CONTEXTPrevious studies assessing adapted physical activity (APA) in Parkinson's disease (PD) have been very heterogeneous regarding methodology and intervention, and have generally not addressed the question of combining various types of physical activity with a long-term evaluation.OBJECTIVESTo evaluate the immediate and long-term effect of a 3-month APA program, combining endurance, resistance training and stretching on motor symptoms, body composition, cardiorespiratory function and metabolic profile in PD patients.METHODSIn this controlled trial, we randomly assigned forty-four PD patients in a 1:1 ratio to receive a 3-month APA program (APA + group, n = 22), or freely practice physical activity (APA- group, n = 22). The patients were evaluated for parkinsonian symptoms (UPDRS-III), body composition, cardiorespiratory function and metabolic profile at baseline, immediately after the end of the program (M3) and six months later (M9).RESULTSBetween baseline and M3, the mean UPDRS-III score decreased in PD patients from the APA + group whereas it increased in the APA- group (-1.2 ± 6.6 vs. +1.9 ± 8.9; p = 0.04), regardless of age, sex, disease duration, dopaminergic treatment, UPDRS-III and axial score at baseline, but these between group differences waned at M9. No between group difference was observed regarding the evolution of body composition, metabolic profile or cardiorespiratory function between baseline, M3 and M9.CONCLUSIONA 3-month APA program combining endurance and resistance training plus stretching is more efficient for improving motor symptoms in PD compared to an unstructured engagement in non-specific physical activities. However, the benefits fade away six months after discontinuation of the program.STATISTICAL ANALYSIS CONDUCTED BYLAMBERT Céline, CHU Clermont-Ferrand, DRCI, Biostatistics Unit, Clermont-Ferrand, FRANCE.REGISTRATION NUMBERclinicaltrials.gov number NCT02816619.