The Effectiveness and Safety of Cannabidiol in Non-seizure-related Indications: A Systematic Review of Published Randomized Clinical Trials

Abstract

Background:

Legislative changes have fueled the global availability of cannabis and cannabis-derived compounds, such as cannabidiol (CBD). Little is known about the effectiveness and safety of CBD for treating health conditions other than seizure disorders.

Purpose:

A systematic review of the literature was performed to investigate other health conditions, characteristics of the studied populations, and effectiveness of CBD in randomized clinical trials (RCTs).

Methods:

Seven publication databases were searched from Feb-March 2021. The inclusion criteria for studies were: 1) utilized a RCT design; 2) published in a peer-reviewed journal or thesis/dissertation; 3) published in English; 4) investigated either prescription (i.e., Epidiolex) or non-prescription CBD that was derived from Cannabis sativa plant with <3% Δ9-tetrahydrocannabinol; 5) reported at least one outcome. This review excluded seizure-related disorders as several previous reviews have been done on this topic; it also excluded published protocols, other systematic reviews, or meta-analyses of RCTs that investigated CBD. Independent reviewing, risk of bias assessment, and data abstraction were performed by two authors.

Results:

Fifty-eight studies from eight countries were included in this review. Twenty-seven studies (47%) were conducted in healthy populations, 14% were restricted to males (N=8), and 72% had sample sizes <40 participants. Dosages of CBD used in these studies ranged from 400 μg to 6000 mg. CBD’s effect on mental health was the most studied topic (53%), which focused mainly on anxiety, psychosis, schizophrenia, and substance use disorders. The remaining studies investigated neurological conditions (19%) and a myriad of other health conditions or outcomes. While CBD appears to be anxiolytic, its effectiveness for other conditions was highly variable.

Conclusions:

This review highlights the inconsistencies of CBD as a treatment for non-seizure-related health conditions or outcomes. Studies incorporating larger sample sizes in more diverse populations are encouraged. While CBD was generally safe and well-tolerated even in high doses among the included studies, clearer dosing guidelines and increased regulation of CBD products are also needed.

1. Introduction

Cannabis is the most extensively used illicit drug in the world, and cannabis use and dependence have increased globally over the past two decades [1]. In the United States (US) specifically, cannabis is the most commonly used illicit drug among younger generations. There were 49.6 million US residents 12 years of age and older reported using marijuana in the last 12 months in 2020 National Survey on Drug Use and Health (NSDUH), with young adults (18-25 years old) accounting for the greatest number with 11.6 million people (34.5%) [2]. The Controlled Substance Act of 1970 made cannabis possession and use illegal in the US; however, State laws that reduce or remove penalties for cannabis-related convictions are becoming more common and 36 States recently enacted legislation to decriminalize cannabis for medical use [3]. Additionally, 18 States, 2 US Territories, and the District of Columbia have decriminalized cannabis for recreational use [3]. To date, the US Food and Drug Administration (FDA) has approved one cannabis-derived drug product: Epidiolex (cannabidiol), and three synthetic cannabis-related drug products: Marinol (dronabinol), Syndros (dronabinol), and Cesamet (nabilone) [4]. These authorized drug products are only available with a prescription from a licensed healthcare provider. Even though several states have approved cannabis for medical use, the US Drug Enforcement Administration still classifies cannabis as a Schedule I drug, which means that it has no officially acknowledged medical use and a high potential for misuse on a federal level.

In addition to the US, legislation regulating marijuana use has been enacted in Canada, Australia, New Zealand, and some European countries. The Canadian Parliament enacted Bill C-45 on June 19, 2018, known as the Cannabis Act, to legalize and regulate cannabis cultivation. In October 2018, Canada became the second country in the world after Uruguay to legalize the sale, possession, and non-medical use by adults [5]. In 2016, the Australian federal government approved laws allowing certified healthcare professionals to prescribe a wide range of cannabis-based medications to patients [6]. In New Zealand, where medical marijuana use is approved, 50.7% of voters voted against a proposal legalizing and regulating recreational marijuana use and supply in a public referendum [7]. Individual European nations have varied levels of decriminalization, with some still maintaining prohibition. Many countries in Europe may only permit “pharmaceutical” cannabis, such as Epidolex, as a prescription medicine. Recreational cannabis is illegal in many European Union countries and the United Kingdom (UK) [8].

Global regulation of cannabis has stymied research of this biologically complex plant, which has over 144 naturally occurring compounds known as cannabinoids. Two cannabinoids, Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD), are the most prevalent cannabinoids found in cannabis and have received the most attention in epidemiological and medical research. THC is the primary psychoactive component of cannabis whereas CBD is considered non-psychoactive [9]. Also, THC has potential therapeutic value of pain relief, reduction and control of nausea and vomiting, and appetite stimulation [10]. The balance of THC and CBD is essential for safety and therapeutic effects because THC and CBD have contrasting mechanisms of action on the endocannabinoid system [11]. When CBD and THC are combined, CBD may alleviate some of THC’s side effects, such as memory impairment and paranoia. [12]. Some cannabis-derived compounds have been made available for medical use. For example, Sativex, which is an oromucosal or sublingual spray composing nabiximols with 1:1 ratio of THC and CBD, has been approved in 29 countries, including the UK, Israel, Canada, Brazil, and Australia, for the treatment of spasticity associated with multiple sclerosis [9], neuropathic pain associated with multiple sclerosis [13], and pain in patients with advanced cancer [14]. Epidiolex, an oral CBD solution derived from the Cannabis Sativa plant with < 0.3% THC (colloquially known as hemp), has been approved by the US FDA for treatment of seizures associated with Lennox-Gastaut syndrome, Dravet syndrome, or tuberous sclerosis complex in patients 1 year of age and older [4, 15]. Consequently, CBD was reclassified in the US as a Schedule III drug with a minimal misuse potential, which allowed it to be used in clinical trials [16]. Additionally, recent legislation in the US has also permitted CBD oils and tinctures extracted from hemp to be sold in non-prescription form as a food/dietary supplement [17].

These legislative changes have fueled CBD’s global availability making it a multibillion-dollar industry [18, 19]. Unfortunately, CBD’s availability has outpaced research on the drug and its effectiveness. While there is scientific evidence regarding prescription CBD’s efficacy for treating Dravet and Lennox-Gastaut syndromes [20–24] and tuberous sclerosis complex [20], the evidence for its effectiveness for other health conditions or outcomes is limited [9, 18]. A survey of non-prescription CBD consumers (N=2,409) revealed that many individuals purchase the product for general wellness or recreation (38%), but most use CBD to treat a health condition (62%), such as anxiety, depression or insomnia [25]. However, research has shown that non-prescription CBD is poorly regulated and many products are mislabeled or may contain THC [9].

Previous systematic reviews investigated the dosage [26, 27] and adverse effects [28] of CBD in clinical trials. Several systematic reviews focused on THC and CBD in human [29–32], only THC in humans [33], cannabinoids in both humans and animals [34], or cannabis in animals only [35, 36]. Additionally, several reviews only focused on specific neurological or psychopathological conditions, such as seizure-related disorders [20–23, 37], or psychiatric disorders [38–41]. However, no published studies to date has reported the effectiveness and safety of CBD in all existing randomized clinical trials (RCTs) excluding seizure-related indications. In order to fill in this research gap, the purpose of this review was to comprehensively summarize all current, published, randomized clinical trials that investigated the effectiveness and safety of CBD; this review does not include seizure-related disorders as several previous reviews have been done on this topic [20–23, 37]. Special focus is given to the medical conditions and characteristics of the populations studied, effectiveness of CBD on study outcomes, and current gaps in knowledge which can inform future medical research directives.

2. Methods

2.1. Search Strategy

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines [42] (see supplementary information in the electronic supplementary material). This review was not registered on PROSPERO because searches started before registration could be initiated, which is a requirement for registration. The following literature databases were searched for eligible studies: 1) PubMed; 2) Web of Science (WOS); 3) Scopus; 4) CINAHL Complete (within EBSCO host); 5) Cochrane Database; 6) Academic Search Complete (within EBSCO host); 7) ProQuest Database. All searches were performed by two reviewers (YT and KT) with the assistance of a third reviewer (TMR). Searches were conducted between February and March 2021. Each search contained the phrases, “cannabidiol”, “CBD”, “clinical trial”, and “human”. The detailed search strategies for each database are included in Supplementary material.

The criteria for studies to be included in the review were as follows: 1) utilized a RCT design; 2) published in a journal, dissertation, or thesis; 3) published in English; 4) investigated either prescription (i.e., Epidiolex) or non-prescription CBD that was derived from Cannabis sativa plant with <3% ∆9-tetrahydrocannabinol; 5) reported at least one outcome. As previously noted, studies were excluded if they were conducted among those with seizure disorders as previous systematic reviews have focused on this population specifically [20–23, 37]. This review did not include published protocols, other systematic reviews, or meta-analyses of RCTs that investigated CBD. This review did not include conference abstracts as these studies could not be fully assessed for bias. Likewise, while the US requires certain clinical trials to be registered (i.e., www.clinicaltrials.gov), this database was not included as studies listed could not be fully assessed for bias.

