Antidopaminergic Medications and Clinical Changes in Measures of Huntington's Disease: A Causal Analysis

Abstract

Background: Antidopaminergic medications (ADM) are often used for symptom management of Huntington's disease (HD). Evidence from past research suggests that ADMs are associated with worse clinical outcomes in HD, but their impact on various domains remains underexplored.

Objective: We used causal inference analysis to understand the impact of ADM use on measures of clinical progression in HD across multiple domains over 2 years.

Methods: We used the Enroll-HD database with a new-user design, which compared a cohort that initiated ADM use after the first visit with an unexposed cohort that remained off ADMs. To control for 27 covariates, we used a doubly robust targeted maximum likelihood estimation and conducted two analyses. First, we analyzed ADM treatment 2 years post-baseline and separately for 12 outcome measures. Second, we examined the association of ADM dose with measures of clinical outcomes.

Results: The ADM-exposed group exhibited faster change in measures of clinical outcome compared with the off-ADM group, which was statistically reliable in cognitive and functional outcome measures, and the composite Unified Huntington's Disease Rating Scale (cUHDRS). Motor domain analyses showed faster change in bradykinesia in the ADM-exposed group versus off-ADM but no difference in chorea or total motor score (TMS). Higher ADM doses also showed greater differences compared to the off-ADM group.

Conclusions: ADM use was associated with more rapid change in clinical measures, particularly in cognitive and functional domains. However, assumptions required to establish causation between ADM use and disease progression may not have been fully met, and further research is warranted. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Keywords: Huntington's disease; antidopaminergic; antipsychotic; targeted maximum likelihood estimation; vesicular monoamine transporter 2.

FIG. 1

Study design and causal analysis framework. (A) The new‐user design in which both groups are off ADM (antidopaminergic medication) at baseline. The exposure group participants are on ADMs beginning sometime after baseline (not under researcher control). (B) The causal diagram for the analysis. The exposure mechanism is adjusted for the observed covariates (C‐to‐A arrow) and so is the outcome mechanism (C‐to‐Y arrow). [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 2

TMLE (targeted maximum likelihood estimation) results by medication class. (A) The estimated average treatment effect (vertical bar) and 99% CI (confidence interval) for the nonmotor variables; a negative value indicates the exposed group had a smaller mean than the unexposed group at 2 years. (B) The results for the motor variables; a positive value indicates the exposed group had a larger mean than the unexposed group. A CI is black if it contains 0 (no mean difference) and red if it does not contain 0. [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 3

Dose group results by ADM (antidopaminergic medication) class. (A) The estimated average treatment effect (vertical bar) and 99% CI (confidence interval) for the nonmotor variables for low‐ and high‐dose exposure groups; a negative value indicates the exposed group had a smaller mean (ie, was worse) than the unexposed group at 2 years. (B) The results for the motor variables; positive values indicate the exposed group had a larger mean (ie, was worse) than the unexposed group. A CI is black if it contains 0 (no difference) and red if it does not contain 0. [Color figure can be viewed at wileyonlinelibrary.com]