INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic obstructive pulmonary disease (COPD) affects at least 15 million Americans, is the fifth leading cause of death in the United States and mortality continues to rise (1). In spite of progress in the medical management of COPD (2, 3), an inability to forestall the progression of disease has led to a resurgence in surgical attempts to treat emphysema. Lung volume reduction surgery (LVRS) (4) has been resurrected as a means of treating patients with severe emphysema (5). Several surgical modifications to LVRS have been documented to provide significant improvements in surgical morbidity and mortality and improve pulmonary function (6). Until recently, little has been reported about the impact of LVRS on exercise performance, beyond simple measures of walk distance (10). Two recent reports demonstrate that exercise performance assessed by maximal exercise testing does improve shortly after LVRS (15, 16). The purpose of this study was to further define the effect of LVRS on exercise performance, to evaluate the physiologic mechanisms responsible for improvements in exercise performance after LVRS, to assess the relationship of changes in patient function and quality of life following LVRS to exercise performance, and to determine if preoperative pulmonary function and exercise testing can predict outcomes, including postoperative exercise improvements and survival.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Selection

Bilateral LVRS via a median sternotomy approach was performed on patients referred for the procedure who met the following inclusion criteria:

1. Clinical diagnosis of COPD.
2. Age < 80 yr.
3. Evidence of severe emphysema based on:

(a) Postbronchodilator FEV₁ between 15% and 35% of predicted

(b) FEV₁/FVC ratio  60%

(c) Minimal reversibility with bronchodilators (FEV₁ improvement  20% after ₂ agonist inhalation)

(d) diffusing capacity of the lungs for carbon monoxide (DL_CO) < 50% of predicted

(e) TLC (by plethysmography) > 120% of predicted

(f) Pa_O2 < 55 mm Hg at rest on room air or the need for supplemental oxygen therapy with activity (at Denver's altitude, blood gas pressure [PB] = 630 mm Hg).

1. Upper lobe predominance of emphysema on computed tomographic (CT) scan.
2. Limitation in function and/or quality of life while receiving optimal medical care.
3. Stable clinical condition for at least 1 mo.
4. No smoking for at least 3 mo before the baseline evaluation and agreement to continue to refrain from smoking.
5. Completion of 3 wk of a formal outpatient rehabilitation program and subsequent continuation of the rehabilitation at home on an unsupervised basis.

Patients were excluded from surgery if they had (1) evidence of pulmonary hypertension (mean pulmonary artery pressure > 40 mm Hg), (2) chronic respiratory failure (Pa_CO2  55 mm Hg at rest), (3) other respiratory diseases, (4) cardiac disease with evidence of congestive heart failure or ischemia, (5) a major disease in another organ system, (6) a prior history of cardiothoracic surgery or any other surgical contraindication, (7) a labile medication schedule (including the regular use of greater than 20 mg of prednisone/day), (8) an expected survival of  3 mo or (9) would not agree to return for clinical follow-up.

Exercise, pulmonary function, and functional/quality of life testing was performed after patients had completed 3 wk of monitored outpatient rehabilitation. Although instructed and encouraged to continue with rehabilitation postoperatively, no formal participation or monitoring of adherence was undertaken. All subjects were retested 3 to 4 mo after discharge from the hospital as a part of clinical follow-up.

Spirometry/Lung Volumes

FVC and maximal expiratory flow rates, including FEV₁, were measured using a Krogh spirometer attached to a plethysmograph. Thoracic gas volume (TGV), total lung capacity (TLC), and residual volume (RV) were measured in a standard fashion using a volume displacement, pressure-compensated body plethysmograph (17). Testing was performed before and after inhaling a short-acting ₂ agonist via nebulizer, and all values were compared with standard reference values (18).

Diffusing Capacity

Diffusing capacity was determined by the single breath method (19) and corrected for hemoglobin. Results were related to alveolar volume (VA) and expressed in absolute terms and as a percent predicted (20).

