L'écart entre ce que le CBD peut réellement faire et ce pour quoi il est vendu est large. Les preuves solides sont étroites et précises. Le cannabidiol sur ordonnance purifié, l'Epidiolex, réduit les crises dans trois épilepsies rares lors d'essais randomisés. Des doses uniques diminuent aussi l'anxiété aiguë dans des expériences, à une dose bien précise.
L'allégation populaire veut que l'huile de CBD de chanvre en vente libre soulage de façon fiable la douleur chronique et les troubles du sommeil. Les données sur ce point sont contrastées et maigres. Le marché grand public ajoute trois problèmes, chacun méritant d'être connu avant de se fier à un produit du commerce. Les étiquettes peuvent ne pas refléter la teneur réelle, certains produits contiennent du THC, et les deux formes présentent un risque d'atteinte des enzymes hépatiques et d'interactions médicamenteuses.
Findings & Outcomes
What It Is
CBD, or cannabidiol, is one of the two main compounds in the cannabis plant. Unlike THC, the other one, it does not make you high. It is sold in two very different forms that often get discussed as if they were the same product:
- Prescription cannabidiol (Epidiolex) is a purified pharmaceutical liquid approved for specific rare epilepsies, made to a known strength and tested in randomized trials.
- Over-the-counter hemp CBD is the oils, gummies, capsules, and creams sold as supplements. It is held to no such manufacturing standard, and it carries almost none of the trial evidence people assume it does.
Almost all of the seizure trials, the anxiety experiments, and the liver-enzyme findings used the purified form at a controlled dose, often far higher than a supplement delivers. A strong result for Epidiolex does not transfer to a retail gummy of uncertain content.
What It Does
At the strong end are three rare epilepsies, where a purified prescription form reduced seizures in randomized trials. In Dravet syndrome, a severe childhood epilepsy, a trial of 120 children cut monthly convulsive seizures by about 39%, against about 13% on placebo. In Lennox-Gastaut syndrome, two trials reduced drop seizures by roughly 42 to 44%, against 17 to 22%. In tuberous sclerosis complex, a trial of 224 patients cut seizures by about 48%, against about 27%. These results made cannabidiol a medicine. They are narrow by design, these specific syndromes, at a controlled dose. They do not show a general sedative or brain-calming effect.
Below that is acute, situational anxiety, on thinner evidence. In small experiments, people took a single dose before a simulated public-speaking test. A 300 mg dose lowered the anxiety spike, while 150 mg and 600 mg did not, an inverted-U where more is not better. This says nothing yet about a daily gummy for an anxiety disorder over months, which has barely been tested. Still, the single-dose signal is an early but real result, worth watching as the research matures.
At the weak end are chronic pain and sleep, the two most-marketed uses. The best systematic review of cannabis-based products rated any pain benefit small, and the CBD-alone data thinner still. For sleep, most of what exists is uncontrolled case reports, and Western trials have not shown a reliable effect.
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
Seizure Control
Le CBD purifié a réduit les crises convulsives mensuelles d'environ 39 % dans le syndrome de Dravet, contre 13 % sous placebo
Dans un essai rigoureux mené chez des enfants atteints d'une forme sévère d'épilepsie, une forme prescrite et purifiée de CBD a réduit les crises convulsives d'environ 39 %, tandis que le placebo les a réduites d'environ 13 %. C'est l'usage du cannabidiol le plus solide et le mieux établi.
Cet essai randomisé en double aveugle contrôlé contre placebo a ajouté une solution buvable de cannabidiol (20 mg/kg/jour) ou un placebo au traitement antiépileptique standard. Au-delà du décompte des crises, davantage de participants sous cannabidiol ont été jugés améliorés sur une échelle d'impression globale. Les effets indésirables étaient plus fréquents sous cannabidiol et comprenaient somnolence, diarrhée et élévation des aminotransférases hépatiques, cette dernière plus fréquente chez ceux prenant aussi du valproate. Le résultat a été reproduit et sous-tend l'approbation réglementaire du médicament pour le syndrome de Dravet.
Who this may not transfer to:Dravet syndrome affects both sexes and the trial enrolled both; the anticonvulsant effect is not sex-specific.
Il s'agit d'un médicament sur ordonnance utilisé sous la supervision d'un neurologue pour une épilepsie rare spécifique, avec surveillance hépatique et attention aux interactions médicamenteuses, pas une raison de s'attendre à ce qu'un produit de CBD de vente au détail contrôle les crises. La dose est élevée et la population restreinte.
