CBD Oil for Sleep - how many drops and when to use 2026

CBD oil for sleep has been surrounded by numbers that are not present in the works cited as sources. This text was created by rechecking each citation at the original source, not by repeating what circulates in guides. The picture has changed significantly: some numbers have disappeared because the work discusses something entirely different, while others remain but in a narrower form. Here you will find the provisional safe dose calculated by EFSA in 2026, along with descriptions of studies by Shannon, Linares, Walsh, and Corroona, including the number of participants and duration. You will not find a dosage table based on body weight or a titration plan. Below we explain why they are not included.

KEY INFORMATION
- EFSA calculated in 2026 a provisional safe dose of cannabidiol at 0.0275 mg per kilogram of body weight per day, which is about 2 mg for a person weighing 70 kilograms.
- The safety of cannabidiol cannot be established in individuals under 25 years of age, in pregnant and breastfeeding women, and in individuals taking medications.
- The statement about 1500 mg per day attributed to WHO has been removed from this entry: the source document does not correspond, and the value diverges from the EFSA calculation by three orders of magnitude.
- The Shannon study from 2019 is a retrospective analysis of the records of 72 patients, and its abstract does not provide any dosage.
- There is no published study on cannabinol with polysomnography or a validated sleep questionnaire.

How much CBD per day is considered a safe dose today?

The EFSA panel calculated in 2026 a provisional safe dose of cannabidiol at 0.0275 mg per kilogram of body weight per day, which is about 2 mg for a person weighing 70 kilograms. This value was derived using the benchmark dose method with an uncertainty factor of 400 and applies only to supplements with a purity of cannabidiol of at least 98 percent, without nanoparticles (EFSA, 2026).

This single number changes the way the entire rest of the text is read. The doses that appear in clinical studies on sleep are hundreds of milligrams administered at once under the supervision of a researcher. A research protocol and a supplement purchased in a store are two different situations, and the number from one does not transfer to the other. For comparison: one drop of oil with a concentration of 5 percent contains roughly as much cannabidiol as the entire daily reference value.

The panel also recorded a statement that appears least frequently in dosage guides but carries the most weight. The safety of cannabidiol cannot be established in individuals under 25 years of age, in pregnant and breastfeeding women, and in individuals taking medications. This is not a precautionary statement added at the end, but a conclusion drawn from a lack of sufficiently quality data. The panel lists gaps concerning the liver, reproductive, nervous, and hormonal systems.

Therefore, there is no dosage table based on body weight or a plan for increasing doses every few days. Such tables circulate on Polish internet and are not backed by any study that would derive them. The number of milligrams appears further only when describing a specific work: who conducted it, how many participants it had, how long it lasted, and how much they received. The statement about 1500 mg per day attributed to WHO, which appeared in this entry in two places, has also been removed.

What are the EFSA's reservations regarding cannabidiol?

The list is longer than the word supplement suggests. The assessment from 2022 indicated significant data gaps, and a search of animal and human literature up to June 2024 confirmed that gaps remain. Many new studies have methodological limitations: non-standardized protocols, short duration, and concurrent pharmacological treatment of participants.

The liver is most frequently mentioned in this assessment. Animal studies have shown consistent hepatotoxicity, with liver mass and histopathological changes being the most sensitive endpoints. It is based on such subchronic studies, conducted according to good laboratory practice principles, that the panel established the toxicological reference point. Human studies indicated a potential hepatotoxicity, especially with concurrent use of other medications, and gastrointestinal effects were reported at higher doses.

The second group of reservations concerns reproduction and development. Animal studies have reinforced concerns about reproductive toxicity, and after prenatal exposure, neurodevelopmental effects suggesting long-term consequences dependent on sex were observed. Hormonal disturbances were also noted, including altered thyroid hormone levels and histopathological changes in the adrenal glands. No study has addressed immunotoxicity, although cannabidiol interacts with immune pathways.