Respiratory Muscle Strength

Global inspiratory and expiratory muscle strength was determined by measuring pressures at the mouth during airway occlusion while inspiring maximally from RV (PI_max) and expiring maximally from TLC (PE_max) using standard methods (21).

Trapped Thoracic Gas

Trapped thoracic gas was considered to be the volume of gas in the lungs that has no significant communication to the airways and does not contribute to gas exchange during ventilation. The total volume of gas in the thorax at end-exhalation was measured by plethysmography (TGV) and the volume of gas at end-exhalation which communicates to the airways and contributes to gas exchange was measured by the helium dilution technique (FRC). Trapped thoracic gas was defined as the difference between TGV and FRC.

Pressure-Volume Curve

The pressure-volume relationship of the lungs was determined from transpulmonary pressures and lung volumes measured during transient occlusion of a mouthpiece at several lung volumes during slow exhalation from TLC to functional RV (22). Lung volumes were determined by measuring TGV followed by inspiration to TLC and subsequent exhalation while continuously assessing airflows connected at the mouthpiece and deriving volumes from the integrated airflow signals. Simultaneously measured transpulmonary pressures were determined using an esophageal balloon catheter referenced to atmosphere and placed in the distal esophagus using standard positioning techniques. A pressure/volume curve was subsequently derived from pressure and lung volume data points obtained during three to five expiratory maneuvers. Maximal transpulmonary pressure (Pstat_max) was determined at peak inspiration / TLC, the coefficient of retraction was derived from Pstat_max/TLC, and a computer-generated curve fitting of the pressure/volume measurements allowed for calculation of the curvature (K value of the pressure/volume curve).

Exercise Testing

Maximal incremental and steady-state cardiopulmonary exercise testing was performed using an upright electrically braked cycle ergometer (Ergometrics 900, Yorba Linda, CA), a metabolic cart (Sensor Medics 2900, Yorba Linda, CA), a 12-lead electrocardiogram (Sensor Medics Max-1), and arterial blood gases drawn from an indwelling arterial catheter. Maximal exercise performance was measured using a 1-min incremental work protocol, with uniform increases in work every 1 min until the patient was unable to continue (5- to 25-watt increments depending on the patient's ability with a goal of exercising for a total of 6 to 8 min). A fraction of inspired oxygen (FI_O2) sufficient to maintain adequate oxygenation throughout exercise was estimated from walking oxygen saturations and provided throughout baseline, exercise, and recovery periods. FI_O2s higher than room air were provided via a high-flow flowmeter connected to an oxygen blender calibrated to provide an FI_O2 of ± 0.1% which filled a 120-L Douglas bag connected to a two-way nonrebreather valve. Flow to the reservoir bag was adjusted to match patient ventilation. The closed system provided a stable FI_O2 which was directly measured at the mouthpiece on a breath-by-breath basis using a phased oxygen sensor calibrated to 0.1% O₂ using known calibration gases at the FI_O2s being studied. FI_O2 was continuously monitored to ensure uniform inhaled gas delivery and utilized for calculation of oxygen consumption O₂. Nose clips and mouth seal were closely monitored throughout testing to ensure no accidental air inspiration or airflow loss during testing. Workload, metabolics, ventilatory function, cardiovascular function, and gas exchange were measured at baseline, at the end of each minute of testing, at maximum exercise, and during recovery (23). After a rest period (30 to 60 min), the patient was retested at a steady-state work load equaling one-half of the maximum workload achieved for a total of 6 min, with exercise parameters again assessed at baseline, at the end of each minute of steady-state exercise, and during recovery. Postoperative maximal exercise testing was performed at workload increments and an FI_O2 identical to the individual's preoperative increments, but with the freedom to perform to maximum exercise tolerance. Workloads and an FI_O2 identical to preoperative steady-state testing were applied during postoperative steady-state exercise testing.