The study · 1
Devinsky et al., Trial of Cannabidiol for Drug-Resistant Seizures in the Dravet Syndrome · N Engl J Med 2017;376(21):2011-2020
Le CBD sur ordonnance a réduit les crises de chute de 42 à 44 % dans le syndrome de Lennox-Gastaut, contre 17 à 22 % sous placebo
Dans deux essais solides sur une autre épilepsie infantile rare, le CBD sur ordonnance a réduit d'environ moitié les crises de type chute pour de nombreux patients, clairement plus que le placebo. C'est le deuxième usage bien établi contre les crises.
Les deux essais ont ajouté une solution buvable de cannabidiol aux antiépileptiques existants chez des patients atteints du syndrome de Lennox-Gastaut. La dose de 10 mg/kg dans l'essai Devinsky a fonctionné presque aussi bien que 20 mg/kg avec moins d'effets secondaires, suggérant qu'une dose plus faible peut suffire. Comme dans le syndrome de Dravet, la somnolence, la diminution de l'appétit, la diarrhée et l'élévation des enzymes hépatiques étaient les effets indésirables notables, et les élévations des enzymes hépatiques étaient plus fréquentes avec la prise concomitante de valproate.
Who this may not transfer to:Both trials enrolled both sexes and the anticonvulsant effect is not sex-specific.
Comme pour le résultat du syndrome de Dravet, il s'agit d'un usage sur ordonnance sous soins spécialisés avec surveillance, et la dose plus faible de 10 mg/kg est souvent préférée. Ce n'est pas une preuve en faveur du CBD de vente au détail dans les troubles épileptiques courants.
The studies · 2
Devinsky et al., Effect of Cannabidiol on Drop Seizures in the Lennox-Gastaut Syndrome · N Engl J Med 2018;378(20):1888-1897
Thiele et al., Cannabidiol in patients with seizures associated with Lennox-Gastaut syndrome (GWPCARE4) · Lancet 2018;391(10125):1085-1096
Le CBD sur prescription a réduit les crises associées à la STB d'environ 48 %, contre 27 % sous placebo
Dans une troisième épilepsie rare, la sclérose tubéreuse de Bourneville, le CBD sur prescription a réduit les crises d'environ 48 % contre environ 27 % sous placebo dans un essai solide. C'est le troisième usage bien établi pour les crises, aux côtés des syndromes de Dravet et de Lennox-Gastaut.
L'essai GWPCARE6 a ajouté une solution orale de cannabidiol à 25 ou 50 mg/kg/jour ou un placebo aux antiépileptiques existants chez des patients âgés de 1 à 65 ans atteints de STB. Les deux doses de cannabidiol ont réduit les crises significativement plus que le placebo, la dose plus faible de 25 mg/kg/jour étant préférée pour sa meilleure tolérabilité. Les effets indésirables les plus fréquents étaient la diarrhée et la somnolence, plus fréquentes à la dose plus élevée, et des transaminases hépatiques élevées sont survenues chez environ un patient sur cinq sous cannabidiol et aucun sous placebo. Ce résultat a soutenu l'approbation du cannabidiol pour les crises associées à la STB.
Who this may not transfer to:The trial enrolled both sexes (about 42% female) and the anticonvulsant effect is not sex-specific.
Comme pour les résultats de Dravet et de Lennox-Gastaut, il s'agit d'un usage sur prescription sous surveillance spécialisée avec suivi hépatique, et la dose plus faible de 25 mg/kg/jour est généralement préférée. Ce n'est pas une preuve pour le CBD en vente au détail dans les troubles épileptiques ordinaires.
The study · 1
Thiele et al., Add-on Cannabidiol Treatment for Drug-Resistant Seizures in Tuberous Sclerosis Complex: A Placebo-Controlled Randomized Clinical Trial (GWPCARE6) · JAMA Neurol 2021;78(3):285-292
Anxiety And Stress
Une dose unique de 300 mg de CBD a réduit l'anxiété avant un test de prise de parole en public, alors que 150 et 600 mg n'ont pas fonctionné
Prendre une seule dose modérée de CBD avant une tâche stressante de prise de parole en public a réduit l'anxiété dans de petits essais, une dose d'environ 300 mg fonctionnant mieux que des doses plus élevées ou plus faibles. Il s'agit d'un résultat précoce et à court terme sur l'anxiété situationnelle.
