CBD for Sleep: What the Evidence Shows

Verdict: Weak. Limited clinical evidence for primary insomnia. No large RCTs. Significant product quality and dosing variability.

The claim: CBD promotes relaxation and helps you fall asleep naturally. The products are everywhere: oils, gummies, capsules, tinctures, sprays, teas. The evidence is not.

The claim, disaggregated

The marketing language around CBD and sleep conflates several different propositions. The first is that CBD reduces anxiety, and reduced anxiety makes sleep easier. The second is that CBD is directly sedating. The third is that CBD improves sleep quality through some mechanism related to the endocannabinoid system. These are different claims with different evidence. Most CBD products do not distinguish between them, because the distinction would make the marketing harder.

The anxiety-reduction claim has the most theoretical support. The direct sedation claim may be dose-dependent and applies, if at all, only at very high doses. The sleep-quality claim via the endocannabinoid system is the vaguest and the least tested.

The mechanism

CBD interacts with the endocannabinoid system, serotonin receptors (5-HT1A), and possibly GABA receptors [2]. The endocannabinoid system modulates neurotransmitter release throughout the brain, including in sleep-wake circuits. The proposed pathway to sleep involves anxiolytic effects reducing hyperarousal, rather than direct sedation. This is plausible: hyperarousal is a core maintaining factor in insomnia, and a compound that reduces it could theoretically improve sleep onset.

CBD at very high doses (300mg+) may have direct sedative properties, while lower doses may actually be mildly alerting. This biphasic dose-response pattern is documented in animal studies and is consistent with how CBD interacts with different receptor systems at different concentrations [2]. The implication for consumers is that the dose they are taking may not do what they expect: a low-dose CBD gummy marketed for sleep could, in theory, have the opposite effect.

The dose-response relationship for sleep in humans is not well characterized. No dose-finding study has been published for CBD specifically targeting sleep as a primary outcome.

What the trials show

The most cited evidence is the Shannon case series: 72 adults presenting with anxiety and poor sleep received 25mg CBD capsules as an adjunct to their usual treatment at a psychiatric clinic [1]. Sleep scores (measured by the Pittsburgh Sleep Quality Index) improved in 66.7% of participants at one month, then declined at month two, suggesting either tolerance, regression to the mean, or the waning of a placebo effect. This was a retrospective case series with no control group, no randomization, and no blinding. The participants were receiving other treatments simultaneously. It is the kind of study that generates a hypothesis, not the kind that tests one.

The systematic review by Suraev et al. is the most thorough assessment available. It identified 34 studies on cannabinoids and sleep and concluded the evidence is “limited and of low quality” [3]. Critically, most of the studies in the review examined THC-containing products, not isolated CBD. The few CBD-specific studies were small, heterogeneous in design, and typically enrolled participants with conditions other than primary insomnia (chronic pain, anxiety disorders, PTSD). The review’s conclusion was not that CBD does not work for sleep, but that the right trials have not been done.

The survey by Kuhathasan et al. found that 1,216 adults reported using cannabis products for insomnia and most reported subjective benefit [5]. Survey data on subjective benefit is the weakest form of evidence for efficacy. People who continue to use a product are more likely to report that it helps; those who stopped are underrepresented. The survey cannot separate pharmacological effect from expectation, ritual, or selection bias.

No large, placebo-controlled RCT has tested CBD for primary insomnia in otherwise healthy adults. This is the gap that matters most.

The product quality problem

Even if the evidence for CBD and sleep were stronger, a consumer would face a separate problem: knowing what they are actually taking. The JAMA study by Bonn-Miller et al. tested 84 CBD products sold online and found that only 31% were accurately labeled for CBD content [4]. Some contained substantially more CBD than stated, some substantially less, and 21% contained detectable THC not declared on the label.

This is not a niche finding. It means a consumer buying a “25mg CBD” product has roughly a one-in-three chance that the label matches the contents. The mismatch works in both directions: a product with less CBD than labeled may be inert, while one with more may produce unexpected side effects. The presence of undeclared THC is a separate concern for drug testing, medication interactions, and for people who specifically chose CBD over THC.

