To the Editor:
We read with great interest the article by Steffen et al titled “Age- and Gender-Related Test Performance in Community-Dwelling Elderly People: Six-Minute Walk Test, Berg Balance Scale, Timed Up & Go Test, and Gait Speeds” in the February 2002 issue of Physical Therapy. The effort to give insight into normal performances on frequently used clinical tests such as the Six-Minute Walk Test, the Berg Balance Scale, the Timed Up & Go Test, and measures of comfortable and maximal gait speed is very much appreciated.
The authors investigated a group of patients that is highly relevant to physical therapist practice–community-dwelling elderly people who are 60 to 90 years of age–using different easy-to-perform tests, including the Six-Minute Walk Test. The Six-Minute Walk Test is an interesting test in patients with chronic diseases. Although often regarded as a submaximal test, it results in a steady-state oxygen uptake (Vo2), at near peak Vo2 in patients with severe obstructive lung disease.1 This test also has been shown to be related to muscle function rather than cardiac or ventilatory impairment, even in patients with severe lung disease.2
We agree with the authors that normative values are of utmost importance in interpreting the results of the Six-Minute Walk Test. In addition, standardization issues are critical for comparison of test results.3 Although we believe that age and sex are important in the interpretation of the measurements, other factors also may be important. Our research has indicated that height was the most important marker of walking distance.4 In addition, age, weight, and sex all together explained approximately 60% of the variance in the 6-minute walking distance.4 The variables contributing to our model were similar to those proposed in other studies of subjects without known impairments5–7 and are those that generally explain variability in exercise capacity. Anthropometric characteristics, therefore, help explain variability in the results of tests such as the Six-Minute Walk Test.
Second, the number of walking tests performed needs attention. In subjects without known impairments, we4 and others7 found a systematically increased second 6-minute walking distance (+47 m). Although this increase represents barely 7% of the predicted value, this is an important difference when one is interested in treatment effects. For the 6-minute walking distance, the minimally clinically important difference is estimated to be 54 m,8 which approaches the difference that may be due simply to a learning effect. As pointed out by the authors, the relatively low walking distance (compared with other normal values, as discussed below) reported by Enright and Sherrill5 could be attributable to the lack of practice. Interestingly, by applying a practice walk and giving rigorous encouragement, the variance explained by our regression equation was substantially higher (66% versus 40% of variance explained). Our findings were replicated by an independent group.7 These authors also confirmed the underestimation of the normal walking distance, using the regression equation proposed by Enright and Sherrill,5 by 108 m (SD=74).
A last issue that, in our opinion, merits further discussion is the issue of encouragement. This topic is discussed in depth by Guyatt and coworkers9 in one of the first publications on the Six-Minute Walk Test. We believe, based on the article by Guyatt et al, that encouragement should be standardized. Typically, encouragement is given every 30 seconds.10 Steffen and colleagues gave encouragement after 1 minute, 3 minutes, and 5 minutes. Using the regression equation proposed by our group, the male subjects studied by Steffen et al achieved a 6-minute walking distance of roughly 85% of the predicted value, and the female subjects achieved a 6-minute walking distance of 93% of the predicted value (calculated from their mean values). Besides maybe local differences and a clear difference in patient selection (we did select subjects without known impairments who were not involved in competitive sports), encouragement could have affected the results.
The article by Steffen and coworkers contributes to our understanding of what we consider a functional test. Comparing their data with ours suggests that patient encouragement during the testing procedure may importantly affect the results of the test. To allow comparability of results across trials and to avoid the development of several variants of the same test,3 we would propose having subjects perform a practice walk and giving standard encouragement (every 30 seconds) as a routine procedure.
- Physical Therapy
We appreciate the response of Dr Troosters, Dr Gosselink, and Dr Decramer. This response expands upon our discussion of the Six-Minute Walk Test (6MWT).
We agree with Troosters and colleagues regarding the importance of standardization of the 6MWT. We also agree that a practice trial is necessary to produce optimal walking distances in the 6MWT, which is why a practice trial was included in our study. We did, however, find a more modest increase in distance (3%) on the second trial than that found by Troosters et al1 (7%). We are also in agreement that encouragement is critical to produce maximum walking distances. Although Guyatt and coworkers2 have shown that encouragement every 30 seconds improves walking distances versus no encouragement, there has been no research indicating that encouragement every 30 seconds is more helpful than encouragement every few minutes. Indeed, careful examination of the review by Solway et al3 demonstrates that there is little consensus on the use of encouragement or the amount. Of the 30 6MWT references shown in the tables in the review by Solway et al,3 in 13 references, there was no use of encouragement or there was encouragement given at intervals greater than 30 seconds; in 8 references, encouragement was given but the authors did not state how often; and in 9 references, encouragement was given at 30-second intervals. Additionally, using data that we are preparing for publication, we found that peak heart rates during the walk test were on average 77% (SD=12%) of the predicted maximum heart rate. This exactly matches the average effort put forth by the subjects in the study by Troosters et al.1 Thus, it is unlikely that frequency of encouragement was the cause of our shorter walking distances.
Troosters and colleagues point out that subject height is the most important marker in explaining variance in 6MWT results. Although this may be true in the subjects in their study,1 it was not true for our data. Multiple regression analysis of our data (using age, height, and weight) yielded an r2 value of .54, with age being the most important variable. There are several differences in the sample characteristics of the 2 studies. In our study, the average age of the subjects was 73 years, the mean body mass index was 28 for male subjects and 29 for female subjects, and there were few exclusions for health status. In the study by Troosters et al,1 the mean age of the subjects was 65 years, the mean body mass index was 26, and only subjects without health problems were included. Indeed, we believe that, if Troosters et al1 selected only subjects that fit into our age and sex categories, these subjects would likely fit into our confidence intervals. One problem with predicting reference values from a regression equation obtained from a population that does not include these subjects is that the regression line may no longer fit to the extrapolated data.
Finally, our major intent was to provide simple tables for therapists to be able to quickly determine whether a given patient's test results match those of a cohort matched by age and sex. We hope that, over the next few years, additional data on older individuals will be presented in the literature so that true normal values are available for therapist use.
We would like to thank Troosters and colleagues for their interest in our study and for the continuing dialogue related to the 6MWT.