Use of non-medical cannabis in epilepsy: A scoping review

Objective The use of medical cannabis among people with epilepsy (PWE) has been garnering increasing interest. In this scoping review, we aimed to summarize the literature on recreational/non-medical cannabis (NMC) use in PWE, focusing on the experience, habits, and beliefs of PWE regarding NMC. Methods Four databases (OVID Medline, OVID Embase, Ovid APA PsycInfo, and Web of Science) were searched for studies describing NMC use in PWE. NMC was defined as cannabis products procured from sources other than by prescription. Studies that consisted in original research and that detailed the experience, habits, and/or beliefs of PWE regarding NMC use were included in the analysis. Data pertaining to study identification, demographics, NMC use, and epilepsy characteristics were extracted. Descriptive statistical analyses and reflexive thematic analyses were performed to map these data. Results In total, 3,228 records were screened, and 66 were included for analysis: 45 had mainly adult samples, whereas 21 had mainly pediatric samples. Most studies were published after 2010, originated from the USA, and were cross-sectional. The median number of PWE using cannabis in these studies was 24.5 (1–37,945). No studies showcased elderly PWE, and most had predominantly Caucasian samples. The lifetime prevalence of NMC use in PWE was variable, ranging between 0.69 and 76.8%. Factors frequently associated with NMC use in PWE were male sex, younger adult age, and lower education status. Children with epilepsy took NMC primarily for seizure control, using high CBD/THC ratios, and only orally. Adults with epilepsy took NMC for various reasons including recreationally, using variable CBD/THC ratios, and predominantly through smoking. The majority of PWE across all studies perceived that NMC aided in seizure control. Other aspects pertaining to NMC use in PWE were rarely reported and often conflicting. Conclusion The literature on NMC use in PWE is sparse and heterogeneous, with many salient knowledge gaps. Further research is necessary to better understanding the experience, habits, and beliefs of PWE pertaining to NMC.


. Introduction
Cannabis, also commonly referred to as marijuana, is most often derived from the Cannabis sativa plant and has been used for millennia for recreational, medicinal, and spiritual purposes (1). Historically, cannabis was categorized as an illegal substance in many Western countries (2). In the last decades, however, cannabis has seen a rise in interest in developed countries due to its potential medicinal qualities (2,3). Research on the medicinal properties of cannabis has pushed many administrative bodies to authorize prescribing medical marijuana for certain disorders, such as epilepsy (3). In the last years, an increasing number of jurisdictions have either decriminalized or legalized non-medical cannabis (NMC) use, making cannabis more available for the general population (2).
Although there is debate on how many active compounds are present in cannabis, trans--9-tetrahydrocannabinol (THC) and cannabidiol (CBD) are the most described. The psychoactive properties of cannabis are mostly mediated by THC (1). While CBD has clearly been shown to have anti-seizure properties, contradictory pro-convulsant and anti-seizure effects have been reported for THC (4,5). Four landmark randomized controlled trials (RCTs) have notably demonstrated the antiseizure benefits of CBD in Lennox-Gastaut syndrome, Dravet syndrome, and tuberous sclerosis. In response, the US Food and Drug Administration (FDA) approved Epidiolex, an oral CBD solution, for the treatment of these three syndromes (3,(6)(7)(8)(9). Although Epidiolex is not yet approved in many parts of the world, the evidence surrounding the anti-seizure properties of CBD has prompted an increased demand for high CBD and low THC cannabis oils among individuals with refractory epilepsy (4).
Although medical cannabis is now more readily available in many developed countries, its price point limits its use by many people with epilepsy (PWE). In the US, for instance, Epidiolex is estimated to cost tens of thousands of dollars yearly (4). As recent research on cannabis in epilepsy has mostly focused on medical cannabis despite it not being readily available for PWE in many countries, the following question arises: what is known about the use of NMC in PWE? The prevalence of PWE consuming NMC is not well-described, although some studies report that more than half of PWE consume cannabis, with many basing their consumption on cannabis' role in seizure control (10, 11). This reasoning is not trivial since NMC (with often non-negligible THC content) has not been shown to be an efficacious antiseizure agent and, while it is used without significant negative consequences by many, it has been associated with various harmful health effects (12). Achieving a better understanding of the various factors underpinning NMC consumption in PWE should be a research priority, especially given its potentially hazardous effects on health and the currently limited availability of medical cannabis in many countries.