2.2. Study Selection Process

All reviewers received training prior to both title/abstract and full-text screenings to ensure understanding of the review’s purpose and study inclusion criteria. All abstracts and titles derived from the initial searches were compiled using the reference management software EndNote, version X9. Studies that appeared in multiple databases (i.e., duplicates) were deleted. YT and KT, whom each had prior experience with systematic review methodology, screened titles and abstracts independently. The study team met, compared results, and resolved any discrepancies regarding the inclusion of studies. Any outstanding disagreements were settled by a third reviewer (TMR). Decisions on inclusion and exclusion were recorded in EndNote. Based on the initial searches of titles/abstracts, any studies that appeared to meet the inclusion criteria were retrieved as full-text articles. The full-text articles were then read by YT and KT in their entirety to determine eligibility, with decisions on inclusion and exclusion being recorded in EndNote. Any discrepancies regarding the inclusion of studies were resolved by discussion, and TMR acted as the arbitrator making the final decision for inclusion.

2.3. Study Quality Assessment

After completing the full-text review, eligible studies were assessed for quality. This study used the Cochrane Tool for Assessing Risk of Bias in RCTs, which was developed by Cochrane Collaboration Groups, to assess study quality and risk of bias. The Cochrane Tool contains a total of 7 items from 5 domains: 1) Selection bias: random sequence generation and allocation concealment; 2) Performance bias: blinding for participants and personnel and blinding for outcome assessment; 3) Attrition bias: incomplete outcome data; 4) Reporting bias: selective reporting; and 4) Other bias: other sources of bias [43]. The Cochrane Tool utilizes a scale to rank the presence of these 7 items in a study (e.g., high risk, low risk, unclear). The included studies were evaluated for quality by YT and KT independently. Once the quality assessments were completed and documented in a Microsoft Excel spreadsheet, YT and KT met and compared their results for accuracy and/or precision. Any disagreements were resolved through discussion. If a consensus was not reached, TMR acted as the arbitrator.

2.4. Data synthesis

Another data abstraction form was created in Microsoft Excel and was used to facilitate the synthesis of included studies’ findings. The form included the year of publication, country in which the study was conducted, study design, study population, total sample size, sample sizes by CBD group, dosage of CBD, study purpose, and effectiveness. No meta-analysis was performed given the heterogeneity of sample sizes, different locations of studies conducted, and different health conditions studied. YT and KT met to compare their results for accuracy and/or precision. Any disagreements were resolved through discussion. If a consensus was not reached, TMR acted as the arbitrator.

3. Results

3.1. Study characteristics

A flow diagram documenting the search process and reasons for excluding studies is shown in Figure 1. Of the 4,579 results initially obtained, 58 studies were included in this review. The characteristics of those studies are presented in Table 1. The included studies were conducted between 1979 and 2020. Of these studies, 24 (41.4%) were conducted in the UK [12, 44–66], 12 (20.7%) in the US [67–78], 11 (19.0%) from Brazil [79–89], 5 (8.6%) from Australia [90–94], 3 (5.2%) from Netherlands [95–97], 1 (1.7%) from Germany [98], 1 (1.7%) from Israel [99], and 1 (1.7%) from Canada [100]. Twenty-seven (46.6%) of these studies were conducted healthy populations, while the rest were from study populations with certain medical conditions. Nine (16%) of included studies were limited to males [46, 51, 61, 63, 75, 83, 86, 89, 96], and one study was limited to children aged 8-16 years [91]. Among studies conducted in healthy populations, six studies enrolled those with prior or current exposure of cannabis use [12, 52, 67, 71, 77, 94, 95], and one study with prior exposure to opioids [74]. Four of the studies investigated the effects of CBD, both alone and in combination with THC [58, 59, 94, 95]. Most included studies used a constant CBD dosage ranging between 400 μg and 6000 mg, with 600 mg being the most common dosage. Seven studies (12%) dosed CBD based on bodyweight [49, 69, 70, 87, 90, 91, 93].

Figure 1. Flow chart for the selection of cannabidiol (CBD) randomized clinical trial (RCT) studies*.

*Studies could have been excluded for multiple reasons

Table 1.