Pharmacokinetics completes the picture. The panel confirms that the bioavailability of cannabidiol is variable and depends on the carrier and the food consumed, and the substance itself crosses the placenta and accumulates in the body. The cautious final conclusion about groups in which safety cannot be established arises from all these premises, not from a single concerning result. For the reader, this means one thing: a lack of data is not the same as evidence of safety.

How can CBD oil affect sleep?

The most honest answer is: indirectly and less effectively than advertised. The 2017 review by Babson and colleagues summarizes that preliminary work on cannabidiol and insomnia indicates therapeutic potential, but studies on cannabis and sleep are in the early stages and have yielded mixed results (Current Psychiatry Reports, 2017).

The same review distinguishes cannabinoids that the industry usually does not separate. THC may shorten the time to fall asleep, but with prolonged use, it worsens sleep quality. Synthetic cannabinoids, nabilone and dronabinol, provide short-term benefits in sleep apnea. The authors primarily associate cannabidiol with REM behavior disorder and excessive daytime sleepiness, while nabilone is linked to reducing nightmares in post-traumatic stress disorder.

The pathway with the most data leads through anxiety, not directly through sleep. The 2015 review by Blessing and colleagues states that preclinical evidence strongly supports the anxiolytic effect of cannabidiol when administered acutely, and human data also support such action, but are limited to acute administration and a few studies in clinical populations (Neurotherapeutics, 2015).

Receptor descriptions that circulate in guides come from animal studies and cell models. There is no study showing the entire chain from receptor to polysomnography outcome in humans. Therefore, the paragraph about inhibiting the FAAH enzyme and increasing anandamide levels as an explanation for the sedative effect has been removed from this entry. The mechanism may be true, but it has not been measured in humans and cannot justify expectations regarding a specific dose of oil.

What did the Shannon study from 2019 really show?

It was a retrospective analysis of documentation from a single psychiatric clinic, not a study with a control group. Records of 103 adult patients were reviewed, and 72 individuals were included in the final sample: 47 primarily reported anxiety, and 25 had sleep issues. Cannabidiol was administered as an adjunct to existing treatment (Permanente Journal, 2019).

The results should be read exactly as they are recorded. Anxiety scores decreased in the first month for 57 patients, which is 79.2 percent, and remained at a lower level throughout the observation period. Sleep scores improved in the first month for 48 patients, or 66.7 percent, but fluctuated in the following months. The authors state that the preparation was well tolerated by all except three patients, and conclude that controlled clinical trials are needed.

Three things that this entry previously mentioned are not supported by the work. The abstract does not mention a single dosage, so the range of 25-175 mg, repeated five times in the text, disappeared along with the table that justified it. There is no basis for the statement that after three months the effect remained stable, as the authors explicitly write about fluctuations. The statement about four patients who withdrew due to adverse effects replaced what is in the work: three patients did not tolerate the preparation well.

The significance of this work should be properly contextualized. The analysis of records from a single clinic shows what happened to patients who received cannabidiol, but it does not indicate how much of this was due to the preparation itself. There was no placebo or random assignment, and patients continued their existing treatment and psychotherapy. This is a premise for further investigation, not proof of efficacy.

Does CBD change sleep architecture?

In healthy volunteers, it did not change it at all. In a randomized crossover study by Linares and colleagues from 2018, 27 healthy individuals were selected, of which one withdrew during the study. Each participant received cannabidiol at a dose of 300 mg or a placebo on the first night, and the substance they had not received previously on the second night (Frontiers in Pharmacology, 2018).

The preparation was administered 30 minutes before the start of an eight-hour polysomnographic recording. Immediately after the recording, cognitive and subjective measurements were taken to capture residual effects. None of the measurements showed a significant difference. The authors conclude that the acute administration of an anxiolytic dose of cannabidiol does not interfere with the sleep-wake cycle in healthy volunteers, and that the result supports the thesis of no impact on normal sleep architecture.