Bergamaschi 2011 a donné à 24 patients naïfs de traitement souffrant d'anxiété sociale 600 mg de cannabidiol ou un placebo avant un test simulé de prise de parole en public et a constaté une réduction de l'anxiété, une altération cognitive et un inconfort sous le médicament actif. Linares 2019 a testé 150, 300 et 600 mg contre placebo chez 57 hommes en bonne santé et a constaté que seuls 300 mg réduisaient significativement l'anxiété, décrivant une réponse dose-effet en U inversé. Ce sont des expériences à dose unique et à court terme dans des contextes spécifiques, pas des essais d'usage quotidien pour un trouble anxieux.
Who this may not transfer to:The primary social-anxiety trial enrolled both sexes; the confirmatory dose-response study was in men only, so the exact optimal dose in women is less certain.
Les données appuient une seule dose modérée avant un facteur de stress spécifique, pas une dose quotidienne élevée, puisqu'une dose plus élevée n'a pas aidé. Elle devrait s'ajouter au traitement établi d'un trouble anxieux, pas le remplacer, et les interactions médicamenteuses doivent être vérifiées au préalable.
The studies · 2
Bergamaschi et al., Cannabidiol reduces the anxiety induced by simulated public speaking in treatment-naive social phobia patients · Neuropsychopharmacology 2011;36(6):1219-1226
Linares et al., Cannabidiol presents an inverted U-shaped dose-response curve in a simulated public speaking test · Braz J Psychiatry 2019;41(1):9-14
Pain
CBD on its own did not clearly relieve chronic pain beyond placebo
Malgré un marketing intense du CBD pour la douleur, la meilleure revue actuelle trouve les preuves faibles, tout bénéfice éventuel étant faible, et les données sur le CBD seul étant trop limitées pour conclure qu'il soulage de manière fiable la douleur chronique.
The review synthesized randomized trials and observational studies of cannabis-based products for chronic pain. Products high in THC showed at most small short-term benefits with meaningful side effects, and the subset of evidence on cannabidiol-only products was sparse, of low certainty, and did not establish a clear analgesic effect. Much of the popular case for CBD in pain rests on preclinical work and testimonials, not controlled human outcomes.
Who this may not transfer to:Pooled across mixed-sex chronic-pain populations; CBD-specific subgroups were too small to assess sex differences.
A reader should treat CBD-for-pain claims with caution: the shelf implies reliable relief that the controlled evidence does not support. This is a place where the popular version runs well ahead of the research.
The study · 1
Chou et al., Cannabis-Based Products for Chronic Pain: An Updated Systematic Review · Ann Intern Med 2026;179(2):230-241
Sleep
Sleep scores improved for about two-thirds of adults on CBD in an uncontrolled clinic series
The main clinical report on CBD and sleep saw sleep scores improve for about two-thirds of people at first, but with no comparison group and gains that came and went. It is too thin to call CBD an established sleep aid.
Shannon 2019 reviewed charts from a psychiatric clinic where 72 adults received CBD (about 25 to 175 mg/day) alongside usual care, mostly for anxiety or poor sleep. Anxiety scores fell and stayed lower in most patients; sleep scores improved initially in about two-thirds but were more variable month to month. As an open, uncontrolled case series it cannot separate a true CBD effect from placebo, regression to the mean, or the knock-on effect of lower anxiety.
Who this may not transfer to:Both sexes were included in this uncontrolled clinic series; no sex-specific analysis was possible.
The durable, free groundwork for sleep does more for most people than a supplement, and CBD-for-sleep should be seen as unconfirmed. If tried, it belongs after the basics and after checking drug interactions.
The study · 1
Shannon et al., Cannabidiol in Anxiety and Sleep: A Large Case Series · Perm J 2019;23:18-041
How It Works
Cannabidiol acts in the brain and in the liver. How far each action carries into an effect you would notice is what separates the strong uses from the marketed ones.
Anatomy of the Practice
1In the brain, in epilepsy
Cannabidiol changes how certain neurons fire, partly through targets other than the usual cannabis receptor, and in these three epilepsies this translates into measurably fewer seizures in controlled trials. This is the only use where the mechanism produces a measured clinical benefit.
2In the stress and fear circuits
At a single dose, cannabidiol dampens activity in brain regions tied to the acute anxiety response. That fits the finding that it blunts the spike of anxiety before a stressful event. The response peaks at a moderate dose and falls off at higher ones.
3In the liver
Cannabidiol is broken down by, and also inhibits, several liver enzymes in the CYP450 family. Those enzymes clear many other drugs, so the same activity can strain the liver and shift the blood levels of medicines taken alongside it.
Using CBD Well
Three things govern whether retail CBD does anything for you: the goal, your medicines, and the product's quality.
