Mobility and balance rehabilitation in multiple sclerosis: A systematic review and dose-response meta-analysis.

Balance problems are common in multiple sclerosis and affect walking and daily safety, so understanding whether rehabilitation actually helps is important for planning care. This systematic review and meta-analysis pooled randomised trials testing balance-focused rehabilitation against either no specific intervention or another active control in people with multiple sclerosis, drawing on 71 studies overall and individual-level data from 20 studies covering just over a thousand participants. The pooled result showed a medium-sized improvement in balance scores favouring the rehabilitation interventions, and a related analysis using a common balance scale found a measurable point improvement in favour of treatment. A further analysis suggested that sessions with higher intensity, meaning more minutes per session, and task-specific training were associated with somewhat better outcomes.

These dose-response findings come from meta-regression across a limited number of studies, so they should be read as an association rather than a proven causal recipe, and the review notes considerable variability in how trials were designed and controlled.

If you live with multiple sclerosis, balance difficulties can make everyday movement and walking feel less secure, so it matters whether rehabilitation exercises genuinely help. Researchers pulled together many small and medium-sized randomised trials of balance and mobility rehabilitation in people with multiple sclerosis, comparing those who did these exercises against people who received no specific balance training or a different type of care. Across studies with detailed data on about a thousand people, those who did balance rehabilitation showed a moderate improvement in balance test scores compared with the control groups.

A closer look at trials using a standard balance test also found a noticeable improvement favouring the exercise groups. The review further found a pattern where longer training sessions and exercises that mimicked real daily tasks tended to go along with better results, though this part of the analysis is based on fewer studies and shows an association rather than a guarantee.

Overall, this adds reasonably solid evidence that structured balance rehabilitation can improve balance scores in multiple sclerosis, but differences between trials in how programmes were designed and controlled mean individual results may vary.

This is a summary of research, not medical advice. Talk to your own healthcare provider about anything affecting your care.

This review speaks directly to the clinical question of whether balance-specific rehabilitation produces measurable gains in people with multiple sclerosis, and whether intensity or task specificity influences the size of that benefit. Pooling individual-level data from 20 randomised trials and 1016 participants, the authors report a medium effect size (standardised mean difference 0.41, 95% confidence interval 0.22 to 0.59) favouring balance intervention over no specific intervention or an active comparator.

A separate analysis of 14 trials with 696 participants using the Berg Balance Scale as the primary outcome found a mean difference of 3.58 points (95% confidence interval 1.79 to 5.38), reinforcing the direction of effect on a widely used clinical measure. The meta-regression across the same 20 studies found that higher session intensity (minutes per session) and task-oriented training were each associated with better outcomes, which supports general principles of neurological rehabilitation but represents an exploratory dose-response analysis rather than a tested protocol comparison. Study comparators were mixed, with roughly half of trials having no active control, so part of the observed benefit may reflect attention or participation effects rather than the specific content of balance training.

The review draws on 71 included trials overall but pooled only a subset for formal dose-response modelling, and the abstract does not report certainty ratings or detailed risk-of-bias findings, so confidence in the precise dose-response relationship remains moderate.