Cannabinoids added about half an hour to total sleep time

Insomnia is one of the most common health complaints there is, and the standard drug options carry well-known problems with tolerance and dependence. That has pushed a lot of people towards cannabinoids as a self-managed alternative, often well ahead of the evidence. Individual trials have been small and used different products, outcomes and comparators, which makes it hard for anyone to judge whether the effect is real or how large it might be. A pooled picture also matters for the harms side, since mild but frequent side effects change the calculation. This review set out to assess both efficacy and safety together.

The authors searched PubMed, EMBASE, the Cochrane Library, Scopus and Web of Science from inception to January 2025 for randomised controlled trials comparing cannabinoids against either placebo or melatonin in adults with insomnia or poor sleep quality. The primary outcomes were changes in the Insomnia Severity Index and the Pittsburgh Sleep Quality Index, both standard questionnaires. Secondary outcomes included total sleep time, sleep efficiency, time taken to fall asleep and adverse events. Ten trials with 2134 participants met the criteria, and results were combined using random-effects models reporting mean differences or risk ratios.

Insomnia severity fell by 3.73 points and sleep quality scores improved by 3.94 points, both statistically significant. Total sleep time rose by 33.99 min and sleep efficiency by 4.31%. Time to fall asleep improved more modestly, by 12.0 min, with a range wide enough that the true effect could be small. Adverse events were nearly four times more frequent with cannabinoids, at a risk ratio of 3.96, but these were mostly mild dizziness and dry mouth and no serious harms were observed across the included trials.

This moves cannabinoids from anecdote towards measurable effect for sleep, and the size of the change in questionnaire scores is not trivial. It does not resolve which cannabinoid, which dose or which delivery method is responsible, because the pooled trials varied and the comparators included both placebo and melatonin. Nor does it speak to long-term use, tolerance or dependence, which were the very problems that drove interest in alternatives in the first place. The authors themselves call for larger and longer trials to define optimal dosing and whether benefits hold over time.

This is a reason to take cannabinoids seriously as a sleep intervention rather than dismiss them, while noting that the evidence is short-term and the side effect rate is not negligible. The improvements reported were in self-reported severity and quality scores plus modest objective gains, not a cure for insomnia. Time to fall asleep improved least, so anyone whose main complaint is lying awake at the start of the night should temper expectations. Legality, product quality and interactions all sit outside what this review examined.