Abstract
BACKGROUND: Lung volume reduction surgery (LVRS) has been shown to improve the clearance of carbon dioxide and minute ventilation (V̇E) in select patients with COPD. One variable often assessed in COPD is ventilatory efficiency (V̇E/V̇CO2).
METHODS: We compared 55 LVRS subjects with 25 controls from the National Emphysema Treatment Trial. V̇E/V̇CO2 was calculated from cardiopulmonary exercise testing at baseline and 6-months. We sought to assess V̇E/V̇CO2 changes with LVRS compared with controls who only received standard medical care.
RESULTS: At 6 months, the LVRS group significantly increased peak V̇O2, work load, V̇E, V̇CO2, and tidal volume while lowering peak and lowest V̇E/V̇CO2 (improved ventilatory efficiency) and end-tidal carbon dioxide pressure. The control group did not display these changes. The changes were greatest in the LVRS subjects who improved their exercise capacity after surgery (> 10 W).
CONCLUSIONS: The changes were greatest in the LVRS subjects who showed the most functional improvement, indicating an association of improved ventilation with improved functional outcome.
Introduction
One variable that is often used to assess the exercise capacity in patients with COPD is ventilatory efficiency, calculated by the relationship between minute ventilation (V̇E) and the rate of carbon dioxide production (V̇CO2): V̇E/V̇CO2. Peak V̇E/V̇CO2 has been found to predict mortality better than oxygen consumption (V̇O2) in elderly patients with heart failure.1 Additionally, V̇E/V̇CO2 slope has been shown to outperform V̇O2 as a predictor of respiratory complications and mortality in patients after pulmonary resection,2 increased mortality in patients with COPD undergoing surgery for lung cancer,3 and mortality prognostication in heart failure.4–8 In COPD patients, the V̇E/V̇CO2 slope was found to be negatively related to peak V̇O29,10 Furthermore, V̇E/V̇CO2 was shown to be inversely proportional to COPD severity, indicating altered gas exchange between patients with various severities of COPD.11 Even the intercept of V̇E/V̇CO2 has been found to be associated with the ventilatory response during exercise10,12 However, there has been no thorough investigation regarding the change in ventilatory efficiency before and after most surgical interventions in patients with COPD since the National Emphysema Treatment Trial (NETT).13
COPD is now the third leading cause of death in the United States14 and by 2030 is expected to rank third in worldwide burden of disease, according to a study published by the World Bank/World Health Organization.15 Lung volume reduction surgery (LVRS) has been shown to be an effective surgical option for a subset of patients with severe emphysema. In appropriately selected patients, LVRS can decrease morbidity, mortality,13,16 and dyspnea,17–19 while improving exercise performance,20,21 endurance,22 pulmonary function,20 and quality of life.21 The NETT used improvement in exercise capacity as one of the main outcomes in their federally sponsored, multicenter, randomized clinical trial that evaluated the effectiveness and safety of LVRS versus standard treatment.13
Cardiopulmonary exercise testing (CPET) measures a broad range of variables related to cardiorespiratory function and links metabolic, cardiovascular, and pulmonary responses to exercise.23 One of these variables is ventilatory efficiency, which represents the effectiveness of pulmonary clearance of carbon dioxide during exercise.24,25 During CPET, ventilatory efficiency is usually calculated as the ratio of V̇E to V̇CO2. In COPD, abnormal lung mechanics often limit the capacity for V̇E. Because it is known that LVRS improves the clearance of CO226–28 and V̇E in patients with COPD,26,27 we hypothesized that there would be an improvement in ventilatory efficiency after LVRS. A previous study has assessed the impact of ventilatory efficiency's association with weight gain29 and showed that weight gain after LVRS was related to improved lung function and ventilatory efficiency. Exercise capacity and gas exchange (not including V̇E/V̇CO2) after LVRS have been studied.13,26,27,30 Additionally, NETT trial papers have only reported arterial levels of CO2 instead of V̇CO2,26 so ventilatory efficiency could not be assessed directly. Because of this lack of thorough investigation regarding the change in ventilatory efficiency before and after LVRS in COPD, we sought to assess V̇E/V̇CO2 in our LVRS subjects.
