Medical cannabis for insomnia: very low certainty of benefit, more adverse events

Design

Cochrane systematic review and random-effects meta-analysis of randomised trials, GRADE-rated

Population

537 adults with acute or chronic insomnia across 9 trials

Primary outcome

Sleep quality, total sleep time and adverse events

Effect

Sleep quality SMD -0.49 (95% CI -1.06 to 0.08); total sleep time +7.68 minutes (-23.04 to 38.39); adverse events OR 11.25 (4.46 to 28.35)

A Cochrane review included nine randomised trials with 537 adults with insomnia, searched to April 2025. Eight compared medical cannabis with placebo and one with amitriptyline. Seven used magistral preparations rather than licensed products, and most participants were women.

Against placebo, medical cannabis may have little to no clinically meaningful effect on sleep quality (SMD -0.49, 95% CI -1.06 to 0.08), total sleep time (mean difference 7.68 minutes, 95% CI -23.04 to 38.39) or daytime functioning (SMD -0.34, 95% CI -0.83 to 0.16). All were very low certainty with high inconsistency (I² 65% to 91%). Adverse events may be more frequent (OR 11.25, 95% CI 4.46 to 28.35; three studies, 98 participants; low certainty).

The fair reading is that the evidence cannot show benefit, and it hints at harm. It does not show that cannabis cannot help any individual, but it gives no basis to recommend it over established insomnia approaches, particularly cognitive behavioural therapy for insomnia.

• Offer cognitive behavioural therapy for insomnia and sleep hygiene before any drug
• If a patient asks about cannabis, explain that trials show little clear benefit and more side effects
• Ask about current cannabis use when assessing insomnia, mood and driving
• Review other contributors: alcohol, caffeine, sleep apnoea, pain, depression and medicines
• Note that most trials used unlicensed preparations, so results may not apply to a given product

Why it matters

Patients increasingly ask for cannabis as a sleep aid, and this separates popularity from evidence.

Don't overread it

Very low certainty does not prove no effect; the trials were small and inconsistent, and it cannot speak to specific licensed products.

The statistics, in plain English

An SMD is a standardised difference that lets different sleep scales be combined; -0.49 would be a moderate effect, but the interval (-1.06 to 0.08) includes zero. I² above 60% means the trials disagreed a lot, which, together with small samples and high risk of bias, is why certainty was rated very low. The large odds ratio for adverse events comes from only 98 participants in crossover designs, so its size is uncertain.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.