Department of Physiotherapy, Maastricht University Medical Center, Maastricht, The Netherlands

Original Scientific Paper Six-minute walking test done in a hallway or on a treadmill: how close do the two methods agree? Antoine F. Lenssen, Lambert C.A.M. Wijnen, Dion G. Vankan, Bart H. Van Eck, Danielle P. Berghmans and George M. Roox Department of Physiotherapy, Maastricht University Medical Center, Maastricht, The Netherlands Received 21 December 2009 Accepted 21 March 2010 Background The 6-min walking test (6-MWT) is probably the most widely used test to measure the functional capacity in cardiac rehabilitation. Although the American Thoracic Society recommends testing on a flat surface, treadmills are also used for testing. Therefore, we want to investigate the interchangeability of results of treadmill and hallway 6-MWT in a population of patients participating in a cardiac rehabilitation programme. Design Preexperimental design. Setting University hospital Department of Cardiology and Physiotherapy. Participants Patients entering the cardiac rehabilitation programme of the Maastricht University Cardiology Department. Main outcome measure Agreement in 6-min walking distance between the hallway and treadmill test results were calculated by taking the mean difference between the two methods and the 95% confidence interval of the difference and plotting this against the average of the two test results. A Bland and Altman plot was constructed, showing the mean difference and the 95% limits of agreement between the two methods. Results Sixty-nine patients participated in this study. Mean difference between walking on a treadmill and walking in a hallway was 9 m in favour of the hallway test. The 95% limits of agreement were ± 118 m. Conclusion Results of the 6-MWT conducted in a hallway or on a treadmill are not interchangeable, because of large between-test variations in the distances walked by individual participants. Eur J Cardiovasc Prev Rehabil 17:713 717 c 2010 The European Society of Cardiology European Journal of Cardiovascular Prevention and Rehabilitation 2010, 17:713 717 Keywords: cardiac rehabilitation, 6-minute walking test, treadmill and hallway Introduction The 6-min walking test (6-MWT) is probably the most widely used test to measure the functional capacity in cardiac rehabilitation [1,2]. It can measure status, and can also be used as a measure of change over time, that is, assessing the effects of rehabilitation. The potential advantage of the 6-MWT is that it is inexpensive, easy to perform, and does not require specialized equipment [3 8]. However, it does require a physical space that allows the test to be done. The American Thoracic Society (ATS) guidelines for the 6-MWT states that the Correspondence to Antoine F. Lenssen, PhD, Department of Physiotherapy, Maastricht University Medical Center, P. Debijelaan 25, 6202 AZ Maastricht, Limburg, The Netherlands Tel: + 31 433 875 145; fax: 31 433 875 142; e-mail: AF.Lenssen@mumc.nl test is best done indoors, along a long, flat, straight, enclosed corridor with a hard surface that is seldom travelled. The walking course should be at least 30-m long [9]. The degree of reproducibility of the 6-MWT has been examined to some extent. Most investigators [4,6,10] have reported high test retest reliability scores (intraclass correlation coefficient 0.75 0.98). However, agreement between the results of two tests has been far less examined. Nogueira et al. [10] reported mean differences and 95% limits of agreement between the two tests of 6 ± 41 m. Reproducibility of an outcome measure influences its ability to detect change. Redelmeier et al. [11] reported 1741-8267 c 2010 The European Society of Cardiology DOI: 10.1097/HJR.0b013e32833a1963

714 European Journal of Cardiovascular Prevention and Rehabilitation 2010, Vol 17 No 6 that the smallest difference in test results associated with a noticeable clinical difference is 54 m, whereas O Keeffe et al. [12] reported that a difference of 43 m is associated with clinical change. Opasich et al. [13] calculated that the minimum variation in the distance walked that can be considered an expression of a real variation of functional capacity is approximately 10% of the average result of the two tests. This means that if the 6-MWT is used as an endpoint in intervention studies, the clinical relevance of the results must be considered with some caution. If the variation in distance is less than 10% or 43 54 m in individuals, interpretation of the results as indicating real changes might be incorrect, even if the difference could be regarded as statistically significant. Although the ATS [9] does not