Abstract
The 6-minute walk test is used in clinical practice and clinical trials of lung diseases; however, it is not clear whether replicate tests need to be performed to assess performance. Furthermore, little is known about the impact of walking course layout on test performance. We conducted 6-minute walks on 761 patients with severe emphysema (mean ± SD FEV1% predicted = 26.3 ± 7.2) who were participants in the National Emphysema Treatment Trial. Four hundred seventy participants had repeated walks on a separate day. The second test was improved by an average of 7.0 ± 15.2% (66.1 ± 146 feet, p < 0.0001, by paired t test), with an intraclass correlation coefficient of 0.88 between days. The course layout had an effect on the distance walked. Participants tested on continuous (circular or oval) courses had a 92.2-foot longer walking distance than those tested on straight (out and back) courses. Course length had no significant effect on walking distance. The training effect found in these patients with severe emphysema is less than in previous reports of patients with chronic obstructive pulmonary disease. Furthermore, the layout of the track may influence the 6-minute walk performance.
The 6-minute walk test is used as an outcome measure in clinical trials of lung disease. Although the test has been somewhat standardized, there are differences in the testing technique in different clinical centers and laboratories (1). In particular, there is not agreement on the length of the test course, the shape of the test course (straight versus continuous circle or oval), whether a practice walk should be done before the final test, and whether the better of two walks or the second of two walks should be the reported value. These questions are of particular importance in longitudinal multicenter clinical trials in which test results may vary between centers and over time.
The National Emphysema Treatment Trial (NETT) is a multicenter trial of lung volume reduction surgery compared with medical treatment for advanced emphysema. The study investigators conduct 6-minute walk tests in participants at entry into the study and periodically thereafter. In this report, we analyze results of 6-minute walks for the first 761 randomized participants. Of these, 470 participants undertook two 6-minute walks on successive days. This afforded us the opportunity to examine the reproducibility of repeated 6-minute walks in this patient population. Because the length of the walking course differed among the 17 clinical centers, we were also able to examine what effects this would have on 6-minute walking distance. Furthermore, because some walking courses were continuous (circular or oval) and others were straight, requiring turns, we were able to evaluate the effect of walking course layout on performance.
Additional details on the methods of this research are available in the online supplement.
The NETT is a multicenter clinical trial comparing lung volume reduction surgery with medical treatment in patients with emphysema. The details of the trial design have been published (2, 3).
Major enrollment criteria include bilateral emphysema judged suitable for lung volume reduction surgery documented by computed chest tomography, FEV1 less than or equal to 45% predicted (4), total lung capacity of at least 100% predicted (5), residual volume of at least 150% predicted (5), and PaCO2 less than 60 mm Hg (or less than 55 mm Hg in Denver). Enrollees had to be validated nonsmokers for 4 months and had to be free of important comorbid conditions.
Participants had a baseline evaluation before enrollment, including a medical history and examination, lung function testing, blood testing, chest computed tomography, dobutamine cardiac stress testing, and echocardiography. If necessary, additional studies such as cardiac consultation and cardiac catherization were performed to exclude people with coronary artery disease or moderate to severe pulmonary hypertension. The data that are presented here are taken from the baseline evaluation in participants ultimately enrolled in the trial, representing those with severe emphysema who are free of important comorbid conditions. We analyzed data from 761 randomized participants who had 6-minute walk tests at the 17 clinical centers. In 470 of the initial enrollees, the 6-minute walk test was repeated on the day after the first test to account for the effect of repeated testing.
On a day before the first 6-minute walk, participants underwent cardiopulmonary exercise testing with a cycle ergometer to exclude important cardiac rhythm disturbances or ischemic electrocardiographic changes. Before the first 6-minute walk, participants also undertook a 1- to 2-mile-per-hour treadmill test to determine whether they needed oxygen during the 6-minute walk test.
Each clinical site used the same walking course for all participants, although the length and shape of the walking course differed among centers. Participants took a short-acting bronchodilator within 2 hours before testing.
Clinic staff gave identical scripted instructions to participants and explained the Borg category ratio scale (CR-10) (6, 7). A staff member walked behind the participant and carried the oxygen delivery system if required. At each minute during the walk, the staff member told the participant how much time had elapsed and the remaining time and gave scripted encouragement. At the end of the 6 minutes, the participant was told to stop, and the distance walked was recorded. Predicted values were calculated using the normative values of Enright and Sherrill (8).
