Which exercise intervention is most promising for Parkinson's balance? A network meta-analysis.

Objective: Parkinson's disease (PD) is a chronic, progressive, and disabling neurodegenerative disorder. While bradykinesia, resting tremor, and rigidity are the cardinal motor symptoms, balance impairments and postural instability frequently emerge during disease progression, substantially increasing the risk of falls and diminishing quality of life. Although exercise is a well-established, effective non-pharmacological intervention for improving balance and motor function in individuals with PD, the comparative efficacy of different exercise modalities remains unclear. Therefore, this network meta-analysis aims to evaluate and compare the effects of various exercise therapies on balance outcomes in patients with PD, providing evidence-based guidance for optimizing exercise prescriptions. Method: We searched databases such as PubMed, EMBASE, Cochrane Library, Web of Science, and CNKI from establishing the database until August 2025. We screened literature based on inclusion and exclusion criteria and compared the effects of 30 different exercise interventions on balance ability in people with PD. Results: Our study included a total of 135 articles and 5948 People with PD. Our study found that Dance exercise (DE) is more effective than Control group (CON) (MD, 6.25; 95% CI, 2.81 to 9.69) and Traditional Rehabilitation (TR) (MD, 7.56; 95% CI, 3.89 to 11.23) in improving Berg Balance Scale (BBS) in Parkinson's disease patients. In terms of reducing the Timed Up and Go Test (TUG) score of people with PD, Biofeedback Balance and Gait Training (BGT) is superior to CON (MD, -10.47; 95% CI, -20.13 to -0.81) and TR (MD, -10.60; 95% CI, -20.05 to -1.15), Stretch exercise (SE) (MD, -14.30; 95% CI, -26.75 to -1.85), Virtual reality balance training (VRB) (MD, -12.85; 95% CI, -23.40 to -2.29) and Yoga (YG) (MD, -13.29; 95% CI, -23.97 to -2.61). Pilates exercise (PE) (MD, -13.47; 95% CI, -25.98 to -0.96).is superior to CON in improving Unified Parkinson's disease rating scale III (UPDRS-III) scores. Qigong (QG) is superior to TR (MD, -6.24; 95% CI, -10.87 to -1.60) and Resistance training (RT) (MD, -8.21; 95% CI, -16.10 to -0.33) and CON (MD, -8.89; 95% CI, -14.35 to -3.42) in improving UPDRS-III. Conclusion: Compared with the CON and TR groups, DE (as measured by the BBS) and BGT (as measured by the TUG) demonstrated superior efficacy in improving balance function in patients with PD. PE and QG showed greater benefits than CON in improving global motor symptoms, as assessed by the UPDRS-III. These findings may provide evidence-based guidance for clinicians, rehabilitation professionals, and patients when selecting exercise interventions tailored to specific therapeutic goals in PD. Nevertheless, the mechanisms underlying these modality-specific effects remain unclear and warrant further investigation. In addition, the observed superiority of BGT for TUG outcomes should be interpreted with caution, as this finding was derived from a single study and requires confirmation in future high-quality randomized controlled trials.