. Methods
The protocol for this review was registered on the Open Science Framework platform (doi: 10.17605/OSF.IO/C5E74) and follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) reporting guideline (13). A scoping review is defined as "a form of knowledge synthesis that addresses an exploratory research question aimed at mapping key concepts, types of evidence, and gaps in research related to a defined area or field by systematically searching, selecting, and synthesizing existing knowledge" (14). The aim of broadly mapping a body of literature is one of the key aspects differentiating a scoping review from a systematic review, which rather aims to answer a set of defined questions by critically appraising the literature. Though both types of review require a similar rigorous, transparent methodology, a scoping review will usually screen for multiple different study designs and will not task itself with quality assessment (15).

. . Objectives
This scoping review sought to answer the following research question: "What is known about the consumption of NMC in PWE?" More broadly, we aimed to summarize the nature, extent, and range of the available research on the subject, identify relevant knowledge gaps, and offer recommendations for future research (16). NMC was defined as all cannabis products, in any form (e.g., whole plant or oil), that are procured from sources other than by prescription from a medical professional (e.g., legal market, if applicable, or illicit vendors). As such, even if a cannabis product was used by an individual for a "self-medicating" purpose (e.g., for seizure control), if this product was not prescribed by a medical professional, it was still considered NMC. The term "recreational cannabis, " which can be thought of as interchangeable with NMC, was not employed in this review, as it may infer that the cannabis product is used only recreationally, whereas it may in fact be used mainly for its perceived health benefits.

. . Search strategy
All authors revised and contributed to the search strategy. A health sciences librarian reviewed the search strategy and offered adjustments, and a consensus was eventually reached to search the following databases: OVID Medline, OVID Embase, OVID APA PsycInfo, and Web of Science (17). Free-text terms were combined in various manners with controlled vocabulary terms (when applicable) to search for research reporting on humans. These search terms were inspired from previous research investigating strategies for identifying citations on epilepsy and cannabis (18,19). The full search strategy, including sources of gray literature, is available in Supplementary Table S1. No restrictions were placed on language or time of publication. Online translation services were employed when feasible for studies written in languages other than English, French, and Spanish. We employed a "snowballing" method by which the reference lists of up to 20 selected literature reviews on cannabis use in epilepsy were manually reviewed for additional works of interest. All citations were imported in the Covidence online platform, which was used for the study selection phase. Duplicate citations were automatically removed by the . /fneur. .
Covidence platform, and duplicates that remained were manually identified during the study selection phase.

. . Study selection
Two reviewers (J.L. and C.C.A.) independently screened titles and abstracts on the Covidence platform for works pertaining to the use of NMC in PWE. At this stage, citations were included if they explicitly described NMC consumption in PWE or if they described cannabis consumption in PWE without enough information to determine if the cannabis was medical or nonmedical. Citations explicitly only describing medical cannabis use in PWE (e.g., studies in which patients with epilepsy are offered CBD products by the investigators), only describing the drug mechanisms of cannabis, only involving non-human subjects, or consisting in duplicates were excluded. No restriction was placed on study type. Full-text articles were then independently reviewed by the same two reviewers (J.L. and C.C.A.) using the Covidence platform. At this stage, citations were included if they described the experience, habits, and/or beliefs of PWE in relation to NMC use. Citations were excluded if they only pertained to medical cannabis use, if they only described the drug mechanisms of cannabis in epilepsy, if they consisted in duplicate citations, or if they were not original research. Any disagreement between the two reviewers were resolved through discussion, with third party (D.H.T.) intervention as necessary.

. . Data extraction
Data were independently extracted from the works having passed the selection phase by one reviewer (J.L.) using an Excel spreadsheet. A second reviewer (C.C.A.) verified the data extraction form, and disagreements were handled through discussion with a third party (D.H.T.) when needed. The following information was extracted when available:

. . Data synthesis
Data were synthesized using Excel. Our analysis comprised a numerical component, wherein data were synthesized using descriptive statistical methods, and a thematic component, wherein data were qualitatively analyzed for themes using a reflexive approach (20). Descriptive statistical analyses were used to synthesize findings from adult studies and pediatric studies separately. Numerical data are presented as medians (range) and count (proportions), where appropriate. All descriptive statistical analyses were performed using R.