Results are presented as means and SDs. Changes in 6-minute walk distances were compared using a paired t test. Trends in differences between the first and second walk distances were evaluated using Bland-Altman analysis (9). Changes in the Borg scores were compared with zero using the Wilcoxon signed rank test. The effect of course layout and length was estimated using multiple linear regression adjusting for age, sex, height, and FEV1 percentage predicted.
The characteristics of the study population are presented in Table 1
Characteristics* | Participants with at Least One Test | Participants with Two Tests |
|---|---|---|
| n | 761 | 470 |
| Age, yr | 67.2 ± 5.9 | 67.2 ± 5.8 |
| Female, % | 39 | 39 |
| Height, cm | 168.9 ± 9.61 | 168.5 ± 9.5 |
| Weight, kg | 70.3 ± 13.8 | 70.1 ± 13.7 |
| BMI, kg/m2 | 24.6 ± 3.85 | 24.6 ± 3.78 |
| FEV1, L BTPS | 0.76 ± 0.24 | 0.75 ± 0.24 |
| FEV1 % predicted | 26.3 ± 7.2 | 26.3 ± 7.4 |
| FVC, L BTPS | 2.45 ± 0.77 | 2.45 ± 0.77 |
| FVC % predicted | 65.7 ± 15.3 | 66.0 ± 15.4 |
| FEV1/FVC % | 31.6 ± 6.6 | 31.4 ± 6.4 |
| MVV, L/min | 30.4 ± 10.2 | 30.4 ± 10.4 |
| TLC, L BTPS | 7.75 ± 1.52 | 7.78 ± 1.49 |
| TLC % predicted | 129.2 ± 13.9 | 130.4 ± 13.4 |
| RV % predicted | 227.6 ± 48.0 | 230.0 ± 45.9 |
| RV/TLC % | 65.7 ± 8.0 | 66.0 ± 7.7 |
| DCO % predicted | 27.5 ± 9.6 | 27.8 ± 9.7 |
| Required oxygen at rest, % | 9.9 | 10.6 |
| Required oxygen during test, % | 76.8 | 78.1 |
| SGRQ, total score | 56.1 ± 12.8 | 56.2 ± 13.3 |
| 6MW distance % predicted† | 70.1 ± 19.2 | 67.4 ± 18.1 |
| Dyspnea score† | 5.1 ± 2.0 | 5.0 ± 2.0 |
| Muscle fatigue score† | 3.5 ± 2.3 | 3.4 ± 2.2 |
In general, we found that the 6-minute walk test was reproducible on subsequent days, with an intraclass correlation coefficient of 0.88 (p < 0.0001), but with a definite learning effect on the second test (Figures 1 and 2)
Figure 1. Scatter plot of first-day and second-day 6-minute walk distance. The bold line shows the line of identity. The dashed line is the regression of the second test to the first. The intraclass correlation between the two tests is 0.88, with the majority of points above the line of identity.
[More] [Minimize]
Figure 2. Scatter plot of the difference between the two walks plotted against the mean value (Bland-Altman plot). The dashed line is the regression of change in walking distance against mean value, indicating that those with higher values tended to improve more on the second day.
[More] [Minimize]
Figure 3. The distribution of differences between the second-day walk distance and the first-day walk distance. The mean ± SD improvement was 66.1 ± 146 feet, which is significantly larger than zero (p = 0.0001).
[More] [Minimize]People with greater distances on the first test had a tendency to improve more on the second test, suggesting that there was not simply regression to the mean. The Bland-Altman plot confirms that the second-day test was consistently and significantly longer (Figure 2). The effect of using different strategies for test selection is shown in Table 2
First Test | Second Test | Best Test | |
|---|---|---|---|
| Distance walked, feet | 1,124.0 ± 298.6 | 1,190.1 ± 312.8 | 1,216.1 ± 307.5 |
| Distance, % predicted | 67.4 ± 18.1 | 71.4 ± 19.1 | 72.9 ± 18.6 |
| Borg score dyspnea | 5.01 ± 2.02 | 5.34 ± 2.05 | 5.26 ± 2.04 |
| Borg score leg fatigue | 3.45 ± 2.20 | 3.45 ± 2.28 | 3.51 ± 2.30 |
Linear regression models adjusting for baseline measures and including demographic, anthropometric, physiologic, previous rehabilitation, and quality-of-life measures showed that participants with higher maximum inspiratory pressure had more improvement on the second walk (0.99 ± 0.34 feet/cm H2O, p < 0.005). There were no other clinical features that predicted which participants increased walking distance.
