Idiopathic Parkinson's disease is a progressive neurodegenerative disorder with motor and non-motor symptoms that may remain inadequately controlled by conventional therapies. Cannabinoid-based interventions have emerged as potential symptomatic treatments, but clinical evidence remains limited and heterogeneous. We assessed their efficacy and safety using complementary frequentist and Bayesian meta-analyses of randomized controlled trials.

This systematic review followed PRISMA 2020. PubMed, Embase, and Cochrane Central were searched from inception through December 2025. Eligible studies were randomized controlled trials comparing cannabinoid-based interventions with placebo in adults with Parkinson's disease. Random-effects models pooled motor and non-motor disease severity, cognition, anxiety, depression, quality of life, and adverse events. Bayesian random-effects models were applied in parallel to characterize posterior treatment effects and uncertainty.

Seven randomized controlled trials were included. Frequentist analyses did not identify clear differences between cannabinoid-based interventions and placebo across efficacy outcomes. Cannabinoid-based interventions increased the risk of confusion (risk ratio, 4.44; 95% confidence interval, 1.17–16.87) and dizziness (risk ratio, 2.78; 95% confidence interval, 1.65–4.68). Bayesian analyses showed posterior probabilities of any benefit of 74.0% for UPDRS III and 71.0% for cognition; probabilities of exceeding the prespecified standardized effect threshold of 0.20 were 42.8% and 54.8%, respectively, with substantial uncertainty across outcomes.

Current randomized evidence is insufficient to establish clinically meaningful benefits of cannabinoid-based interventions in Parkinson's disease and suggests increased risk of some adverse events. Interpretation is limited by few trials and participants, short follow-up, intervention heterogeneity, and imprecise estimates.