Amphetamine use and Parkinson’s disease: integration of artificial intelligence prediction, clinical corroboration, and mechanism of action analyses

Abstract

Parkinson’s disease (PD) is an increasingly prevalent neurologic condition for which symptomatic, but not preventative, treatment is available. Drug repurposing is an innovate drug discovery method that uncovers existing therapeutics to treat or prevent conditions for which they are not currently indicated, a method that could be applied to incurable diseases such as PD. A knowledge graph artificial intelligence prediction system was used to select potential drugs that could be used to treat or prevent PD, and amphetamine was identified as the strongest candidate. Retrospective cohort analysis on a large, electronic health record database of deidentified patients with attention deficit hyperactive disorder (the main diagnosis for which amphetamine is prescribed) revealed a significantly reduced hazard of developing PD in patients prescribed amphetamine versus patients not prescribed amphetamine at 2, 4, and 6 years: Hazard Ratio (95% Confidence Interval) = 0.59 (0.36, 0.98), 0.63 (0.42, 0.93), and 0.55 (0.38, 0.79). Pathway enrichment analysis confirmed that amphetamine targets many of the biochemical processes implicated in PD, such as dopaminergic synapses and neurodegeneration. Together, these observational findings suggest that therapeutic and legal amphetamine use may reduce the risk of developing PD, in contrast to previous work that found the inverse relationship in patients using amphetamine recreationally.