Borg scores for breathlessness and leg fatigue were also reproducible on subsequent tests on average (Figure 4)
Figure 4. Scatter plots are shown for Borg scores for breathlessness (A) and for muscle fatigue (B) at the end of the 6-minute walk (6MW) comparing the first and second test days. Data points are offset to allow better display of the individual data points. The bold diagonal line is the line of identity.
[More] [Minimize]
Figure 5. The distribution of differences between the second-day walk Borg scores for breathlessness (A) and for leg fatigue (B). The distributions are skewed in opposite directions. The mean change in the Borg dyspnea score is 0.33 ± 1.74 U, which is significantly greater than zero (p < 0.0001). The mean change for the leg fatigue score on the second test is 0.002 ± 1.70 U, which is not statistically different from zero.
[More] [Minimize]Among the 14 clinics with straight walking courses, the mean length of the course ranged from 50 to 164 feet (mean 99.9 ± 34.1 feet). Three clinics had continuous layout courses (either oval or square) that ranged in length from 185 to 397 feet in circumference (mean 298 ± 75.4 feet). The 6-minute walk distance was greater in those individuals who completed the test on a continuous track (1,156 ± 302 feet for straight courses and 1,266 ± 360 feet for continuous courses, p = 0.003, t test). The effect of course layout persisted when adjusted for FEV1 percentage predicted, age, sex, and height (adjusted difference 92.2 feet, p < 0.001). Among the 14 clinics with straight course layout, there was no statistically significant effect of track length on 6-minute walk distance in bivariate as well as multivariate analyses.
The 6-minute walk test is widely used for the evaluation of functional status in patients with lung disease in both clinical practice as well as clinical trials (12–16). The test is widely used because it involves a familiar daily activity and involves the use of minimal technical resources (17).
Because verbal encouragement can improve test performance, most protocols use standardized verbal prompts that are similar to those that we have used (18). Less well documented, however, is the short-term variability of the test in patients with advanced chronic obstructive pulmonary disease (COPD). Because the test requires some degree of strategy and pacing, some laboratories require a practice walk before the test walk, whereas others do not. There is also little known about the effects of the walking course on test results. This analysis was performed on enrollees in the NETT to address these questions.
On the second day of testing, we found a statistically significant improvement of approximately 66 feet. This was likely due to familiarity with the walking course, better pacing, or motivational factors. It is also possible that the patients were less fatigued on the second test day because the test had not been preceded by an oxygen titration test requiring treadmill walking at 1 to 2 miles per hour. Similar short-term improvement in 6-minute walking distance and treadmill exercise testing has been reported in cardiac patients, which has been attributed to a learning effect (19, 20). Gibbons and colleagues performed four 6-minute walks on the same day in 79 healthy adults. They found that the longest distance occurred on the first walk 14% of the time and that the longest of the four walks averaged 6.6% longer than the initial test, similar to the magnitude of improvement in our COPD patients (21).
There have been several studies of repeated walking tests in COPD patients. Leach and colleagues found in 30 patients with hypoxemia from COPD and restrictive lung diseases that the learning effect of subsequent 6-minute walks was 14.9% (22). This is a greater learning effect than we found in our more impaired patients who were given supplemental oxygen to prevent hypoxemia. McGavin and colleagues found that a second 12-minute walking test was associated with a 7% improvement, closer to what we found in this study (23). Swinburn and colleagues found a 16% improvement on four successive 12-minute walks in 17 COPD patients over the course of 1 week (24). Knox and colleagues performed 12 5-minute walks over 3 consecutive days in 36 COPD patients. He found that there was a 33% improvement in walking distance, with half of the improvement occurring in the first three walks on Day 1. If the 12 walks were repeated over a 4-week interval, however, the total learning effect was only approximately one-fourth as large (25). Stevens and colleagues performed three 6-minute walk tests in 21 COPD patients on separate days and found a mean increase of 10% on the second test and an additional 3% on the third test (26).