QUICK LOOK
Current knowledge
COPD represents a significant worldwide public health burden and is a leading cause of morbidity and mortality. Lung volume reduction surgery has been demonstrated to reduce hypercarbia and minute ventilation in select subjects with the disease.
What this paper contributes to our knowledge
Following lung volume reduction surgery, ventilatory efficiency improves as measured by the peak minute ventilation/carbon dioxide production and lowest minute ventilation/carbon dioxide production during cardiopulmonary exercise testing. The greatest improvements were seen in subjects who increased their peak wattage by greater than 10 W from baseline to 6 months after surgery. These changes suggest an increase in alveolar ventilation after lung volume reduction.
Methods
Subjects
The institutional review board of the New York Presbyterian-Columbia University Medical Center (New York, New York) approved this study. A retrospective chart review was performed on all patients who underwent LVRS at Columbia University Medical Center from January 1998 to October 2009. Subjects who had CPET and pulmonary function tests (PFT) within 2 months (37 ± 27 d) before LVRS and repeat testing approximately 6 months (191 ± 26 d) after were included. PFTs concurrent with CPET were chosen. The subjects underwent 16–20 sessions (minimum) of pulmonary rehabilitation over 10–12 weeks, performed the first CPET 2 months before LVRS (baseline), and underwent a repeated 10–12 weeks of 16–20 sessions after surgery rehabilitation. These subjects then performed a second CPET 6 months after LVRS. These rehabilitation protocols are still used at this center for surgical patients, even after discontinuation of the NETT. The selection criteria after the completion of the NETT are according to National Coverage Determination Centers for Medicare & Medicaid Services guidelines for LVRS.31 In brief, subjects in this study of LVRS have Global Initiative for Chronic Obstructive Lung Disease level 3 or 4 COPD with severe exercise limitation (peak wattage less than 40 W for men and 25 W for women, with preference for upper lobe disease in Centers for Medicare & Medicaid Services guidelines, severe emphysema in NETT). Additionally, subjects had no pulmonary hypertension and no evidence of active cardiac disease.
To rule out the possible effects of the intensive pre- and post-LVRS pulmonary rehabilitation given to the subjects, and to have a control group of comparable subjects eligible for LVRS, we analyzed the NETT subjects who were randomized to the non-surgical group (they underwent only pulmonary rehabilitation both before and after randomization) at our center between 1998 and 2003. Full exercise programs, as prescribed in the NETT, consisted of 16–20 sessions over 10 weeks, with a similar program after randomization.32
Medications, specifically cardioactive medications such as beta agonists, beta blockers, anticholinergics, digitalis, and calcium channel blockers, were reviewed in each subject's chart for both the first and second CPET. Additionally, a dobutamine stress echocardiogram or thallium testing was performed to assess for ischemic heart disease. Diagnosis for subjects with any suspicious changes on the dobutamine or thallium test was confirmed with cardiac catheterization. Left ventricular ejection fraction (> 45%) and absence of significant valvular disease were confirmed by review of the echocardiogram closest to the time of the first CPET.