The regulatory environment contributes to this problem. CBD products are sold as dietary supplements in the US, which means they are not subject to the same testing and labeling requirements as pharmaceutical products. This makes it difficult for even a well-informed consumer to match the product they buy to the product studied in any trial.

The regulatory and legal landscape

CBD’s regulatory status adds complexity for consumers trying to make informed choices. In the United States, CBD derived from hemp (containing less than 0.3% THC) was federally legalized by the 2018 Farm Bill, but it remains in a regulatory gray area. The FDA has not approved CBD as a dietary supplement, and has issued warning letters to companies making health claims about their CBD products. This means that CBD products are sold without the quality controls that apply to either pharmaceuticals or approved dietary supplements.

The legal status varies internationally and by US state. Some states permit CBD sales with minimal oversight; others restrict or regulate them. This patchwork means that a CBD product purchased in one jurisdiction may not be legally equivalent to one purchased in another, even if the labels appear identical.

For consumers, the practical implication is that the regulatory system provides no assurance that a CBD product is what it claims to be, contains what it says it contains, or is free from contaminants. Third-party testing certificates (COAs) are the closest thing to quality assurance available, but many products do not provide them, and the certificates themselves vary in reliability.

The dose-response problem

The absence of dose-response data for CBD and sleep is a more fundamental gap than it might appear. Most published work uses either very low doses (25mg in the Shannon case series) or very high doses (300-600mg in anxiety studies). These are different pharmacological contexts. At low doses, CBD may act primarily through 5-HT1A serotonin receptor agonism and mild anxiolysis. At high doses, additional receptor systems come into play, including more direct GABAergic effects and possible effects on adenosine signaling.

Without a dose-finding study, neither clinicians nor consumers can determine the optimal dose for sleep, whether a threshold exists below which CBD has no sleep-relevant effect, or whether higher doses produce diminishing returns or qualitatively different effects. The marketing standard of “start low and go slow” is practical advice in the absence of data, but it is not a substitute for the data itself.

How it compares to CBT-I

CBD, at its most optimistic reading, might reduce sleep onset difficulty through anxiety reduction. CBT-I, the first-line treatment for chronic insomnia recommended by the AASM, produces durable improvement in sleep onset latency, total sleep time, and sleep efficiency in 70-80% of people who complete it. The effect sizes are larger, the evidence is from hundreds of trials, and the results last beyond the treatment period without continued use.

The comparison is not intended to dismiss CBD but to provide proportion. A supplement with a handful of small, uncontrolled studies cannot be placed alongside a treatment with decades of RCTs, multiple meta-analyses, and guideline-level recommendations. They are not in the same category of evidence. For someone whose insomnia has an anxiety component, and who is drawn to CBD specifically for its anxiolytic reputation, the comparison with evidence-based anxiety-targeted interventions (CBT-I with a cognitive emphasis, or a conversation with a prescriber about pharmacotherapy) is also worth considering. The evidence for CBD’s anxiolytic effects is stronger than for its sleep effects, but still limited compared to established treatments.

The bottom line

CBD is popular for sleep, plausibly connected to sleep-relevant pathways, and largely untested in the kind of trials that would actually tell us whether it works. The one frequently cited study is a retrospective case series with no control group. The systematic review calls the evidence limited. The product quality problem adds a layer of uncertainty even for people willing to try it on weak evidence: you may not be taking what the label says, and even if you are, the dose-response relationship for sleep has not been mapped.

The evidence does not support CBD as a first-line sleep intervention. If you are considering it, the honest answer is that we do not yet have the data to say whether it works, and the products you can buy do not reliably contain what they claim.

If you’re struggling with sleep regularly, speak with a GP or pharmacist before adding supplements or medications. For persistent insomnia, AASM guidelines recommend CBT-I as the first-line treatment. The American Academy of Sleep Medicine has a provider finder at sleepeducation.org.