Gait tests in multiple sclerosis: Reliability and cut-off values

Background: Gait limitation is one of the most common disabilities in people with multiple sclerosis (MS). Several studies have used gait parameters to determine the effects of different therapies. However, few studies have determined their reproducibility, also the therapeutic effects could be overestimated. Research question: To examine the reproducibility in gait measurements during short and long distances. Methods: In this cross-sectional study we recruited a group of MS patients and compare it to a control group. The participants performed the following tests in a fixed order: a 25-foot walk at a comfortable speed, at a fast speed and during a dual task, a timed up-and-go test (TUG) and a six- minute walk test (6MWT). Two measurements were conducted a week apart. Systematic error was evaluated by the Student t-test, reliability by the infra-class correlation coefficients (ICC) and agreement by the minimum detectable change (MDC95). Results: A total of 58 people with MS and 19 healthy people were included. The absence of systematic error was only found for the fast speed condition. The reliability of the gait parameters had moderate to high ICC values (ICC > 0.7) except for the dual task cost (DTC) which was 0.45. The MDC95 was higher in people with MS compared to healthy people, and it was higher in people with MS for gait speeds in all conditions ( > 34%). For the TUG and 6MWT, the MDC95 were 51.5% and 31.7% respectively. For people with MS the smallest MDC95 was found for the stance time for all conditions (6.8%), whereas the highest was found for the dual task cost (158.7%). Significance: The MDC 95 values were higher than the cut-off point based on the minimally important clinical difference (MICD) proposed in previous studies. Thus, the MDC95 should be used as a cut-off rather than MICD values.