recommend the use of a treadmill, as it is more difficult for participants to pace themselves on a treadmill, the advantage of testing on a treadmill is that it saves space and allows constant monitoring during the exercise [5,14,15]. In the Netherlands most outpatient physiotherapy is provided by therapists working in private practices, in which a hallway complying with the recommendations of the ATS protocol is not always available. This is the main reason why 6-MWTs are often conducted on a treadmill. As the ATS does not recommend testing on a treadmill, and in most hospitals the first tests may be conducted in a hallway, there is a problem of interchangeability of tests, that is, it is unclear whether the method by which the measurement was made can be ignored. The reliability of the 6-MWT conducted on a treadmill has been examined [14,15] and seems to be quite similar to that of the results of tests in free space, with reported intraclass correlation coefficient values of 0.86 0.88. Most studies into the reproducibility of the 6-MWT have been based on correlation. However, there are two problems with the use of correlation [16] to examine reproducibility. First, correlation depends on the range and distribution of the variables in the target population. Second, correlation looks at the degree of association, not agreement, and is thus more suitable for measuring reliability than for agreement between tests [16,17]. Unfortunately, not a single study has so far included an analysis to investigate how closely two 6-MWT measurements agree. Only one study has compared testing in a hallway with treadmill testing. Stevens et al. [5] reported that standardized treadmill testing is feasible and reliability parameters are similar to those of hallway test results. However, the test results themselves were significantly different and therefore not interchangeable. Participants walked farther during hallway testing than in treadmill tests. Stevens et al. argued that one of the possible reasons for the difference found could be the unfamiliarity of participants with the treadmill test, so that test results might be better if participants were given time to become comfortable with walking on a treadmill and if a strict protocol was used to encourage participants to adjust their walking speed on the treadmill as appropriate. The objective of this study was to investigate the interchangeability of results of treadmill and hallway 6-MWT and present this with data on agreement rather than reliability. Our hypothesis was that differences between the two methods should not exceed the limits of agreement (43 54 m [11,12]) found in test retest experiments in a hallway by more than 10% to allow the results of these two tests to be used interchangeably. Methods Participants Between 1 December 2008 and 1 July 2009, consecutive eligible patients entering the cardiac rehabilitation programme of the Maastricht University Cardiology Department, after having suffered an acute myocardial infarction or after coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, valve replacement or a combination of the above, were invited to participate in the study. Patients with comorbidity affecting their gait pattern were excluded, as were patients who had major problems walking on a treadmill. The local medical ethics committee reviewed and approved the study procedures. Measurements For the hallway test, patients were instructed to walk a 44-m rectangular course at their own preferred pace. They were instructed to cover as much ground as possible in 6 min. Patients were not encouraged to increase their speed during the 6 min of testing. The only information that was given during the first 2 min of the test was the time that had elapsed, and patients were told that they were doing well every minute. After 3 min, patients were additionally told that they were now halfway through the test, and for the next 2 min they were given information on the remaining test duration, that is, 2 and 1 min, respectively. At the conclusion of each walk, the total distance covered by the patient was recorded along with a Borg Fatigue scale rating (Borg scale ranging from 7 to 20). Heart rate and oxygen saturation were measured using a portable oxygen saturation monitor (Nonin, Onyx 9550; Nonin Medical Inc. Plymouth, Minesota USA). The treadmill test was carried out on a Biodex gait trainer 2 (Biodex Medical Systems, Shirley, New York, USA). The total distance covered was stored in the equipment s software programme, and the distance covered was not

