A systematic review of cannabidiol dosing in clinical populations

Aims Cannabidiol (CBD) is a cannabis‐derived medicinal product with potential application in a wide‐variety of contexts; however, its effective dose in different disease states remains unclear. This review aimed to investigate what doses have been applied in clinical populations, in order to understand the active range of CBD in a variety of medical contexts. Methods Publications involving administration of CBD alone were collected by searching PubMed, EMBASE and ClinicalTrials.gov. Results A total of 1038 articles were retrieved, of which 35 studies met inclusion criteria covering 13 medical contexts. Twenty‐three studies reported a significant improvement in primary outcomes (e.g. psychotic symptoms, anxiety, seizures), with doses ranging between <1 and 50 mg/kg/d. Plasma concentrations were not provided in any publication. CBD was reported as well tolerated and epilepsy was the most frequently studied medical condition, with all 11 studies demonstrating positive effects of CBD on reducing seizure frequency or severity (average 15 mg/kg/d within randomised controlled trials). There was no signal of positive activity of CBD in small randomised controlled trials (range n = 6–62) assessing diabetes, Crohn's disease, ocular hypertension, fatty liver disease or chronic pain. However, low doses (average 2.4 mg/kg/d) were used in these studies. Conclusion This review highlights that CBD has a potential wide range of activity in several pathologies. Pharmacokinetic studies as well as conclusive phase III trials to elucidate effective plasma concentrations within medical contexts are severely lacking and highly encouraged.

low risk. 2 There is no evidence for dependency or abuse potential with CBD use, as concluded by the World Health Organisation Expert Committee on Drug Dependence. 1 The purported effects of CBD include analgesic, anti-inflammatory, antioxidant, anxiolytic, anticonvulsant and cytotoxic effects, which are mediated through signalling mechanisms including the cannabinoid receptor 1 (weak agonist), the cannabinoid receptor 2 (inverse agonist), the serotonin 1a receptor (5-HT 1A ), G protein-coupled receptor 55 (GPR55), G protein-coupled receptor 18 (GPR18) and the transient receptor potential cation channel subfamily V member 1 (TRPV1) receptors, amongst others. 3 Clinically, CBD is being investigated in multiple disease states including neurodegeneration, anxiety disorder, orphan childhood diseases with a prevalence of <5 in 10 000 individuals (e.g. tuberous sclerosis complex) and addiction (ongoing trials in cannabis and cocaine craving). [4][5][6] Epidiolex has recently become the first Food and Drug Administration-approved CBD medicine, indicated for use in Lennox-Gastaut or Dravet syndrome (childhood epilepsy) by oral administration. Sativex is an oromucosal spray containing both CBD and δ-9-tetrahydrocannibinol, which is licenced in the EU and Canada Hemp-derived CBD is commercially available and is currently used as a health and food supplement commonly for anxiety and pain relief. This market represents a flourishing industry expected to rise financially and globally. 7 However, the blurred lines between CBD as a licensed medicine and CBD as an over-the-counter remedy contribute to the overall lack of understanding of what dose of CBD may be considered therapeutic. This is further hampered by the lack of standardisation in over-the-counter CBD products and their unregulated labelled doses.
Despite the prevalence of CBD use and current hype, guidance on dose recommendations has not advanced and is not clear, additionally hampered by the striking lack of accessible pharmacokinetic and bioavailability data of CBD in humans. 8

| Eligibility criteria
The titles and abstracts of retrieved studies were examined by 2 independent researchers, and inappropriate articles were rejected.
Inclusion criteria were as follows: an original, peer-reviewed published paper that involved administration of CBD to a clinical population, or reported on clinicaltrials.gov, and included an outcome measurement to assess the efficacy of CBD i.e. improvement in disease. Exclusion criteria were: administration in healthy participants only; CBD administered in combination with other cannabinoids such as with δ-

| Data acquisition and analysis
The included articles were analysed, and the following data extracted: sample size, clinical population/medical context; study design and length; administration route of CBD; source of CBD; dose of CBD; side effects; and primary outcome results. All data entry was checked by an additional independent researcher. Risk of bias of the 15 randomised controlled trials was assessed using the 2011 Cochrane Collaboration's tool for assessing risk of bias.
As this review included studies of participants of all ages (from infants to adults), dosing is reported in mg/kg of body weight to allow for comparison. Where not available as mg/kg (24 studies), dose was converted for adults using an average adult body weight of 62 kg. 9 In only 1 publication, a case report on a child, an average child weight of 40 kg had to be used to convert reported mg/d dose into mg/kg/d. 10 A positive effect of CBD was determined by the presence of a significant improvement in primary end points(s) or outcomes reported compared to placebo or baseline. A lack of positive effect was determined if no significant improvements were reported. Mixed findings were reported for example in case reports wherein some patients improved, others did not, or where a primary outcome was not specified (exploratory study) and in which some endpoints improved while others worsened (1 study) or remained unchanged.

