CBD and sleep: what the trials show, dose by dose
No health authority recognises CBD as a sleep aid, and the controlled trials give no grounds to do so. In the most rigorous insomnia trial, 150 mg of CBD every night for two weeks left insomnia severity no lower than on placebo. The doses studied are 12 to 150 times the provisional reference of 2 mg a day set by EFSA in 2026.
Key takeaways
- In a sleep lab, 300 mg of CBD changed none of the measures recorded in 27 healthy volunteers.
- In 30 adults with insomnia, 150 mg a night improved how well they said they felt; insomnia severity did not change.
- The most cited study followed 72 patients without a placebo group, so it cannot show any effect of CBD.
- Drowsiness is a side effect of the CBD medicine, reported in 23% of patients on high doses.
- Cognitive behavioural therapy is still the first-line treatment for lasting insomnia.
What the controlled trials say
A controlled trial compares a group taking the product with a group taking a placebo, and participants do not know which one they receive. This is the only way to separate the effect of a molecule from the effect of expectation, which weighs heavily on sleep. For CBD alone, such trials are rare and small.
| Study | Participants | CBD per dose | Result against placebo |
|---|---|---|---|
| Linares et al., 2018 | 27 healthy volunteers, one night | 300 mg | No effect on sleep architecture |
| Narayan et al., 2024 | 30 adults with insomnia, 2 weeks | 150 mg | No difference in insomnia severity; better reported well-being |
| Wang et al., 2025 | 125 adults with insomnia, 56 analysed | 300 mg + terpenes | +1.3 points of deep and REM sleep, wrist-worn device |
The Linares trial, run in Brazil, recorded the sleep of 27 volunteers with polysomnography, the reference test that tracks brain activity through the night. A 300 mg dose of CBD had no significant effect on sleep, or on the cognitive tests taken on waking.
Published in the Journal of Clinical Sleep Medicine, the Australian Narayan trial enrolled 30 adults with moderate to severe insomnia. After two weeks at 150 mg a night, insomnia severity, reported time to fall asleep and night-time waking were the same as in the placebo group. The CBD group reported better well-being, and sleep efficiency measured at the wrist improved in the second week. The authors conclude that CBD performed like placebo on most sleep outcomes.
The third trial paired 300 mg of CBD with eight terpenes. Of 125 participants enrolled, 56 were analysed; the rest had dropped out or had not worn the sensor. Compared with placebo, time in deep and REM sleep rose by 1.3 percentage points, and total sleep time did not change. The company Defined Research funded the study, and six of the eight authors hold shares in it.
Why the most cited study proves nothing
Many articles lean on a case series published in 2019 in The Permanente Journal. A US clinic followed 72 adults who had come in for anxiety or sleep problems, most of them given 25 mg of CBD a day. After one month, 66.7% had better sleep scores, though those scores fluctuated afterwards.
The authors list the limits themselves: treatment given openly, no comparison group, and patients on other psychiatric medicines. Without a placebo, nobody can tell what comes from CBD, from the consultation or from the passage of time. A study of this kind can raise a hypothesis; testing it takes a controlled trial.
The EFSA reference: 2 mg a day
On 9 February 2026, EFSA published a provisional safe intake of 0.0275 mg of CBD per kilo of body weight per day, about 2 mg for a 70 kg adult. It covers CBD at least 98% pure and excludes under-25s, pregnancy and breastfeeding, and people on medication. The figure is a safety limit and recommends no dose. The conversion into drops is in how much CBD per day.
The trials sit far above that: 150 mg is 75 times the reference, and 300 mg is 150 times. In 2025, the food safety commission of the German Research Foundation reviewed the data available at food-level doses. Its review in Nutrients found no evidence that would support a health claim, sleep, relaxation and stress included.
Drowsiness: a side effect, not a promise
Epidyolex, the only CBD-based medicine, is prescribed for rare epilepsies at 10 to 25 mg per kilo per day, or 700 to 1,750 mg for 70 kg. Its summary of product characteristics lists drowsiness among the most common side effects, in 23% of patients. Drowsiness and sedation affected 43% of patients who also took clobazam, another anti-epileptic, against 14% of the others. The document advises against driving until you know how the treatment affects you.
Two lessons follow. At these doses, the drowsiness described is a side effect and says nothing about sleep quality. CBD also interacts with other molecules acting on the brain: the liver enzymes CYP3A4 and CYP2C19 break it down, and they process many medicines too. If you take a sleeping pill, an anti-anxiety drug or an antidepressant, ask your doctor or pharmacist. CBD's grapefruit effect explains the mechanism.
What health authorities advise for difficult nights
French adults sleep 7 h 04 min on average on weekdays, according to the 2025 survey by the French National Institute for Sleep and Vigilance, carried out among 1,000 people. Nearly a quarter sleep less than 6 hours, and 73% wake at least once a night.
The French National Authority for Health (HAS) says that non-drug treatments should come before sleeping pills, meaning sleep hygiene rules and cognitive behavioural therapy. It reserves benzodiazepines for acute insomnia, for 4 weeks at most. Updated in 2023, the European insomnia guideline makes cognitive behavioural therapy for insomnia (CBT-I), in person or online, the first-line treatment in adults. It advises against antihistamines, immediate-release melatonin and herbal remedies.
Two products close to CBD are worth a mention. CBD and melatonin compares CBD with melatonin, the hormone of the sleep-wake rhythm. "Sleep" gummies combining CBN and THC are illegal in France, for reasons set out in our article on CBN.
CBD is not a medicine. See a doctor if insomnia lasts more than three months, or if drowsiness disrupts your day.
Frequently asked questions
No. No French or European authority recognises it as one, and in the most rigorous insomnia trial, 150 mg a night made no difference to insomnia severity compared with placebo. A CBD product may not be sold with a sleep claim.
Doses range from 25 mg a day, in a case series without placebo, to 300 mg per dose in two controlled trials. EFSA's provisional reference is 2 mg a day for 70 kg. Depending on the dropper, a 10% oil gives about 5 mg per drop.
Only with medical advice. The liver enzymes that break down CBD also handle many medicines, and the Epidyolex leaflet reports more drowsiness with some combined treatments. Your doctor or pharmacist can review your treatment.
Roadside tests do not look for CBD; they do look for the THC in a full-spectrum product, with no tolerance threshold. In an Australian trial from 2024, saliva tests no longer detected an oil with 10 mg of THC and 200 mg of CBD 10 and 18 hours after a single dose. Regular use was not studied. The details are in the saliva test guide.
Written by the Phytogrammes team
Each journal article draws on cited primary sources: ANSM, EFSA, EUR-Lex and peer-reviewed publications. Product profiles are distinguished from documented laboratory results. Our approach.
Article published on . CBD is not a medicine.