In the previous version of this entry, the study had a third arm with flunitrazepam and stated that the benzodiazepine shortened the REM phase by 19 percent. Such an arm does not exist in the work. Only cannabidiol was compared to placebo in a crossover design, and that concludes the protocol. The figure of 19 percent had no source and was removed along with the sentence that carried it.

A negative result can be misleadingly read as evidence of effect. The study states that for a person who sleeps well, a single dose does not disrupt sleep. It says nothing about whether it helps someone who sleeps poorly. The authors themselves conclude the work with a call for research in patient populations and on the chronic administration of various doses, as this was not covered by their protocol.

Does CBN really help with falling asleep?

There is no published evidence for this. Corroon published a review in 2021 specifically to answer this question. He sifted through the abstracts of 99 studies on humans, qualified eight full texts for detailed analysis, and formulated the conclusion in one sentence: published evidence is insufficient to support claims about the influence of cannabinol on sleep (Cannabis and Cannabinoid Research, 2021).

The details are stronger than the conclusion itself. Studies on cannabinol are outdated and few, with most human studies coming from the 1970s and 1980s, based on small samples with narrow socio-demographic characteristics. Studies measuring sedation or fatigue are rare. The author found not a single published clinical study that compared cannabinol with a validated sleep questionnaire or with polysomnography.

Additionally, there is the observation about quantity. Cannabis products sold with the promise of supporting sleep usually contain 5 mg of cannabinol or less, and the author notes that future studies should use significantly higher doses. Even if an effect existed, the amounts present in a typical oil would likely fall below the threshold at which anything could be observed.

This entry previously claimed that cannabinol prolongs deep sleep and that broad-spectrum oils work better for sleep than isolates precisely because of it. Both statements have been removed. There is no supporting work behind them, and the review that sought this question at the source ends with a recommendation to remain skeptical regarding manufacturers' claims.

Five studies on CBD and sleep along with their limitations

The data looks best where poor sleep is secondary to another problem, but we are still talking about small samples. The following summary collects the works cited in this entry along with the number of participants and duration, because without these two columns, each of these studies sounds stronger than it is.

Three disclaimers should be read together with the table. The study by Walsh and colleagues was the only randomized and placebo-controlled trial in people with chronic insomnia, but it tested a sublingual cannabinoid extract labeled ZTL-101, not pure cannabidiol, so its results cannot be attributed to CBD itself. The work by Elms and colleagues is a case series with open dosing, and its endpoint was the severity of post-traumatic stress symptoms, not sleep quality. Chagas's description includes four patients.

The statement about 91 percent, which was in this entry earlier, referred to sleep improvement. In the study, this percentage describes a decrease in symptom scores on the PCL-5 questionnaire, and regarding sleep, the authors write more cautiously: the preparation provided relief to some patients reporting frequent nightmares. This sentence was corrected, not removed, because the work exists and says something meaningful, just about something else. A broader discussion of the disorder itself is in a separate text about insomnia from the perspective of sleep medicine.

How much of the drops under the tongue is actually absorbed?

No one has measured this in humans. Millar and colleagues reviewed 792 articles in 2018 and found 24 with pharmacokinetic data in humans. Absolute bioavailability could only be provided for one route of administration, inhalation, which is about 31 percent. For the oral and sublingual routes, no study has undertaken such trials (Frontiers in Pharmacology, 2018).

This is why the four-field table with values of 6-10 percent for ingestion, 13-19 percent for sublingual administration, and 30-40 percent for nanoemulsions has disappeared from this entry. These numbers have been circulating in the Polish internet for years, and the work cited as a source is a review on terpenes and the entourage effect in mood disorders. It contains no data on bioavailability. This table has no correct version, so it was replaced by a sentence about what has not been measured.

The practical conclusion is more modest than is often stated. Comparing drops with capsules in percentages today is guesswork. Holding oil under the tongue for a minute is a recommendation from manufacturers, not a result of measurement in humans. It makes more sense to observe the effect on yourself with a consistent method of administration than to calculate how many milligrams reached the bloodstream.

When to take the oil and should you eat before the dose?

