The inclusion criteria of the review were clear. Relevant databases were searched. Efforts were made to find both published and unpublished studies, minimising the potential for publication bias. Publication bias was assessed where possible, but use of a funnel plot to assess publication bias in the small number of included studies might not have been appropriate. Only English-language studies were sought, introducing the potential for language bias. Steps were taken to minimise reviewer biases and errors by having more than one reviewer undertake the study selection and data extraction, but it was unclear whether the process of validity assessment was performed in duplicate.
Relevant criteria were used to assess the quality of clinical controlled trials, but no formal validity assessment was performed for observational studies. Given the diversity of included studies, pooling the results from studies with different types of study design might not have been appropriate. The authors reported that there was insufficient data to pool mean change in Expanded Disability Status Scale; however, some data were presented in tables, but no narrative was provided.
Given the limited quality and small number of included studies, and high levels of statistical heterogeneity between studies for some outcomes, the authors' conclusions should be interpreted with caution.