Predictors of the placebo response in clinical trials on Parkinson's disease: A meta-analysis

Abstract

Introduction: Previous studies have assessed the placebo response in clinical trials on PD using the individual data of participants from the placebo-assigned group. The aim of this study was to examine the group predictors of the placebo response in randomized placebo-controlled trials on PD using a meta-analysis with meta-regression models.

Methods: The placebo response was defined as the mean change in the UPDRS part III score from baseline to the primary efficacy end point in the placebo group. The impacts of the predictors were assessed with meta-regression analyses, and significant predictors were used in a multivariable analysis. Subgroup analyses were conducted in studies that enrolled PD patients with or without motor fluctuations.

Results: Forty-eight studies (consisting of 5618 participants on placebo) were included. Motor fluctuation and baseline UPDRS part III score were significant predictors in the univariable analyses. The high baseline UPDRS part III score (β = -0.21, 95% CI -0.34, -0.08; p = 0.005) significantly increased the magnitude of the positive placebo response in the multivariable analysis. In the subgroup analyses, the positive placebo response was significant only in studies that enrolled patients with motor fluctuations; high baseline UPDRS part III score and low baseline daily levodopa dose increased the positive placebo response independently in the subgroup with motor fluctuations.

Conclusion: Researchers should consider the positive placebo response when they design clinical trials in advanced PD patients with motor fluctuations and severe motor symptoms. Baseline daily levodopa dose may be the independent predictor in studies that enrolled fluctuating patients.

Keywords: Meta-analysis; Parkinson’s disease; Placebo; Randomized clinical trials.