution computed tomography of the chest and quantitative perfusion lung scanning were used preoperatively to identify the worst regions of emphysematous lung for targeted resection. These sites were augmented by visual inspection by the thoracic surgeon at the time of LVRS, as lung regions that were slow to deflate, or never totally deflated with the isolated bronchus opened to atmosphere were resected. At the conclusion of the operation, chest tubes were placed on waterseal, and suction avoided to minimize air leaks.

Pulmonary rehabilitation consisted of twenty-four, 2-h sessions over an 8-wk period in all patients, and then an additional 3 mo in patients randomized to the medical arm. Rehabilitation included education, physical and respiratory care instruction, psychosocial support; and supervised exercise training by an exercise physiologist. After baseline exercise tests, all subjects received an individualized exercise prescription based on prior symptom-limited maximal exercise test results. Patients used a motor-driven treadmill, performed arm cycling, and lifted arm and leg weights under supervision. The intensity of the program was increased on an individual basis.

QOL Assessment

The Sickness Impact Profile (SIP) scoring system is a sensitive, behaviorally based measure of sickness-related dysfunction (15). The SIP is composed of 136 items that reflect the patient's perception of his or her activities of daily living. SIP scores are inversely proportional to the level of function and QOL (i.e., a higher score identifies poorer function and more impaired QOL). The SIP evaluation was administered at baseline, following rehabilitation, and then 3 mo after pulmonary rehabilitation or LVRS.

Data Analysis

All data are presented as mean ± SD. Paired two-tailed t tests were used to compare baseline and 3 mo post-LVRS data or rehabilitation data. Data that were not normally distributed were subjected to a Mann-Whitney rank sum test and are reported as median values (QOL scores). A value of p < 0.05 was considered statistically significant. All statistical analysis were performed using a commercially available computer-based software program (Sigmastat, version 2.0; Jandel Co., San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Physiologic Data before and 8 wk after Pulmonary Rehabilitation

Baseline characteristics are shown in Table 2. There was no difference in any baseline demographic data between the medical or LVRS study arms. Thirty-six patients completed 8 wk of pulmonary rehabilitation. One patient developed a pelvic fracture and was unable to complete the rehabilitation program.

Table 3 shows lung function data before and after 8 wk of rehabilitation. After 8 wk of comprehensive pulmonary rehabilitation, values for spirometry, lung volumes, and diffusion capacity remained unchanged when compared with baseline data. Similarly, there was no significant difference in gas exchange parameters between baseline values and 8 of rehabilitation (Table 4). There was a trend to a higher 6-min walk distance after 8 wk of rehabilitation compared with baseline values (285 ± 96 versus 269 ± 91 m, p = 0.14).

Total exercise time on maximal exercise test was significantly longer after 8 wk of rehabilitation compared with baseline (7.4 ± 2.1 versus 5.8 ± 1.7 min, p < 0.001) (Table 4). However, peak oxygen consumption and breathing pattern during maximal exercise were essentially unchanged after 8 wk of rehabilitation compared with baseline study.

Sickness Impact Profile before and after 8 wk of Pulmonary Rehabilitation

Figure 2A shows the overall, physical, and psychological SIP scores at baseline values and after 8 wk of pulmonary rehabilitation in 29 patients. The overall (18.9 versus 22, p < 0.001), physical (11 versus 13, p < 0.01), and psychological (7 versus 11, p < 0.02) SIP scores decreased significantly after 8 wk of pulmonary rehabilitation in comparison to baseline values.

View larger version (11K): [in this window] [in a new window]   Figure 2.   (A) QOL scores before and after 8 wk of pulmonary rehabilitation (n = 29). (B) QOL at 8 wk and after 3 additional months of pulmonary rehabilitation (n = 15). *p < 0.05 compared with baseline values.

