Effects of a cannabidiol/terpene formulation on sleep in individuals with insomnia: a double-blind, placebo-controlled, randomized, crossover study

Abstract

Study Objectives:

Cannabidiol (CBD) is increasingly used as a health supplement, though few clinical studies have demonstrated benefits. The primary objective of this study was to evaluate the effects of an oral CBD-terpene formulation on sleep physiology in individuals with insomnia.

Methods:

In this double-blind, placebo-controlled, randomized clinical trial, 125 individuals with insomnia received an oral administration of CBD (300 mg) and terpenes (1 mg each of linalool, myrcene, phytol, limonene, α-terpinene, α-terpineol, α-pinene, and β-caryophyllene) for ≥ 4 days/wk over 4 weeks using a crossover design. The study medication was devoid of Δ⁹-tetrahydrocannabinol. The primary outcome measure was the percentage of time participants spent in the combination of slow-wave sleep (SWS) and rapid eye movement (REM) sleep stages, as measured by a wrist-worn sleep-tracking device.

Results:

This CBD-terpene regimen marginally increased the mean nightly percentage of time participants spent in SWS + REM sleep compared to the placebo (mean [standard error], 1.3% [0.60%]; 95% confidence interval, 0.1–2.5%; P = .03). More robust increases were observed in participants with low baseline SWS + REM sleep, as well as in day sleepers. For select participants, the increase in SWS + REM sleep averaged as much as 48 minutes/night over a 4-week treatment period. This treatment had no effect on total sleep time, resting heart rate, or heart rate variability, and no adverse events were reported.

Conclusions:

Select CBD-terpene ratios may increase SWS + REM sleep in some individuals with insomnia and may have the potential to provide a safe and efficacious alternative to over-the-counter sleep aids and commonly prescribed sleep medications.

Clinical Trial Registration:

Registry: ClinicalTrials.gov; Name: Evaluation of an Oral Cannabidiol (CBD)-Terpene Formulation on Sleep Physiology in Participants With Insomnia; URL: https://clinicaltrials.gov/study/NCT05233761; Identifier: NCT05233761.

Citation:

Wang M, Faust M, Abbott S, et al. Effects of a cannabidiol/terpene formulation on sleep in individuals with insomnia: a double-blind, placebo-controlled, randomized, crossover study. J Clin Sleep Med. 2025;21(1):69–80.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Physicians are increasingly asked by their patients about the merits of using cannabidiol for insomnia and other ailments, but any rigorous clinical research to support recommending its use is lacking. The current study represents the first double-blind, placebo-controlled, randomized, crossover clinical trial to investigate how an oral formulation of cannabidiol and terpenes influences sleep physiology in individuals with insomnia.

Study Impact: In contrast to many over-the-counter sleep aids and commonly prescribed sleep medicines, the cannabidiol-terpene formulation may increase slow-wave and rapid eye movement sleep, which are critical for the immune system, tissue regeneration, cognition, and memory. These results, if confirmed in larger clinical trials, suggest that cannabidiol might offer a promising alternative to other prescription sleep medications and over-the-counter sleep aids.

INTRODUCTION

Insomnia is the most common sleep disorder and an established risk factor for anxiety, depression, and other diseases.^1,2 Approximately 30–40% of the adult population in the United States alone (∼63–84 million) report symptoms of insomnia, and the prevalence of insomnia increases with age.^1,2 Insomnia is defined clinically as the perception or report of inadequate or poor-quality sleep due to a number of factors, such as difficulty initiating or maintaining sleep, waking up too early in the morning, or having nonrestorative sleep.^1,2 It causes significant distress and/or impairment in daytime functioning.^1,2 A clinical diagnosis of insomnia is typically obtained by patients’ reports about their sleep.^2,3 Objective testing is not usually recommended unless another disorder is suspected, yet researchers often use objective testing in sleep studies.^2,3

Cannabidiol (CBD) is a bioactive ingredient produced by Cannabis and hemp and was approved by the Food and Drug Administration in 2018 for the treatment of seizures associated with two rare and severe forms of epilepsy, Lennox–Gastaut syndrome and Dravet syndrome.⁴ Significantly, somnolence was the most frequent adverse event (AE) reported in previous clinical studies on Lennox–Gastaut syndrome and Dravet syndrome.⁵ Preclinical and early human studies provide support that CBD may modulate sleep physiology. A study in rats showed that CBD increases slow-wave sleep (SWS) in a dose-dependent manner.⁶ In a placebo-controlled clinical trial, participants with insomnia (n = 15) that received 160 mg of CBD self-reported an increase in their sleeping time relative to those in the placebo control group in a sleep survey, but this study did not track objective measures of sleep physiology.⁷ In an open-label study in adults with clinically diagnosed anxiety (n = 72), CBD (25–75 mg/d) improved sleep quality in ∼66% of the patients as determined by a sleep-quality questionnaire.⁸ More recently, a randomized controlled pilot trial (n = 15) of CBD (150 mg/night) over a period of 2 weeks in participants with moderate to severe insomnia showed no effect on subjective and objective measures of sleep but showed an improvement in participants’ well-being.⁹ Finally, a recent phase III, prospective, randomized, double-blind, parallel group, multicenter comparative study in 178 patients showed that an oral formulation of CBD (150 mg/ml) improved self-reported measures of sleep quality and anxiety.¹⁰

Terpenes are a class of small molecules produced by most plants, including Cannabis and hemp. Although we know that select terpenes are sedating in mice,^11–15 their effects on sleep physiology in humans have not been studied.

The objective of this interventional study was to evaluate the effects of oral administration of a CBD-terpene formulation on sleep physiology, as measured by a wrist-worn sleep-tracking device over a treatment period of 4 weeks in 125 participants with insomnia.

