Ayla Pauwels ^1,2 Albert Phan ^3,4 Catherine Ding ^1,4 Thanh G. Phan ^1,4 Peter Kempster ^1,4*

• ¹ Department of Neurology, Monash Health, Melbourne, Australia
• ² Center for Neurosciences, Vrije Universiteit Brussel, Brussels, Belgium
• ³ Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia
• ⁴ Department of Medicine, School of Clinical Sciences at Monash Health, Monash University, Melbourne, Australia

Background: The search for neuroprotective treatments for Parkinson's disease (PD) still relies largely on motor disability scales. A limitation of these tools is the strong influence of symptomatic dopaminergic treatment effects. Drawing on a wealth of published information, we conducted a systematic review and meta-analysis of motor progression in PD and its relationships with dopaminergic therapy. Methods: We searched MEDLINE, EMBASE and CENTRAL to identify 84 publications with adequate serial motor scores to calculate progression, expressed as increase in percentage of maximum disability. Results: A random-effects model showed motor progression at 2.0% p.a. (95% CI 1.7-2.4%). There were no significant differences by baseline age, sample size or observation period. However, untreated patients, in 8 publications, progressed at 4.5% p.a. compared with 1.6% p.a. in 76 studies containing individuals on dopaminergic drugs (p = 0.0004, q = 0.003). This was supported by research on phenoconversion in prodromal PD, where motor progression exceeded 5% p.a. in the 2 years before diagnosis. Starting levodopa improved pre-treatment disability by 40.3 ± 15.2%. Practically defined off state measurements increase faster than on scores by a modest degree (p = 0.05). Conclusions: This survey suggests that accurate long-term measurements of motor progression to assess disease-modifying therapies can be conducted despite sequential commencement of dopaminergic drugs and sample attrition over time. While study designs involving prodromal or untreated PD avoid confounding effects of symptomatic treatment, different assumptions about motor progression may be needed. A defined off state with levodopa test dose method maximises information about medication cycle once dopaminergic therapy has begun.