Physiologic Data after 8 wk and 3 Additional Months of Pulmonary Rehabilitation

Tables 5 and 6 show the effects of maximal medical therapy and 3 additional months of pulmonary rehabilitation on lung function, gas exchange, and maximal exercise testing. Fifteen patients from the medical arm completed 3 mo of additional pulmonary rehabilitation. Spirometry, lung volumes, diffusion capacity, and gas exchange parameters were unchanged after an additional 3 mo of pulmonary rehabilitation in comparison to 8 wk of therapy (Table 5). A modest improvement in 6-min walk distance was observed at 3 mo postrehabilitation compared with 8-wk data but did not reach statistical significance (p = 0.11) (Table 6). No further improvement in total exercise time was noted after 3 additional months of pulmonary rehabilitation (Table 6). The O₂ peak trended to be higher at 3 mo postrehabilitation when compared with 8 wk postrehabilitation (13.3 ± 3.0 versus 12.6 ± 3.3 ml/kg/min, p < 0.08). Breathing pattern during peak exercise was unchanged at 3 mo compared with 8 wk after pulmonary rehabilitation.

Sickness Impact Profile at 8 wk and after 3 mo of Additional Pulmonary Rehabilitation

Figure 2B shows the overall, physical, and psychological SIP scores of 15 patients in the medical arm after 8 wk and 3 additional months of pulmonary rehabilitation. No additional improvement in all categories measured after 3 additional months of pulmonary rehabilitation was noted.

Physiologic Data 8 wk after Pulmonary Rehabilitation and 3 mo Post-LVRS (Excluding Crossover Patients)

Tables 7 and 8 show complete lung function, gas exchange, and exercise data before and 3 mo after LVRS in 15 of 19 patients randomized to that arm (three died postsurgery, one had incomplete 8-wk postrehabilitation data). At 3 mo after LVRS, there were significant improvements in lung function as measured by spirometry, lung volumes, and diffusion capacity studies (Table 7). At 3 mo post-LVRS, the Pa_O2/FI_O2 remained unchanged, but Pa_CO2 was significantly lower compared with 8 wk postrehabilitation (43 ± 7 versus 47 ± 8 mm Hg, p < 0.01) (Table 8). There was a trend toward a longer 6-min walk distance at 3 mo post-LVRS compared with 8 wk of pulmonary rehabilitation. On symptom-limited exercise test, both total exercise time (9.1 ± 1.9 versus 8.1 ± 2 min, p = 0.25) and O₂ peak were higher after LVRS but were not statistically significant (Table 8). There was no significant change in oxygen pulse at peak exercise (O₂ pulse_max). Breathing pattern changed significantly after LVRS as shown by significant increases in both maximal E (30 ± 10 versus 24 ± 5.7 L/min, p < 0.02) and VT (1.1 ± 0.3 versus 0.91 ± 0.21 L, p < 0.01) achieved during peak exercise after LVRS when compared with postrehabilitation values.

Sickness Impact Profile after 8 wk of Pulmonary Rehabilitation and 3 mo after Post-LVRS (Excluding Crossover Patients)

Figure 3A shows the overall, physical, and psychological SIP scores after 8 wk of pulmonary rehabilitation and 3 mo after LVRS in 11 patients. In contrast to 3 additional months of pulmonary rehabilitation, there was a further improvement in overall (13 ± 8 versus 16.5 ± 11, p < 0.008), physical (5 ± 8 versus 13 ± 11, p < 0.01), and psychological (0 versus 7, p < 0.008) SIP scores 3 mo post-LVRS in contrast to SIP scores after 8 wk of pulmonary rehabilitation.

View larger version (11K): [in this window] [in a new window]   Figure 3.   (A) QOL scores 8 wk postrehabilitation and after 3 mo post-LVRS (n = 11), excluding crossover patients. Panel B includes crossover patients (n = 21). *p < 0.01 compared with 8 wk postrehabilitation.

Physiologic Data 8 wk after Pulmonary Rehabilitation and 3 mo Post-LVRS (Including Patients Who Crossed over from the Medical Arm)

Tables 9 and 10 show complete lung function, gas exchange, and exercise data before and 3 mo after LVRS in 28 patients. The magnitude of improvement in spirometry and lung volume observed in this larger group of patients is similar to the surgical arm as previously described. A decrease in Pa_CO2 at 3 mo post-LVRS was observed as compared with pre-LVRS values (42 ± 6 versus 45 ± 7 mm Hg, p < 0.01) whereas Pa_O2/FI_O2 remained unchanged. In contrast to our findings in Table 8, the improvement in 6-min walk distance after LVRS becomes highly significant when 13 patients who crossed over from the medical to the surgical arm were included in the statistical analysis (337 ± 99 versus 282 ± 100 m, p < 0.002) (Table 10).