If you take an anticoagulant, an anti-seizure drug, or any medicine with a narrow safe range, do not start cannabidiol without checking the interaction with a pharmacist or clinician.
Ways to Do It
Match the goal to the evidence first, sort out your medications, then treat product quality as the main filter. In an unregulated market the label may not match the bottle.
The strong evidence is for three rare epilepsies, a medical route. There is early evidence for acute, situational anxiety. For chronic pain and sleep the evidence is weak, so the free basics in the anxiety and insomnia guides are a solid first step.
Cannabidiol acts on the CYP450 liver enzymes that also clear many common drugs, so before starting anything, list what you take and check it with a pharmacist or clinician. This is free, takes one conversation, and prevents most dangerous interactions.
Labels often miss the real content. Buy only products that publish a recent third-party lab test, a Certificate of Analysis. It should show the actual CBD amount and confirm THC is within legal limits, and it is the one document that verifies what is in the bottle.
Isolate is CBD alone. Broad-spectrum adds other hemp compounds but aims to remove THC. Full-spectrum keeps a trace of THC. That is legal below a set limit, but it can still matter for drug testing and for sensitive users. For anyone who must avoid THC entirely, isolate or a THC-free broad-spectrum with a Certificate of Analysis is the safer pick.
The epilepsy trials used 10 to 20 mg per kilogram of body weight a day, hundreds of milligrams for a child. A typical gummy delivers 10 to 25 mg, so a labeled retail dose is a fraction of what the strong evidence tested.
The relevant experiment used a single oral dose around 300 mg, taken before a specific stressor. Treat that as short-term and early, keep it alongside proper care for an anxiety disorder, and re-check interactions first.
The strong seizure evidence is for a prescription product used under a neurologist, for Dravet syndrome, Lennox-Gastaut syndrome or tuberous sclerosis complex. It comes with liver-enzyme monitoring and dose adjustment of any interacting drugs. This is a medical decision made with a specialist, separate from a retail CBD oil.
Go Deeper
- Anxiety: where the early single-dose CBD evidence fits among approaches with a longer track record, and what to build first for an anxiety disorder.
- Insomnia: the durable, free groundwork for sleep, and the practices the evidence supports first.
- Low back pain: a common reason people reach for CBD, and where the controlled pain evidence actually points.
- Interaction checker: test CBD against the prescriptions you take before you start.
The Chinese Medicine View
Chinese medicine has no concept of CBD, and it would be an invention to claim otherwise. The isolated compound does not map onto the tradition. What the tradition holds is a long, cautious relationship with the cannabis plant itself, dosed to the person and the purpose and used sparingly in its potent parts.
The seed, Huo Ma Ren, or fire hemp seed, is a recognized herb. Classed as sweet and neutral, it moistens the intestines and is used gently for dry constipation in the depleted or the elderly. That is a mild, food-like laxative, a different substance from a concentrated cannabinoid extract, and pointing to it does not lend CBD a tradition it lacks. The flowering and resinous parts of the plant were known too, and classical sources treat them warily, warning that in excess they can disturb the spirit and the mind.
Cautions For This Practice
Extra restraintEverything to be aware of is here, in one place. This practice suits most healthy people; a few situations call for real care.