METHODS

Aims, objectives, and hypotheses

The goal of this trial was to determine the effects of an oral CBD-terpene formulation on sleep physiology in individuals with insomnia. The primary end point of the study was to determine whether the CBD-terpenes formulation increases SWS + rapid eye movement (REM) (percent total sleep time, % TST) sleep over a treatment period of 4 weeks, as quantified with an objective wrist-worn sleep-tracking device. The null hypothesis was that there was no difference in SWS + REM (% TST) sleep in participants that received CBD-terpenes and the placebo control; the alternative hypothesis was that there was a significant difference among the 2 treatment groups.

The rationale to use SWS + REM (% TST) as a primary end point for the current study was based on 2 clinical studies we conducted previously with the same oral CBD-terpene formulation used in this study. These prior studies, which were agnostic in design, measured the following sleep physiology parameters as tracked by a wrist-worn sleep-tracking device: TST, light sleep, SWS sleep, and REM sleep; these studies provided the preliminary evidence that SWS + REM (% TST) would be a suitable primary end point for the current study (data not shown). The combination of SWS and REM sleep stages is commonly referred to as “restorative” sleep and is thought to strengthen the immune system, increase cell/tissue regeneration and brain metabolite clearance, and replenish energy stores.^16,17 SWS and REM sleep are also critical for learning, memory, attention, and executive function.^16,17 As such, we posited that using SWS + REM (% TST) as a primary end point for the current study would have potential clinical significance for those who have insomnia.

Study design and participant population

This study involved a double-blind, placebo-controlled, randomized, crossover design in which participants cycled through 2 independent treatment groups (Group 1: treatment A, placebo and treatment B, CBD-terpenes; Group 2: treatment A, CBD-terpenes and treatment B, placebo), each for 4 weeks (Figure 1). The study included a 2-week run-in/baseline period during which participants were monitored for protocol adherence to ensure they were wearing the sleep-tracking device and collecting sleep data correctly. The run-in/baseline period was followed by 2 independent 4-week treatment periods (A/B and B/A), which were each followed by a 1-week wash-out period (Figure 1). A 1-week wash-out period was chosen to rule out any carryover effects, because the elimination half-life of CBD is 2–5 days after chronic oral administration.¹⁸

Figure 1. Crossover study design.

CBD = cannabidiol.

This study was not performed in sleep clinics but rather was conducted in a decentralized manner with participants recruited from across the United States who completed the study in the comfort of their own homes. Although studies performed in sleep clinics offer several clear advantages, including a carefully controlled sleep environment, we wanted to conduct this research study in a real-world setting that would be representative of the majority of current CBD users.

This study was approved by Allendale Institutional Review Board, Lyme, Connecticut. Recruitment, enrollment, informed consent, and distribution of study materials to participants was managed by a contract research organization (83 Bar, Austin, Texas). Recruitment was initiated on February 18, 2022. A total of 13,385 potential study leads were recruited from across the United States via social media advertising (Figure 2). After a prospective study participant clicked on a social media ad they were directed to a dedicated landing page that included an embedded survey. The survey was designed to prescreen prospective candidates. The presence of chronic insomnia in all participants was determined in this online survey as a self-reported difficulty initiating (latency to persistent sleep > 30 minutes) and/or maintaining sleep (> 30 minutes awake during the middle of the night or waking > 30 minutes before desired waking time on 3 or more nights per week) for at least 3 months. In addition, all participants scored as having severe insomnia after taking a clinically validated insomnia severity index¹⁹ that was also part of the survey. The study identified 422 qualified leads from survey respondents, who were further screened via phone interviews conducted by registered nurses to ensure that potential study participants met all of the inclusion and exclusion criteria, resulting in the enrollment of 125 participants (101 females and 24 males) in the study (Figure 2). Each participant provided written informed consent prior to participation. The average insomnia severity index score of participants in our study was 26 ± 2 (severe clinical insomnia) on a scale from 0–28.¹⁹

Figure 2. Flow diagram of study participation.

Allocation and blinding

A total of 125 participants with chronic insomnia were randomly assigned in a 1:1 allocation ratio to 1 of 2 treatment sequences (n = 63 A/B sequence group and n = 62 B/A sequence group) using a computer-generated randomization code (https://www.randomizer.org) that was implemented by an independent researcher (Figure 2). Allocation was concealed and medications were numbered to mask the randomization code. Study participants and research personnel were blinded to the randomization sequence and group allocation throughout the study. Every participant cycled through both treatment groups. Participants were instructed to take the treatment on a minimum of 4 nights/wk. Throughout the study, participants responded to a text-based survey on a daily basis to record whether they had taken the study medication on the previous evening.

Study medication

The study medication was composed of capsules made from vegetable cellulose that contained > 99.9% purity hemp-derived CBD (300 mg) and 1 mg each (> 98% purity, food grade) of the terpenes linalool, myrcene, phytol, limonene, α-terpinene, α-terpineol, α-pinene, and β-caryophyllene dissolved in organic coconut oil. The CBD and terpene doses were established from a series of in-house dose-ranging studies (data not shown). The CBD and terpenes were purified in Good Manufacturing Practice–certified facilities. Participants took capsules with a glass of water 1 hour before going to sleep. Of note, Δ⁹-tetrahydrocannabinol was undetectable in the study medication, as measured by an independent ISO-10725–accredited analytical testing laboratory (Figure S1 in the supplemental material). Furthermore, capsules were free from potential contaminants (pesticides, residual solvents, or heavy metals) (data not shown). Placebo-control capsules were identical to those that contained the treatment medication but contained only organic coconut oil.