. Results
A total of 4,182 records were identified using the search strategy in Supplementary Table S1. With 953 records being flagged as duplicates by Covidence, 3,228 records underwent the title/abstract screening phase. A total of 121 records passed this phase and underwent full-text screening. Of these records, the reference lists of 20 selected literature reviews were manually examined for additional works of interest. No additional works of interest were detected. With 55 records being excluded from the full-text screening phase, 66 records were included in the final analysis (2,4,10,11,. The data selection flowchart is presented in Figure 1. Table 1 lists the studies included in this review and provides a summary of their sample and primary aim. Table 2 presents in detail the publication identification and study characteristics of all 66 included studies whilst providing separate information for adult and pediatric studies. Supplementary Table S2 presents how many times each variable of interest (demographic, epilepsy, and NMC data) is explored across all included studies and between adult and pediatric studies. Figure 2 consists in a visual summary of the topics/variables touched upon by each study, taking the form of a "heatmap." An in-depth exploration of all variables of interest measured across all included studies is provided in Supplementary Table S3.

. . Publication identification and study characteristics
The following section is summarized in

. Discussion
In the last years, there has been a steady rise in interest in the use of cannabis in epilepsy, with many studies reporting that CBD extracts may provide beneficial effects for seizure control (2,3). In several countries, medical, pharmaceutical-grade cannabis may be prescribed by health professionals for seizure control in specific populations of PWE (2). Notwithstanding, personal experience and anecdotal accounts support the notion that a significant portion of PWE acquire cannabis by other means and that many of these PWE use cannabis in the hopes that it may diminish their seizure frequency and severity. This phenomenon is compounded by the fact that recreational cannabis itself is also being legalized in many parts of the world, and cannabis may therefore be more easily accessible in these areas regardless of its medical or recreational use (2). In addition, where cannabis is still bought from illicit vendors, one may question the quality and components of these cannabis products, as their production and distribution are not standardized (85). In this scoping review, we summarized what is known in the literature about the use of NMC (i.e., cannabis not obtained by prescription nor through special access programs) in PWE, focusing primarily on the lived experience, habits, and beliefs of these people regarding NMC. In the following sections, we will further analyze the literature and highlight knowledge gaps that may benefit from more research.
The evidence surrounding NMC use in PWE can be characterized as somewhat sparse and heterogeneous, with studies presenting varying samples, diverse aims, mostly low sample sizes, and differing outcome measures. The literature can naturally be divided into studies with adult samples and those with pediatric samples. Few relevant studies were published before the year 2000 (30,33,58,59,67,68); no relevant pediatric studies were published before 2010, which may reflect how interest in cannabis use in children is more recent than in adults. Western countries were predominantly represented in the literature (10, 11, 21, 23, 25-34, 36-46, 48-54, 58-66, 68-75, 77-81, 84), and very little is known about NMC use in PWE from African, Middle Eastern, Central/Southern American, and Asian countries. Efforts should be made to obtain more data on the subject from non-Western countries. Regarding sample size, only two studies, which were nationwide patient registry studies (that unfortunately shared to some extent the same sample), had sample sizes in the 100,000s (50, 61). Otherwise, two studies had sample sizes of ∼900 (75,83), and every other study had samples sized at <500 individuals. Oftentimes, PWE who used NMC only consisted in a subset of the overall sample and were therefore in even lower numbers. The majority of studies were cross-sectional by design (10, 11, 21 (35,45,65,70,74). Although cross-sectional studies appear to be a quick, cost-efficient way of evaluating PWE's experience with NMC, they may also be plagued by significant selection biases and response biases, especially when considering that studies may have been conducted in locations or in time periods where/when cannabis use was more stigmatized. Longitudinal or qualitative studies may provide different types of information whilst subverting some of the biases inherent to cannabis surveys, and more studies using these designs should be performed.
In terms of demographics, there seems to be a male preponderance for NMC use in adults with epilepsy (10, 11, 22,30,38,39,48,50,57,58,60,61,63,71,76,78,80). Such a finding, if confirmed, would probably reflect a general preponderance that men have for cannabis use vs. women (86). There are also slightly more cases of epilepsy in men than in women (87). Similarly, several adult studies suggested that a younger age may be associated with cannabis use in epilepsy (10,11,33,38,50,53,61,78), though this may once more reflect an association that is independent of epilepsy itself (88). Interestingly, there is no literature focusing specifically on NMC use in elderly PWE. This constitutes a major knowledge gap, especially given the high prevalence of epilepsy in the elderly and the effects of cannabis on cognition (89, 90). In addition, little is known about NMC use in non-Caucasian PWE, as almost all studies that reported their sample's ethnicity reported a Caucasian predominance (25,41,45,48,50,61,(70)(71)(72). The education .