Maximum Voluntary Ventilation (MVV), Ventilatory Ceiling, and Ventilatory Reserve

Maximum voluntary ventilation (MVV) was measured with patients encouraged to hyperventilate to their maximum for 12 s while breathing though a mouthpiece attached to a metabolic cart (Sensor Medics 2900) with nose clips in place (24). Respiratory rates and tidal volumes were maintained in ranges judged to achieve maximal ventilation throughout the effort. MVV was then determined by multiplying the total exhaled ventilation for 12 s by 5. Calculated MVV was derived by the equation FEV₁ × 35 (25). MVV (measured and calculated) was considered to represent the maximum ventilation a patient could achieve during exercise and was defined as the patient's ventilatory ceiling. Ventilatory reserve at any point in time was derived by measuring minute ventilation (E) and calculating ventilatory reserve = (MVV  E/MVV) × 100. 

Dyspnea

Dyspnea was determined using a standardized dyspnea questionnaire (26) with each patient questioned with regard to the type and amount of exertion required to precipitate dyspnea.

Six-minute Walk

Distance walked in 6 min was determined according to standard walk testing procedures while being continuously monitored by a therapist specifically trained in standardized walk testing (27). Patients were all provided adequate oxygen supplementation to maintain saturations greater than 90% throughout the walk.

Quality of Life

Quantitative assessment of health-related quality of life was achieved using a standardized tool (the Health Status Index SF-36) (28). Physical, social, and role functioning scales capture behavioral dysfunction caused by health problems. The SF-36 also incorporates measures of mental and emotional health and general well-being.

Statistical Analysis

All data were analyzed utilizing JMP statistics package for Macintosh (version 3.15; SAS Institute, Inc., Cary, NC). All groups analyzed were tested for normal distribution using Shapiro-Walk W testing. Data sets with normal distributions were analyzed using univariate analysis of variance techniques. Between-group comparisons were then performed utilizing Tukey-Kramer (honest significant difference [HSD]) methodologies, if significance was determined. Comparisons between pre- and postoperative variables were performed using paired two-tailed Student's t test with Bonferonni's adjustment. Data sets without normal distribution were analyzed using nonparametric methods. Comparisons between groups were determined using Wilcoxon/Kruskal-Wallis rank sum testing, and comparisons between pre- and postoperative variables were performed using Wilcoxon signed rank methods. Correlations between variables were determined using linear regression analysis methods. A p value of 0.05 or less was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 27 patients (18 men, 9 women) with an age range of 44 to 78 yr and a mean of 64 yr of age underwent bilateral LVRS via median sternotomy during the study period. All were ex-smokers. None of the patients had -1-antitrypsin deficiency. Eighteen patients completed pre- and postoperative exercise performance and pulmonary function studies a median of 4 mo (range 3 to 6 mo) after LVRS. Five patients (18%) died prior to follow-up evaluation. Four patients (15%) refused to return for follow-up testing. Eleven of the patients completing follow-up exercise performance and pulmonary function also completed pre- and postoperative dyspnea scales, 6-min walks, and quality-of-life questionnaires.

There were no significant differences in preoperative pulmonary function or exercise performance in the subjects who completed follow-up and those who refused follow-up. In addition, there were no significant differences in pre- and postoperative exercise performance or pulmonary function in the subgroup of patients who did and did not complete the functional/quality-of-life measurements. The effects of LVRS on pulmonary function are shown in Table 1.

Effect of LVRS on Maximum Exercise Performance

Exercise performance was significantly increased after LVRS (Figure 1 and Table 2). Ventilatory capacity/ceiling and resting ventilatory reserve (whether measured by maximum ventilatory volume or calculated based on airflows) were significantly increased following LVRS. Resting Pa_CO2 and pH (measured while on oxygen supplementation for exercise) were also improved (Table 3). All other resting exercise/metabolic parameters were unchanged.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Maximum workload (left) and O2max (right) achieved before and after LVRS. *p < 0.01.