Efficacy assessments

The study was conducted in a decentralized manner in which all participants took treatments and slept in their own homes. Objective sleep data were gathered throughout the study from a noninvasive sleep-tracking wrist-worn device (WHOOP, https://www.whoop.com) that electronically collected and transmitted sleep data from study participants in the comfort of their own beds. The wristband collects hundreds of data points per second from a 3-axis accelerometer, 3-axis gyroscope, and heart-rate sensor. The wristband can accurately measure TST (minutes) and the time spent in sleep stages (defined as light sleep, SWS [deep] sleep, and REM sleep). The device also collects data using photoplethysmography, a technology that quantifies blood flow by measuring superficial changes in blood volume. Heart rate, heart rate variability, and respiratory rate are derived from photoplethysmography data, and these metrics are processed by WHOOP’s sleep detection and staging algorithms. Two recent validation studies, published independently from WHOOP, indicate that data on sleep stages collected from the WHOOP device correlate reasonably well with polysomnography (PSG), the gold standard of sleep tracking used in clinical studies conducted in sleep clinics.^20,21 Specifically, in one study the WHOOP device had low bias (13.8 minutes) and precision (17.8 minutes) errors for measuring sleep duration and measured REM sleep and SWS accurately (intraclass coefficient, 0.74 ± 0.28 and 0.85 ± 0.15, respectively).²⁰ In addition, the accuracy of WHOOP-based measurements of heart rate, respiratory rate, and heart rate variability were excellent compared with the gold-standard PSG.²⁰ In a second independent study that compared the WHOOP device to PSG, the sensitivity (ie, the percentage of PSG-determined sleep epochs correctly identified by the WHOOP device) to light sleep, SWS, and REM sleep was 62%, 68%, and 70%, respectively.²¹

Participants were blinded to the sleep data collected by the WHOOP device and responded to a daily text-based questionnaire throughout the duration of the study to report whether or not they had taken the study medication on the previous night. Thus, objective sleep data were analyzed only for nights when participants took the study medication.

Questionnaires

A modified version of a clinically validated questionnaire/survey entitled the patient’s global impression (PGI)²² was used to assess perceptions of sleep by the study participants after each 4-week treatment period. The PGI is a four-item participant self-report that assesses treatment benefit to sleep quality (item 1), sleep induction (item 2), sleep duration (item 3), and appropriateness of study medication strength (item 4).²² The PGI was used in previous a clinical study that led to the approval of Zolpidem.²² Each item in the PGI is presented as a survey to patients that consists of a 3-point categorical scale, with a score of 1 representing a treatment benefit/advantage on items 1–3 (“too strong” on item 4), a score of 2 representing no effect/change on items 1–3 (“just right” on item 4), and a score of 3 representing worsening/disadvantage on items 1–3 (“too weak” on item 4).²² We modified the PGI by removing item 4 (appropriateness of study medication strength) from the survey. Our modified PGI also included 2 additional items to assess a potential treatment benefit to sleep depth (item 4) and treatment effect on the presence of vivid dreams (item 5). Moreover, each item in our modified PGI was presented as a survey that consisted of a 10-point categorical scale, with scores of 6–10 representing a treatment benefit/advantage, a score of 5 representing no effect of treatment, and a score of 1–4 representing a treatment worsening/disadvantage. These data were collected from study participants in the form of a 5-minute survey that was completed on their mobile phones at the end of each 4-week treatment period.

Safety assessments

A medical monitor (JGB Biopharma, Belmont, California) was used to assess AEs for causality and severity and for final review and confirmation of accuracy of event information and assessments. An AE was defined as any untoward medical occurrence in a study participant during the trial period. A serious AE was defined as an event that resulted in death, was life-threatening, required patient hospitalization, resulted in a persistent or significant disability/incapacity, and/or resulted in a medically important event or reaction. The following guidelines were used to describe the severity of AEs using the Common Terminology Criteria for Adverse Events grading system: Grade 1 Mild – Events that required minimal or no treatment and did not interfere with the participant’s daily activities; Grade 2 Moderate – Events that resulted in a low level of inconvenience or concern with the therapeutic measures. Moderate events may cause some interference with functioning; Grade 3 Severe – Events that interrupted a participant’s usual daily activity and may have required systemic drug therapy or other treatment. Severe events are usually potentially life-threatening or incapacitating. The term “severe” does not necessarily equate to “serious”; Grade 4 Life Threatening – Events in which urgent medical intervention was indicated; Grade 5 Death – If related to the event. The medical monitor was responsible for assessing the relationship to study treatment using clinical judgment and the following considerations: Not related: There was not a reasonable possibility that the administration of the study intervention caused the AE, there was no temporal relationship between the study intervention and event onset, or an alternate etiology had been established; Related: The AE was known to occur with the study treatment, there was a reasonable possibility that the study intervention caused the AE, or there was a temporal relationship between the study intervention and event. Reasonable possibility means that there is evidence to suggest a causal relationship between the study intervention and the AE. Participants in our trial were instructed to respond to a daily text-based survey to determine whether they had taken the study treatment on the previous evening and whether they had experienced any AEs. In the event that the participant reported an AE, the medical monitor was instructed to follow up directly with the participant to assess the AE for causality and severity.

Power analysis

Based on the results from 2 smaller clinical trials we conducted previously with the CBD-terpenes formulation, we calculated that a total of 100 participants would provide at least 90% power to detect a 15% difference in the percentage of time participants spend in SWS + REM sleep, as quantified using the WHOOP sleep-tracking device after the 4-week treatment phase. Based on our previous clinical studies, we anticipated a dropout rate of 33%.

