Exploring Potential Predictors of Treadmill Training Effects in People With Parkinson Disease.

Abstract

To explore the potential predictors of people with Parkinson disease (PD) who would benefit the most from treadmill training. A cohort study. Medical university rehabilitation settings. Seventy participants diagnosed of idiopathic PD. Twelve sessions of treadmill training. Hierarchical logistic regression models were used to explore significant predictors of the treadmill training effect with respect to 3 health domains: Unified Parkinson's Disease Rating Scales part III (UPDRS III); gait speed; Parkinson's Disease Questionnaire-39 (PDQ-39). A receiver operating characteristic (ROC) curve analysis was conducted to identify proper cut-off points for clinical use. Male sex (adjusted odds ratio [OR]: 3.73, P=.036) significantly predicted the improvement of UPDRS III. Individuals with a slower baseline gait speed (cut-off: 0.92 m/s, adjusted OR: 14.06, P<.001) and higher baseline balance confidence measured by the Activity-specific Balance Confidence scale (cut-off: 84.5 points, adjusted OR: 4.66, P=.022) have greater potential to achieve clinically relevant improvements in gait speed. A poorer baseline PDQ-39 score (cut-off: 23.1, adjusted OR: 7.47, P<.001) predicted a greater quality of life improvement after treadmill training. These findings provide a guideline for clinicians to easily identify suitable candidates for treadmill training. Generalization to more advanced patients with PD warrants further investigation.