During symptom-limited maximal exercise testing, the improvement in O₂max was significant when 23 of the 28 patients who underwent LVRS were analyzed (13.8 ± 4 versus 12.0 ± 3 ml/kg/min, p < 0.01). Additionally, there was a trend toward a longer total exercise time after LVRS (8.6 ± 1.8 versus 7.8 ± 2 min, p = 0.11) and greater O₂ pulse_max (6.8 ± 2 versus 6.1 ± 2, p = 0.15). Breathing pattern had higher VT at peak exercise in all patients after LVRS when compared with postrehabilitation values.

Sickness Impact Profile before and after 3 mo Post-LVRS (Including Patients who Crossed over from the Medical Arm)

Figure 3B shows the overall, physical, and psychological SIP scores of all patients who underwent LVRS after 8 wk postrehabilitation and 3 mo after LVRS. The overall (10 versus 21, p < 0.001), physical (3 versus 11, p < 0.001), and psychological (3 versus 8, p < 0.001) SIP scores were significantly lower 3 mo after LVRS, suggesting further substantial improvement in QOL compared with pulmonary rehabilitation alone.

Morbidity and Mortality before and after LVRS

Table 11 shows morbidity and mortality before and after LVRS. Before surgery, seven of 37 (18.9%) patients required hospitalization for exacerbations of chronic airflow obstruction, whereas two of 32 (6.2%) required hospitalization in the 3 mo after LVRS. Two of the seven patients who were hospitalized before LVRS for respiratory failure required intubation and mechanical ventilation; two of the three patients who died postoperatively required prolonged mechanical ventilation. One patient before LVRS had three separate emergency room visits; two patients post-LVRS each had one emergency room visit. Two patients had an episode of pneumonia before surgery and one patient in the LVRS group developed pneumonia during the follow-up period. All patients who required systemic steroid therapy before surgery were totally weaned off the drug in the months after LVRS. In contrast, seven of 37 patients (19%) required increasing doses of steroids in the medical arm, or prior to LVRS in the surgical arm.

The overall mortality rate (early and late) for the surgical procedure was 9.4% (3 of 32). Table 12 shows the characteristics of the three patients who died postoperatively. One patient randomized to the LVRS arm died preoperatively during the period while receiving rehabilitation and medical therapy secondary to pneumonia and respiratory failure (mortality 3%, 1 of 37).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data show that in patients with severe, nonbullous emphysema, intensive pulmonary rehabilitation has no effect on spirometry, lung volumes, or gas exchange. However, total exercise time during maximal exercise performance was enhanced, and 6-min walk test tended to increase. After 8 wk of pulmonary rehabilitation, and an additional 3 mo of maintenance pulmonary rehabilitation, no improvement was observed in the patient's ability to perform 6-min walk and/or maximal exercise tests. A significant improvement in QOL was observed after 8 wk of pulmonary rehabilitation. There were no further perceived improvements in the patient's QOL after 3 additional months of pulmonary rehabilitation. In contrast, however, bilateral LVRS in addition to medical therapy and intensive pulmonary rehabilitation, resulted in substantial increases in FVC and FEV₁, reductions in static lung volumes, and a reduction in resting Pa_CO2. Moreover, after LVRS, breathing pattern during maximal exercise was altered, such that a higher E_max was achieved owing to greater VT generation. In contrast to 3 mo of additional pulmonary rehabilitation, there was a further improvement in QOL 3 mo after LVRS superimposed on the initial improvement seen after 8 wk of pulmonary rehabilitation.

Effects of Pulmonary Rehabilitation in Severe COPD

Functional capacity. Several studies have shown that comprehensive exercise-based pulmonary rehabilitation improves 6-min walk test (16, 17), increases O₂max (18, 19), decreases dyspnea (20, 21), improves QOL, and may even decrease the number of hospitalizations. However, tests of static lung function remain unchanged after comprehensive pulmonary rehabilitation (22, 23). Most of the aforementioned studies, however, were performed in patients with mild to moderate COPD; the effectiveness of pulmonary rehabilitation in patients with severe COPD as reported in our study is not clear (24, 25).