Characteristics of included studies of cannabidiol in randomized clinical trials

First author (Publication year/Country	Study Design	Study Population	Total Sample Size / CBD Sample Size	Dosage of CBD	Route of administration / Dosage form	Study Purpose	Reported adverse events	Key findings
Mental Health
Anxiety, fear, or paranoia
Bergamaschi, Queiroz [79] (2011) / Brazil	Double blind, RCT	SAD and HC	36 / 24 (two groups with 12 subjects each in the SAD)	600 mg/d	Oral / Dissolved in corn oil and sealed into capsules	To compare effects of a public speaking test on HC and treatment naive SAD who received a single dose of CBD or placebo	Not reported	Pretreatment with CBD significantly reduced anxiety, cognitive impairment, and discomfort in speech performance, and significantly decreased alertness in their anticipatory speech. No significant differences were observed between CBD and HC in SSPS-N scores or in the cognitive impairment, discomfort, and alert factors of VAMS.
Das, Kamboj [47] (2013) / UK	Placebo-controlled between-subjects RCT	Healthy volunteers between 18-35 of age, no mental health history, no substance problems, and not pregnant	48 / 32 (prior to extinction; 32 (following extinction)	32 mg	Inhalation (via a vaporizer) / Vaporized at 210 °C	To assess if CBD has potential to help in treatment of pathological fear memories	Not reported	CBD can enhance consolidation of extinction learning; CBD may be adjunct to extinction-based therapy for anxiety disorders
De Souza Crippa, Zuardi [89] (2004) / Brazil	Double blind, within-subject RCT	Healthy males who were postgraduate students	10 / 5	400 mg	Oral / Dissolved in corn oil and packed in gelatin capsules	To investigate CBD effects on regional cerebral blood flow using functional neuroimaging	Not reported	CBD significantly decreased subjective anxiety and increased mental sedation, and had greater activity in the left parahippocampal gyrus, which suggest that CBD has anxiolytic properties, and these effects are mediated by an action on limbic and paralimbic brain areas
Hundal, Lister [53] (2018) / UK	Between-groups, double-blind, RCT	Non-clinical high paranoia group aged 18-50 years	32 / 16	600 mg	Oral / Capsules	To assess the effects of CBD on anxiety and paranoia in those at risk of major mental illness	CBD-treated group had mild adverse events including tiredness/sedation, lightheaded/dizziness, nausea, abdominal discomfort, and increased appetite/hunger	There was no evidence of benefits on anxiety or persecutory ideation in healthy volunteers with high trait paranoia
Linares, Zuardi [85] (2019) / Brazil	Parallel-group RCT	Healthy volunteers	57 / 42	150 mg, 300 mg, and 600 mg	Oral / Dissolved in corn oil and packed in gelatin capsules	To compare the acute effects of different doses of CBD and placebo under a simulated public speaking test	Not reported	300 mg of CBD pretreatment can significantly reduce anxiety, cognitive impairment, and distress
Zuardi, Shirakawa [87] (1982) / Brazil	Placebo-controlled RCT	Healthy volunteers aged 20-38 years	8 / 8	1 mg/kg	Oral / Stored in alcohol solution (100 mg/ml) and taken from the storage solution, made up to 1.5 ml with alcohol (99% ethanol) and then added to 200 ml of artificial lemon juice	To see if CBD reduced the anxiety caused by THC in normal volunteers, and to see if this impact is due to a general block of THC’s activity or a specific effect on anxiety	Volunteers in CBD group had sleepiness at 30-60 mins and 60-12 mins. However, more effects reported among volunteers in THC group	CBD inhibits the anxiety generated by 9- THC, but it also inhibits marijuana-like effects and other subjective changes induced by 9- THC. CBD’s actions may play a role in the antagonistic effects of the two cannabinoids
Zuardi, Cosme [88] (1993) / Brazil	Placebo-controlled RCT	Healthy volunteers aged 20-30 years	40 / 10	300 mg	Oral / Dissolved in corn oil (100 mg/ml) and packed in gelatin capsules	To test the acute effects of ipsapirone and CBD in healthy volunteers subjected to a simulated public speaking test		CBD significantly reduced post-stress anxiety, whereas ipsapirone reduced the test-induced increase in systolic blood pressure. In human volunteers exposed to a stressful condition, ipsapirone and CBD have anxiolytic properties
Psychosis or Schizophrenia
Appiah-Kusi, Petros [44] (2020) /UK	Randomized, placebo-controlled, between-groups, double blind study	CHR patients	58 /16	600 mg/d	Oral / Capsules	To determine whether CBD would normalize the neuroendocrine and anxiety responses to stress in high-risk psychosis patients	Not reported	The cortisol change associated with experimental related stress was greatest in HC and least in CHR-placebo patients, with CHR-CBD patients exhibiting an intermediate response. Changes in anxiety and experience of public speaking stress was greatest in CHR-P and least in HC
Bhattacharyya, Wilson [45] (2018) / UK	Parallel group, double blind, placebo controlled RCT	Antipsychotic medication-naïve and at clinical high risk for psychosis patients	33 / 15	600 mg	Oral / Capsules	To investigate neurocognitive mechanisms that underlie the therapeutic effects of CBD in psychosis	No adverse or serious events were observed in any of the CHR participants following single dose of study drug	No significant difference in task performance among study groups. However, a single dose of CBD may partially normalize alterations in parahippocampal, striatal, and midbrain function associated with the CHR state
Boggs, Surti [68] (2018) / US	Parallel group, double blind, placebo controlled RCT	Stable antipsychotic-treated patients diagnosed with chronic schizophrenia, aged 16-65 years	36 / 18	600 mg daily (300 mg BID)	Oral / Not reported	To compare cognitive, symptomatic, and side effects of CBD vs placebo	Sedation more prevalent in CBD group and one participant in the CBD arm withdrew early due to sedation	No improvement in MCCB or PANSS scores. At the dose tested, CBD augmentation for 6 weeks did not improve MCCB performance or psychotic symptoms
Davies, Wilson [48] (2020) / UK	Double-blind, placebo-controlled, three-arm parallel group design	CHR for psychosis patients aged 18-35 years and healthy controls	52 / 16	600 mg	Oral / Capsules	To assess the effects of CBD on regional brain activation in CHR patients while viewing faces with fearful expressions	Not reported	CHR participants with placebo showed greater activation than controls in the parahippocampal gyrus but less activation in the striatum. Also, activation in the CHR group that received CBD was intermediate between that of the CHR placebo and control groups. Therefore, CBD modulates brain function in regions showed in psychosis risk and emotion processing
Englund, Morrison [12] (2013) / UK	2×3 mixed RCT	Healthy volunteers aged 21-50 with at least one previous cannabis use	48 / 22	600 mg	Oral / Capsules	To assess the pretreatment with CBD inhibited THC-elicited psychosis and cognitive impairment	Not reported	CBD is effective in decreasing THC elicited positive psychotic symptoms and protecting hippocampal dependent memory from the impact of THC
Leweke, Piomelli [98] (2012) / Germany	Phase II, double blind, monocenter, parallel group, controlled, RCT	Aged 18-50 years who had schizophrenia or schizophreniform psychosis diagnosis	42 / 20	Starting 200 mg per day each and increased stepwise by 200 mg/d to a daily dose of 200 mg four times daily (total 800 mg per day)	Not reported	To evaluate the clinical relevance of cannabidiol vs. amisulpride in acute schizophrenia	Well-tolerated and did not significantly affect hepatic or cardiac functions	Patients undergoing either cannbidiol or amisulpride treatment showed significant clinical improvement, but comparison of the clinical effects of amisulpride and cannabidiol revolved no relevant difference. There was a statistically significant association between increase in anandamide levels and decrease in psychotic symptoms in patients treated with CBD
McGuire, Robson [56] (2018) / UK	Multicenter, placebo-controlled, parallel-group RCT	Age 16-65 years with schizophrenia or a related psychotic disorder defined by DSM-IV	88 / 43	1000 mg	Oral / Solutions	To explore the safety and effectiveness of CBD as an adjunctive treatment in schizophrenia	Well-tolerated but gastrointestinal adverse events are mostly reported, and a serious adverse event was reported	CBD group had lower levels of positive psychotic symptoms and greater improvements in cognitive performance and overall functioning. CBD was well tolerated, and rates of adverse events were similar between the CBD and placebo groups
Morgan, Freeman [58] (2018) / UK	4 sessions, cross-over RCT	Self-reported abstinence from cannabis	48 / 48	16 mg CBD; 8mg THC+16 mg CBD	Inhalation / Dissolved in ethanol and then vaporized within a 10-s inhalation cycle	To assess the effect of cannabinoids on psychotomimetic symptoms and memory function	Not reported	The THC increased overall PSI, negative symptoms on the BPRS, and robustly impaired episodic and working memory. CBD did not attenuate these effects
O’Neill, Wilson [60] (2020) / UK	Placebo-controlled, repeated-measures, within-subject, crossover RCT	Patients with diagnosis of psychotic mental illness and within 5 years of onset of illness	34 / 15	600 mg	Oral / Capsules	To investigate the acute neurophysiological effects of CBD in psychosis patients	Not reported	CBD attenuated dysfunction in prefrontal and mediotemporal activation, greater mediotemporal-striatal functional connectivity, and caused trend-level symptom reduction in psychosis patients
Solowij, Broyd [94] (2019) / Australia	Double blind RCT	Healthy current cannabis users and non-naïve nonusers	36 / 36	400 mg (CBD alone), 4 mg (THC+CBD low), 400 mg (THC+CBD high)	Inhalation / Dissolved in a 10% ethanol solution and vaporized at a higher temperature for administration	To investigate the effects of vaporized high-dose CBD, as well as low and high doses of CBD administered concurrently with THC, on a variety of parameters in order to better understand the links between cannabis and cannabinoid chemicals and psychotic-like outcomes	Well-tolerated but some with anhedonia and amnesia	When CBD is combined with THC at high levels, it can help to mitigate the negative effects of THC. Low amounts of CBD mixed with THC, on the other hand, exacerbated THC’s intoxication effects
Wilson, Bossong [66] (2019) / UK	Placebo-controlled, parallel-arm fMRI, randomized design	Aged 13-35 years with clinical high risk of psychosis	33 / 16	600 mg	Oral / Capsules following a standard light breakfast	To investigate if there is a pattern of aberrant activation in clinical high-risk psychosis patients compared to healthy controls during the processing of motivationally salient stimuli, and if a single dosage of CBD can reduce the abnormal function in clinical high-risk psychosis patients	Not reported	CBD reduced the enhanced activation in the left insula/parietal operculum, which was linked to a slower reaction time overall
OCD
Kayser, Haney [73] (2019) / US	Within-subject, double blind, RCT	Aged 21-55 with principal DSM-IV diagnosis of OCD over 1 year and near-constant symptoms	14 / 5	~800 mg of cannabis on CBD (0.4% THC/10.4% CBD)	Inhalation / Cued-smoking procedure	To assess the acute effects of smoked cannabis varying in THC and CBD concentrations in adults with OCD	Well-tolerated but the most common self-reported side effects were nervousness and dry mouth	THC increased heart rate, blood pressure and intoxication compared with CBD and placebo. Either primary THC or CBD has little acute impact on OCD symptoms and yields smaller reductions in anxiety
Substance use or substance use disorders and CBD
Consroe, Carlini [81] (1979) / Brazil	Cross-over RCT	Healthy volunteers with an age range of 21-33 years and a weight range of 52-85 kg	10 / 10	200 mg	Oral / Incorporated into opaque gelatin capsules	To assess interaction between CBD and alcohol in humans	Not reported	CBD alone does not produce decrements of motor and cognitive responses
Babalonis, Haney [67] (2017) / US	Randomized, within-subject, double blind, placebo controlled	Healthy, frequent marijuana users	31 /31	200, 400, 800 mg	Oral/ Capsules	To assess abuse liability of a range of oral CBD doses compared to placebo and smoked marijuana in a sample of heavy marijuana users	Well-tolerated	Active marijuana reliably produced abuse-related subjective effects. However, CBD did not display any abuse liability symptoms
Freeman, Hindocha [50] (2020) / UK	Double blind, RCT	16-60 years old cannabis use disorders with a desire to stop use in next month	48 / 12 for each CBD group	200 mg, 400 mg, 800 mg	Oral / Capsules	To investigate efficacious doses of CBD to treat cannabis use disorder	CBD 200mg group: 42 mild and 4 moderate adverse events (n=12) CBD 400mg group: 96 mild and 8 moderate adverse events (n=24) CBD 800mg group: 78 mild and 8 moderate adverse events (n=23)	CBD 400 mg and 800 mg were safe and effective than placebo at reducing cannabis use
Hindocha, Freeman [52] (2020) / UK	Double blind, placebo-controlled, 4-way cross-over RCT	Recreational and daily cannabis users aged between 16 and 23 years	48 / 48	16 mg (vaporized)	Inhalation / Dissolved in ethanol and then vaporized within a 10-s inhalation cycle	To understand the genetic factors that contribute to cannabis addiction	Not reported	After acute administration of cannabinoids, genetics of the endocannabinoid system can alter addiction endophenotypes in healthy people
Manini, Yiannoulos [74] (2015) / US	Placebo-controlled, cross-over, RCT	Aged 21-65 years healthy volunteers with prior opioid exposure	17 / 12 (6 high-dose CBD group and 6 low-dose CBD group)	400 mg and 800 mg	Oral / Dissolved in corn oil and encapsulated in gelatin capsules	To determine the safety and pharmacokinetics of CBD when administered concomitantly with opioids in healthy human subjects	Well-tolerated but with minor adverse events such as dizziness/drowsiness, itching or rash, headache, abdominal discomfort, nausea/vomiting, and diarrhea	CBD and opioids were safe and well tolerated. CBD does not exacerbate adverse effects associated with intravenous fentanyl administration
Meneses-Gaya, Crippa [86] (2020) / Brazil	Placebo-controlled RCT	Aged 18 years or older males with a DSM-IV diagnosis of crack-cocaine dependence	31 / 14	150 mg	Oral / Dissolved in corn oil and placed in gelatin capsules	To assess the efficacy of CBD in the treatment of crack-cocaine craving and the frequent withdrawal symptoms	Well-tolerated but some experience mild or moderate sleepiness and increased sleep duration, nausea, and headache	There are no differences between the CBD and placebo groups of craving levels and induction, anxiety, depression, and sleep
Morgan, Das [57] (2013) / UK	Placebo-controlled RCT	Aged 18-35 years males and females smoked >10 cigarettes per day and expressed an intention to quit smoking	24 / 12	400 µg	Inhalation / The solution aerosol in the inhaler administered CBD dissolved in absolute ethanol ≈5%	To assess the impact of low dose of CBD inhaler on nicotine addiction to tobacco smokers who wanted to quit smoking	Greater sedation in both groups on the second testing day	The number of smoked cigarettes across a 7-day period was a dramatic reduction among the individuals using the CBD inhaler compared to the placebo group. CBD can also have craving reduction and reduced anxiety
Schoedel, Szeto [100] (2018) / Canada	Phase 1, single-dose, double-dummy, placebo- and active-controlled, crossover, randomized design	Aged 18-55 years of healthy male and female adults with BMI between 19.0 and 30.0 kg/m² and a weight of ≥50 kg	95 / 43	750 mg, 1500 mg, and 4500 mg	Oral / Oral solutions	To evaluate the abuse potential of single dose of highly purified CBD compared with the single dose of alprazolam, dronabinol, and placebo in healthy, recreational polydrug users	Well-tolerated but some experience mild or moderate adverse events, such as somnolence	750 mg CBD showed significantly low abuse potential in a highly sensitive population of polydrug users
Severe behavioral problems
Efron, Freeman [91] (2020) / Australia	Two-arm parallel design RCT	Children with 8-16 years of age who had ID and severe behavioral problems	8 / 4	Up to 20 mg/kg/d in 2 divided doses	Oral / Oil solution of 98% CBD mg/ml in grapeseed oil	To estimate the feasibility of conducting a RCT of CBD to reduce SBP in kids with ID	No serious adverse events reported	The study protocol is feasible and acceptable and severe behavioral problems and their families
Autism
Pretzsch, Freyberg [62] (2019) / UK	Placebo-controlled, repeated-measures, crossover, case-control RCT	Healthy males including typically brain developing individuals and diagnosed with autism spectrum disorder, and with an IQ over 70	34 / 34	600 mg	Oral / A liquid dose of pharmacological probe	To test the CBD impacts in vivo glutamate and GABA levels and examine the potential differences in these response in ASD individuals	Not reported	This study did not directly test the efficacy. CBD changes glutamate and GABA+ that may affect glutamatergic excitation and GABAergic inhabitation, and thereby impact on brain function. However, the direction of a functional response to CBD may be distinct in autistic individuals