FIGURE
Heatmap of topics discussed in each included study. This figure consists of a heatmap summarizing which topics were discussed in each included study. The x-axis represents the studies that were included in this scoping review, as identified with the first author. The y-axis represents all the topics that we searched for in each study (see Section . . Data extraction for more information). A black square at the junction between a study (x-axis) and a topic (y-axis) indicates that that topic was discussed in that study. A light gray square at that same junction would indicate that that topic was not discussed in that study. The last column represents the sum of the times each topic was discussed by all authors. In this last column, a square takes on a darker shade of gray as the topic is more "popular" (i.e., it was discussed in more publications). For instance, the perceived e ect of seizures was a topic that was discussed in many studies, whereas the age at first use of cannabis was rarely discussed. ASM, anti-seizure medication; NMC, non-medical cannabis; PWE, people with epilepsy.
To confirm these associations would require larger, more robust epidemiological studies.
Regarding the age at seizure onset, the type of epilepsy (focalvs. generalized-onset), and the epilepsy syndromes at play, these data should always be disclosed in future research, mostly for better between-study comparability. We note no pattern from the literature regarding these factors, other than the fact that pediatric studies often presented samples of individuals with epileptic encephalopathies (21,29,40,45,51,65,69,77,84). No study specifically mentioned if children with tuberous sclerosis complex were included. How NMC use in individuals with epileptic encephalopathies evolves through time and particularly through adulthood remains poorly understood. Finally, data on ASM use was heterogeneously reported; how many failed ASMs were used before NMC was begun, which ASMs are most often taken with NMC, and how does the use of other ASMs evolve once NMC is begun all remain open questions. More data on ASMs would also allow for a better evaluation of the proportion of people with pharmacoresistant vs. pharmacosensitive epilepsy using NMC.
The prevalence of NMC use in PWE was extremely variable and depended upon the study populations at hand and methodological factors. Most studies reported lifetime prevalence of NMC use in PWE grossly between 10 and 40% (25,31,33,38,43,44,49,53,54,59,66,67,75,76,(78)(79)(80)82). The two largest studies (with sample sizes several 100,000s larger than every other study) reported prevalence of NMC use in PWE of <10%, though these were prevalence for "cannabis use disorder" in people hospitalized with epilepsy (50, 61). As a matter of fact, in many studies, the type of prevalence being reported (e.g., lifetime cannabis use, active cannabis use, cannabis use only after epilepsy diagnosis) was ambiguous. Future studies exploring the prevalence of NMC use in PWE should carefully distinguish the type of prevalence they are studying. Very few studies reported the prevalence of epilepsy in cannabis consumers, and this may represent an avenue for future research (26,27,37). The consequences of NMC use on mental, physical, and social spheres were rarely explored, though recurrent themes included feelings of being stigmatized (32, 43) and depressed mood (41,80). Given the prevalence of mood disorders in PWE, how cannabis consumption interacts with this association would be important to clarify (91,92).

