Kirsten Lykke Knak, Linda Kahr Andersen, Nanna Witting, John Vissing
Rigshospitalet, Department of Neurology, Copenhagen Neuromuscular Center, Rigshospitalet, DK-2100 Copenhagen, Denmark. E-mail: kirsten.lykke.knak@regionh.dk
DOI: 10.2340/16501977-2222

Objective: The 2- and 6-minute walk tests are used to evaluate walking capacity, but reliability has been sparsely investigated in patients with neuromuscular diseases. The aim of this study was to investigate the relative and absolute reliability of the 2- and 6-minute walk tests in patients with neuromuscular diseases.
Design: Each patient performed a 2- and a 6-minute walk test on 2 test days separated by 1–2 weeks.
Subjects: A total of 93 adult patients (mean age 53 years, age range 22–83 years) with 12 different neuromuscular diseases were included.
Results: The mean walking distance increased by 4. 3 and 11. 2 m (p < 0. 001) in repeated 2- and 6-minute walk tests, respectively. Intraclass correlation coefficient in the 2- and 6-minute walk tests was 0. 99 (p < 0. 001). Standard error of measurement was 4. 9 m in the 2-minute walk test and 14. 0 m in the 6-minute walk test. Minimal detectable difference was 13. 7 m in the 2-minute walk test and 38. 8 m in the 6-minute walk test.
Conclusion: These findings show good relative reliability of the 2- and 6-minute walk tests in patients with neuromuscular diseases. However, absolute reliability demonstrated variability in neuromuscular diseases. This should be considered when interpreting a change in walking distance.

Reliability of the two- and six-minute walk tests in neuromuscular diseases
To accurately assess disease progression and efficacy of treatment, endpoints to monitor this need to be reliable. We studied the reliability of the two- and six-minute walk tests in 93 patients with neuromuscular diseases. In conclusion, the two- and six-minute walk tests showed excellent relative reliability. Thus, the tests can be used as endpoints to follow neuromuscular diseases, but the change in walking distance must exceed the presented measurement error level in order to indicate a genuine change and not random error.