Objective: To investigate whether genetic variation in the GLP1R gene modifies treatment response to the GLP1 agonist exenatide in the Exenatide-PD3 trial, and to characterise the power constraints of post hoc Pharmacogenomic (PGx) analysis in clinical trials.

Background: The biological heterogeneity of Parkinson’s disease (PD) contributes to successive failures of disease-modifying clinical trials. To address this, stratified approaches are warranted. PGx offers a framework for identifying individuals who respond differentially to a therapy based on genetic factors and may be applied to clinical trial datasets in post hoc analyses.

Method: The Exenatide-PD3 trial (N=194) showed no benefit for its primary endpoint, MDS-UPDRS Part III. We performed post hoc PGx analyses using genotype-by-treatment (G×T) interaction models. Two candidate GLP1R polymorphisms (rs10305420, rs6923761) were tested for interaction with treatment on change in MDS-UPDRS Part III and 124 plasma biomarkers. Simulation-based power analyses quantified the minimum detectable G×T interaction across a range of sample sizes and variant frequencies.

Results: No significant G×T interaction was demonstrated for the primary motor outcome, and no G×T interaction survived FDR correction across 124 plasma biomarkers. The strongest signal was for rs10305420, where each additional variant allele was associated with a greater rise in plasma neurofilament light chain on exenatide compared to placebo (β= 0.234, nominal p = 0.033, FDR = 0.94). Exploratory follow-up suggested a per-allele effect in the exenatide arm (p = 0.015) but not placebo (p = 0.321), though this did not reach significance after covariate adjustment (p = 0.092). Power simulations demonstrated that a trial of N=194 can only detect G×T effects exceeding 5.8 UPDRS points per allele for a common variant and >10 points at 10% carrier frequency. Detecting a clinically meaningful interaction of 3.25 UPDRS points requires N≈500 for common variants and N>3,000 at GBA1-like carrier frequencies.

Conclusion: Post hoc PGx analyses in clinical trials are constrained by insufficient power to detect meaningful G×T interactions. Larger clinical trials, genotype-stratified designs and consortium-level meta-analysis offer alternatives to overcome this limitation.

References: Exenatide once a week versus placebo as a potential disease-modifying treatment for people with Parkinson’s disease in the UK: a phase 3, multicentre, double-blind, parallel-group, randomised, placebo-controlled trial. Vijiaratnam, Nirosen et al. The Lancet, Volume 405, Issue 10479, 627 – 636.

« Back to 2026 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/how-large-must-a-parkinsons-disease-trial-be-to-detect-pharmacogenomic-interactions-power-benchmarks-derived-from-the-exenatide-pd3-clinical-trial/