Ventilatory function during maximum exercise testing was significantly improved, with patients achieving greater tidal volumes (VT) and E (Table 3). Associated with this, the ratio of dead space volume to tidal volume (VD/VT) during exercise was lower following LVRS when compared with preoperative measurements. The combination of increased VT and reduced dead space ratio resulted in a significant increase in VA and a significant decrease in Pa_CO2. In spite of this, exercise performance remained ventilatory limited after LVRS, with patients reaching their ventilatory ceiling at maximum exercise (i.e., having no remaining ventilatory reserve) and developing significant acute increases in Pa_CO2 compared with rest. Alveolar-arterial oxygen pressure difference (A-a gradient) during exercise worsened following LVRS. However, increases in ventilation offset reductions in A-a gradient, such that Pa_O2 during exercise was unchanged after LVRS. Heart rate, O₂ pulse, and lactate at maximum exercise did increase significantly after LVRS in a manner appropriate for the increase in workload performed.

Effect of LVRS on Exercise Performance at Isoworkloads and During Steady-state Exercise

Table 4 provides detailed assessment of physiologic performance during exercise at equivalent workloads, comparing function at isoworkloads during maximum exercise (equal to 50% and 100% of maximum workload during baseline testing) and during steady-state exercise testing. All three measures revealed an ability to achieve significantly higher VT with associated reductions in VD/VT and increases in E and VA following LVRS. No change in breathing frequency was observed. These changes were associated with a significantly lower Pa_CO2 and higher pH following LVRS. In spite of the increases in E, ventilatory reserve at each of the submaximal workloads was significantly greater. Similar to maximum exercise testing, there were no changes in Pa_O2 and an increase in oxygen A-a gradient at equivalent workloads following LVRS. Cardiac function and lactate levels at equivalent workloads were unchanged following LVRS.

Mechanisms of Improvement in Exercise Performance Following LVRS

Linear regression analysis revealed that changes in exercise performance after LVRS were significantly related to changes in maximum minute ventilation (E_max) achieved (r² = 0.634, p < 0.01), VT_max achieved (r² = 0.513, p < 0.01), and VA_max volume achieved (r² = 0.522, p < 0.01) during exercise. Consequently, exercise performance following LVRS was inversely correlated with the maximum Pa_CO2 achieved (r² = 0.453, p < 0.01). Importantly, changes in exercise performance correlated significantly with percent changes in the ventilatory ceiling or resting ventilatory reserve determined by either measured (r² = 0.437, p < 0.01) or calculated (r² = 0.343, p = 0.01) E_max, with changes in maximum ventilation achieved during exercise closely correlating with changes in ventilatory ceiling and resting ventilatory reserve (r² = 0.615, p < 0.01). No other changes in physiologic parameters during exercise correlated with changes in exercise performance following LVRS.

Analysis of changes in exercise performance relative to changes in pulmonary function after LVRS revealed that changes in maximal exercise performance were significantly related to changes in FEV₁ (r² = 0.475, p < 0.01), changes in FVC (r² = 0.403, p < 0.01), and changes in PI_max (r² = 0.437, p < 0.01). No other changes in pulmonary function correlated with changes in maximal exercise performance following LVRS.

Predictors of Improved Exercise Performance Following LVRS

None of the preoperative pulmonary function measurements correlated with changes in exercise performance after LVRS. On the other hand, change in exercise performance following LVRS was inversely related to ventilatory reserve determined at rest prior to surgery (r² = 0.289, p = 0.02; Figure 2). Change in exercise performance following LVRS was also correlated with preoperative dead space/tidal volume ratio (VD/VT) at rest (r² = 0.294, p = 0.02; Figure 2) and at maximum exercise (r² = 0.365, p < 0.01). None of the other preoperative physiologic parameters measured during exercise, including cardiac, gas exchange, and oxygen delivery/utilization measurements, correlated with changes in exercise performance following LVRS.

View larger version (11K): [in this window] [in a new window]   Figure 2.   Relationship between resting ventilatory reserve (left) and VD/VT (right) prior to LVRS versus change in exercise performance after LVRS.