Pretzsch, Voinescu [61] (2019) / UK	Placebo-controlled, double blind, repeated-measures, crossover RCT	Healthy men (half with ASD)	34/ 34	600 mg	Oral / A liquid dose of pharmacological probe	To test whether the CBD has a different fMRI response among ASD individuals	Not reported	CBD significantly altered fALFF in the cerebellar vermis and the right fusiform gyrus. Also, CBD significantly altered vermal FC with several of its subcortical and cortical targets
Behavioral and cardiovascular effects
Haney, Malcolm [71] (2016) / US	Double blind, within-subject, RCT	Healthy volunteers aged 19-50 years, non-treatment seeking healthy cannabis smokers	31 / 31	200 mg, 400 mg, 800 mg	Oral / Capsules	To assess the influence of CBD pretreatment on the reinforcing, subjective, cognitive, and physiological effect of smoked cannabis	Nausea, vomiting, constipation, gas, and headache were the most frequently cited sided effects	CBD alone had no significant psychoactive or cardiovascular effects and did not significantly alter any of these outcomes. Oral CBD does not reduce the reinforcing, physiological, or positive subjective effects of smoked cannabis
CBD’s Impact or Interaction with Other Drugs
Belgrave, Bird [90] (1979) / Australia	Double blind, RCT	Healthy, paid, volunteers aged 18-24 years old with body weights of 56-98 kg	15 /15	320 ug/kg (51-58kg: 17.5mg, 59-66kg: 20.0 mg, 67-74kg: 22.5 mg; 75-82kg: 25.0mg, 95kg: 30.0mg)	Oral / Dissolved in sesame oil and sealed into capsules	To determine if CBD could influence the behavioral effects of ethanol	Well-tolerated and no adverse reactions to ethanol, CBD or the combination were noted	No combined effect of CBD, either alone or ethanol + CBD. Prior administration of CBD did not significantly affect the blood ethanol levels
Hallak, Dursun [83] (2011) / Brazil	Double blind, placebo controlled RCT	Healthy males with aged 20-36	10 / 5	600 mg	Oral / Capsules	To investigate whether CBD attenuates behavior effects of ketamine in healthy human subjects	Not reported	CBD significantly augmented the activating effects of ketamine but showed non-significant trend to reduce ketamine-induced depersonalization
Hurd, Spriggs [72] (2019) / US	Placebo-controlled, RCT	Aged 21-65 years abstinent volunteers with heroin use disorder	42 / 17	400 mg, 800 mg	Oral / Solution	To investigate the potential of CBD to reduce critical features of addiction that continue to relapse and continued drug use in drug-abstinent individuals with heroin use disorder	Well-tolerated but some experienced mild diarrhea, headache, and tiredness or fatigue	CBD decreased cue-induced desire and anxiety, as well as physiological markers of stress reactivity. CBD also exhibited a sustained impact on cue-induced desire and anxiety
Neurological or Movement
Chagas, Zuardi [80] (2014) / Brazil	Double blind, RCT	PD patients without dementia, aged above 45 years, use of stable dose of anti-Parkinson medication at least 30 days before the trial	21 / 14	75 mg and 300 mg	Oral / Dissolved in corn oil and place in gelatin capsules	To assess effects of CBD in PD	No significant side effects were recorded in any of the groups	There are no statistically significant differences in UPDRS scores, plasma BDNF levels or H1-MRS measures. However, possible effect of CBD 300mg/day in improving QOL measures in PD patients
Consroe, Kennedy [70] (1991) / US	Cross-over, double blind, RCT	HD patients	14 / 14	10 mg/kg/d	Oral / Dissolved in sesame oil and incorporated into soft gelatin amber-colored capsules	To analysis of plasma levels of CBD measured by gas chromatography/mass spectroscopy	Not reported	CBD might serve as a useful cannabinoid to model THC in long-term, high-dose pharmacokinetic studies
Consroe, Laguna [69] (1991) / US	Double blind, RCT	HD patients	15 / 15	10 mg/kg/d	Oral / Dissolved in sesame oil and incorporated into soft gelatin amber-colored capsules	To evaluate symptomatic efficacy and safety in patients with HD	No significant differences of variables related to safety between the CBD and placebo	CBD with 700 mg/d for 6 weeks was neither symptomatically effective nor toxic in neuroleptic-free patients with HD
de Faria, de Morais Fabrício [82] (2020) / Brazil	Cross-over RCT	Idiopathic PD patients, absence of marked cognitive alterations, not on benzodiazepines or antidepressants	24 / 24	300 mg	Oral / Dissolved in corn oil and packed in gelatin capsules	To evaluate impacts of acute CBD administration at dose of 300 mg on anxiety measures and tremors in individuals with PD	No side effects during or after the experiments	Effective in decreasing anxiety in patients with PD with acute CBD administration at dose of 300 mg and decrease in tremor amplitude under an anxiogenic situation
Freeman, Adekanmi [49] (2006) / UK	Double blind, RCT	Aged 18-64 years with clinical definite or laboratory-supported multiple sclerosis at least 6 months with problematic spasticity	630 /219	Dose was based on bodyweight (30-49 kg: 4×1.25 mg; 50-69kg: 6×1.25mg; 70-89kg: 8×1.25mg; >89kg: 10×1.25mg) The maximum possible dose: 25 mg/d	Oral / Capsules	To estimate whether cannabinoids reduce urge incontinence episodes without affecting voiding in patients with multiple sclerosis	Well-tolerated but with some minor adverse events such as dizziness or light-headedness and dry of mouth	All three groups, including CBD group, showed a significant reduction (38%) in incontinence from baseline to the end of treatment
Fusar-Poli, Crippa [51] (2009) / UK	Double blind, placebo controlled RCT	Healthy, English native, right-handed men with lifetime exposure to cannabis of less than 15 times or less, no personal or family history of psychiatric illness or substance abuse	15 / 15	600 mg	Oral / Capsules	To investigate effects of CBD and THC on regional brain function during emotional processing	Not reported	CBD may reduce autonomic and subjective anxiety whereas THC increased anxiety, as well as levels of intoxication, sedation, and psychotic symptoms The anxiogenic effects of THC may be related to effects in other brain regions
Notcutt, Price [59] (2004) / UK	An ‘N of 1’, placebo-controlled, cross-over RCT	Aged over 18 years with chronic, stable pain	34 / 24	2.5 mg CBD, 2.5 mg CBD+2.5 mg THC	Oral / Sublingual spray (aerosol delivery system or pump action spray)	To study the effect of medical extracts of cannabis for chronic pain	Well-tolerated but some experienced dry mouth, drowsiness, and dysphoria/euphoria	CBD alone did not have effective in pain symptom control, but THC and THC:CBD were effective in relieving pain and improving sleep
Skosnik, Hajós [76] (2018) / US	Placebo-controlled, 2×2, counterbalanced, randomized design	Healthy volunteers	15 / 15	5 mg	Oral / Dissolved in 95% ethanol at a concentration of 10 mg/ml	The effect of acute intravenous THC, CBD, the combination of THC +CBD, and placebo on P50 sensory gating and neural oscillations in healthy humans	Not reported	Both THC and THC+CBD disrupted P50 gating ratios compared to placebo while CBD alone had no effect. Also, THC and THC+CBD also decreased first click evoked theta power
Wade, Robson [65] (2003) / UK	Placebo-controlled, crossover, randomized design	Neurological diagnosis and to be able to identify symptoms (neuropathic pain, spasticity, muscle spasms, impaired bladder control and tremor) which were stable and unresponsive to standard treatment	24 / 21	2.5-120mg	Oral / Applied sublingually with a pump-action oromucosal spray with a 100 µL actuator value	To see if cannabis medicinal extracts obtained from plants can help with neurogenic symptoms that aren’t responding to usual treatment and to measure side effects.	Well-tolerated but 67% of patients in THC: CBD open label group with one or more adverse events	CBD had a substantial impact on pain and spasticity ratings. Cannabis medical extracts can help with neurogenic symptoms that aren’t responding to conventional therapy
Xu, Cullen [78] (2020) / US	Placebo-controlled, single-center, randomized trial	Aged 18 or older males and females with at least 3 months of symptomatic peripheral neuropathy	29 / 15	250 mg	Oral / topical oil solution	To estimate if topical CBD-enriched emu oil could help with the symptoms of chronic pain caused by peripheral neuropathy in the lower limbs.	Well-tolerated and no adverse events reported	When compared to the placebo group, the CBD group saw a statistically significant reduction in acute pain, sharp pain, cold, and itching feelings. CBD was well tolerated and might be a more effective therapeutic option for peripheral neuropathy than current medications
Safety, Tolerability, and Pharmacokinetics of CBD
Hosseini, McLachlan [92] (2020) / Australia	Single-dose study: Open-label, randomized, 4-way crossover study Multiple-dose study: Double blind, randomized, placebo-controlled study	Healthy volunteers aged between 18 and 50 years with no clinically significant findings upon physical examination	12 / 12	Single-dose study:25 mg or 50 mg Multi-dose study: 50 mg	Oral or oromucosal/Single dose study: sublingual safer, oil solution, or nabiximols oromucosal spray; multi-dose study: sublingual water	To assess the safety, tolerability of 2 different doses of CBD, and pharmacokinetics of CBD compared to nabiximols	Mild or moderate somnolence, sedation and altered mood	Both oil solution and sublingual wafer formulations of CBD were well tolerated and achieved equivalent concentrations of CBD when compared to nabiximols
Perkins, Butler [93] (2020) / Australia	Placebo-controlled, single-dose escalation, randomized clinical trial	Aged 18-48 years healthy volunteers with BMI 23-30 kg/m²	24 / 18 (6 for each CBD group)	5 mg/kg, 10mg/kg, and 20 mg/kg	Oral / Lipid-based solution comprising soybean oil and flavoring agents	To assess the safety, tolerability, and pharmacokinetics of cannabidiol with a single-dose in healthy volunteers	Well tolerated but some experienced at least one TEAE, infections, nervous system disorders	CBD was well tolerated in healthy volunteers who were given a single dosage of CBD; there were no safety issues with increasing the dose, and the CBD and placebo groups had identical safety profiles
Taylor, Gidal [97] (2018) / Netherlands	Placebo-controlled RCT	Aged 18 to 45 years healthy adult males or females with BMI between 18 and 28 kg/m²	Single dose arm (n=32); multiple dose arm (n=24)/ 6 per group in single dose arm; 9 per group in multiple dose arm	Single dose arm: 1500 mg, 3000 mg, 4500 mg, 6000 mg Multiple dose arm: 750 mg, 1500 mg Food effect arm: 1500mg	Oral / Supplied as 100 ml in 105 ml amber glass bottles and delivered orally using a syringe	To identify the maximum tolerated dose, researchers looked at the safety, tolerability, and pharmacokinetics of CBD when given as a single and multiple oral doses to healthy people	Well-tolerated but with mild or moderate severity of diarrhea, nausea, headache, and somnolence that were the most common reported adverse events	CBD was generally well tolerated but with mild severity of adverse events. The safety and pharmacokinetic profile support twice-daily CBD administration.
GI Tract Impacts or GI Disorders
Couch, Cook [46] (2019) / UK	Randomized, placebo-controlled double-blind study	Healthy male subjects between the age of 18 and 50	30 / 10	600 mg	Oral / Aspirin dose of 600 mg was dissolved in 400ml water with CBD	To assess the effect of CBD and PEA on permeability of human GI tract in vitro, ex vivo, and in vivo	No side effects reported in experimental sessions	Decreased the inflammation-induced flux of dextrans and permeability reducing effects of CBD in gut
De Bruijn, De Graaf [96] (2017) / Netherlands	Cross-over RCT	Healthy male subjects with normal weight (BMI=21.7±1.2kg/m²)	10 / 10	2 divided doses (25 mg then 10 mg)	Inhalation / Vaporized at 230 °C	To assess if THC or CBD affects sweet taste perception or liking in humans	Not reported	The administration of cannabis preparations at low doses tested does not affect sweet taste intensity perception and liking and does not influence food preferences
Irving, Iqbal [54] (2018) / UK	Placebo-controlled, parallel-group, RCT	18 years or older with left-sided or extensive ulcerative colitis, Mayo scores of 4-10, and on stable 5-amino-salicylic acid dosing	60 / 29	250 mg (A 2-week dose escalation period with maximum tolerated dose twice daily, 5 capsules)	Oral / Capsules	To investigate the efficacy and safety of CBD-rich botanical extract treatment among patients with mild to moderate ulcerative colitis	Over 90 percent of patients received CBD-rich botanical extract treatment with at least one adverse event	CBD-rich botanical extract may have provided therapeutic benefit to patients who tolerated it
Naftali, Mechulam [99] (2017) / Israel	Placebo-controlled RCT	Aged 18-75 years with Crohn’s disease	19 / 10	20 mg	Oral / Dissolved in olive oil	To evaluate the effect of low-dose CBD on active Crohn’s disease	Well-tolerated but some experienced headache, sleepiness, nausea, and dizziness	There was no significant difference between the CBD and placebo groups in CDAI score before and after the treatment
Pharmacokinetics / Pharmacodynamic of CBD
Spindle, Cone [77] (2019) / US	Placebo-controlled, double-dummy dosing, within-subject, cross-over RCT	Aged 18-45 years and healthy males and females with BMI between 18 and 36 kg/m², and have prior experience inhaling cannabis	6 / 6	100 mg (orally and vaporized)	Oral / Placed CBD into a size 00 gelatin capsule and filled the remaining space with microcrystalline cellulose Inhalation/ Vaporized using desktop vaporizer with 204 °C	To evaluate the urine drug-testing outcomes of acute CBD administration via both oral ingestion and vaporization	Well-tolerated and no adverse events occurred	Under current federal occupational drug testing criteria, acute dose of pure CBD will not result in a positive urine drug test. However, using CBD and THC at the same time may result in a positive THC urine test
Sultan, O’Sullivan [63] (2020) / UK	Randomized, placebo-controlled, parallel group trial	Healthy men	26/ 13	600 mg	Oral / Not reported	To assess the effects of repeated CBD dosing on haemodynamics	Lack of appetite, headache, insomnia, hyperactivity and dysuria	Volunteers who had taken CBD had lower systolic BP after acute and repeated dosing. CBD reduces BP at rest after a single dose, but the effect is lost after 7 days of treatment
Other Topics
Arkell, Vinckenbosch [95] (2020) / Netherlands	Double blind, within-participants, RCT	Healthy, occasional users of cannabis aged 20-50 years with valid driver’s license at least 2 years’ driving experience and driving more than 2,000 km per year	26 /26	13.75 mg CBD (THC/CBD condition) and 13.75 mg CBD (CBD condition)	Inhalation / Vaporized at 200 °C	To determine the driving impairment caused by vaporized cannabis containing THC and CBD	One participant had panic attack (THC), some tests terminated due to driving safety concerns, terminations due to patient concerns about driving safety	No significant difference between CBD dominant cannabis and placebo in driving performance. However, the effect size for CBD-dominant cannabis may not have excluded clinically important impairment
Linares, Guimaraes [84] (2018) / Brazil	Cross-over RCT	Healthy volunteers	26 / 26	300 mg	Oral / Capsules	To assess the effect of the acute administration of an anxiolytic dose of CBD on sleep in health volunteers	Well-tolerated	Acute administration of an anxiolytic dose of CBD does not seem to interfere with the sleep cycle of healthy volunteers. CBD do not alter normal sleep architecture
Jadoon, Ratcliffe [55] (2016) / UK	Placebo-controlled, parallel-group, phase IIa proof-of-concept randomized clinical trial	Aged 18 years or older with type 2 diabetes and HbA1c ≤10% (86 mmol/mol), HDL-C ≤1.3 mmol/L in females and ≤1.2 mmol/L in males, and plasma TGs ≤10 mmol/L	62 / CBD=13; 1:1 CBD: TCHV=11; 20:1 CBD: TCHV=12	100 mg	Not reported	To investigate the effects of CBD and THCV on dyslipidemia and glycemic control in individuals with type 2 diabetes	Well-tolerated but with the majority of subjects experiencing mild or moderate adverse events	CBD did not appear to have any metabolic effects. However, THCV improved glycemic control and might be a new treatment option for those with type 2 diabetes
Matsuyama and Fu [75] (1981) / US	Within-subject RCT	Healthy adult males	27 / 7	1200 mg	Oral / Capsules	To investigate the effects of subacute oral administration of THC, CBN, and CBD on the frequencies of chromosome aberrations and sister chromatid exchanges	Not reported	The effects of subacute oral administration of the THC, CBN, and CBD did not show a statistically significant increase in chromosome breaks
Tomida, Azuara-Blanco [64] (2006) / UK	Randomized, placebo-controlled, 4 way cross-over study	Ocular hypertension or early primary open angle glaucoma with mild field defect and untreated IOP> 24 and < 36 mm Hg in at least one eye	6 /6	20 mg or 40 mg CBD	Oral / Supplied sublingually with a pump-action oromucosal spray	To assess the effect on IOP and the safety and tolerability of oromucosal administration	Generally tolerated; 3 participants on THC, 2 on CBD 20 mg, 5 on CBD 40 mg, and 2 on Placebo with at least one adverse event. No serious or severe adverse events occurred, but all but 2 (nausea and hypotension after administration of 5 mg THC, rated moderate) were rated mild	20 mg CBD did not reduce IOP at any time, whereas 40 mg CBD produced a transient increase IOP rise after 4 hours administration

