Evaluation of causality between ADHD and Parkinson's disease: Mendelian randomization study

Abstract

In a retrospective cohort study, patients with attention-deficit hyperactivity disorder (ADHD) and psychostimulant prescription were associated with increased risk of Parkinson's disease (PD). It is unclear whether ADHD per se or psychostimulant prescription is associated with PD. We aim to determine if genetic correlation or/and causal association exists between ADHD and PD using summary statistics obtained from the largest meta-analysis of genome-wide association studies of ADHD (20,183 cases; 35,191 controls) and PD (26,421 cases; 442,271 controls). Genetic correlation was tested between ADHD and PD by linkage disequilibrium score regression. Causal estimate was assessed by inverse-variance weighted (IVW) method as the main mendelian randomization analysis, with sensitivity analyses to detect horizontal pleiotropy. Weak and inverse genetic correlation existed between ADHD and PD (r=-0.100;SE=0.045;P = 0.026). Univariable IVW analysis with 10 and 77 genetic instruments respectively revealed null association for ADHD with PD (OR=0.930 per doubling in odds of ADHD; 95% CI:0.792-1.092) and PD with ADHD (OR=0.986 per doubling in odds of PD; 95% CI:0.956-1.015). Multivariable IVW analyses adjusted for BMI/smoking also revealed null association of ADHD with PD. Using 58 PD-associated genetic instruments, multivariable IVW analysis with/without adjustment for BMI/smoking suggested a weak and inverse causal association for PD on ADHD, but cautious interpretation is required. This well-powered study did not support causality between ADHD and PD. The observed positive association between ADHD and PD is more likely to be caused by unmeasured confounders. As psychostimulant use is associated with high risk of early-onset PD, future research should focus on this area.