Design: Systematic review and meta-analysis of 9 randomised trials
Population: 334 people with Alzheimer's disease or dementia
Primary outcome: Agitation (CMAI) and neuropsychiatric symptoms (NPI-NH)
Effect: CMAI SMD −0.58 (−1.71 to 0.55); somnolence RR 2.03 (1.29–3.20)

This meta-analysis pooled nine randomised trials with 334 people with Alzheimer's disease or other dementia, testing cannabinoid-based treatments against placebo.

Cannabinoids did not improve agitation on the Cohen-Mansfield Agitation Inventory (SMD −0.58, 95% CI −1.71 to 0.55), total neuropsychiatric symptoms (SMD −0.02, −1.00 to 0.96) or cognition. Bayesian estimates were close to zero. Certainty was moderate for behavioural outcomes and low for cognition. Overall adverse events were similar to placebo, but somnolence was twice as common (RR 2.03, 1.29–3.20).

Families increasingly ask about cannabis products for agitation, often sold without prescription. These data give a clear answer: no benefit shown, and more sedation in a group already at high risk of falls and aspiration. Non-drug approaches — identifying pain, constipation, infection, sensory loss and environmental triggers — remain first line.

Advise families against cannabinoid products for dementia-related agitation
Ask directly whether any cannabis-based or herbal sedative products are being used
Look for pain, constipation, urinary retention, infection and sensory deprivation as causes of agitation
Use structured behavioural approaches before any sedating medication

Why it matters

Cannabis products are being marketed to families of people with dementia, and the trial evidence shows harm without benefit.

The statistics, in plain English

A standardised mean difference of −0.58 sounds moderate, but its interval runs from a large benefit to a moderate harm, and heterogeneity (I² 84%) means trials disagreed. The Bayesian analysis, which weighs all the evidence together, centred near no effect. Somnolence doubling is the more reliable finding.