Statistics

The primary end point of the study was to determine whether a CBD-terpenes formulation increases SWS + REM (% TST) sleep, as quantified with an objective wrist-worn sleep-tracking device. The null hypothesis was that there was no difference in SWS + REM (% TST) sleep in participants that received CBD-terpenes and the placebo control; the alternative hypothesis was that there was a significant difference among the two treatment groups. The mean nightly difference in SWS + REM sleep (% TST) in participants that received CBD-terpenes and the placebo control at the end of the 2 4-week treatment phases was analyzed using a 2-sided test with an alpha level at 0.05 to evaluate superiority. The P value and 95% confidence interval for the estimate of treatment difference in percentages was estimated and constructed using the above-mentioned method.

Objective sleep data were also computed for treatment effects, period effects, and carryover effects by the method reported by Hills and Armitage for 2-period crosssover clinical trials.²³ A pretest was used to assess potential carryover effects, whereby the sum of the values in the 2 periods was calculated for each participant and compared across the 2 sequence groups. If results were statistically significant (P < .05), then there was evidence of relevant carryover effects and further between-groups difference tests were not undertaken. Statistical analysis was performed with Prism 10.0.2 (GraphPad, San Diego, California).

RESULTS

Participants

A total of 13,835 potential participants were recruited via social media advertising and screened for isomnia via an online survey (Figure 2). This resulted in 422 potential participants that qualified for further lead screening via a phone interview with registered nurses, and 125 participants that fulfilled the inclusion criteria and were randomly allocated to the active or placebo group.

During the 2-week baseline period, 63 participants were excluded from the study, either voluntarily (n = 22), lost to follow-up (n = 18), or due to deviation from the study protocol because they did not collect sleep data using the WHOOP device (n = 23) (Figure 2). This high and unanticipated attrition rate resulted in a modified intent-to-treat/per-protocol approach, and 62 participants completed the study. Furthermore, after the data were unblinded, data from 6 participants (4 participants from Group 1 [A/B] and 2 participants from Group 2 [B/A]) were excluded from the final data analysis due to protocol deviations (eg, due to a lack of reporting whether or not they took the study treatment and/or insufficient sleep data collection). Thus, the final data analysis included complete data sets from 56 study participants (33 participants from Group 1 [A/B] and 23 participants from Group 2 [B/A]) (Figure 2).

A CBD-terpene formulation marginally increases SWS + REM sleep (% TST) and significantly decreases light sleep (% TST) in individuals with insomnia

Study participants with insomnia orally self-administered capsules that contained CBD (300 mg) and 8 individual terpenes (1 mg each) or a placebo-control capsule, on a minimum of 4 nights/wk over a 4-week treatment period (Figure 1). After a 1-week wash-out period, participants crossed over and self-administered the second treatment (ie, capsules with CBD-terpenes or the placebo control) for a second 4-week treatment period (Figure 1).

Objective sleep physiology data were collected from a wrist-worn sleep-tracking device.^20,21 The CBD-terpenes treatment marginally increased the mean nightly time participants spent in SWS + REM sleep (% TST) and SWS sleep (% TST) and significantly decreased the mean nightly light sleep time (% TST) compared with the placebo control (Figure 3A, Figure 3B, Figure 3G, and Table 1, respectively). The CBD-terpenes treatment also increased the absolute mean nightly time participants spent in SWS sleep (minutes) and SWS + REM sleep (minutes) (Figure 3D, Figure 3E and Table 1). However, the CBD-terpenes treatment did not change the mean relative (% TST) or absolute (minutes) time participants spent in REM sleep (Figure 3C, Figure 3F, and Table 1), nor did it change the mean absolute time (minutes) participants spent in light sleep (Figure 3H and Table 1). The treatment had no significant effect on TST (minutes) (Figure 3I and Table 1). No significant effect of sex was observed for any outcome measure (data not shown).

Figure 3. Treatment of study participants with CBD-terpenes marginally increased SWS + REM (% TST) sleep in patients with insomnia.

Treatment with CBD-terpenes increased SWS + REM (% TST) (A), SWS (% TST) (B), SWS + REM (minutes) (D), and SWS (minutes) sleep (E) and decreased light sleep (% TST) (G). This treatment did not change REM (% TST) (C), REM (minutes) (F), or light sleep (minutes) (H), nor did it change TST (minutes) (I). Shown is the mean nightly value (mean ± standard error) for each outcome measure over a treatment period of 28 days. Participants were required to take the treatments ≥ 4 times/wk and averaged 22 days of treatment for each treatment arm over the course of the 28-day treatment period. n = 56 study participants. *P < .05; **P < .01 (t test). CBD = cannabidiol, ns = not statistically significant, REM = rapid eye movement, SWS = slow-wave sleep, TST = total sleep time.

Table 1.

Mean difference in SWS + REM (% TST), SWS (% TST), REM (% TST), light sleep (% TST), SWS + REM (minutes), SWS (minutes), REM (minutes), light sleep (minutes), and TST.

