Abstract

Introduction
Exercises have been proposed as adjuvants for the treatment of Parkinson's disease (PD); however, responses to exercise interventions have shown inconsistent results. We investigated clinical factors associated with improvements in motor deficits after exercise.
Methods
85 PD patients were enrolled from five tertiary hospitals and classified into four exercise groups: home exercises, strength training, Tai Chi, and yoga. Clinical measurements of the motor and non-motor features of PD were assessed at baseline and 12 weeks after the exercise intervention. We employed principal component analysis (PCA) to reduce variables into ten factors and then examined associations of baseline characteristics with percentage improvement in the Movement Disorder Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III) using the Bayesian regression model.
Results
In the multivariate Bayesian regression model including ten PCA-driven factors, the percentage improvement of the MDS-UPDRS III was associated with factors including prominent motor deficits (posterior interval; mean ± SD, – 3.9 ± 1.7) and non-motor symptoms such as depression, anxiety, and subjective memory impairment (2.5 ± 1.5). Another factor related to functional impairments in gait and postural control was associated with less improvement after the exercise intervention (3.3 ± 1.7). In the subgroup analyses, motor features were associated with improvement in the home exercise and strength training groups, whereas mood disturbance, fatigue, and subjective cognitive impairment were related to changes in the home exercise and Tai Chi groups.
Conclusions
Our results suggest that individual phenotypes of PD patients may be associated with clinical improvement following exercise.

• Keywords: Parkinson’s disease; exercise