Patient Function/Quality of Life

Indices of patient function and quality of life improved significantly following LVRS. Dyspnea scores improved (pre-LVRS: 14 ± 2 versus post-LVRS: 10 ± 1, p < 0.01: lower score equals less dyspnea) as did 6-min walk (1,081 ± 109 versus 1,273 ± 101 feet, p = 0.02). Analysis of quality-of-life indices revealed significant improvements in social (pre-LVRS: median 55, range [12-100] versus post-LVRS: median 88, range [32-100], p < 0.01: higher score equals improved symptoms), physical (median 20, range [0-70] versus median 35, range [10-90], p < 0.01), emotional role (median 67, range [0- 100] versus median 100, range [67-100], p < 0.05), physical role (median 0, range [0-100] versus median 50, range [0-100], p < 0.05), energy (median 45, range [5-65] versus median 65, range [25-80], p < 0.01), general health (median 47, range [0- 82] versus median 47, range [15-92], p < 0.05), and health change (median 25, range [0-100] versus median 100, range [25-100], p < 0.01) scores. There were no significant changes in pain, mental health, or depression scores.

Relationship of Exercise Measurements to Patient Function/Quality of Life

Linear regression analysis revealed no significant relationship between changes in exercise performance determined by maximum oxygen consumption O_2max or workload achieved to changes in 6-min walk distance following LVRS. Changes in exercise performance correlated with changes in quality-of-life measurements of physical function (r² = 0.312, p < 0.05) and inversely correlated with changes in dyspnea (r² = 0.594, p < 0.01; Figure 3) after LVRS. Changes in 6-min walk distance correlated with changes in the quality-of-life social index (r² = 0.702, p < 0.01), but otherwise did not correlate with changes in dyspnea or any other indices of quality of life.

View larger version (13K): [in this window] [in a new window]   Figure 3.   Relationship between changes in dyspnea and changes in exercise performance following LVRS.

Preoperative Pulmonary Function, Exercise Performance, and Short-term Survival

Although nonsurvivors tended to have a lower exercise tolerance, no significant difference was found. All patients had reduced exercise tolerance owing to ventilatory limitation. Short-term nonsurvivors had a significantly higher resting VD/VT ratio (median 53, range [50-57] in nonsurvivors versus median 43, range [30-48] in survivors, p < 0.01; Figure 4) with a higher breathing frequency (median 23, range [21-30] versus median 19, range [13-21], p < 0.05) and a lower VA (median 260, range [216-300] versus median 378, range [275-533], p < 0.05). In addition, nonsurvivors had a higher resting Pa_CO2 (median 52, range [34-55] versus median 40, range [36-49], p < 0.05) after oxygen supplementation was provided to ensure adequate oxygenation throughout exercise. Short-term nonsurvivors had significantly lower expiratory muscle strength when compared with survivors (median 65, range [60-70] versus median 110, range [83- 155], p < 0.01; Figure 4). No other differences in pulmonary function or physiologic parameters during exercise were found when comparing survivors with short-term nonsurvivors.

View larger version (9K): [in this window] [in a new window]   Figure 4.   PEmax (left) and resting VD/VT (right) prior to LVRS in all-cause short-term nonsurvivors and in survivors of LVRS. *p < 0.01.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Until recently, there has been little evidence of improved exercise performance in patients undergoing LVRS other than reports of increased walk distances (10). However, recent reports document an increase in maximum work and O_2max in COPD patients shortly after thoracoscopic or open sternotomy LVRS (14). Our results support these new findings, documenting a significant increase in exercise performance in patients undergoing median sternotomy with bilateral LVRS, as measured by maximum workload and O_2max achieved during cycle ergometry. Importantly, our results demonstrate improvements in specific areas of physiologic function during maximum exercise and also demonstrate significant physiologic improvements throughout exercise (at various isoworkloads and during submaximal steady-state exercise), further defining the physiologic mechanisms for improved exercise performance in patients following LVRS.