Six-min walking test performed Lenssen et al. 715 visible to the patients during the test. Only the start, stop and speed up or slow down buttons were accessible to the patients. The initial speed was set at 3 km/h. The patients were given the same information about the goal of testing as in the hallway test protocol. Patients were instructed how to increase and decrease the speed of the treadmill and how to stop if they needed a rest. The rest of the test procedure was similar to that of the hallway test, except for the inclusion of a 3-min trial walk to get comfortable with walking on a treadmill. A 3-min rest was inserted between the trial walk and the 6-MWT on the treadmill. Patients were excluded if, during or after the 3-min trial walk, they reported great difficulty with treadmill walking. All tests were supervised by one of the four therapists who were experienced in conducting 6-MWT on a treadmill and in a hallway. Therapists conducting the hallway tests were unaware of the distance covered during the treadmill test. The treadmill test was always performed first, and the hallway test was taken within 1 week of the treadmill test to ensure stability in the patient s functional capacity. Table 1 Patient characteristics Number of patients 69 Age (years) (mean SD) 61 (16.0) Men/women 62/7 Diagnosis CABG 37 ACS (infarction) 18 PTCA 5 AVR 3 Combination 6 Results 6 MWT hw + % norms a (mean + SD) 547 (103.6) 78.2 (12.4) 6 MWT tm + % norms a (mean + SD) 538 (124.2) 77.3 (15.6) ICC 0.89 Limits of agreement (metres) 9 ± 118 m Heart rate after hallway test (mean + SD) 90 (19) Heart rate after treadmill test (mean + SD) 84 (16) Borg after hallway test (mean + SD) 12 (1.3) Borg after treadmill test (mean + SD) 12.5 (1.2) BP after hallway test (mean + SD) 147/82 (22.2/11.3) BP after treadmill test (mean + SD) 143/82 (23.1/10.4) ACS, acute coronary syndrome; AVR, aortic valve replacement; BP, blood pressure; CABG, coronary artery bypass graft; ICC intraclass correlation coefficient; 6-MWT hw, 6-min walking test in hallway; 6-MWT tm, 6-min walking test on treadmill; PTCA, percutaneous transluminal coronary angioplasty; SD, standard deviation. a Percentage of predicted value derived from hallway test [18]. Analyses Results of the tests are reported as means and standard deviations of walked distances. Agreement between the two different methods of measurement was examined by constructing a 95% limits of agreement plot, by taking the difference between the results obtained with the two methods for each patient and plotting this difference against the average of the two methods. The mean and standard deviations of the difference provided information on the difference between the results of the two methods used on the same patient. The mean difference with 95% limits of agreement is presented in a Bland and Altman plot [17]. A histogram of the differences was added to check for normality of the difference between the two measurements [16]. Results Between 1 December 2008 and 1 July 2009, 74 patients were recruited for participation in the study. Two patients were excluded because they were incapable of being tested twice within 1 week. Three patients were excluded because they had reported difficulties with treadmill walking during the 3-min pretest training, leaving 69 patients for analyses (Table 1). The majority of the participants (62) were men, and their mean age was 61 years, ranging from 37 to 81. Most patients were attending the programme after coronary artery bypass grafting (34 patients) or after having had an acute myocardial infarction (24 patients). Functional capacity deficits were moderate in this population, with walking distances being at least 20% below the normative values for healthy volunteers described by Troosters et al. [19]. Fig. 1 150 100 50 0 200 50 300 400 500 600 700 800 100 150 200 Bland and Altman plot: difference in walking distance in metres in a hallway and on a treadmill (vertically) plotted against the mean walking distance in metres, with 95% limits of agreement (horizontally). Mean walking distance in the hallway test (547 ± 103 m) was marginally greater than the distance measured on the treadmill (538 ± 124 m). Borg rating of perceived exertion was comparable in both test situations (12.0 vs. 12.5). Mean 6-MWT distance was 78% (range: 41 104%) of the walking distance calculated from 6-MWT standards [19]. In 45 cases, the difference between hallway and treadmill distance was less than 54 m (Fig. 1). The mean difference and limits of agreement for these two measurements was 9 ± 118 m (Fig. 2). Figure 2 shows that the difference between hallway and treadmill testing seems to show a normal distribution, which is a requirement for Bland and Altman plots to be used [16].