The timing is less important than the meal. In Millar's review, the time to reach maximum concentration ranged widely from zero to four hours, and after ingestion and sublingual administration, the peak came slower than after inhalation. No one has designated a good hour before sleep.

The meal changes the picture the most. Birnbaum and colleagues administered a single dose of capsules containing 99 percent pure cannabidiol to eight adults with drug-resistant epilepsy in 2019, once on an empty stomach and once after a high-fat breakfast of 840-860 kilocalories. The maximum concentration was on average fourteen times higher after the meal, and the area under the curve was four times larger. No adverse effects were reported (Epilepsia, 2019). Millar's review confirms the trend: maximum concentration increases in a fed state and in fat formulations.

The practical conclusion sounds different than it is usually formulated. The same dose taken on an empty stomach and after a fatty meal is practically two different exposures. If you want to assess whether something works at all, maintain a consistent routine: the same time in the evening and the same relation to dinner. Changing both things at once makes it impossible to distinguish a good night from good digestion.

The limits of this data must be recognized. Eight participants with drug-resistant epilepsy are not the population of the reader of this text, and a single dose says nothing about chronic use. However, the direction of the effect is consistent in both studies, and that is enough to treat the meal as a variable, not an insignificant detail.

Does more CBD mean better sleep?

The dose-response relationship is not a simple increasing one here. Linares and colleagues administered cannabidiol orally to 57 healthy men in 2019 at doses of 150 mg to fifteen individuals, 300 mg to fifteen, 600 mg to twelve, or placebo to fifteen, in a double-blind setup, before a simulated public speaking event (Revista Brasileira de Psiquiatria, 2019).

Anxiety during the presentation was significantly reduced only by the 300 mg dose. The groups receiving 150 mg and 600 mg did not differ from placebo. The authors write that the result confirms the bell-shaped curve described in animals and that optimal therapeutic doses must be determined rigorously before being transferred to practice. The caveat is significant: healthy men were studied, the endpoint was anxiety before the presentation, not sleep, and the administration was a one-time event.

The second work in this thread concerns cortisol and is often summarized in a way that is the opposite of what it says. Zuardi and colleagues administered placebo or cannabidiol at doses of 300 mg to seven individuals and 600 mg to four in 1993 to eleven volunteers. The sessions took place in the morning. Cortisol levels significantly decreased in sessions with placebo, following the normal diurnal rhythm, and after cannabidiol, this decrease was significantly blunted (Brazilian Journal of Medical and Biological Research, 1993).

In other words, cannabidiol did not lower cortisol but blunted its natural morning drop. Prolactin and growth hormone did not change at all, and a calming effect was noted on self-assessment scales. A previous version of this entry attributed the bell-shaped curve thread to Zuardi's 2017 work in the Journal of Psychopharmacology, leading to a study on empathy after MDMA. The citation was removed, and the claim was rewritten to reflect what both studies actually say.

Czy olejek CBD wchodzi w interakcje z lekami?

Yes, and this is today the best-supported part of all the caution surrounding cannabidiol. Nasrin and colleagues checked in 2021 under laboratory conditions, on microsomes of cells with overexpressed enzymes, how cannabinoids inhibit liver cytochrome P450 enzymes. Cannabidiol competitively inhibited CYP3A4, CYP2B6, CYP2C9, CYP2D6, and CYP2E1 (Drug Metabolism and Disposition, 2021).

The authors add that simple modeling of these results indicates the possibility of pharmacokinetic interactions with drugs that are extensively metabolized by CYP2B6, CYP2C9, or CYP2D6. This is an ex vivo measurement, so it does not establish a dose threshold in humans. Therefore, the statement about enzyme inhibition only above 100 mg per day, which was in this entry in two places, has been removed. It did not come from any of the cited works.

The second source says the same from a different angle. The EFSA panel writes that studies in humans indicated the potential hepatotoxicity of cannabidiol, especially when used together with other medications, and that safety cannot be established in individuals taking medications. If you are taking anything by prescription, talking to your doctor is not a formality but a prerequisite. This also applies to situations where the drug seems harmless, as the interaction occurs at the metabolic level, not at the action level.