Outcome Measure	All Study Participants (n = 56)					Low Baseline SWS + REM (% TST) (n = 21)					Day Sleepers (n = 7)
	Placebo	CBD-Terpenes	MD (SEM)	95% CI	P	Placebo	CBD-Terpenes	MD (SEM)	95% CI	P	Placebo	CBD-Terpenes	MD (SEM)	95% CI	P
SWS + REM (% TST)	41.8	43.1	1.3 (0.6)	0.1 to 2.5	*	32.0	35.8	3.8 (0.9)	2.0 to 5.5	****	38.4	45.5	7.1 (2.0)	4.4 to 12.1	****
SWS (% TST)	18.1	19.1	1.0 (0.3)	0.4 to 1.6	**	14.4	15.8	1.4 (0.5)	0.4 to 2.3	**	14.2	17.9	3.7 (0.9)	1.9 to 5.5	****
REM (% TST)	23.7	24.0	0.3 (0.4)	−0.6 to 1.2	n.s.	17.6	20.0	2.4 (0.7)	1.0 to 3.7	***	24.2	27.6	3.4 (1.5)	0.5 to 6.3	*
Light sleep (% TST)	57.7	56.0	−1.7 (0.6)	−2.8 to −0.5	**	68.0	64.3	−3.7 (0.9)	−5.5 to 2.0	****	62.5	54.5	−8.0 (2.0)	−4.2 to 11.9	****
SWS + REM (minutes)	157.7	162.6	4.9 (2.6)	−0.6 to 9.5	*	119.9	129.8	9.9 (4.3)	1.8 to 18.6	*	146.4	169.1	22.7 (10.0)	0.0 to 39.3	*
SWS (minutes)	67.7	71.2	3.3 (1.3)	0.7 to 6.0	*	53.7	56.4	2.7 (2.0)	−1.2 to 6.6	*	54.6	64.8	10.2 (3.7)	2.9 to 17.5	**
REM (minutes)	90.0	91.4	1.3 (2.0)	−2.7 to 5.3	n.s.	66.2	73.4	7.2 (3.0)	1.3 to 13.0	*	91.8	104.3	12.5 (6.7)	−0.6 to 25.7	n.s.
Light sleep (minutes)	212.9	206.2	−6.7 (0.6)	−2.8 to −0.5	n.s.	252.7	235.9	−16.8 (6.7)	−30.0 to −3.6	*	242.3	207.4	−34.9 (12.0)	−11.3 to 58.3	**
TST	373.4	366.5	−6.9 (4.7)	−16.0 to 2.4	n.s.	372.1	353.1	−19.0 (8.5)	−35.7 to −2.3	*	387.8	376.5	−11.3 (13.8)	15.8 to 38.4	n.s.

Shown is the mean difference between placebo and CBD-terpenes groups for the total population of study participants (n = 56), participants with low baseline SWS + REM (% TST) (n = 21), and day sleepers (n = 7). Participants were required to take the treatments 4 times/wk and averaged 22 (total participants), 21 (low baseline SWS + REM [% TST]), and 25 (day sleepers) days of treatment for each treatment group over the course of the 28-day treatment period. *P < .05; **P < .01; ***P < .001; ****P < .0001 (t test). CBD = cannabidiol, CI = confidence interval, n.s. = not statistically significant, MD = mean difference, REM = rapid eye movement, SEM = standard error of the mean, SWS = slow-wave sleep, TST = total sleep time.

Objective sleep data were also computed for treatment effects, period effects, and carryover effects by the method reported by Hills and Armitage for 2-period crosssover clinical trials.²³ This analysis confirmed that the CBD-terpenes treatment significantly increased SWS + REM sleep (% TST) and SWS sleep (% TST), and decreased light sleep (% TST) compared with the placebo control (Table S1 in the supplemental material). No evidence of relevant carryover or period effects was observed for the CBD-terpenes treatment on these outcome measures (data not shown). Using this statistical analysis, the CBD-terpenes treatment did not significantly change REM sleep (% TST), nor did it change the absolute time participants spent in SWS + REM (minutes), SWS (minutes), or REM (minutes) sleep, although there was a trend toward increasing absolute SWS + REM sleep (minutes) (Table S1). The CBD-terpenes treatment did not significantly change TST (minutes) (Table S1). No significant effect of sex was observed for any outcome measure (data not shown).

Analysis of the raw SWS + REM (% TST) data suggested that the CBD-terpenes treatment may have skewed the data distribution toward a decrease in the time participants spent in low levels (0–24%) of SWS + REM sleep (% TST) and an increase in the time participants spent in high levels (25–54%) of SWS + REM sleep (% TST) (Figure S2A in the supplemental material). To assess, retrospectively, whether there was an association between treatment with CBD-terpenes and SWS + REM (% TST) sleep, these data were analyzed as a contingency table. A Fisher’s exact test confirmed that treatment with CBD-terpenes significantly shifted the distribution of data away from low levels and toward high levels of SWS + REM sleep (% TST) (P = .004; odds ratio, 1.4; 95% confidence interval, 1.1–1.8) (Figure S2A). Conversely, analysis of the raw light sleep (% TST) data suggested that the CBD-terpenes treatment may have skewed the data distribution toward an increase in the time participants spent in low levels (0–54%) of light sleep (% TST) and a decrease in the time participants spent in high levels (55–100%) of light sleep (Figure S2B). To assess, retrospectively, whether there was an association between treatment with CBD-terpenes and light sleep (% TST), these data were analyzed as a contingency table. A Fisher’s exact test confirmed that treatment with CBD-terpenes significantly shifted the distribution of data away from high levels and toward low levels of light sleep (% TST) (P = .048; odds ratio, 0.8; 95% confidence interval, 0.7–1.0) (Figure S2B).

Treatment with CBD-terpenes increases SWS + REM sleep (% TST) and other objective sleep outcome measures in the majority of study participants

For numerous Food and Drug Administration–approved and over-the-counter drugs, the response rate to drug treatment in clinical trials is highly variable, due in part to genetic variants in drug metabolism.^24,25 In the current study we observed that SWS + REM sleep (% TST and minutes) increased in 55% and 61% of study participants, respectively, treated with CBD and terpenes (Table 2). In contrast, light sleep (% TST and minutes) decreased in 57% and 61% of study participants, respectively (Table 2). For some objective outcome measures evaluated in this study, including SWS + REM (% TST), SWS (% TST), and SWS + REM (minutes), the percentage of participants that responded positively to treatment with CBD-terpenes was ∼2-fold higher than those with a negative response and ∼4-fold higher than those with a neutral response (Table 2). For comparison, clinical studies of CBD treatment in children with Lennox–Gastaut syndrome and Dravet syndrome (for which CBD is Food and Drug Administration–approved) indicate that 32% and 43% of patients, respectively, have a positive response to drug treatment.^26,27

Table 2.