716 European Journal of Cardiovascular Prevention and Rehabilitation 2010, Vol 17 No 6 Fig. 2 12 8 4 0 10 000 0 10000 Histogram of the difference between hallway and treadmill results (hallway treadmill). Discussion This study investigated the interchangeability of results of 6-MWT on a treadmill and in a hallway. The results show that there was considerable variation in the agreement between the two tests. The limits of agreement greatly exceeded those reported for test retest agreement in hallway testing [10]. Our 118 m against the 41 or 66 m reported by Nogueira et al. [10] represents a far greater difference in test retest variability than the 10% increase in variability, which we assumed to be acceptable for the two test procedures to be used interchangeably. Redelmeier et al. [11] and O Keeffe et al. [12] reported that a noticeable clinical difference between the patients would be indicated by a difference in the walking distance of between 43 and 54 m. Using these two potential cutoffs on our data would mean that 34 and 24, respectively, of the 69 participants would show a clinically significant difference between hallway and treadmill distances, even though the protocol was designed to ensure functional stability. Unlike Stevens et al. [5], we found no significant differences between mean walking distances in the hallway and treadmill tests. Nevertheless, our results support their conclusion that results of the two test methods are not interchangeable. In addition to the outcomes reported above, we also measured blood pressure and heart frequency before and immediately after testing. We found that the mean heart rate after treadmill testing was lower compared with the heart frequency than that after hallway testing, with values of 84 and 90, respectively. The mean Borg Fatigue rating was comparable in both situations: 12 and 12.5, respectively. The difference in heart rate may be explained by the nature of walking on a treadmill, where propulsion is triggered by movement of the tread, which may save energy. There are several explanations for the main results of our study. As there was no difference in mean walking distance, we cannot conclude that treadmill walking is more difficult and requires more skill than hallway walking. Chang et al. [18] reported that there is no difference in walking dynamics while walking on a comfortable paced walk. However, the greater variability in walking distance in the treadmill 6-MWT might stem from unfamiliarity with walking on a treadmill, making the participants less likely to change the initially set walking speed. The difference might also be caused by the use of an uncomfortable walking speed. Chang et al. [18] reported no difference in walking dynamics while walking on a comfortable paced walk; walking with maximum speed may also have led to a greater variability in distances walked. Reports on the reproducibility of 6-MWT on a treadmill and in a hallway show that this seems to be similar. Unfortunately, most researchers have reported on reliability, and failed to report on agreement between the tests. Although it is clear that the agreement between our test results from hallway and treadmill 6-MWT is far smaller than the reported agreement between repeated hallway tests, analyses of reliability using correlation yield results similar to those reported by others. Analysing our dataset with correlation would yield a significant correlation between hallway and treadmill test results of 0.89. However, as stated in the Introduction section, we tend to follow Bland and Altman plot [16] in the belief that correlation is not the right technique to analyse agreement between two methods of measurement. We decided to adapt the ATS protocol [9] to our local situation by having participants walk a rectangular 44 m course instead of walking back and forth in a hallway. This may have led to somewhat larger walking distances in our hallway-walking test, as participants did not need to turn 1801 but only 901. As participants do not turn on a treadmill at all, walking a rectangular course may make both applications of the test more comparable. However, Stevens et al. [5] still reported larger walking distances for hallway compared with treadmill testing, which we could not affirm. We did not use a full exercise test before testing on the treadmill; instead, we used a 3-min warm-up test.