Pulmonary Function and Exercise Testing
All PFTs were performed according to the NETT protocol.13 All values are reported post-bronchodilator, as per NETT protocol. This includes FVC, FEV1, total lung capacity (TLC), and residual volume (RV). Percent of predicted PFTs were calculated for males and females as described in prior publications.33–35
CPET was performed on an electronically braked cycle ergometer (Ergometrics 800, SensorMedics, Yorba Linda, California) with a Viasys SensorMedics Encore metabolic cart (Viasys Corporation, Loma Linda, California); before 2005, a Vmax 229 series workstation (SensorMedics) was used. The equipment was calibrated before every test. Continuous 12-lead telemetry was monitored via CardioSoft electrocardiogram software (GE/CardioSoft, Houston, Texas); before 2005, a model Max-1 electrocardiogram was used (Marquette Medical Systems, Milwaukee, Wisconsin). Oxygen saturation was recorded with a N595 pulse oximeter (Nellcor, Boulder, Colorado); before 2005, a SensorMedics Sat-Trak (SensorMedics) was used. All subjects were tested on 29.99 ± 0.25% fractional inspired oxygen via a closed system both before and after LVRS; all calibrations were 2-point calibrations with hyperoxic and hypoxic points and included hypercarbic and absent carbon dioxide levels of calibration. The metabolic system was validated for use with hyperoxic testing and was stable for measurement of FeCO2 at hyperoxic situations. The exercise protocol was ramping and followed the NETT13 and American Thoracic Society guidelines.36,37 Even after termination of the NETT, NETT guidelines were followed at this institution for CPET. An individualized ramping protocol was determined by the maximal voluntary ventilation (MVV) test. The MVV was performed over 12 s. Individuals who achieved ≤ 40 L/min MVV performed a 5-W/min ramping protocol, whereas those attaining > 40 L/min MVV performed a 10-W ramping protocol. The same ramp protocol was used pre- and post-LVRS or at both sessions in the non-surgical group. CPET variables were collected breath-by-breath, and included rate of carbon dioxide production (V̇CO2), maximal work load (peak watts), volume of oxygen consumption by weight (V̇O2 in mL/kg/min), percent-of-predicted oxygen consumption attained (V̇O2[% pred]), V̇E, tidal volume (VT), breathing frequency, partial pressure of end-tidal CO2 (PETCO2), HR at baseline (HRrest), and heart rate at peak exercise (HRmax). Peak variables averaged over 20 s include V̇O2, V̇E, V̇CO2, PETCO2, and VT. Ventilatory efficiency was measured in 4 ways: (1) the V̇E/V̇CO2 slope, taken as the linear relationship between V̇E and V̇CO2; (2) the intercept of the V̇E/V̇CO2 regression line; (3) the lowest V̇E/V̇CO2 ratio during exercise; and (4) the V̇E/V̇CO2 at peak exercise. The V̇E/V̇CO2 slope was calculated using the breath-by-breath data averaged over 20 s. Rest was not included in the slope calculation; the 3-min warm-up was included.
Age-predicted peak HR was calculated using 208 − (0.7 × age).38 The percent-of-predicted heart rate reserve (%HRR) was calculated as %HRR = (HRmax − HRrest)/([age predicted HR]) − HRrest) × 100.39 Breathing reserve was calculated as 100 × (MVV − peak V̇E)/MVV. V̇O2(% pred) was calculated from the 1983 Jones equation for males as 4.2 − (0.032 × age) and for females as 2.6 − (0.014 × age).
Statistics
Data were analyzed with SPSS 20 (SPSS, Chicago, Illinois). Variables are presented as mean ± SD. Continuous variables were analyzed with a paired-sample t test to evaluate whether there were differences between baseline and 6 months. Only subjects with complete data were used. An independent-sample t test was used to assess differences between those included and not included and between the non-surgical group and the surgical group. Simple linear regression was used to determine the association between variables. Chi-square test was used to determine differences between categorical variables. Statistical significance was set a priori at P ≤ .05.
It was decided, in addition, to compare the LVRS subjects who significantly improved their exercise capacity after surgery with those who did not. An improvement in maximal exercise capacity was defined as an increase in the maximal work load of > 10 W from the baseline level, as defined from the results in the NETT.13 We explored whether the changes in functional improvement would be associated with the changes in ventilatory improvement. Furthermore, an iso-work load analysis was used to compare the same work loads at both exercise tests. The maximal work load achieved by the subject during either the baseline or 6-month time period was chosen for comparison. For example, if the maximal work load at baseline was 15 W, then 15 W was the level used for both tests to compare CPET variables. Looking at iso-work load allowed us to control for the increase in functional capacity.