Relative percentage of study participants who showed an increase, decrease, or no effect of treatment with CBD-terpenes on objective sleep outcome measures.

Outcome Measure	All Participants (n = 56)	Low Baseline SWS + REM (% TST) (n = 21)	Day Sleepers (n = 7)	All Participants (n = 56)	Low Baseline SWS + REM (% TST) (n = 21)	Day Sleepers (n = 7)	All Participants (n = 56)	Low Baseline SWS + REM (% TST) (n = 21)	Day Sleeper (n = 7)
	Increase	Increase	Increase	Decrease	Decrease	Decrease	No Effect	No Effect	No Effect
SWS + REM (% TST)	55%	67%	86%	30%	19%	0%	14%	14%	14%
SWS (% TST)	54%	52%	86%	32%	33%	14%	14%	14%	0%
REM (% TST)	48%	67%	86%	38%	19%	0%	14%	14%	14%
Light sleep (% TST)	32%	24%	0%	57%	67%	86%	11%	9%	14%
SWS + REM (minutes)	61%	62%	86%	35%	33%	14%	4%	5%	0%
SWS (minutes)	48%	43%	86%	48%	57%	14%	4%	0%	0%
REM (minutes)	50%	71%	86%	48%	29%	14%	2%	0%	0%
Light sleep (minutes)	39%	29%	0%	61%	71%	100%	0%	0%	0%
TST	57%	52%	43%	41%	43%	57%	2%	5%	0%

CBD = cannabidiol, REM = rapid eye movement, SWS = slow-wave sleep, TST = total sleep time.

A subset of study participants show robust beneficial responses to treatment with CBD-terpenes

This clinical study used a crossover design in which each participant served as their own control, and therefore the data from each study participant could be quantified individually (equivalent to single-patient [n-of-1] clinical trials).²⁸ Analysis of the top 5 responding participants’ sleep data individually (5/56 or ∼9% of all study participants) showed that the CBD-terpenes treatment robustly increased the relative and absolute mean nightly time these participants spent in SWS + REM sleep (% TST and minutes, respectively) (Figure S3 and Figure S4 in the supplemental material and Table 3). In these participants, the nightly mean increase in SWS + REM sleep (minutes) was 48 minutes over a 4-week treatment period (Figure S4 and Table 3). These participants also averaged 32 minutes less light sleep per night over the 4-week treatment period (data not shown). Of note, here too the CBD-terpenes treatment did not affect TST (minutes) (data not shown).

Table 3.

Mean difference in SWS + REM (% TST) and SWS + REM (minutes) for top-responding individual study participants (n = 5).

	SWS + REM (% TST)	SWS + REM (% TST)	SWS + REM (% TST)	SWS + REM (% TST)	SWS + REM (% TST)	SWS + REM (minutes)	SWS + REM (minutes)	SWS + REM (minutes)	SWS + REM (minutes)	SWS + REM (minutes)
	Placebo	CBD-Terpenes	MD (SEM)	95% CI	P	Placebo	CBD-Terpenes	MD (SEM)	95% CI	P
DR24	34.9	44.0	9.1 (3.8)	1.2 to 16.9	*	121.3	169.6	48.3 (20.1)	(7.0 to 89.5)	*
DR45	38.1	45.7	7.6 (2.9)	1.6 to 13.5	*	144.1	192.7	48.6 (18.6)	(11.2 to 85.9)	*
DR51	11.5	24.2	12.7 (3.8)	4.9 to 20.5	**	45.3	96.7	51.4 (14.2)	(22.6 to 80.0)	***
DR111	43.2	48.2	5.0 (2.3)	0.3 to 9.6	*	157.7	198.2	40.5 (14.3)	(11.5 to 69.5	**
DR128	38.5	56.0	17.5 (3.4)	10.6 to 24.4	****	159.3	207.7	48.4 (19.8)	(8.3 to 88.4)	*

Shown is the mean difference between control and treatment groups for top-responding individual study participants. Participants were required to take the treatments 4 times/wk and averaged 19 days of treatment for each treatment group over the course of the 28-day treatment period. *P <.05; **P < .01; ***P < .001; ****P < .0001 (t test). CBD = cannabidiol, CI = confidence interval, DR = defined research, MD = mean difference, REM = rapid eye movement, SEM = standard error of the mean, SWS = slow-wave sleep, TST = total sleep time.

Study participants with low baseline SWS + REM (% TST) sleep (≤ 40%/night) show robust beneficial responses to treatment with CBD-terpenes

The mean nightly SWS + REM sleep (% TST) in hundreds of thousands of users of the WHOOP sleep-tracking device is ∼42%.²⁹ We next analyzed the effects of the CBD-terpenes treatment on objective sleep outcome measures in a subset of study participants with low baseline SWS + REM sleep (% TST), which was chosen arbitrarily as ≤ 40%/night. In these participants (n = 21/56, or 38% of study participants), treatment with CBD-terpenes increased the mean nightly SWS + REM sleep (% TST) and other objective sleep readouts by a greater magnitude and higher level of statistical significance than in the total study population (Table 1, Figure S5 in the supplemental material, and Table S2). Surprisingly, these beneficial responses were observed despite these participants’ having potentially lower TST (Figure S5). Moreover, in this subset of study participants with low baseline SWS + REM sleep (% TST), we observed that the relative percentage of participants who responded positively to treatment with CBD-terpenes increased relative to the response rate observed in the total study population (67% in participants with low baseline SWS + REM [% TST] sleep vs 55% in total study population), whereas the relative percentage of participants who responded negatively decreased (19% in participants with low baseline SWS + REM [% TST] sleep vs 30% in total study population) (Table 2).