We are not certain why we found a smaller learning effect than other series, but it may relate to the severity of disease, the scripted encouragements, the instructions to perform a maximal test, the prior treadmill and cycle exercise testing, or the familiarity of the patients with the testing staff and environment. A possible explanation of our findings is that some of the NETT patients had previous experience with pulmonary rehabilitation and therefore were not truly “naive” to the test procedures. We could not, however, find any difference in learning effect based on prior pulmonary rehabilitation exposure.
The obvious implication of our results and the results of others is that clinical trials that rely on 6-minute walks before and after an intervention should include appropriate control groups, or repeated tests, to account for this learning effect. If a single 6-minute walk test is used, there should be a contemporaneous control group and the tests should be spaced several weeks apart to minimize the learning effect. The selection of reference equations also depends on the testing method. Normative values for 6-minute walk tests have been derived from single test sessions (8). Reference equations from healthy volunteers who performed multiple tests on a single day predict significantly longer distances if the best of multiple tests is chosen as the study value (21, 27). There can be statistically and potentially clinically important differences depending on whether the first, second, best, or mean value is reported. Because of the small difference and the burden of a second day of exhaustive exercise testing in patients with advanced obstructive lung disease and the little additional information that such testing provides, we have eliminated the second 6-minute walk from the NETT protocol. The recent American Thoracic Society guidelines for 6-minute walk tests do not require a practice walk. They do recommend, however, that if a practice walk is done, the repeat test should be done on the same day at least 1 hour later and that the larger of the two values should be reported (28).
The data presented here may be useful in interpreting changes in walking distance after an intervention in individual patients. If the study is repeated within a short period such that a learning effect may be involved, an individual patient would need to improve by more than 352 feet (66 feet + 1.96 × 146 feet) to be 95% confident that there had been improvement. If the short-term learning effect is discounted, for example in tests done 4 weeks apart, then an improvement of 286 feet (1.96 × 146) is necessary to be 95% confident that the change was not random variation. This threshold is similar to the upper 95% confidence interval of clinically important changes in 6-minute walk distance of 280 feet reported by Redelmeier and colleagues (11).
The impairment of 6-minute walk distance (70.1% predicted) on average was not as severe as one might predict from the severity of impairment of FEV1 (26.3% predicted) or quality of life (St. George's Respiratory Questionnaire, 56.2 out of 100). It is possible that this reflects the selected population enrolling in NETT or may reflect the benefit of oxygen supplementation, which was used by 78% of the patients. For clinical purposes, it is important to consider other factors than walking distance. Van Stel and colleagues found that half of 53 COPD patients experienced declines in 6-minute walk distance after rehabilitation, whereas 83% reported improved exercise capacity that could be attributed to cardiovascular conditioning, less oxygen desaturation, and less dyspnea with everyday activities (29). Thus, the 6-minute walking distance is only one dimension of functional assessment of COPD treatments.
The Borg score for dyspnea at the end of the 6-minute walk was slightly higher on the second walk, perhaps corresponding to greater effort expenditure to achieve the longer distance. This is different from the findings of Belman and colleagues who found lower dyspnea scores on four successive treadmill tests in nine patients with COPD despite comparable maximal exercise achievement (30). The Borg score for leg fatigue was not greater on the second test day. The range of Borg scores was quite large. There was a weak negative correlation between distance walked and the dyspnea score (r = −0.22) but none for the leg fatigue score. The average degree of dyspnea on study Day 1 was 5.01 ± 2.01, corresponding to a descriptor of “severe.” This is slightly lower than the value of 5.2 reported by Hamilton and colleagues during symptom-limited ergometer exercise testing in untrained patients with pulmonary disease (31). We found a lower mean Borg score for leg fatigue than dyspnea in our study population 3.45 ± 2.2. This is in contrast to two previous studies that have found similar or higher scores for leg fatigue at maximal exercise in pulmonary disease patients performing symptom-limited cycle ergometer exercise (31, 32). The difference between these findings may be related to the greater severity of disease in our patient population or the different type of exercise challenge.
