Twenty-two studies (n=600) were included the meta-analysis. Ten studies had a control condition. Sample sizes varied from six to 111. The included studies were generally of poor to moderate quality (quality score ranged from 9 to 15).
When the studies were pooled, exercise training intervention was associated with a significant improvement in walking mobility in patients with multiple sclerosis (weighted mean ES 0.19, 95% CI 0.09 to 0.28; 22 studies). No evidence of heterogeneity was observed for this outcome.
For subgroup analyses, there were larger effects associated with supervised exercise training (weighted mean ES 0.32, 95% CI 0.19 to 0.44; 17 studies), exercise programmes with a duration of less than three months (weighted mean ES 0.28, 95% CI 0.15 to 0.41; 14 studies) and a mixed group of patients with relapsing/remitting and progressive multiple sclerosis (weighted mean ES 0.52, 95% CI 0.08 to 0.96; three studies). The type of study design (with or without a control condition) did not influence the overall effect of exercise training on walking mobility.
No evidence of publication bias was found according to the visual scanning of the funnel plot and the fail-safe N analysis.