It is of interest that the percent increase in exercise performance as determined by O_2max was less than the percent increase in maximum work achieved following LVRS in our patients. Similar results were observed in other recently published reports (15, 16). A likely explanation for this may relate to the impact of basal O₂. Indeed, the percent change in O₂ from basal to maximum exercise following LVRS is very similar to the percent change in workload. Alternatively, the greater improvement in workload achieved could be the result of improved exercise mechanics over time leading to a lower O₂ for a given workload. If this is so, it does not detract from the positive impact of LVRS on exercise performance as determined by O_2max, especially since all other physiologic analysis is compared against O₂. It is also possible that the lesser increase in O₂ following LVRS is a manifestation of ongoing rehabilitation and improved O₂ utilization, a result that would not be due directly to LVRS. However, O₂ and lactates at isoworkloads and during steady-state testing after LVRS was not different, making this unlikely.

The improvement in exercise performance in our patients was due predominantly to improvements in ventilatory function, allowing for greater levels of exercise prior to developing a ventilatory limitation to exercise. A significant component of this ventilatory improvement was caused by the increase in ventilatory ceiling and resting ventilatory reserve, allowing for greater increases in E during exercise. The precise mechanism for the increase in ventilatory ceiling remains uncertain, although improved expiratory airflows likely play a role. LVRS also improved the ability to recruit VT during exercise, possibly related to reductions in end-expiratory lung volumes (29). Although having no direct effect on dead space ventilation, the ability to recruit VT resulted in significant increases in VA and alveolar ventilation, which when coupled with the increased ventilatory reserve resulted in improved carbon dioxide clearance and exercise performance. Importantly, all of the subjects studied had ventilatory limitation to exercise prior to and after LVRS, with the key physiologic impact of LVRS being to increase the level of work before which ventilatory limitation occurred.

Recent studies have also demonstrated that E at maximum exercise following LVRS increases in association with increases in VT (15, 16). Our results confirm these findings. Our results also demonstrate that the effects of LVRS on VT and E are evident throughout exercise, with increases in E and VT observed at isoworkloads and during steady-state exercise. In addition, our results demonstrate that decreases in the dead space-to-alveolar ventilation ratio occur throughout exercise and when coupled with the above ventilatory changes result in increases in alveolar ventilation and reductions in Pa_CO2 throughout exercise. The improved ventilatory function in our study at levels of exercise other than at maximum exercise as well as the improvement in maximal ventilatory function suggests that LVRS improves ventilatory function at all levels of work and likely reduces the work of breathing and dyspnea associated with any given task. In addition, the higher ventilatory thresholds and delay in achieving these thresholds (i.e., improved ventilatory reserves at a given level of activity) following LVRS allows for increases in maximum activity levels, in spite of continued ventilatory limitation.

Factors such as the development of respiratory muscle fatigue may also contribute to limitations in maximum ventilation during exercise in COPD patients (30). Improvements in diaphragm strength and recruitment have been suggested following LVRS, questionably related to improvements in end-expiratory lung volumes (31). Although no significant changes in inspiratory muscle function were noted in our patients after LVRS, there was a significant correlation between changes in exercise performance and changes in inspiratory muscle force. Because factors other than muscle strength may influence muscle endurance/fatigue, and because diaphragm strength and respiratory muscle endurance were not measured in our study, the lack of change in global inspiratory muscle strength following LVRS in our patients does not eliminate the possibility that improved respiratory muscle function contributed to the improvements in VT, E_max, and exercise performance.

In general, the effects of LVRS on pulmonary function in our patients were similar to those reported in the literature (5, 31). Although there was no significant change in lung elastance in our patients, there was a tendency toward increased elastic properties similar to those reported previously (34). Our data do suggest that preoperative pulmonary function and exercise measurements may be of some value as predictors of improvement in exercise function following LVRS. The lower the ventilatory ceiling/ventilatory reserve at rest, the greater the likelihood for improvement in exercise performance after surgery. In addition, the higher the dead space-to-alveolar ventilation ratio at rest and with exercise prior to surgery, the greater the improvement in exercise performance after LVRS. These predictors of improved exercise performance following LVRS fit well with the above findings of ventilatory limitation and improved ventilatory function as the physiologic etiology of exercise limitation.