ASD=autism spectrum disorder; BDNF=brain-derived neurotrophic factor; BMI=Body Mass Index; BP=blood pressure; BPRS=Brief Psychiatric Rating Scale; CBD=Cannabidiol; CBN=Cannabinol; CDAI=Crohn’s Disease Activity Index; CHR=Clinical High Risk; DSM-IV= Diagnosis and Statistical Manual for Physical Disorders, 5^th Edition-Research Version; fALFF=fractional amplitude of low-frequency fluctuations; FC=functional connectivity; fMRI=function magnetic resonance imaging; GABA= γ-aminobutyric acid; HC= healthy controls; HD= Huntington’s Disease; H¹-MRS=magnetic resonance spectroscopy; ID= Intellectual Disability; IQ=Intelligence Quotient; IOP= intraocular pressure; NA=not applicable; MCCB=MATRICS Consensus Cognitive Battery; OCD=Obsessive-compulsive disorder; PANSS=Positive and Negative Syndrome Scale; PD= Parkinson’s Disease; PEA= palmitoylethanolamide; PSI=Psychotomimetic States Inventory; QOL=Quality of Life; RCT=Randomized Clinical Trial; SAD=Social Anxiety Disorder; SSPS-N=Negative Self-Statement Scale; THC= Δ9-tetrahydrocannabinol; THCV= Δ⁹-tetrahydrocannabivarin; TG=Triglyceride; UPDRS=Unified Parkinson’s disease rating scale; VAMS=Visual Analogue Mood Scale

