By Dr. Benjamin Caplan, MD | Board-Certified Family Physician, CMO at CED Clinic | Evidence Watch
A small but carefully controlled pilot trial found that a single oral dose of THC and CBD dramatically reduced REM sleep in insomnia patients, cutting dream-stage sleep by roughly 34 minutes while delaying its onset by over an hour. Despite these pronounced changes in sleep architecture, objective next-day alertness, cognitive performance, and driving ability appeared unaffected at nine or more hours after the dose, though the study’s 20-person sample size means these findings remain firmly preliminary.
Cannabis Pill Disrupts REM Sleep but Leaves Next-Day Performance Largely Intact, Small Trial Finds
A pilot crossover study in 20 insomnia patients shows that a single dose of 10 mg THC combined with 200 mg CBD significantly suppresses REM sleep and delays its onset, while objective measures of next-day alertness, cognitive function, and simulated driving reveal no measurable impairment at nine or more hours post-dose, highlighting both the acute sleep-architecture effects of oral cannabinoids and the critical need for larger, longer-duration trials before any clinical conclusions can be drawn.
#72
Strong Clinical Relevance
One of very few rigorous RCTs measuring cannabinoid effects on objective sleep architecture in diagnosed insomnia patients, though its pilot size limits actionable conclusions.
REM Sleep
Insomnia Treatment
THC/CBD Pharmacology
Next-Day Impairment
Millions of patients already use cannabis products for sleep, yet the field has operated with remarkably little rigorous polysomnographic data from clinical insomnia populations. Most existing evidence comes from healthy volunteers, recreational users, or subjective self-reports, leaving clinicians to advise patients based on inference rather than direct measurement. This pilot trial, while small, represents one of the first efforts to pair gold-standard overnight polysomnography and high-density EEG with next-day functional safety assessments in people who actually carry an insomnia diagnosis. Understanding how cannabinoids reshape sleep architecture in this population is foundational to any future therapeutic strategy.
Insomnia is among the most commonly cited reasons patients pursue medicinal cannabis, yet the evidence base supporting this use has been dominated by subjective surveys and observational data. The present study recruited 20 adults with DSM-5-diagnosed insomnia disorder and administered a single oral capsule containing 10 mg THC plus 200 mg CBD in a double-blind, placebo-controlled crossover design. Each participant served as their own control, completing both treatment and placebo arms with in-laboratory polysomnography (PSG) and 256-channel high-density EEG recording. The mechanistic rationale centers on THC’s well-characterized agonism at CB1 receptors in REM-generating brainstem nuclei, with the high-dose CBD component hypothesized to modulate THC side effects, though the combined formulation precludes isolating individual cannabinoid contributions.
The dominant finding was a striking suppression of REM sleep: participants spent approximately 34 fewer minutes in REM (p less than 0.001, Cohen’s d of negative 1.5) and experienced a 66-minute delay in REM onset (p=0.008, d=0.7) under the THC/CBD condition. Total sleep time was modestly reduced by about 25 minutes (p=0.05, d of negative 0.5), while wake after sleep onset remained unchanged. High-density EEG revealed regional reductions in gamma activity during N2 and delta activity during N3, alongside increased beta and alpha activity during REM. At nine or more hours post-dose, the Maintenance of Wakefulness Test, psychomotor vigilance, and simulated driving showed no significant impairment compared to placebo, though self-reported sleepiness was marginally elevated (d=0.22). The authors emphasize that all findings are preliminary given the 20-person sample and single-dose design, and they call for larger trials examining repeated dosing and individual cannabinoid contributions.
This is exactly the kind of study the field needs more of. The investigators used proper blinding, gold-standard sleep measurement, and next-day functional assessments in actual insomnia patients rather than healthy college students. The REM suppression finding is dramatic and unsurprising to anyone familiar with THC pharmacology, but seeing it quantified with a Cohen’s d of negative 1.5 in a clinical population is genuinely useful. What the study cannot tell us, and this is where I urge caution, is whether this matters for people using cannabis nightly for months or years. REM rebound, tolerance, and the long-term cognitive consequences of chronic REM reduction are entirely unaddressed here.
In my practice, I see many patients who report better sleep with cannabis, and I take those reports seriously. But I also counsel them that “falling asleep faster” or “sleeping through the night” may come at the cost of altered sleep architecture in ways we do not yet fully understand. For patients who are using THC-containing products for sleep, I monitor for signs of REM-related issues like vivid rebound dreaming during breaks, I favor lower THC doses when possible, and I never treat cannabis as a long-term monotherapy for insomnia without also addressing behavioral strategies and underlying contributors. One pilot does not change that approach, but it does sharpen the conversation.
This study sits at an early but important position in the research arc for cannabinoid-based sleep therapeutics. We have long known from recreational-use studies and preclinical models that THC suppresses REM sleep, but translating that observation into a clinical insomnia population with rigorous measurement had not been done with this level of methodological care. The crossover design and within-subject comparisons are well suited to a pilot of this size, and the inclusion of high-density EEG offers spectral detail that standard polysomnography cannot. However, the single-dose protocol captures only acute pharmacology.
