Cannabis Products in Parkinson’s Disease: Why Reported Benefit Does Not Establish Causation
A small Parkinson’s disease study found both perceived benefits and cognitive concerns among cannabinoid-product users, but its cross-sectional design could not determine what treatment caused. Baseline differences and treatment effects remain competing explanations.
Who was compared?
Sousa and DiFrancisco-Donoghue assessed 15 people with Parkinson’s disease: eight using CBD, THC, or combined products and seven nonusers. Testing included global cognition, more specific neuropsychological measures, motor assessments, and symptom questionnaires.
Products and administration methods varied. This was not random assignment to one standardized cannabis formulation, and participants were not followed from a common pretreatment baseline.
Different outcomes gave different answers
Users had lower Montreal Cognitive Assessment scores and reported more nonmotor symptoms. More specific cognitive measures and motor outcomes did not show corresponding significant group differences.
Some users described improved pain, sleep, or anxiety, while others reported sleepiness, forgetfulness, or concentration difficulty. These accounts are clinically relevant observations, but they are not blinded demonstrations of efficacy or adverse-event incidence.
Why direction matters
People with greater symptom burden may be more likely to seek cannabinoid products. Conversely, an exposure could contribute to symptoms. A one-time comparison cannot distinguish those explanations reliably, especially with such a small sample and heterogeneous treatment.
The study supports monitoring cognition and documenting actual products when evaluating use in Parkinson’s disease. It does not establish that cannabis caused a participant’s lower score, that it slowed disease progression, or that subjective improvement guarantees safe driving. Those questions require designs and endpoints matched to the proposed conclusion.
A lower group-average global screening score is not a diagnosis for every user in the group. Equally, a nonsignificant difference on a more specific test does not establish equivalence between users and nonusers. With eight exposed participants and seven comparison participants, estimates are too limited to settle a general treatment-benefit or treatment-harm claim.
Related reading: limits of simulator evidence in older medical-cannabis patients.
Primary research
Amber Sousa; Joanne DiFrancisco-Donoghue. Cannabidiol and Tetrahydrocannabinol Use in Parkinson’s Disease: An Observational Pilot Study. Cureus (2023). DOI: 10.7759/cureus.42391.