Although not a mechanism directly contributing to improved exercise performance, heart rate, O₂ pulse (an estimate of cardiac stroke volume), and lactate levels at maximum exercise were significantly higher following LVRS in our patients. The ability to increase heart rate and O₂ pulse, at maximum exercise has also has been observed in another recent study of LVRS (16). The lack of change in heart rate, O₂ pulse, or lactate at similar loads suggests that no acute changes in oxygen delivery or utilization had occurred. Nevertheless, these findings suggest additional potential for cardiovascular conditioning in these patients and may make these patients candidates for even more aggressive rehabilitation after LVRS (37).

Importantly, the observed improvements in exercise physiology and pulmonary function in our patients are of clinical relevance. Our data suggest that patients had significant improvements in dyspnea, activity level, and quality of life following LVRS. Many factors can be proposed to account for the improved dyspnea, including reduced work of breathing and improved ventilatory patterns (34, 35). The improved activity levels can be attributed to the ventilatory changes observed during exercise testing as well as to the improved dyspnea. The improved quality of life in these debilitated patients is likely a consequence of the reduced symptomatology and improved activity, although doing something to "fix" a chronic "irreversible" disease may itself have contributed to the improvements in social, emotional, and general health symptoms.

Analysis of changes in maximal cycle ergometry and 6-min walk testing following LVRS revealed that the two measures of exercise performance do not correlate. Importantly, there was a good correlation between maximum exercise performance during cycle ergometry and physical function as assessed by the Quality of Life questionnaire. In addition, there was a significant inverse relationship between changes in maximal exercise performance and patient dyspnea following LVRS. On the other hand, 6-min walk results did not correlate with either of these measurements. However, changes in 6-min walk following LVRS did correlate well with social performance as assessed by the Quality of Life questionnaire. These findings suggest that exercise performance and 6-min walk results are important, but measure different outcomes, with maximal cycle ergometry better assessing physiologic and physical performance while the 6-min walk may better estimate a patient's ability to perform in a social environment.

Certain preoperative physiologic findings may be capable of defining which patients are at greatest risk for short-term nonsurvival following LVRS. Nonsurvivors had significant resting hypercapnia when hypoxemia was corrected with oxygen supplementation and an increased resting respiratory rate when compared with survivors. However, no discrete cutoff or predictive value could be determined on the basis of confidence intervals from the data separating nonsurvivors from survivors, likely due to our small numbers. On the other hand, nonsurvivors had significantly lower expiratory muscle strength, a measurement which had discriminatory capabilities based on distinct confidence intervals separating the survivor groups. COPD is typically associated with significant inspiratory muscle dysfunction due to hyperinflation (38), with minimal impact on expiratory muscle function. On the other hand, reduced expiratory muscle forces may be more of a marker of global muscle weakness. This suggests that malnutrition and other factors influencing global muscle function may be key markers for survival following LVRS. Indeed, malnutrition has previously been shown to be a marker for increased mortality in COPD patients (39). Nonsurvivors also had extremely high resting dead space-to-alveolar ventilation ratios (> 50%) compared with survivors, with some discriminatory capabilities based on confidence intervals. However, since increasing levels of dead space ventilation also predicted better exercise outcomes in survivors following LVRS, more detailed evaluation of the predictive value of baseline dead space ventilation is required.

In conclusion, LVRS in selected patients with severe hypoxemic COPD results in significant improvements in exercise performance, due predominantly to improved ventilatory function related to increased ventilatory reserves, enhanced VT recruitment, and improved alveolar ventilation. Improved exercise performance is associated with preoperative measurements of reduced resting ventilatory reserve and increased dead space ventilation during exercise testing, which may predict patients likely to have positive outcomes following LVRS. Our data also demonstrate that improvements in exercise performance following LVRS are of clinical relevance, with patients undergoing LVRS showing improvements in dyspnea, functional capacity, and quality of life. LVRS was performed safely in the majority of patients with end-stage COPD, although all-cause short-term mortality in COPD patients with severe disease may be somewhat higher than expected, based on other studies. Long-term outcomes of LVRS related to exercise performance remain to be determined.