The Iowa Gambling Task in Parkinson's disease: A meta-analysis on effects of disease and medication

Background: Decision-making under uncertainty as measured by the Iowa Gambling Task has frequently been studied in Parkinson's disease. The dopamine overdose hypothesis assumes that dopaminergic effects follow an inverted U-shaped function, restoring some cognitive functions while overdosing others. The present work quantitatively summarizes disease and medication effects on task performance and evaluates evidence for the dopamine overdose hypothesis of impaired decision-making in Parkinson's disease. Methods: A systematic literature search was performed to identify studies examining the Iowa Gambling Task in patients with Parkinson's disease. Outcomes were quantitatively combined, with separate estimates for the clinical (patients ON medication vs. healthy controls), disease (patients OFF medication vs. healthy controls), and medication effects (patients ON vs. OFF medication). Furthermore, using meta regression analysis it was explored whether the study characteristics drug level, disease duration, and motor symptoms explained heterogeneous performance between studies. Results: Patients with Parkinson's disease ON dopaminergic medication showed significantly impaired Iowa Gambling Task performance compared to healthy controls. This impairment was not normalized by short-term withdrawal of medication. Heterogeneity across studies was not explained by dopaminergic drug levels, disease durations or motor symptoms. Discussion: While this meta-analysis showed significantly impaired decision-making performance in Parkinson's disease, there was no evidence that this impairment was related to dopamine overdosing. However, only very few studies assessed patients OFF medication and future studies are needed to concentrate on the modulation of dopaminergic drug levels and pay particular attention to problems related to repeated testing. Furthermore, short- vs. long-term medication effects demand further in-depth investigation. (C) 2016 Elsevier Ltd. All rights reserved.