Repeatability of the Six-Minute Walk Test and Relation to Physical Function in Survivors of a Critical Illness

Discussion

The main findings were that a small but consistent and significant average increase was observed in the second 6MWT distance at each assessment time point, that the reported change in 6MWT distance at weeks 8 and 26 was greater when the longer of the two 6MWT distances at these time points was used for analysis rather than the first 6MWT distance, and that the 6MWT distance was reflective of self-reported limitations to the SF-36 PF domain items of “climbing a flight of stairs” and “walking 100 m.”

Performing a second 6MWT in participants who survived a critical illness resulted in a significant average increase in walk distance at all assessment time points. The increase in walk distance when a second 6MWT was performed was similar to the findings for other patient cohorts, particularly those with chronic lung disease^1,2,12 and heart disease¹³; however, this is the first study that assessed repeatability of the 6MWT in people who survived a critical illness. This cohort may differ from cohorts with chronic respiratory or cardiac diseases due to younger mean age and potential for full recovery after a critical illness. The limits of agreement between test 1 and test 2 were similar to those reported in people with chronic obstructive pulmonary disease,¹⁴ demonstrating the inherent within-participant variability in the 6MWT results. Although the majority of the cohort walked farther on the second 6MWT at each time point, most likely due to a familiarization or learning effect, a substantial minority walked a shorter distance at the second test (Fig. 1). The reason for this variability is unclear, but it has been reported in other cohorts.¹⁴ Some contributing factors may have been lack of motivation, fatigue, and psychological issues.

The guidelines for the 6MWT⁵ recommend that a practice walk test (equivalent to test 1 in our study) should be performed in people naive to the test. Therefore, two 6MWTs should be performed at initial assessment. However, there are no recommendations for whether 2 tests should be performed at follow-up assessments for the estimate of group change in intervention research. Despite the smaller mean difference in distance walked in a second 6MWT at week 8 and week 26 compared with the difference at week 1, the difference that persisted had a significant impact on the estimate of change in 6-minute walk distance over time. If only one 6MWT had been performed at week 8 and week 26, the reported change in 6-minute walk distance (compared with week 1) would have been approximately 19 m less at week 8 and week 26 than if the longer of two 6MWT distances was used. Thus, the change in 6-minute walk distance during recovery from a critical illness would have been underestimated. This underestimation could potentially represent a substantial proportion of the effect of an intervention and result in concluding that an effective intervention was ineffective, especially if effectiveness was based on reaching a minimum clinically important difference in 6MWT distance, such as 35 m, as has been reported for people with chronic obstructive pulmonary disease.¹⁵ Our findings suggest that for research evaluating the effects of an intervention over time, it is important to perform two 6MWTs at each subsequent assessment and to use the longer test distance for analysis in reporting change in 6-minute walk distance.

A novel aspect of this study was that the 6MWTs were performed at participants' homes or in their local environment. This home environment meant that walking tracks for the 6MWT differed among participants. For example, some indoor tracks were straight “shuttle” tracks along corridors and others were circular tracks through rooms, and some outdoor tracks were straight shuttle tracks along flat pavements and others were circular tracks in flat parks. Within-participant variability was controlled by duplicating the track used for each assessment.

When examining the effect of different track types, the main finding was that only track length affected the 6-minute walk distance, with longer tracks resulting in greater distance walked. The impact of slowing down to turn more frequently on shorter tracks may explain this finding.⁵ Our findings differed from those of a study of people with severe chronic obstructive pulmonary disease,² where straight track length had no effect on walk distance and a circular track resulted in greater 6-minute walk distance than a straight shuttle track. However, those findings could conversely support our finding that track length affected walk distance, as, by its nature, a circular track is longer than any straight track.

The SF-36 PF score correlated significantly with 6-minute walk distance at each time point (strongest at week 1). Similar positive correlations between self-reported physical function and 6-minute walk distance have previously been reported in other cohorts.^16,17 The week 1 SF-36 PF score was a significant predictor of 6-minute walk distance, adding 26% to the variance in 6-minute walk distance explained by other variables (age, sex, APACHE II score, and days in hospital). The closer relationship between the SF-36 PF score and 6-minute walk distance at week 1 compared with weeks 8 and 26 was likely due to the greater limitation in walk distance and physical functioning at this time point. At subsequent time points, recovery from a critical illness may have resulted in a ceiling effect in scoring of the SF-36 and the 6MWT. Both factors would reduce the strength of the relationship between SF-36 PF score and 6-minute walk distance. An interesting finding was that the SF-36 PF domain items of “walking one flight of stairs” and “walking 100 m” were able to distinguish clear differences in 6-minute walk distance at week 1. This may be a useful finding, as it suggests that these items may act as a surrogate for the 6MWT at this early stage of recovery from a critical illness.