In a recent prospective randomized controlled trial comparing 8 wk of an inpatient multidisciplinary rehabilitation program plus 16 wk of a supervised outpatient program in 89 patients with severe COPD (FEV₁ < 1 L), Goldstein and coworkers (26) reported mean increases in 6-min walking distance, submaximal leg cycle time, dyspnea index, and emotional function after pulmonary rehabilitation. Similarly, Lake and coworkers (16) reported an increase in 6-min walk distance in 28 patients with severe COPD (FEV₁ 31% of predicted) after 8 wk of a supervised outpatient walking exercise program. Weiner and coworkers (18) showed similar improvement in the 12-min walk in 12 COPD patients (FEV₁ 32% of predicted) after 6 mo of an outpatient cycle exercise program. In a recent randomized controlled trial of pulmonary rehabilitation in severe COPD (FEV₁ < 40% predicted), Wedzicha and coworkers (27) showed that patients who had severe dyspnea scores showed no improvement in functional capacity after a supervised pulmonary rehabilitation program in contrast to patients with moderate dyspnea scores. This study implies that COPD patients with severe dyspnea during initial evaluation may not benefit from a supervised pulmonary rehabilitation program because they are too physiologically disabled to realize any substantial physiological improvement.

In our study, all patients had severe COPD (FEV₁ < 30% of predicted). We were unable to show a significant improvement in 6-min walk test after 8 wk of pulmonary rehabilitation; only the improvement in total exercise duration on incremental exercise test was statistically significant. The differences in our results compared with the other studies may be, in part, explained by the following: (1) many of our patients were already actively undergoing pulmonary rehabilitation at the time they were referred to our institution and thus may not have been as deconditioned as those reported in previous trials; (2) our patients were more severely obstructed compared with prior studies; (3) there were variations between studies in the methods, duration, and intensity of exercise training; (4) results of timed walking test are greatly influenced by both effort and practice.

Cardiopulmonary Exercise Testing

Similar to the effect of pulmonary rehabilitation on functional capacity in severe COPD, the effect of rehabilitation on O₂max during maximal exercise testing in patients with COPD is also unclear. Whereas the majority of investigators report no significant increase in O₂max after pulmonary rehabilitation, others report a modest increase in O₂ peak or a lower O₂ for a given work rate, suggesting an improved exercise efficiency (28).

A recent randomized controlled study comparing comprehensive pulmonary rehabilitation with education alone reported a significant increase in O₂max of 0.11 L/min (8.8%) in the group that completed pulmonary rehabilitation compared with 0.03 L/min in the group that completed education alone (18). This particular study, apart from different study design, is not comparable to ours because their patients had a baseline FEV₁ value of 1.21 L, significantly higher than our patient population. Other studies found a modest increase in exercise duration (17, 29) or a decrease in blood lactate level during incremental exercise suggesting an improved aerobic capacity of the exercising muscles (30). However, as previously discussed, most studies showing an improvement in O₂max include patients with only a moderate degree of airflow obstruction.

Reardon and coworkers (29) randomized 10 patients with COPD (FEV₁ 0.94 L) to receive 6 wk of outpatient pulmonary rehabilitation and 10 patients (FEV₁ 0.79 L) served as control subjects. In patients who received outpatient pulmonary rehabilitation, O₂max increased 10% above baseline but was not statistically significant. In addition, the endurance time on the treadmill was significantly higher compared with control values, suggesting that patients tolerated higher workloads.

Our results on incremental exercise testing mirror those of Reardon and coworkers (29). Three months of additional rehabilitation maintained the modest improvement in exercise duration that was attained after 8 wk of supervised pulmonary rehabilitation. This improvement in exercise duration may be due in part to increased tolerance to dyspnea during exercise or enhanced patient motivation. Alternatively, a true physiologic training effect is possible, but unlikely in the absence of any improvement in the other exercise parameters.

Effect of LVRS on Gas Exchange

The reported effects of LVRS on gas exchange have been variable. Whereas some investigators showed a significant increase in Pa_O2 after LVRS (2, 4), others were unable to show a significant change (6, 9). In the initial study of Cooper and colleagues, Pa_O2 increased from 64 to 70 mm Hg at 6 mo post-LVRS, but no significant change in Pa_CO2 was observed (1). In contrast, Daniel and associates (31) were unable to observe significant differences in Pa_O2 and Pa_CO2 after surgery. More recent reports suggest that LVRS may have limited effects on gas exchange (6). Similarly we were unable to find a significant increase in Pa_O2 after LVRS. However, we found a small but significant decrease in Pa_CO2 3 mo after LVRS. The reason for the differences in the reported response of Pa_O2 after LVRS between our data and prior reports remains unclear. It may be possible that we were unable to detect small changes in Pa_O2 in our study because of the possible small errors introduced by calculating the Pa_O2/FI_O2 ratio.