| RESULTS
The initial search yielded 1038 records, from which 896 abstracts were reviewed, and 35 articles were included in the final analysis, comprising a total number of 1223 participants. A flow chart of article retrieval and selection is presented in Figure 1. Of the 15 RCTs, the range of doses investigated varied from <1 mg/kg up to 20 mg/kg per day (average 9 mg/kg/d). [11][12][13][14][15][16][17][18][19][21][22][23][24][25] Seven RCTs reported CBD efficacy (average dose 14 mg/kg/ d), [11][12][13]16,19,20,24 7 studies describe neutral effects of CBD (average dose 5 mg/kg/d) 14,15,17,[21][22][23]25 and 1 study showed both positive and negative outcomes. 18 In the remaining 8 clinical trials of various    reported beneficial effects of CBD on seizure frequency, duration and severity with an average administered dose of 21 mg/kg/d. [33][34][35][36] Seven studies were conducted in the context of schizophrenia and bipolar disorder. Within the RCTs, 2 conducted with an average dosing   11,19 One of these compared CBD against an active control (amisulpride), and the other as an add-on therapy to usual medication compared to placebo as an add-on therapy. However, a third RCT employing CBD as an add-on therapy did not report any improvements in cognition or symptoms of schizophrenia after a lower average dose of 10 mg/kg/d over 6 weeks (n = 36). 14 An acute dose of 5 or 10 mg/kg/d did not improve selective attention in a placebocontrolled trial of 28 schizophrenia patients. 30  patients. 20 Likewise, in another double-blind placebo-controlled study, a dose of 6.7 mg/kg reduced subjective anxiety in 10 adults with generalised SAD. 5 Additionally, in a case report in a child, 0.6 mg/kg/d increased sleep quality and duration, and decreased anxiety secondary to PTSD. 10 Lastly, it was found that doses of 5 mg/kg/d prevented occurrence of graft-vs-host disease in a phase II clinical trial (n = 48) and 5-10 mg/kg/d doses have been shown in a case report to remove withdrawal symptoms from a patient with cannabis dependency. 29,38 Within studies that compared CBD against a placebo or control Analysis of these data revealed that a greater proportion of studies reported a beneficial effect of CBD in the add-on therapy group compared to the monotherapy group (n = 6 and n = 2 respectively).
However, higher doses were used overall within the add-on therapy group compared to the monotherapy group (average 11 and 6 mg/kg/d, respectively) and, due to such a small data set and heterogeneity of studies, we did not perform any further analysis.

| DISCUSSION
To our knowledge, this is the first study to compile and compare all publications in which CBD was administered to clinical populations.  46,47 ), neutral results may be secondary to subtherapeutic dosing, and dose-escalation trials with embedded pharmacokinetic studies are the next logical step. 15,22 Studies in this review using higher doses concluded that CBD was generally well-tolerated with the most frequent side effects including drowsiness, nausea, somnolence, fatigue and vomiting.
Among the clinical trial records retrieved from clinicaltrials.gov, only 60% of completed trials had results uploaded and available. This may represent a significant publication bias and is suggestive of disregard for the priority of publication of negative results, which is a wellrecognised problem. 48 Unfortunately, this may potentially skew the findings presented in this review and so should be interpreted with caution and is acknowledged as a limitation. We also acknowledge that despite all routes of administration being oral, there may be further bias introduced between studies as one dose cannot be directly compared to another due to lack of standardisation of formulations and pharmacokinetic activity, including differences in bioavailability between an oral spray and an oral capsule.
Future studies should also consider the safety of drug interactions with CBD. CBD is a known inhibitor of the cytochrome P450 (CYP) system 49 and can therefore increase plasma concentrations of medicines already in use, in particular antiepileptic drugs. Indeed, this has been reported in a number of publications investigating concomitant use of CBD and antiepileptic drugs. 50 Similarly, CYP inhibitors are predicted to increase CBD plasma concentrations which should be equally monitored. Where possible, further well designed trials with CBD may disentangle whether CBD offers unique therapeutic potential in addition to benefits seen when used as an add-on treatment.

| CONCLUSION
Although larger confirmatory and efficacy clinical trials examining dosing in more detail for each medical context is required, this review summarises that CBD appears to offer a wide-range of activity between 1 and 50 mg/kg/d, and there was a tendency of studies with positive outcomes to have used higher doses of CBD. We recommend pharmacokinetic dosing schedules in subsequent trials to consider this range along with safety data and individual patient requirements. Finally, we implore all completed trial results to be made readily available so the research community can progress and learn from equally important positive and negative outcomes for the ultimate benefit of patients.