Design: Systematic review and meta-analysis of randomised trials
Population: 20 RCTs, 714 people with Parkinson disease
Primary outcome: UPDRS-III, freezing of gait, Berg Balance Scale, Mini-BESTest
Effect: UPDRS-III SMD −0.80 (−1.02 to −0.59); FOGQ −0.62 (−1.06 to −0.18)

A meta-analysis in PLOS ONE (25 September) pooled 20 randomised trials with 714 people with Parkinson disease comparing structured exercise with control.

Exercise was associated with lower UPDRS-III motor scores (standardised mean difference −0.80, 95% CI −1.02 to −0.59; moderate certainty), less freezing of gait (−0.62) and better Berg balance scores (0.85), but not a significant change in Mini-BESTest. No type or dose of exercise stood out. No trial was at low risk of bias.

The finding matches clinical experience and existing advice: regular exercise helps motor function, and the kind of exercise matters less than doing it. Referral to physiotherapy with gait and balance work is reasonable at any stage.

Recommend regular structured exercise to every patient with Parkinson disease.
Let patients choose a modality they will sustain; none clearly outperformed the others.
Refer patients with freezing or falls to physiotherapy for gait and balance training.
Exercise complements, not replaces, optimised dopaminergic treatment.

Why it matters

It reinforces exercise as treatment with a measurable motor effect, not only general advice.

The statistics, in plain English

A standardised mean difference of −0.80 is a large effect by usual conventions, but small trials at some risk of bias tend to overstate effects. Certainty was moderate only for UPDRS-III; the balance and freezing results are low certainty.