CBD was well tolerated among participants in these studies with few minor or moderate adverse events reported. The most common reported adverse events included headaches, sleepiness, dizziness, and/or diarrhea. One study reported a serious adverse event, but it was deemed unrelated to the clinical trial [56]. In general, sample sizes were small; the majority of studies (i.e., N=42; 72%) had less than 40 participants in total. The majority of included studies used the oral administration route and packed CBD in capsules. Although several studies did not report adverse events during the study, most of studies showed that the dose of CBD tested was generally well-tolerated.

With respect to study quality, 14 studies (24%) were determined to have at least one source of bias with “high risk” (Table 2), and only 9 studies (16%) had “low risk” with all sources of bias. Sixty-two percent (N=35) of the articles were categorized as “unclear” on the procedure for concealing the allocation sequence. In terms of incomplete outcome data, 9 studies were rated as “high risk” because those studies had relatively higher numbers of missing values. Of these studies, one study used listwise deletion when dealing with missing data, which might cause selection bias [95]. Most studies reported different outcomes, and even when the same outcome was reported, it was measured differently.

Table 2.

Risk of bias assessment

First author (Publication year)	Random Sequence Generationa	Allocation Concealmentb	Blinding of Participants and Personnelc	Blinding of Outcome Assessmentd	Incomplete Outcome Datae	Selective Reportingf	Other Sources of Biasg
Appiah-Kusi, Petros [44]	Unclear	Unclear	Unclear	Low	High	Low	High
Arkell, Vinckenbosch [95]	Low	Low	Low	Low	High	Low	Low
Babalonis, Haney [67]	Unclear	Unclear	Low	Low	Low	Low	Low
Belgrave, Bird [90]	Low	Unclear	Low	Low	Low	Low	Low
Bergamaschi, Queiroz [79]	Low	Unclear	Low	High	Low	Low	High
Bhattacharyya, Wilson [45]	Low	Low	Low	Low	Low	Low	Unclear
Boggs, Surti [68]	Unclear	Unclear	Unclear	Unclear	Low	Low	Low
Chagas, Zuardi [80]	Low	Unclear	Low	Low	Low	Low	Low
Consroe, Carlini [81]	Low	Low	Low	Low	Low	Low	Low
Consroe, Kennedy [70]	Low	High	Unclear	Low	Low	High	Low
Consroe, Laguna [69]	Low	High	Low	Low	Low	Low	Low
Couch, Cook [46]	Low	Low	Low	Low	Low	Unclear	Unclear
Das, Kamboj [47]	Unclear	Unclear	Low	Unclear	High	High	High
Davies, Wilson [48]	Unclear	Unclear	Low	Unclear	High	Low	High
De Bruijn, De Graaf [96]	Unclear	Low	Low	Unclear	High	Low	Low
de Faria, de Morais Fabrício [82]	Unclear	Unclear	Unclear	Unclear	Low	Low	Unclear
De Souza Crippa, Zuardi [89]	Low	Unclear	Low	Unclear	Low	Low	Low
Efron, Freeman [91]	Low	Low	Low	Unclear	Low	Low	Low
Englund, Morrison [12]	Low	Unclear	Low	Unclear	Low	Low	Low
Freeman, Adekanmi [49]	Low	Low	Low	Low	Low	Low	Low
Freeman, Hindocha [50]	Low	Low	Low	Low	Low	Low	Low
Fusar-Poli, Crippa [51]	Unclear	Unclear	Low	Unclear	Low	Low	Low
Hallak, Dursun [83]	Low	Unclear	Low	Low	Low	Low	Low
Haney, Malcolm [71]	Unclear	Unclear	Low	Low	Low	Low	Low
Hindocha, Freeman [52]	Low	Unclear	Low	Low	High	Low	Low
Hosseini, McLachlan [92]	Low	Unclear	Unclear	Unclear	Low	Low	Low
Hundal, Lister [53]	Low	Unclear	Low	Low	Low	Low	Low
Hurd, Spriggs [72]	Low	Low	Low	Unclear	Low	Low	Low
Irving, Iqbal [54]	Low	Low	Low	Low	High	Low	Low
Jadoon, Ratcliffe [55]	Low	Low	Unclear	Low	Low	Low	Low
Kayser, Haney [73]	Low	Low	Low	Low	Low	Low	Low
Leweke, Piomelli [98]	Low	Low	Unclear	Low	Low	Low	High
Linares, Guimaraes [84]	Low	Unclear	Low	Low	Low	Low	Low
Linares, Zuardi [85]	Unclear	Unclear	Low	Low	Unclear	Low	Low
Manini, Yiannoulos [74]	Unclear	Unclear	Low	Low	Low	Low	Low
Matsuyama and Fu [75]	Low	Unclear	Low	Unclear	Low	Low	Low
McGuire, Robson [56]	Low	Low	Low	Low	Low	Low	Low
Meneses-Gaya, Crippa [86]	Unclear	Unclear	Low	Low	Low	Low	Low
Morgan, Das [57]	Low	Unclear	Low	Low	Low	Low	Low
Morgan, Freeman [58]	Low	Unclear	Low	Low	Low	Low	Low
Naftali, Mechulam [99]	Low	Low	Low	Low	Low	Low	Low
Notcutt, Price [59]	Unclear	Low	High	Low	Low	Low	Low
O’Neill, Wilson [60]	Low	Low	Low	Low	Low	Low	Low
Perkins, Butler [93]	Low	Unclear	Low	Low	Low	Low	Low
Pretzsch, Freyberg [62]	Low	Low	Low	Unclear	Low	Unclear	Low
Pretzsch, Voinescu [61]	Low	Low	Low	Unclear	Low	Unclear	Low
Schoedel, Szeto [100]	Low	Unclear	Low	Low	Low	Low	Low
Skosnik, Hajós [76]	Low	Unclear	Unclear	Unclear	Low	Unclear	Low
Solowij, Broyd [94]	Low	Unclear	Low	Low	Low	Low	Low
Spindle, Cone [77]	Unclear	Unclear	Low	Low	High	Low	Low
Sultan, O’Sullivan [63]	Low	Low	Low	Low	Low	Low	Low
Taylor, Gidal [97]	Low	Unclear	Low	Unclear	Low	Low	Unclear
Tomida, Azuara-Blanco [64]	Unclear	Low	Low	Low	Low	Low	Low
Wade, Robson [65]	Low	Unclear	Low	Low	Low	Low	Low
Wilson, Bossong [66]	Unclear	Low	Low	Low	High	Low	Low
Xu, Cullen [78]	Low	Unclear	Low	Low	Low	Low	Low
Zuardi, Shirakawa [87]	Unclear	Unclear	Low	Unclear	Low	Low	High
Zuardi, Cosme [88]	Unclear	Unclear	Low	Low	Low	Low	Low