The address that supported this paragraph earlier leads to the work of Udomsinprasert and colleagues from 2019 about the relationship between adiponectin gene polymorphism and the risk of anterior cruciate ligament injury. The work exists and has nothing to do with cytochrome P450. The same identifier appears in the second entry of this pair and in dozens of other texts on the blog, always under the claim of drug interactions.

Higiena snu przed suplementem, nie zamiast niego

The preparation will not fix a disrupted circadian rhythm. If the bedtime changes from night to night, and the last hour before sleep looks like any other hour of the day, then the variable governing sleep quality remains largely untouched. It is worth setting it before starting any testing, because otherwise, it will be impossible to say what worked.

The simplest set includes a consistent wake-up time, even on days off, morning exposure to daylight, a cool and dark bedroom, and an evening that is distinctly different from the day. Caffeine consumed in the afternoon still works in the evening. Alcohol shortens the time it takes to fall asleep at the expense of the second half of the night. Intense training just before sleep raises body temperature when it should be dropping. None of these things require purchasing anything.

If the problem lasts longer than a few weeks and interferes with daytime functioning, it's time to talk to a doctor, not to reach for another supplement. Cognitive-behavioral therapy for insomnia is a method conducted with a therapist, and no oil can replace it. We have elaborated on a set of home habits in our post about sposobach na sen bez tabletek, and a separate discussion of melatonin is in the text about melatoninie i jej dawkowaniu.

There is one more reason to start with habits. All the studies described above were conducted in conditions where participants had a stabilized daily schedule and researcher oversight. Transferring their results to a person who goes to bed at different times is comparing two different situations.

How to recognize an oil worth buying?

By the documents, not by the label description. The certificate of analysis for a specific batch, issued by an independent laboratory, is the only place where the declared cannabidiol content meets the measured one. It also checks for heavy metals, pesticide residues, and solvents used in extraction.

Purity gained additional significance after the EFSA opinion. The provisional safe dose calculated by the panel applies only to supplements with a cannabidiol content of at least 98 percent and without nanoparticles. Products outside this description do not even have such a reference point, as the panel explicitly excluded them from its assessment.

Price alone does not determine anything, but a rough estimate can be useful. In the category of oils in the u Bucha store, prices ranged from 65 to 240 złoty per package on August 10, 2026, with various concentrations and volumes. The stock status changes faster than the article, so treat this range as a reference point, not as a price list.

The choice between broad and full spectrum does not have a definitive answer in sleep data today. The Corroona review showed that the argument based on cannabinol is unsupported, and the entourage effect remains a hypothesis mainly described in review papers. If you undergo drug tests at work, full spectrum carries a real risk due to trace amounts of THC. This criterion is more concrete than speculation about synergy.

How to store oil so it doesn't lose its content?

Cool and dark, because cannabidiol is an unstable compound. Fraguas-Sánchez and colleagues studied its stability in solution in 2020 using high-performance liquid chromatography. Temperature turned out to be one of the most critical parameters, with an activation energy of 92.19 kilojoules per mole (Journal of Chromatography B, 2020).

The numbers from this study speak for themselves. At room temperature, cannabidiol was highly unstable, and the time after which 95 percent of the substance remained was 117 days. At a temperature of 5 degrees, the product remained stable for at least 12 months. In an oxidizing environment, the same indicator dropped to 1.77 days, and the substance was also sensitive to light, with the photolytic reaction appearing to be oxidative.

The type of matrix also matters. Cannabidiol is more stable in ethanol than in an aqueous environment, and under conditions close to physiological, at pH 7.4 and a temperature of 37 degrees, 10 percent of the substance degraded within a day. Kosović and colleagues added in 2021 an observation from a stress study conducted according to ICH guidelines: cannabidiol in powder form was significantly more stable than dissolved in sunflower oil (Pharmaceutics, 2021).