Results
Baseline Demographics and Pulmonary Function
Figure 1 shows a flow chart of the data. Of the 125 subjects who underwent LVRS between 1998 and 2009, 82 subjects had PFT and CPET data before and after LVRS. Despite our usual practice of performing CPET both before and after surgery to assess function and help guide exercise programs, there was variability in practice before the NETT guidelines were established in 2004. Twenty-two subjects had no CPET before LVRS, 17 subjects had no CPET after LVRS due to loss of follow up, and 4 subjects died within 6 months of surgery. Three additional subjects were excluded because the CPET before LVRS was done on room air instead of supplemental oxygen, 4 had ischemic heart disease, and an additional 16 were excluded because the CPET was too far from the time of surgery. Four subjects did not have completed data. Overall, 55 subjects were available for analysis. Additionally, 25 controls (non-surgical group) were analyzed.
Flow chart. LVRS = lung volume reduction surgery, CPET = cardiopulmonary exercise test, PFT = pulmonary function test.
All subjects were former heavy smokers who were confirmed abstinent from smoking for at least 6 months before and after surgery. All subjects had an electrocardiogram at the time of CPET and were in sinus rhythm. None had active ischemic heart disease. In comparison with the surgery group, there were a greater proportion of men in the non-surgical group. All other demographics and pulmonary function variables were comparable between the two groups at baseline (Table 1). Medications did not significantly differ between LVRS subjects or the non-surgical group before and after surgery or rehabilitation. Specifically, regarding cardioactive medications, all subjects were on a beta agonist for first and second CPET, with none on beta blockade, 96% of both LVRS subjects and non-surgical participants were on anticholinergic medications for first and second CPET (P > .99), 9% were on digitalis for first and second CPET: 8% of controls and 9% of LVRS subjects (P > .99), and 24% were on calcium channel blockers for first and second CPET: 20% of non-surgical participants and 26% of LVRS subjects (P = .60).
Baseline Demographics and Pulmonary Function
Maximum Exercise Testing and Ventilatory Efficiency
At baseline, the non-surgical group had a significantly higher V̇O2(% pred) and peak V̇E with a lower breathing reserve than those who received LVRS (P values = .04, .04, and < .01, respectively). All PFTs were comparable at baseline.
From baseline to 6 months, the LVRS group significantly increased their peak V̇O2, V̇O2(% pred), work load, peak V̇E, breathing reserve, peak V̇CO2, peak VT, and %HRR while lowering their peak V̇E/V̇CO2, lowest V̇E/V̇CO2, and peak PETCO2 (Table 2). The non-surgical group only significantly increased their breathing reserve and actually decreased their peak V̇O2, V̇O2(% pred), work load, peak V̇E, and peak V̇CO2. It is interesting to note, however, that the non-surgical group significantly decreased their V̇E/V̇CO2 intercept. This was only by 0.58 ± 1.39, and its relevance as to a clinically meaningful change is questionable.
Cardiopulmonary Exercise Testing Measurements at Baseline and 6 Months
After 6 months, the LVRS group had a significantly higher work load (P = .01), breathing reserve (P = .03), peak V̇CO2 (P < .01), and %HRR (P < .01), compared with the non-surgical group. Additionally, all of the PFTs were significantly better in the LVRS group compared with the non-surgical group (all P < .001).
LVRS Subjects Who Significantly Improved Exercise Capacity
We then compared the LVRS subjects who significantly improved their exercise capacity (> 10 W) after surgery to those who did not (≤ 10 W) (Table 3). The only significant baseline difference between these two groups was that those who improved had lower watts (P = .04). Six months after surgery, those who improved significantly increased their BMI, FEV1(% pred), FVC(% pred), peak V̇O2, V̇O2(% pred), peak watts, peak V̇E, breathing reserve, peak V̇CO2, peak VT, and %HRR while decreasing their RV(% pred), TLC(% pred), RV/TLC, peak V̇E/V̇CO2, lowest V̇E/V̇CO2 (Fig. 2), and peak PETCO2. Those who did not improve failed to increase their peak V̇O2, V̇O2(% pred), peak watts, breathing reserve, while also failing to decrease their peak V̇E/V̇CO2 and lowest V̇E/V̇CO2. Additionally, the correlation between the change in watts and the change in V̇E/V̇CO2 was r = −0.288 with a P value = .033.