Because the NETT study involved centers that used the same 6-minute walk protocol but differed in terms of walking course length and layout, it afforded us the opportunity to examine whether these factors affected the distance walked. We found that the three centers that used continuous walking courses, either oval or square, produced longer 6-minute walk distances than those with straight courses. The difference attributable to the walking course layout was 110 feet, an approximate 10% advantage for the continuous courses. Because the same subjects were not tested on different courses, we cannot be certain that there are not subtle differences in patient or other clinical center characteristics that account for this effect. However, we could not eliminate this difference by adjustment for age, sex, height, and FEV1 percentage predicted in the patient populations at the different clinical centers. Therefore, we believe it is a reasonable inference that the shape of the 6-minute walk course has an impact on the distance walked and supports the need for standardization of this variable on subsequent tests. We speculate that the advantage of continuous walking courses is due to the effort and time required for the test subject to turn around on a straight course. Timed treadmill distance testing, where an individual is free to alter the speed of the treadmill, might be considered comparable to a continuous walking course, but experimental evidence indicates that the treadmill tests result in shorter distances than straight courses, possibly because of difficulty in pacing accurately with a treadmill (26).
We also hypothesized that longer courses could offer an advantage over shorter courses for the same reason. However, the 6-minute walk distance was not longer in clinics with longer walking courses, which does not support this hypothesis. It is possible that the range of course lengths was insufficient to test this adequately or that our hypothesized mechanism for the effect of track layout in incorrect. Whichever explanation is correct, we note that literature reference values derived from a 20-m course are similar to those derived from a 50-m course (21, 27). Thus, it seems less important to standardize the length of the course as long as it exceeds the minimum of 50 feet, which was the shortest distance among the 14 centers with straight courses. The American Thoracic Society guidelines suggest that the minimum course length be at least 100 feet and suggests a straight rather than a continuous course (28).
In summary, we have found that the 6-minute walking test in patients with advanced emphysema is a measure that can be implemented in a large multicenter trial. There is a statistically significant improvement, averaging 7% when the test is repeated on a second day. The shape of the walking course (continuous versus straight) appears to be a determinant of distance walked but not the length of a straight course.
| 1. | Sciurba FC, Slivka WA. Six-minute walk testing. Semin Resp Crit Care Med 1998;9:383–391. |
| 2. | The National Emphysema Treatment Trial Research Group. Rationale and design of the National Emphysema Treatment Trial: a prospective randomized trial of lung volume reduction surgery. Chest 1999;116:1750–1761. |
| 3. | The National Emphysema Treatment Trial Research Group. Rationale and design of the National Emphysema Treatment Trial (NETT): a prospective randomized trial of lung volume reduction surgery. J Thorac Cardiovasc Surg 1999;118:518–528. |
| 4. | Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981;123:659–664. |
| 5. | Crapo RO, Morris AH, Clayton PD, Nixon CR. Lung volumes in healthy nonsmoking adults. Bull Eur Physiopathol Respir 1982;18:419–425. |
| 6. | Borg G. Psychophysical basis of perceived exertion. Med Sci Sports Exerc 1982;14:377–381. |
| 7. | Borg G. Psychophysical scaling with applications in physical work and the perception of exertion. Scand J Work Environ Health 1990;16:55–58. |
| 8. | Enright PL, Sherrill DL. Reference equations for the six-minute walk in healthy adults. Am J Respir Crit Care Med 1998;158:1384–1387. |
| 9. | Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310. |
| 10. | Jones PW, Quirk FH, Baveystock CM, Littlejohns P. A self-complete measure of health status for chronic airflow limitation: the St. George's respiratory questionnaire. Am Rev Respir Dis 1992;145:1321–1327. |