Six-min walking test performed Lenssen et al. 717 This may well be why we had to exclude three participants who were unable to walk on a treadmill at their maximum voluntary speed after the 3 min of warm-up testing. We did not use a 3-min exercise period before testing in the hallway. Although we believe that participants are aware of the task in a 6-MWT on level ground, dropping the pretest training may also contribute to the measurement error found in our study. For practical reasons, the treadmill test was always conducted before the hallway test. This might have introduced bias from a learning effect. However in contrast to Stevens et al. [5] we did not find a major difference in walking distance between both the test procedures. Randomization of test procedures might have diminished the mean between test differences we found. To ensure that the results reflected with the clinical care situation, tests were conducted by all therapists at our department working with the target population. Although this might have led to greater variability between tests, we did not detect significant differences between testers in secondary analyses. Conclusion The 6-MWT on a treadmill and in a hallway are both feasible in a population of cardiac rehabilitation patients. However, results of hallway and treadmill tests are not interchangeable, because of a large individual variability and thus high disagreement between the two test situations in terms of distances walked. We therefore recommend that cardiac rehabilitation teams must discuss in advance which testing protocol will be used, and that such groups should not allow patients to be tested with different testing methods if they wish to evaluate the progress of patients recovery. References 1 Vogels EMHM, Bertram RJJ, Graus JJJ, Hendriks HJM, Hulst van R, Voort van der S, et al. KNGF-guidelines for physical therapy in cardiac rehabilitation. Supplement Nederlands Tijdschrift Voor Fysiotherapie 2005; 3:1 58. 2 Olsson LG, Swedberg K, Clark AL, Witte KK, Cleland JG. Six minute corridor walk test as an outcome measure for the assessment of treatment in randomized, blinded intervention trials of chronic heart failure: a systematic review. Eur Heart J 2005; 26:778 793. 3 Faggiano P, D Aloia A, Gualeni A, Brentana L, Dei Cas L. The 6 min walking test in chronic heart failure: indications, interpretation and limitations from a review of the literature. Eur J Heart Fail 2004; 6:687 691. 4 Demers C, McKelvie RS, Negassa A, Yusuf S. Reliability, validity, and responsiveness of the six-minute walk test in patients with heart failure. Am Heart J 2001; 142:698 703. 5 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 (5 Pt 1):1540 1543. 6 Kervio G, Ville NS, Leclercq C, Daubert JC, Carre, F. Intensity and daily reliability of the six-minute walk test in moderate chronic heart failure patients. Arch Phys Med Rehabil 2004; 85:1513 1518. 7 De Feo S, Tramarin R, Lorusso R, Faggiano P. Six-minute walking test after cardiac surgery: instructions for an appropriate use. Eur J Cardiovasc Prev Rehabil 2009; 16:144 149. 8 Heijblom KG, Aufdemkampe G, Peters RHJ, van der Bolt CLB. Six minute walking test in patients with heart failure, a Dutch study. In: Smits-Engelsman BCM, van Ham I, Aufdemkampe G, editors. Jaarboek fysiotherapie kinesiotherapie. Houten: Bohn Stafleu Van Loghum; 1998, pp. 128 153. 9 Brooks D, Solway S, Gibbons WJ. ATS statement on six-minute walk test. Am J Respir Crit Care Med 2003; 167:1287. 10 Nogueira PA, Leal AC, Pulz C, Nogueira ID, Filho JA. Clinical reliability of the 6 min corridor walk test performed within a week of a myocardial infarction. Int Heart J 2006; 47:533 540. 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 O Keeffe ST, Lye M, Donnellan C, Carmichael DN. Reproducibility and responsiveness of quality of life assessment and six minute walk test in elderly heart failure patients. Heart 1998; 80:377 382. 13 Opasich C, Pinna GD, Mazza A, Febo O, Riccardi PG, Capomolla S, et al. Reproducibility of the six-minute walking test in patients with chronic congestive heart failure: practical implications. Am J Cardiol 1998; 81:1497 1500. 14 Guimaraes GV, Carvalho VO, Bocchi EA. Reproducibility of the selfcontrolled six-minute walking test in heart failure patients. Clinics (Sao Paulo) 2008; 63:201 206. 15 Laskin JJ, Bundy S, Marron H, Moore H, Swanson M, Blair M, Humphrey R. Using a treadmill for the 6-min walk test: reliability and validity. J Cardiopulm Rehabil Prev 2007; 27:407 410. 16 Bland JM, Altman DG. Applying the right statistics: analyses of measurement studies. Ultrasound Obstet Gynecol 2003; 22:85 93. 17 Bland JM, Altman DG. Comparing two methods of clinical measurement: a personal history. Int J Epidemiol 1995; 24 (Suppl 1):S7 S14. 18 Chang MD, Shaikh S, Chau T. Effect of treadmill walking on the stride interval dynamics of human gait. Gait Posture 2009; 30:431 435. 19 Troosters T, Gosselink R, Decramer M. Six min walking distance in healthy elderly subjects. Eur Respir J 1999; 14:270 274.