6-min Walk Distance and Cardiopulmonary Exercise

Our study confirms prior reports of significant improvements in exercise capacity after LVRS as measured by 6-min walk test or by comprehensive cardiopulmonary exercise testing. Three months after LVRS, both O₂max and E_max increased by 16% and 29%, respectively. Ferguson and associates (8) showed that changes in exercise performance were significantly related to changes in E_max (r² = 0.634, p < 0.01) and VT_max (r² = 0.513, p < 0.01). Thus, the increase in O₂max after LVRS appears to be due to in part to the increase in E_max that the patients could achieve during exercise.

After LVRS, the mean VT at peak exercise increased by 190 ml with no change in respiratory rate, suggesting a more efficient breathing pattern. A similar improvement in breathing pattern at isowatt exercise after LVRS was reported by Benditt and coworkers (7). The improved breathing pattern during exercise could be explained in part by a reduction in lung volume and a shift in ventilatory muscle recruitment pattern toward increased diaphragm action during exercise after LVRS (32).

QOL after Pulmonary Rehabilitation and LVRS

Several uncontrolled and controlled studies have shown significant improvements in QOL after pulmonary rehabilitation. Using a disease-specific questionnaire (chronic respiratory disease questionnaire), Goldstein and coworkers (26) demonstrated a significant improvement in QOL after pulmonary rehabilitation in patients with severe COPD compared with patients who received only standard medical therapy. Using the same disease-specific questionnaire but in a home setting and with shorter follow-up duration, Wijkstra and coworkers (19) found a highly significant improvement in the rehabilitation group in the dimensions of dyspnea, emotion, and mastery of breathing compared with a control group. In contrast, Ries and coworkers (18) did not show an improvement in the QOL measures after pulmonary rehabilitation despite a significant improvement in exercise performance.

Our study confirmed earlier reports of significant perceived improvements in the QOL in COPD patients after 8 wk of pulmonary rehabilitation, despite the fact that no significant improvement in exercise performance was noted. Furthermore, in our study, the improvement in QOL was maintained up to 3 mo with continued pulmonary rehabilitation.

Hospitalization and Survival

In most series, the 30- to 90-d mortality rate after LVRS ranges between 4 to 10% (2, 5, 33). In our study, one patient died of respiratory failure owing to a severe exacerbation of airflow obstruction while undergoing medical therapy. Exacerbations of airflow obstruction requiring increased use of bronchodilators, systemic steroids, and hospitalization were observed in the majority of our patients before LVRS. Post-LVRS, the hospitalization rate was lower, no patient developed respiratory failure after discharge from the hospital, and the need for systemic steroids was reduced. A longer follow-up period is needed to ascertain fully the risks and benefits of LVRS.

Benefits and Limitations of the Study

Our study confirms the short-term improvements in lung function, gas exchange, exercise capacity, QOL, and dyspnea in patients with severe COPD after LVRS. In addition, this is the first controlled, randomized study with comprehensive data collection comparing pulmonary rehabilitation with LVRS. Moreover, all our patients received standardized medical care and were operated on by the same surgeon, eliminating variability in management style and surgical technique that may potentially affect outcome.

Our study is limited by the small number of patients and short-term follow-up. The crossover design of the study may potentially bias the results toward LVRS because patients on the rehabilitation arm may not be as motivated as patients on the LVRS arm given the often dramatic account of LVRS in the lay press. The single-site nature of the study is an additional limitation.

In conclusion, bilateral LVRS, in addition to intensive medical therapy and pulmonary rehabilitation, improves static lung function, gas exchange, QOL, and breathing pattern during maximal exercise. In addition, we showed the limited role of pulmonary rehabilitation in improving aerobic exercise performance in patients with severe COPD. LVRS, however, is not without risk for increased perioperative morbidity and mortality compared with standard medical therapy. Further delineation of the risks, benefit, effects on survival, and durability of LVRS over time will hopefully be elucidated by the results of the National Emphysema Treatment Trial (NETT), the large, multicenter, long-term study of LVRS compared with maximal medical therapy being conducted by the National Institute of Health and Health Care Financing Administration.