The practical conclusion is simple and costs nothing. A dark bottle, a tight seal, a cool place away from the kitchen window, and use within a reasonable time after opening do more for the content of the product than the choice between broad and full spectrum. Oil kept for months on a windowsill may contain significantly less cannabidiol than the label claims, even though the certificate of analysis for that batch was correct at the time of testing.

Summary: what do these studies imply in practice

After checking all sources, the picture is more modest and fairer than before. Cannabidiol has the best-documented anxiolytic effect when administered acutely, and its impact on sleep appears mainly in data where poor sleep accompanies anxiety or post-traumatic stress. In healthy volunteers, a single dose of 300 mg did not change the polysomnographic record in any direction.

The reference point for the supplement remains the EFSA calculation, which is about 2 mg per day for a person weighing 70 kilograms, conditioned by the purity of the product and a clear statement about groups where safety cannot be established. Doses from clinical studies, measured in hundreds of milligrams, belong to research protocols and are not recommendations for anyone other than their participants.

What remains to be done in the evening? A fixed time, a consistent relationship to meals, one product with a certificate of analysis, and patience long enough to distinguish effect from coincidence. If you are taking medications, are pregnant, breastfeeding, or under 25 years old, make the decision about cannabidiol with a doctor, as data on which to base it independently simply do not exist.

Frequently Asked Questions

How many drops of CBD oil should I take for sleep?

There is no study from which such a number could be derived. The only available reference point is the provisional safe dose from EFSA from 2026: 0.0275 mg per kilogram of body weight per day, which is about 2 mg for a person weighing 70 kilograms. The dosage tables by weight have no source and have been removed from this entry.

Does CBD oil treat insomnia?

No. In the analysis of Shannon's charts from 2019, sleep improvement was noted in the first month for 66.7 percent of 72 patients, but the results varied in subsequent months, and the study had no placebo. The only randomized study in people with chronic insomnia tested a cannabinoid extract, not pure cannabidiol.

When is the best time to take CBD oil in the evening?

No one has set a specific hour. In the 2018 review by Millar, the time to reach maximum concentration ranged from zero to four hours. Meal timing is more significant: after a high-fat breakfast, the maximum concentration was on average fourteen times higher than on an empty stomach. Maintain a consistent evening routine.

Does CBD disrupt REM sleep?

In a randomized crossover study by Linares in 2018, a dose of 300 mg did not significantly affect any measured sleep parameter in 27 healthy volunteers, including REM sleep. The study did not include individuals with insomnia and did not compare cannabidiol with benzodiazepines, contrary to what was previously stated in this text.

Is it worth looking for CBN oil for sleep?

The 2021 review by Corroona found no published clinical studies comparing cannabinol with a validated sleep questionnaire or polysomnography. The author concludes that the published evidence is insufficient to support claims of sedative effects and advises skepticism towards manufacturers' declarations.

Can CBD be combined with prescription medications?

Not without consulting a doctor. In laboratory studies, Nasrin in 2021 found that cannabidiol competitively inhibited CYP3A4, CYP2B6, CYP2C9, CYP2D6, and CYP2E1, which are enzymes that metabolize many drugs. The EFSA panel adds that the safety of cannabidiol cannot be established in individuals taking medications.

Is it permissible to use CBD during pregnancy or before the age of 25?

The EFSA panel stated in 2026 that the safety of cannabidiol cannot be established in individuals under 25 years of age, in pregnant and breastfeeding women, and in individuals taking medications. This conclusion stems from a lack of sufficiently quality data, not from results indicating harm. Make the decision with your doctor.

You can check the current assortment and concentrations in the category hemp oils, as the stock status changes continuously.

This article is for informational and educational purposes and does not constitute medical advice. Before starting to use cannabis or CBD for therapeutic purposes, consult a doctor, especially if you are taking other medications, are pregnant, or breastfeeding.

Author: Michał Waluk · Opublikowano: 2026-05-11 · Aktualizacja: 2026-08-10