^aRandom sequence generation: The procedure for generating the allocation sequence to determine if it should create comparable groups

^bAllocation concealment: The procedure for concealing the allocation sequence to determine if intervention allocations could have been predicted prior to or during enrollment

^cBlinding of participants and personnel: Blinding participants and personnel from knowledge of which intervention each participant received

^dBlinding outcome assessment: Blinding outcome assessors from knowledge of which intervention each participant received

^eIncomplete outcome data: The completeness of outcome data for each major outcome, including attrition and analytical exclusions.

^fSelective reporting: How the authors investigated the potential of selective outcome reporting and what they discovered

^gOther sources of bias: Any important concerns about bias that were not addressed

3.2. Health Conditions

Mental health was the most studied topic in this review. Among the included studies, 31 studies (53%) investigated the effects of CBD on mental health, and 8 (14%) of these studies investigated the effects of CBD on anxiety, fear, or paranoia. Three studies reported that CBD significantly reduced anxiety, cognitive impairment, and distress among study participants [79, 85, 88]. One study showed that CBD significantly decreased subjective anxiety, increased mental sedation, and had greater activity in the left parahippocampal gyrus, which suggested that CBD has anxiolytic effects, and those effects are likely mediated by the drug’s impact on limbic and paralimbic brain areas [89]. Another study found that CBD may inhibit the anxiety caused by THC among healthy volunteers, and this effect can extend to other subjective effects caused by THC [87]. Additionally, CBD can be adjunct to extinction-based therapy for patients with anxiety disorders [47]. However, CBD had no effect on anxiety or persecutory ideation among healthy volunteers with high trait paranoia [53].

Eight studies examined the effects of CBD on psychosis and three on schizophrenia. Among the studies conducted on psychosis, CBD showed intermediate therapeutic effects in clinical high risk (CHR) psychosis patients by reducing the level of anxiety and stress [44] and modulating brain function while viewing fearful face expressions [48]. Additionally, a single dose of CBD had normalized clinical high risk-related changes in parahippocampal, striatal, and midbrain function [45], and reduced the enhancement activity in the left insula/parietal operculum [66], and normalized mediotemporal and prefrontal dysfunction [60]. However, when adding THC, THC had negative effects among CHR patients, such as elicited positive psychotic symptoms [12], robustly impaired episodic and working memory [58]; CBD did not attenuate these effects caused by THC and can help to mitigate those effects only when CBD is combined with THC at high levels [94]. Among studies focusing on schizophrenia, CBD showed fewer extrapyramidal symptoms [98] and lower levels of positive psychotic symptoms, greater cognitive performance improvements, and overall functioning [56]. However, one study showed that CBD augmentation for 6 weeks did not improve cognitive ability nor psychiatric symptom severity among stable antipsychotic-treated schizophrenia outpatients [68].

Eight studies (14%) investigated effects of CBD on substance use, which included cannabis [50, 52, 67], alcohol [81], crack-cocaine [86], opioids [74], tobacco [57], and recreational polysubstance use [100]. As for cannabis, a secondary analysis of a clinical trial [71] focused on the abuse liability of a range of oral CBD doses alone and in combination with smoked marijuana (5.3-5.8% of THC); this study found that CBD did not display any abuse-related subjective effects (i.e., feeling high, feeling mellow, cravings) unlike active marijuana [67]. Another study suggested that both 400 mg and 800 mg of CBD were safe and effective to reduce cannabis use among individuals with cannabis use disorders compared to placebo [50]. One study suggested that genes producing the CNR1 receptor and the FAAH enzyme modulate the acute effects of different cannabinoids on addiction endophenotypes in healthy people, which was the first to provide a clear understanding that genetic factors contribute to cannabis addiction [52]. Alcohol alone and in combination with CBD produced significant impairments of motor and psychomotor performances, but CBD alone did not produce those effects [81]. CBD might not be an ideal treatment for crack-cocaine withdrawal [86]. However, CBD could be a potential treatment for opioid abuse [74], reduce tobacco smoking consumption [57], and has a low abuse potential among polydrug users [100].

Lastly, the rest of studies investigated the effects of CBD on obsessive-compulsive disorder (OCD) [73], severe behavioral problems [91], autism spectrum disorder (ASD) [61, 62], and behavioral and cardiovascular effects [71]. Efron et al conducted a study to determine whether it was feasible to investigate the effects of CBD in children with intellectual disability and severe behavioral problems and their families; while the study was not powered to discern CBD’s effects, the trial was feasible [91]. Hanley et al determined that CBD alone could not reduce the reinforcement, physiological, or positive subjective effects of smoked cannabis [71]. Kayser el al determined that either THC or CBD had little impact on OCD symptoms and smaller reductions in anxiety, whereas THC increased heart rate, blood pressure, and intoxication compared to CBD and placebo [73]. One of the two ASD studies did not show the efficacy of CBD on ASD, but that CBD can alter the glutamate and GABAergic levels [62]; another study showed that CBD can alter the fALFF and FC in/between brain regions in adults with ASD [61].

Neurological diseases or movement disorders were another commonly studied topic found during this review (N=11) and included Parkinson’s disease (PD) [80, 82], Huntington’s disease (HD) [69, 70], multiple sclerosis [49], brain function [51], chronic pain [59, 78], neural oscillations [76], and neurogenic disorders [65]. CBD with 300 mg/day was shown to improve quality of life measures in PD patients with no psychiatric comorbidities, although there were no statistically significant differences in Unified Parkinson’s disease rating scale (UPDRS) to assess PD symptoms, plasma brain-derived neurotrophic factor (BDNF) levels, or magnetic resonance spectroscopy (H¹-MRS) measures [80]. Among PD patients, CBD can also reduce anxiety and tremor amplitude with an acute dose of 300 mg [82]. In terms of HD, CBD was neither symptomatically effective nor toxic in neuroleptic-free patients [69]. However, CBD might serve as a useful cannabinoid to model THC in long-term, high-dose pharmacokinetic studies [70]. The findings of a study found that both CBD and THC had significant clinical effects on reducing urge incontinence episodes among patients with multiple sclerosis [49]. Another study, which focused on the effects of THC and CBD on neural activation, found that CBD’s effects on activation in limbic and paralimbic areas may be related to its ability to reduce autonomic arousal and subjective anxiety, whereas THC can increase anxiety, levels of intoxication, sedation, and psychotic symptoms, and THC’s anxiogenic effects may be related to effects in other brain regions [51]. In terms of chronic pain, CBD was not effective in chronic pain symptom control [59], but CBD significantly reduced in acute pain, sharp pain, cold, and itching feelings, which can be a more effective therapeutic option for peripheral neuropathy compared to current medications [78]. THC alone and in combination with CBD disrupted sensory P50 gating and decreased first evoked theta oscillations relevant to perception and cognition, while CBD alone had no effects [76]. The THC, CBD, and 1:1 THC:CBD extracts may improve neurogenic symptoms such as impaired bladder control, muscle spasms, and spasticity, and both THC and CBD had significantly superior effect on pain relief compared to placebo [65].