Study participants who sleep during the day show the most robust beneficial responses to treatment with CBD-terpenes

One of the exclusion criteria in this study was if any prospective participant was a night shift worker during the 12 months prior to the study and during the study. Nonetheless, inspection of the raw sleep data after unblinding indicated that a subset of participants who enrolled in our study slept only during the daytime. In this subset of study participants (7/56, or ∼13% of study participants), treatment with CBD-terpenes increased mean nightly SWS + REM sleep (% TST) and other objective sleep readouts in a highly significant manner and by a much greater magnitude than that observed in the total study population (Table 1, Figure S6, and Table S3). Moreover, in these day-sleeping participants we observed that the relative percentage of participants who responded positively to treatment with CBD-terpenes increased dramatically relative to the response rate in the total study population (86% in day sleepers vs 55% in total study population, respectively), whereas the relative percentage of participants who responded negatively decreased to zero (0% in day sleepers vs 30% in total study population) (Table 2).

Effects of treatment with CBD-terpenes on self-reported measures of sleep

A modified version of a clinically validated questionnaire/survey, the PGI,²² was used to assess insomnia symptoms as perceived by the study participants after each 4-week treatment period. This modified PGI is a five-item participant self-report that assesses treatment benefit to sleep quality (item 1), sleep induction (item 2), sleep duration (item 3), sleep depth (item 4), and treatment effect on vivid dreams (item 5).²² Treatment with CBD-terpenes modestly improved all of these measures of sleep in study participants relative to the placebo control, but these improvements did not reach statistical significance (data not shown). Note that participants with low baseline SWS + REM (% TST) and day sleepers showed a trend to perceive improvements in sleep induction and duration (data not shown), despite their objective data indicating that their TST may have slightly decreased. Remarkably, despite the absence of statistically significant difference in these self-reported outcome measures, a higher percentage of study participants perceived a benefit of treatment with CBD-terpenes as opposed to worsening or neutral response compared with the placebo control treatment (Table 4). Furthermore, the percentage of participants with low baseline SWS+REM (% TST) and day sleepers who perceived a benefit of treatment with CBD-terpenes to sleep induction, duration, and depth was higher than that reported by all study participants (data not shown).

Table 4.

Relative percentage of study participants who showed an improvement, worsening, or no effect of treatment with CBD-terpenes on self-reported sleep outcome measures (n = 56).

Metric	Better	Worse	No Effect
Sleep quality	45%	41%	14%
Sleep induction	56%	30%	14%
Sleep duration	52%	34%	14%
Sleep depth	48%	38%	14%
Vivid dreams	48%	30%	22%

Treatment with CBD-terpenes was well-tolerated and was not associated with any AEs

Study participants responded to a daily text message to describe any potential AEs they experienced throughout our study, and these text messages were tracked on a daily basis by a medical monitor. No AEs were reported by any participants throughout this 12-week study. Moreover, treatment with the CBD-terpenes formulation caused no significant effect on resting heart rate or heart rate variability (data not shown).

DISCUSSION

Despite the promising therapeutic potential of CBD and terpenes, sparse human clinical evidence supports their use for the treatment of patients with insomnia or other sleep disorders. Here we report that daily (or near-daily) administration of an oral CBD-terpenes formulation marginally increased the mean nightly time that 56 participants with severe insomnia spent in SWS + REM sleep (% TST and minutes) and significantly decreased the time spent in light sleep, as measured by an objective wrist-worn sleep tracking device.

As with many interventional drug studies, we observed variability in the treatment response to CBD-terpenes in our participant population for the objective outcome measures tracked in our study. Given that 55% and 61% of the total study participants had an increase in SWS + REM (% TST and minutes, respectively), it is not surprising that the magnitude of the treatment benefits across the total study population was marginal and of questionable clinical significance. However, in certain subpopulations in our study (eg, in participants with low baseline SWS + REM sleep and in day sleepers) the magnitude of the beneficial responses was much more robust and achieved higher levels of statistical significance. For example, day sleepers averaged a mean nightly increase of 23 minutes in SWS + REM sleep and a decrease of 35 minutes in light sleep (Table 1), whereas the top-responding participants in our study averaged a nightly increase of 48 minutes in SWS + REM sleep (Table 3) and a decrease of 32 minutes of light sleep over a 4-week treatment period. It is also worth noting that in previous clinical studies of CBD treatment in children with Lennox–Gastaut syndrome or Dravet syndrome (for which CBD is Food and Drug Administration–approved), only 32% and 43% of patients, respectively, had a positive response to drug treatment.^26,27 Numerous studies have shown strong associations between genetic polymorphisms and levels of SWS and/or REM sleep,³⁰ and future studies may elucidate whether such polymorphisms play a role in the response to treatment with CBD-terpenes in individuals with insomnia. There are currently millions of consumers who self-medicate with CBD ostensibly as treatment for insomnia, and our results may have potential clinical relevance for a subset of those individuals, especially those with low SWS + REM sleep and day sleepers.

Participants who slept during the day showed the most robust treatment response to CBD-terpenes. We presume that these participants represent night shift workers and may have shift work sleep disorder, a circadian rhythm sleep disorder that causes insomnia and interferes with falling/staying alseep.³¹ Objective assessment of shift workers by PSG indicates that day sleep is significantly shorter than night sleep, and the sleep loss is primarily taken out of stage 2 and REM sleep.³¹ We hypothesize that day sleepers in our study may efficiently respond to treatment with CBD-terpenes due to lower REM sleep at baseline. Consistent with this scenario, REM sleep (% TST) significantly increased in day sleepers in the current study (Figure S6 in the supplemental material and Table 1), in contrast to our observations in the total study population.