Changes After LVRS: Subjects Who Significantly Improved Their Workload and Subjects Who Did Not
Bar graphs show that lung volume reduction surgery subjects who significantly improved their exercise capacity (> 10 W) also significantly improved their ventilatory efficiency (peak V̇E/V̇CO2) compared to those who did not improve their exercise capacity. Data are shown as mean ± SD. * P ≤ .05.
The 26 subjects who did improve their exercise capacity (improvers) had significantly better FEV1 (P < .01), FEV1(% pred) (P = .001), FVC(% pred) (P < .01), RV(% pred) (P = .01), and RV/TLC (P = .01) at 6-month testing compared with the 29 subjects who did not improve exercise capacity (non-improvers). Additionally, improvers had significantly higher V̇O2(% pred) (P = .03), work load (P = .005), peak V̇E (P = .04), and breathing reserve (P = .04); actual values are found in Table 3.
Iso-Work Load Comparisons
We then collected CPET measurements at the same work load before and after LVRS. Responses to identical levels of exercise before and after LVRS for the overall group and those who improved and did not improve their exercise capacity are shown in Table 4. Only 1 non-surgical participant significantly improved exercise capacity, and none of the variables significantly changed in the overall non-surgical group, in the single non-surgical participant who improved, or in non-surgical participants who did not improve at iso-work load (data not shown). For the LVRS subjects (Table 4), all 3 groups (overall, improvers, and non-improvers) increased their VT significantly. The improved work load group also significantly increased their respiratory exchange ratio while decreasing their breathing frequency and PETCO2.
Comparisons of Metabolic and Ventilatory Variables During Iso-Work Load Exercise
Discussion
In support of our hypothesis, we found that compared with the non-surgical group, LVRS subjects had improved ventilatory efficiency as measured by a small but significant decrease in their peak V̇E/V̇CO2 and lowest V̇E/V̇CO2 from baseline to 6 months. The changes were greatest in the LVRS subjects who showed the most functional improvement measured by peak wattage, indicating an association of improved ventilation with improved functional outcome as measured by CPET. This is supported by the fact that those who significantly improved their wattage also increased their V̇E significantly more than those who did not improve.
The severity of COPD is assessed using resting PFTs, but this does not usually correspond to exercise impairment in individual patients.23 Although multifactorial, one of the biggest reasons for exercise intolerance in COPD is the inability to increase V̇E relative to higher work loads with higher end-expiratory volumes and air trapping. Ventilatory limitation is typically classified by a breathing reserve < 15%.23,36 In patients with COPD, this results from a reduction in MVV, increased dead space ventilation, and the inefficiency of gas exchange, raising the required V̇E at any metabolic rate.23 After LVRS, the elastic recoil increases,28,40 thereby decreasing TLC and RV and decreasing hyperinflation, which could explain the increase in peak V̇E seen in our subjects. Criner et al26 found in their NETT exercise substudy that the improvements in exercise capacity following LVRS were due to less rapid and shallow breathing, reductions in ventilatory dead space, and enhanced carbon dioxide elimination. Our results support these findings, as those who improved their exercise capacity also significantly increased their peak V̇E and V̇CO2. In the subjects who improved their exercise capacity, the RV/TLC ratio was significantly lower, demonstrating less air trapping which leads to more efficient ventilation. Additionally, those who improved their exercise capacity showed a small but significant decrease in their peak V̇E/V̇CO2 and lowest V̇E/V̇CO2.
