BackgroundThe long-term impact of opioid use disorder (OUD) on brain health has been little explored although of potentially high public health importance.ObjectivesTo investigate the potential causal impact of OUD on later life brain health outcomes, including dementia, stroke and brain structure.MethodsObservational and Mendelian randomization (MR) analyses were conducted. Participants included in observational analyses were enrolled in the US Million Veteran Program (MVP). Cox proportional hazards were used to examine the association between electronic health record (EHR)-derived diagnoses of OUD and incident dementia in European and African ancestry populations. Two-sample MR was applied to explore the association between genetic predisposition to OUD and dementia, as well as key endophenotypes including brain structure. Several analyses were used for insight into aetiological pathways, including cis-MR to assess the impact of genetically-proxied opioid receptor perturbation, Bayesian colocalization, and polygenic risk score analyses of longitudinal brain changes in non-opioid users from the Lifebrain project (n=229).ResultsAmongst 222,518 MVP participants, 8397 developed dementia during follow up. Participants with OUD (n=9,399) were younger and more likely to be male. In observational analyses, OUD was associated with a higher risk of incident all-cause dementia (hazard ratio [HR]=1.56, 95% confidence interval [CI] [1.39,1.76];p=2.23×10-13), Alzheimer’s [HR=1.40[1.04,1.87]; p=0.02) and vascular dementia (HR=1.49[1.19,1.86];p=0.0004). In the genetic analysis, genetically-proxied OUD also associated with higher dementia risk. A doubling in genetically-proxied OUD prevalence was associated with a 77% increase in odds of dementia (IVW OR=1.77[1.43,2.19];p=1.69×10⁻⁷). Variation in μ-opioid receptor genes were strongly associated with dementia risk. No significant associations were observed with brain structure in non-opioid users, nor in lower powered non-European ancestry groups.ConclusionsThese findings suggest a potential causal impact of opioid use disorder on dementia. Genetic analyses supported an aetiological role of μ-opioid receptor pathways. Further pharmacovigilance and investigation into the long-term effects of opioids on brain health are warranted.