The current study has several important limitations. First, the study was not conducted in a sleep clinic using PSG but was instead performed as a decentralized study using a wrist-worn sleep-tracking device. Although a decentralized study is more representative of a real-world setting in which people are sleeping in the comfort of their own beds, future studies will be required to determine whether our findings using actigraphy are confirmed by studies performed in a sleep clinic using PSG. Second, the CBD-terpene formulation only marginally increased SWS + REM sleep in the total study population, which calls into question the clinical significance of these findings for individuals with insomnia. On the other hand, the magnitude of benefit from the CBD-terpenes formulation was more pronounced and robust in certain subpopulations in the study (eg, participants with low baseline SWS + REM sleep and day sleepers) and could therefore potentially provide greater clinical benefit in these subpopulations. Future studies will be required to validate the clinical significance of our current findings in specific subpopulations of individuals with insomnia. Third, the wrist-worn sleep-tracking device used in the current study does not measure sleep latency and may not be sufficiently accurate to report sleep staging, so future studies using PSG will be required to determine whether the CBD-terpenes formulation influences the onset of sleep and confirms our current results on sleep staging as reported by actigraphy. Fourth, treatment with CBD-terpenes only modestly improved self-reported measures of sleep in study participants relative to the placebo control, and these results fell short of statistical significance. Because trends toward statistical significance were observed for some of these self-reported outcome measures, we conclude that our study lacked sufficient statistical power to detect such differences and that future studies should address this particular shortcoming. Fifth, the unexpected high dropout rate in the study limited its statistical power. Although stringent measures were put in place to retain participants in the study, future studies will require a larger number of participants and improved retention measures.

The combination of SWS and REM sleep stages is widely referred to as “restorative” sleep and is thought to strengthen the immune system, increase cell/tissue regeneration and brain metabolite clearance, and replenish energy stores.^16,17 SWS and REM sleep are also critical for learning, memory, attention, and executive function.^16,17 SWS is associated with decreased heart rate, blood pressure, sympathetic nervous activity, and cerebral glucose utilization, compared with wakefulness.^16,17 Moreover, during SWS, human growth hormone is released and the stress hormone cortisol is inhibited.^16,17 The percentage of time spent in SWS and REM sleep may decrease with aging,³² and a decreased percentage of REM sleep is associated with a greater risk of all-cause, cardiovascular, and other non-cancer-related mortality.³³ Accordingly, novel and safe therapeutic modalities that selectively increase these sleep stages are urgently needed, and future PSG studies will clearly be required to confirm that CBD and terpenes increase these sleep stages in individuals with insomnia.

The molecular- and systems-level mechanisms that mediate the effects of CBD-terpenes on sleep physiology will also require further studies. One intriguing possibility is that these effects may be mediated by CBD’s modulation of the endocannabinoid system. For example, CBD has been shown to increase levels of the endocannabinoid anandamide in humans,³⁴ and preclinical studies show that anandamide increases SWS and REM sleep.³⁵

The vast majority of ingestible CBD products used by consumers are whole-plant hemp-based extracts that contain a large variety of terpenes. Unfortunately, because extracts are derived from harvests that vary from crop to crop and season to season, generating extracts with a consistent terpene profile is close to impossible. Therefore, in the current study we chose to test a CBD-terpene formulation with a precise formulation of terpenes (1 mg each [> 98% purity, food grade] of the terpenes linalool, myrcene, phytol, limonene, α-terpinene, α-terpineol, α-pinene, and β-caryophyllene). Preclinical studies indicate that some terpenes are sedating,^11–15 but their effects on sleep physiology in humans remains unknown. In a smaller clinical study we obtained preliminary evidence that this terpene combination contributed to the sleep-promoting benefits of CBD (data not shown; Patent Pending PCT/US2023/011720). Numerous studies indicate that terpenes may synergize with cannabinoids to influence physiology by the so-called entourage effect,³⁶ but further studies will be required to determine their role in sleep physiology.

Despite tracking by a medical monitor, no AEs were reported during our trial. These results are consistent with 2 recent studies documenting a good safety profile for chronic dosing of CBD in humans.^5,37 In contrast, AEs that included dry mouth, dizziness, vertigo, and acute tachycardia were reported by 83% of participants in a previous clinical trial with medical cannabis oil that contained Δ⁹-tetrahydrocannabinol and CBD.³⁸ In addition to these AEs, Δ⁹-tetrahydrocannabinol decreases REM sleep,³⁸ is intoxicating, and could create fall risks in older adults. Taken together, our results suggest that select CBD-terpene formulations may be a more effective and safer sleep aid than products that are Δ⁹-tetrahydrocannabinol–based. Notably, some commonly prescribed sleep medications and over-the-counter sleep aids induce significant side effects that limit their utility.⁴⁰ Finally, many commonly prescribed sleep medications decrease sleep onset latency, and they might also decrease SWS and REM sleep, a significant shortfall of these types of medications.⁴¹

In summary, our clinical study indicates that combination therapies based on CBD-terpene ratios may increase SWS and REM sleep and decrease light sleep in subsets of individuals with severe insomnia, particularly those with low baseline SWS + REM sleep and day sleepers. This result provides a solid foundation to perform additional placebo-controlled clinical trials with CBD-terpenes in larger cohorts of patients as a potential treatment for insomnia that may increase restorative sleep if confirmed by PSG.

DISCLOSURE STATEMENT

All authors have seen and approved this manuscript. Funding for this study was provided by Defined Research, Inc. M.W., M.F., S.A., V.P., E.C., and P.J.M. hold equity in Defined Research Inc., a for-profit company. J.I.C. and N.S. are unpaid consultants who hold stock option grants for equity in Defined Research Inc.

ABBREVIATIONS

AE

adverse event

CBD

cannabidiol

PGI

patient’s global impression

PSG

polysomnography

REM

rapid eye movement

SWS

slow-wave sleep

TST

total sleep time