Eight RCTs were included in the review (1,068 participants). Sample sizes ranged from 30 to 311. Physiotherapy Evidence Database scale scores ranged from four to eight with the risk of bias being the greatest in the domains of allocation concealment and blinding. Half the studies did not conceal allocation. None of the studies blinded participants or therapists and only half of them blinded outcome assessors.
Physical function: The exercise group showed a significant increase in gait speed (WMD 0.07m/s, 95% CI 0.02 to 0.11, four trials) and a significant improvement in Berg Balance Scale scores (WMD 1.7 points, 95% CI 0.6 to 2.8, four trials) compared to the control group. There was evidence of statistical heterogeneity for the gait speed analysis (Ι²=60%). There was no significant difference between groups in time taken to get up and go (three trials), There was evidence of significant heterogeneity (Ι²=96%) in this analysis.
Activities of daily living: The exercise group showed a significantly better performance in activities of daily living score than the control group (WMD 5.33, 95% CI 1.01 to 9.64, three trials). There was evidence of significant statistical heterogeneity for this analysis (Ι²=77%).
Quality of life: There were no significant differences between groups regarding both mental (two trials) and physical (three trials) quality of life. No information on heterogeneity was presented for this analysis.