| 11. | Redelmeier DA, Bayoumi AM, Goldstein RS, Guyatt GH. Interpreting small differences in functional status: the six minute walk test in chronic lung disease patients. Am J Respir Crit Care Med 1997;155:1278–1282. |
| 12. | Criner GJ, Cordova FC, Furukawa S, Kuzma AM, Travaline JM, Leyenson V, O'Brien GM. Prospective randomized trial comparing bilateral lung volume reduction surgery to pulmonary rehabilitation in severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999;160:2018–2027. |
| 13. | Miyamoto S, Nagaya N, Satoh T, Kyotani S, Sakamaki F, Fujita M, Nakanishi N, Miyatake K. Clinical correlates and prognostic significance of six-minute walk test in patients with primary pulmonary hypertension: comparison with cardiopulmonary exercise testing. Am J Respir Crit Care Med 2000;161:487–492. |
| 14. | Hoeper MM, Schwarze M, Ehlerding S, Adler-Schuermeyer A, Spiekerkoetter E, Niedermeyer J, Hamm M, Fabel H. Long-term treatment of primary pulmonary hypertension with aerosolized iloprost, a prostacyclin analogue. N Engl J Med 2000;342:1866–1870. |
| 15. | Kadikar A, Maurer J, Kesten S. The six-minute walk test: a guide to assessment for lung transplantation. J Heart Lung Transplant 1997;16:313–319. |
| 16. | Cahalin LP, Mathier MA, Semigran MJ, Dec GW, DiSalvo TG. The six-minute walk test predicts peak oxygen uptake and survival in patients with advanced heart failure. Chest 1996;110:325–332. |
| 17. | Guyatt GH, Sullivan MJ, Thompson PJ, Fallen EL, Pugsley SO, Taylor DW, Berman LB. The 6-minute walk: a new measure of exercise capacity in patients with chronic heart failure. Can Med Assoc J 1985;132:919–23. |
| 18. | Guyatt GH, Pugsley SO, Sullivan MJ, Thompson PJ, Berman L, Jones NL, Fallen EL, Taylor DW. Effect of encouragement on walking test performance. Thorax 1984;39:818–822. |
| 19. | Hamilton DM, Haennel RG. Validity and reliability of the 6-minute walk test in a cardiac rehabilitation population. J Cardiopulm Rehabil 2000;20:156–164. |
| 20. | Tonino RP, Driscoll PA. Reliability of maximal and submaximal parameters of treadmill testing for the measurement of physical training in older persons. J Gerontol 1988;43:M101–M104. |
| 21. | Gibbons WJ, Fruchter N, Sloan S, Levy RD. Reference values for a multiple repetition 6-minute walk test in adults older than 20 years. Cardiopulm Rehabil 2001;21:87–93. |
| 22. | Leach RM, Davidson AC, Chinn S, Twort CH, Cameron IR, Bateman NT. Portable liquid oxygen and exercise ability in severe respiratory disability. Thorax 1992;47:781–789. |
| 23. | McGavin CR, Gupta SP, McHardy GJ. Twelve-minute walking test for assessing disability in chronic bronchitis. BMJ 1976;1:822–823. |
| 24. | Swinburn CR, Wakefield JM, Jones PW. Performance, ventilation, and oxygen consumption in three different types of exercise test in patients with chronic obstructive lung disease. Thorax 1985;40:581–586. |
| 25. | Knox AJ, Morrison JF, Muers MF. Reproducibility of walking test results in chronic obstructive airways disease. Thorax 1988;43:388–392. |
| 26. | Stevens D, Elpern E, Sharma K, Szidon P, Ankin M, Kesten S. Comparison of hallway and treadmill six-minute walk tests. Am J Respir Crit Care Med 1999;160:1540–1543. |
| 27. | Troosters T, Gosselink R, Decramer M. Six minute walking distance in healthy elderly subjects. Eur Respir J 1999;14:270–274. |
| 28. | American Thoracic Society. Guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002;166:111–117. |
| 29. | van Stel HF, Bogaard JM, Rijssenbeek-Nouwens LH, Colland VT. Multivariable assessment of the 6-min walking test in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:1567–1571. |
| 30. | Belman MJ, Brooks LR, Ross DJ, Mohsenifar Z. Variability of breathlessness measurement in patients with chronic obstructive pulmonary disease. Chest 1991;99:566–571. |
| 31. | Hamilton AL, Killian KJ, Summers E, Jones NL. Muscle strength, symptom intensity, and exercise capacity in patients with cardiorespiratory disorders. Am J Respir Crit Care Med 1995;152:2021–2031. |
| 32. | Killian KJ, Leblanc P, Martin DH, Summers E, Jones NL, Campbell EJ. Exercise capacity and ventilatory, circulatory, and symptom limitation in patients with chronic airflow limitation. Am Rev Respir Dis 1992;146:935–940. |