Three studies investigated the interaction between CBD and ethanol [90], opioids [72], and ketamine [83]. CBD can decrease and exhibit a sustained impact on cue-induced desire and anxiety among people who were abstinent from heroin use disorder [72], whereas CBD showed non-significant effect on reducing ketamine-induced depersonalization [90] and blood ethanol levels when combined with ethanol [83]. Four studies investigated the impact of CBD on the gastrointestinal tract [46, 54, 96, 99]. In combination with palmitoylethanolamide, CBD can reduce the permeability in the human colon, which provides the implications in disorders associated with increased gut permeability [46]. Among patients who had mild to moderate ulcerative colitis, although the study showed the poor tolerability of CBD, it might provide therapeutic benefit to patients who tolerated the dosage of CBD [54]. However, CBD had no beneficial effects on Crohn’s disease [99], and no effect on sweet taste intensity at low doses [96]. In terms of other health outcomes, one study suggested that an acute dose of 100 mg CBD would not result in a positive urine test drug test under the current federal occupational drug testing guidelines but using 100 mg CBD and 3.5 mg THC at the same time may result in a positive urine test [77]. Other studies found that CBD did not have effect on impaired-driving performance [95], interfere with the sleep cycle among healthy volunteers or alter normal sleep architecture [84], nor produce any metabolic effects among patients with Type 2 diabetes [55]. However, the effect size for CBD-dominant cannabis may not have excluded clinically significant impairment [95]. An additional study determined that CBD did not increase chromosomal damage [75]. Tomida et al found that CBD did not reduce intraocular pressure when participants received a 20 mg dose at any time, while 40 mg of CBD produced a transient increase in intraocular pressure after 4 hours administration [64]. Lastly, three studies investigated the safety and tolerability of CBD; all studies found that CBD was generally well tolerated even at a single arm dose of 6,000 mg [92, 93, 97].

4. Discussion

This review found 58 RCTs conducted from 8 countries, which documented the effectiveness and safety of CBD. Nearly half of these studies were conducted among healthy individuals while the other half focused on populations with pre-existing medical or health conditions. The included studies utilized wide ranges of dosages (i.e., 400 ug to 6,000 mg), which suggests that CBD displays a wide therapeutic range in different medical conditions. Additionally, most studies (i.e., 72%) had sample sizes less than 40 participants. Almost all studies, including three which focused solely on safety and tolerability [92, 93, 97], determined that CBD was well-tolerated with few adverse events even at extremely high doses. More importantly, this review found that the current state of knowledge regarding CBD’s effectiveness is mixed, and it contains breadth, but not depth. While this review shows that a myriad of health conditions was investigated, the research questions, objectives, and outcomes of these studies were so diverse which makes comparability difficult.

For example, CBD’s effectiveness on mental health-related conditions or outcomes were the most studied topic in this review. The findings of these studies were highly varied. Many studies were interested in determining CBD’s effectiveness at reducing anxiety. Three studies reported that CBD can reduce the anxiety, distress, and discomfort experienced during public speaking for both healthy volunteers and people who had generalized social anxiety disorder [79, 85, 88], as well as patients with PD [82]. However, in other studies CBD did not show any effects on anticipatory or performance anxiety [88], and CBD did not show any significant reduction of some physiological measures of anxiety (i.e., blood pressure and heart rate) [79]. Another study also had a contradictory conclusion that 600 mg of CBD, which was the most studied dose in this review, was not an effective anxiolytic under the condition where anxiety and cortisol levels were elevated among participants during a virtual-reality experiment [53].

Likewise, other discrepancies were observed for other conditions. For example, two studies suggested that CBD may be effective at assisting with the treatment of schizophrenia [56, 98], while another study suggested that CBD did not improve cognitive ability nor reduce psychiatric symptom severity among stable schizophrenia outpatients [68]. Some studies suggested that CBD may be an effective treatment for some substance use problems, but not others. For example, one study determined that both 400 mg and 800 mg of CBD were safe and effective for cannabis use disorder treatment [50], while other studies found CBD may be a potential treatment for opioid abuse [74] and could reduce tobacco cravings [57]. However, another study found that CBD was not effective for crack-cocaine craving treatment [86]. CBD’s effectiveness for pain management was also varied. For example, one study found that CBD alone did not reduce chronic pain symptoms [59]; conversely, two other studies determined that CBD was superior to placebo at improving pain experienced by neurogenic patients [65], while the other suggested that CBD can be a more effective therapeutic option for acute and sharp pain, or other peripheral neuropathy [78].

This incongruence seen between studies looking at similar medical conditions or outcomes is potentially explainable. Most studies were comprised of small sample sizes, which could significantly reduce the statistical power for finding differences between study groups. Some studies were occasionally limited to certain sex or age groups, which could impact the findings. For example, 9 studies were restricted to male participants [46, 61–63, 75, 83, 86, 89, 96]. Therefore, the findings of those RCTs may not be generalizable to women. Also, one study was limited to children aged 8-16 years [52], again possibly limiting generalizability to other age groups.

Several methodological weaknesses were identified in the included studies based on the Cochrane Tool. Approximately 60% of included studies did not provide detailed procedures for concealing the allocation sequence and only specified that study participants were randomized rather than fully describing the procedure of randomization. This is important information in the RCTs for assuring that the random allocation process was carried out without knowledge of which patient would receive which treatment [101]. The differences seen could also be attributed to the dosage of CBD used, the route of administration, drug tolerability, or even physiologic differences due to disease progression. It could also be that some studies were conducted in healthy participants while others were conducted in clinical populations.

From a research perspective, this review shows that more research is needed regarding CBD’s effectiveness as it is still unclear. Additional studies and larger sample sizes are paramount to elucidating these relationships. Also, two recent cross-sectional studies identified that many healthy individuals use CBD to reduce anxiety, stress, and depression and improve sleep [25, 102]. However, very few if any studies were conducted to determine if CBD can effectively treat these symptoms in healthy individuals.

The findings of this review also have several key clinical implications. CBD dosages varied greatly as did the route of administration (i.e., inhalation, oral capsule, sublingual oil). Additionally, CBD has very low oral bioavailability (6%) [103], which might be an important reason why CBD might not be effective in some RCTs as the therapeutic doses were not achieved. No clear guidance exists for many of the conditions studied nor by the route of administration. While there are clear dosing requirements for prescription CBD and the medical conditions it’s known to treat, these may not translate directly to non-prescription forms nor other health conditions. Also, prescription CBD has reported side effects [28, 104]. These side effects likely translate to non-prescription CBD, but this is currently unknown. Clear dosing guidelines are needed to ensure safety of the general public which has unfettered access to the non-prescription form in many countries. This issue is further complicated by the lack of regulation, which often leads to mislabeled products or products which contain more THC than legally permitted [105–107]. Increased regulations regarding non-prescription CBD are likely needed to ensure that customers are aware how much they are taking and if they are taking a product that contains THC, which is psychoactive.

The main strength of this review was that it included over 40 years of data among seven different literature databases. This systematic review also included “grey literature” (i.e., thesis and dissertations), which can make important contributions to systematic reviews, such as reducing publication bias, facilitating a more balanced review of evidence, and disseminating null or negative results [108, 109]. Although this review adhered to the PRISMA guidelines, it is not without limitation. This review was limited to English publications. Thus, studies in other languages were not included. The search strategy was limited to library databases. While the US requires clinical trials to be registered and this database can be searched by the public, it often does not provide the full publication which was needed to assess potential biases. Thus, there were likely other studies that existed, but were not included in this review due to this reason. Also, if full text articles were unavailable, studies were not included. It is also possible that some articles may have been missed despite searching multiple databases. Because our searches began before registration could be initiated, our systematic review was not pre-registered, which might have duplicate or similar reviews.

Conclusions

This systematic review summarized the effectiveness and safety of CBD in RCTs. Among the 58 studies included, this review determined that a myriad of health conditions and outcomes have been studied in both clinical and healthy populations across eight different countries. While numerous conditions have been studied, the effectiveness of CBD is unclear and additional research with larger sample sizes are paramount to elucidate these relationships. Also, while CBD was generally safe and well-tolerated even in high doses among the included studies, additional research examining dosages are necessary for patient safety. This information will help physicians give more appropriate guidance to patients using CBD for specific medical treatments or those taking CBD as a dietary supplement. Also, further regulations are needed to provide correct and transparent labeling of products, especially as CBD is available to the public in non-prescription form.

Key Points:

• Cannabidiol (CBD) is one of the most prevalent cannabinoids found in cannabis and considered a non-psychoactive component of cannabis when compared to Δ9-tetrahydrocannabinol (THC).

• A systematic review was conducted to assess the efficacy and safety of CBD on study outcomes of specific medical problems and features of the population investigated in randomized clinical trials.

• The effectiveness of CBD in different health conditions is inconsistent and still not clear. Additional research with larger sample size and dosages, route of administration, and increased regulations are needed to make sure the safety of CBD products.

Availability of data:

We have made all data that were extracted for this manuscript available in the electronic supplementary material (Supplementary Table)