CBD raised liver enzymes above five times normal in a third of adults at 1,500 mg a day, more so with valproate
Watkins 2021 gave 16 healthy adults a fixed 1,500 mg a day (about 20 mg/kg) for roughly 3.5 weeks; 44% (7 of 16) had ALT above the upper limit of normal and 31% (5 of 16) exceeded five times normal, meeting international drug-induced liver injury criteria and leading to discontinuation, with enzymes recovering on stopping. Dose-dependence comes from the Dravet and Lennox-Gastaut trials, where transaminase elevations were among the more notable adverse effects and concentrated in patients also taking valproate. The signal was strong enough that liver-function monitoring is part of the approved prescribing.Watkins et al., Cannabidiol and Abnormal Liver Chemistries in Healthy Adults: Results of a Phase I Clinical TrialDevinsky et al., Trial of Cannabidiol for Drug-Resistant Seizures in the Dravet Syndrome
CBD raised the active clobazam metabolite about 500%, forcing dose cuts in most children
Geffrey 2015 prospectively measured clobazam and N-desmethylclobazam levels in 13 children as cannabidiol was added. Levels of the active metabolite rose by a mean of about 500%, roughly five- to sixfold, via CYP2C19 inhibition, and 10 of 13 children had their clobazam dose lowered to manage sedation and other effects, while seizure control generally improved. This is a small, uncontrolled prospective series but a well-characterized, mechanistically expected interaction.Geffrey et al., Drug-drug interaction between clobazam and cannabidiol in children with refractory epilepsy
CBD raised one patient's warfarin effect, requiring about a 30% dose reduction
Grayson 2018 reported a 44-year-old man with post-stroke epilepsy (Marfan syndrome, mechanical mitral valve replacement), stable on warfarin, whose INR rose progressively as cannabidiol was titrated up, necessitating serial warfarin dose reductions of about 30% overall to keep the INR in range. Warfarin is metabolized largely by CYP2C9, which cannabidiol inhibits, making the interaction mechanistically expected. As a single case it is low-tier, but it is a documented, plausible, and clinically important warning for a drug where overshoot causes bleeding.Grayson et al., An interaction between warfarin and cannabidiol, a case report
Liver strain, especially with valproate
Cannabidiol can raise liver-enzyme readings, a sign of strain on the liver. This showed up clearly enough in the epilepsy trials that liver monitoring is built into prescription use. In a phase I trial, healthy adults given a fixed 1,500 mg a day saw liver enzymes climb. About a third rose above five times the normal limit, meeting the criteria for drug-induced liver injury. The risk climbs higher again when cannabidiol is taken with valproate, an epilepsy drug. If you use cannabidiol at any meaningful dose, particularly alongside other medicines, liver-function testing is a reasonable precaution to raise with your clinician.
Drug interactions through CYP450
Cannabidiol inhibits several CYP450 liver enzymes that clear many common drugs, so it can push their blood levels up. A prospective study in children showed it raised the active form of clobazam, an anti-seizure drug, about fivefold, so most of the children needed their clobazam dose cut. A documented case showed it raised the effect of warfarin, a blood thinner, enough to need about a 30% dose reduction. That kept the INR, a measure of blood-thinning, in its safe range. If you take an anticoagulant, an anti-seizure drug, or any medicine with a narrow safe range, do not start cannabidiol without checking the interaction with a pharmacist or clinician.
Label accuracy and THC contamination
The over-the-counter market is poorly regulated. When 84 online CBD products were tested, fewer than a third were labeled accurately for CBD content, and THC was detectable in about one in five. That means an uncertain dose, plus the chance of an unwanted psychoactive compound that can show up on a drug test.
Pregnancy, breastfeeding and children
Outside the specific prescription use for rare epilepsies under a specialist, cannabidiol has not been shown safe in pregnancy or breastfeeding, and regulators advise against it in those situations. Keep retail CBD products away from children, both for the unknowns and because of the THC and dosing uncertainty.
Start slow, be smart, read the research, and consult a professional if you have any concerns. This is here to inform your choice, not make it for you.
Common Questions
Does CBD actually work, or is it hype?
It depends entirely on the use, and lumping the uses together is what feeds the hype. The seizure evidence is genuine and strong. Acute, situational anxiety has early single-dose support. For pain and sleep, the evidence is thin and far behind the marketing. So CBD is neither a cure-all nor a scam: it is a compound with a few well-backed uses and many unconfirmed ones.
Is store-bought CBD the same as the prescription kind?
No, and not only in quality. Epidiolex is a specific FDA-approved seizure medicine, dispensed by prescription and taken under a specialist. You cannot buy it on a shelf, and no retail gummy or oil is that product, whatever the label claims. The trials that built CBD's reputation used that pharmaceutical, and no supplement can claim them.
Can CBD interfere with my medications?
Yes, and it is the caution to take most seriously. Bring your full medication list to a pharmacist, and run it through the interaction checker before you start. An anticoagulant or an anti-seizure drug is where it matters most.
What dose of CBD helps with anxiety?
The experiments point to roughly 300 mg taken once, ahead of a one-off stressor such as public speaking. There is no established dose for daily use, or for an anxiety disorder over time, an area with almost no controlled data. If you try it, treat it as short-term, keep real treatment in place, and clear interactions with a pharmacist first.
Does CBD help you sleep?
The evidence is weak and uncontrolled. In one clinic series of 72 adults given CBD mostly for anxiety or sleep, sleep scores improved for about two-thirds in the first month. The improvement was inconsistent from month to month, and there was no comparison group. Much of it tracked falling anxiety, so whether CBD did anything for sleep itself is unclear. The steadier route is the free groundwork in the insomnia guide; if you still try CBD, check the drug interactions before you start.
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 12 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 9, 2026.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.