Patients with emphysema have an increased physiological dead space to tidal volume ratio (VD/VT) that fails to decrease normally during exercise41,42 and contributes to respiratory acidemia and resulting exercise limitation.27 Benditt and colleagues27 found, for their group of LVRS subjects, that the VD/VT decreased slightly at rest and noticeably during exercise. The authors noted that, although a reduction in VD/VT should improve ventilatory efficiency and decrease the V̇E needed for a given work load, they saw no change in the V̇E measured at the point of isowatt exercise. This is most likely because they found the partial pressure of arterial carbon dioxide to be lower at the isowatt level after surgery due to the increased alveolar ventilation.27 Another study by Criner et al26 showed that there was a decrease in ventilatory dead space during unloaded cycling and at maximum exercise after LVRS. With the increase in peak V̇E, the reduced dead space also suggested improvements in alveolar ventilation throughout exercise that resulted in improved carbon dioxide elimination. Similarly, in our LVRS group at identical work loads during the ramped exercise test (iso-work load) and at peak exercise, there was an increase in V̇E. This was coupled with a decrease in PETCO2. This may help to explain why the change in V̇E/V̇CO2 was small. Because the PETCO2 values were lower, this suggests a lower alveolar PCO2. Because our subjects did not increase their V̇E/V̇CO2, despite a lower alveolar PCO2, one can infer that their V/Q distribution improved. Although V̇E/V̇CO2 reflects the total breathing effort required to eliminate carbon dioxide, improved ventilatory function can cause it to go either up or down.
Our findings are in agreement with those of Criner et al,26 who found that the improvement in ventilatory function (improved peak V̇E, reduced dead space, and improved carbon dioxide elimination though alveolar ventilation) is the major benefit of LVRS in terms of exercise performance. In our study, those who significantly increased their work load also significantly improved their peak V̇E, peak V̇CO2, peak VT, breathing reserve, peak V̇E/V̇CO2, and lowest V̇E/V̇CO2. Those who did not increase their work load did not improve these variables, or at least not to the same degree. In the study by Criner et al,26 15% of the subjects improved their exercise capacity by > 10 W; however, in our study, 47% significantly improved their exercise capacity. This is most likely attributed to better selection criteria after the NETT results. The differences between those who improved their exercise capacity and those who did not are most likely due to better pulmonary function outcomes. Those who improved had significantly better FEV1, FEV1(% pred), FVC(% pred), RV(% pred), and RV/TLC at 6-month testing compared with those who did not improve their exercise capacity.
In a study measuring ventilatory efficiency in 474 adults, Sun and colleagues24 showed that the lowest V̇E/V̇CO2 ratio was the most reproducible and stable response accounting for laboratory sites, exercise mode, gender, and age. It has been hypothesized that the V̇E/V̇CO2 ratio may be a more stable marker of ventilatory efficiency, as the slope is subject to transient hyperventilation early in exercise and can be affected by metabolic acidosis.24,25 This supports the use in our study of peak and lowest V̇E/V̇CO2.
Study limitations include the facts that non-surgical and LVRS groups were not perfectly matched and that data were collected retrospectively. Additionally, the non-surgical subject group was not temporally identical to all of the surgery subjects because it is not ethical to delay LVRS to eligible candidates. Finally, most of the subjects assessed after the NETT had video-assisted thoracoscopic surgery, whereas the subjects in the NETT period all had median sternotomy. However, there has not been any indication in published studies that the surgical approach affects long-term functional outcomes.43 Strengths of this study are that it has a relatively large group of unique subjects with data collected on both types of surgeries in a single NETT center with a single exercise lab and using the NETT protocols and rehabilitation guidelines throughout.
Conclusions
These study results show that LVRS improves ventilatory efficiency, as measured by the peak V̇E/V̇CO2 and lowest V̇E/V̇CO2 ratio on CPET. It is further demonstrated that the greatest improvement is seen in those who increase their peak wattage by greater than 10 W from baseline to 6 months after LVRS. The change in V̇E/V̇CO2, although significant, is most likely a small change that agrees with the decrease in PETCO2, which suggests an improvement in alveolar ventilation during exercise and results in improved carbon dioxide elimination. Future investigations are needed to determine whether the change in V̇E/V̇CO2 is associated with the known quality of life, dyspnea, morbidity, or mortality benefits of LVRS.
Footnotes
- Correspondence: Hilary F Armstrong, MA, 180 Fort Washington Avenue, HP 1-169b, New York, NY 10032. E-mail: hfa2104{at}columbia.edu.
The authors have disclosed no conflicts of interest.
This research was funded in part by the Vidda Foundation. This publication was supported by grant UL1 TR000040 from the National Center for Advancing Translational Sciences, National Institutes of Health.
- Copyright © 2015 by Daedalus Enterprises