FIGURE
Lifetime prevalence of NMC use in PWE. This barchart presents the lifetime prevalence of NMC use in PWE for each publication for which this information was available. Gray bars represent prevalence calculated from adult samples, whereas orange bars represent prevalence calculated from pediatric samples. Confidence intervals are provided, as calculated using an α-error of . .
Interestingly, though some studies used self-reported outcome scales to measure mood and quality of life (41,45,71,72,80), no study has used formal neuropsychological tests to evaluate the impact of NMC on PWE. Measures of dependence were almost never addressed in the literature and would be important to evaluate in future studies. The main type of NMC used varied between pediatric and adult studies; children appeared to mostly use oral CBD extracts (21, 31, 40, 45, 47, 51, 65, 69, 70, 72-74, 77, 81), whereas adults used NMC with various CBD-THC compositions (10, 39, 62, 63, 76). The route of administration in adults was also much more varied, although smoking seemed to be a recurrent, predominant route (10, 11, 30,39,41,48,71,76,80). This difference in NMC use between age groups can probably be explained by the fact that children used NMC primarily for seizure control (34,35,44,51,64,70,74,82), whereas adults often used NMC for recreational purposes as well (10,11,32,36,38,39,41,57,66,71,75,79). No clear correlation between NMC type, NMC administration route, and goals of consumption has been established yet, and this potential association could benefit from confirmatory studies. The sources of NMC acquisition included both legal and illicit routes and were highly variable (10, 11, 29,32,34,35,41,44,45,48,57,69,[75][76][77]. More research on the importance and implications of illicit cannabis vendors in PWE should be done, especially given the potential variability in the quality of the cannabis sold by these vendors (85). Information on NMC dosage, frequency of consumption, and pricing was sparse and heterogeneous. The age at which PWE began consuming NMC was rarely explored, representing another significant knowledge gap. In the case of PWE consuming NMC for seizure control, it would be interesting to investigate if they had a history of recreational cannabis use before beginning NMC for seizure control. Ultimately, more research will simply need to be conducted which systematically evaluates the habits of PWE regarding their NMC use. Care will need to be taken to present data in a manner that can be readily compared between studies.
PWE receive information on NMC from various sources, many non-medical. These sources included the internet, social media (e.g., Facebook), traditional media (e.g., TV broadcasts), and friends and family (24,28,32,33,35,41,44,51,57,58,70,74). On a public health standpoint, further investigation of PWE's sources of information-especially social media where information may be largely unregulated (93) This scoping review featured some limitations. Firstly, given the sheer scope of this study, decisions had to be made to not explore certain variables that may have been interesting to explore. For example, the concomitant/past use of ketogenic diet, vagus nerve stimulation, and epilepsy surgeries could have been searched for in this review. Likewise, though this scoping review focused exclusively on PWE who used NMC, other populations could have been studied. For example, in the screening phase of this review, many studies focusing on healthcare professionals' experience with PWE using NMC were excluded. A future review tackling this subset of studies could yield interesting findings. Secondly, certain variables, such as opinions on regulatory policies, were in hindsight a bit vague and yielded wide-ranging results. Separating these variables into sub-topics would perhaps have enhanced the between-study comparability at the expense of feasibility. Finally, some may argue that the inclusion of case reports/series as primary sources of evidence renders our final interpretations less robust; however, the inclusion of these studies also allowed for a more comprehensive portrayal of the literature.

. Conclusion
In this scoping review, we summarized the body of literature surrounding NMC use in PWE, with emphasis on PWE's experience, habits, and beliefs concerning NMC. Overall, the literature can be summarized as heterogeneous and mostly populated by small, cross-sectional studies for which selection, participation, and confirmation biases are of concern. Though some topics are better explored than others, significant knowledge gaps exist throughout the literature and include (but are not limited to) the use of NMC in PWE from non-Western countries, the patterns of use of other ASMs in concomitance with NMC, NMC consumption in elderly PWE, the NMC consumption patterns of PWE (including the influence of illicit vendors), the barriers to the disclosure of NMC use to physicians, the comorbidities and characteristics of consumers vs. non-consumers, and the neuropsychological effects of NMC consumption in PWE. We hope our work may encourage researchers to address these gaps through high-quality research. We equally hope to encourage researchers to explore the landscape of NMC use in other neurological conditions, such as multiple sclerosis, Alzheimer's disease, and amyotrophic lateral sclerosis.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Author contributions
JL: study design, data collection, data analysis, and drafting. CA: data collection and data analysis. DT: data collection and manuscript revision. DC and CD: manuscript revision. DJ-A and MK: study design and manuscript revision. DN: study design, manuscript revision, and principal investigator. All authors contributed to the article and approved the submitted version.

Conflict of interest
DJ-A reports research grants for investigator-initiated studies from government entities (Ministère de la Santé et des Services Sociaux du Québec, Health Canada, Canadian Institutes of Health Research, Fonds de Recherche Québec -Santé). DJ-A receives payment for expert testimony at the Ministère de la justice du Québec. DJ-A is a board member for Grand Chemin and the Mental Health Commission of Canada. DJ-A has received study materials for clinical trials from Cardiol Rx and Tetra Biopharma for clinical trials funded by the Quebec government. DJ-A's salary was supported by the Fonds de Recherche Québec-Santé. MK reports unrestricted educational grants from UCB and Eisai, research grants for investigator-initiated studies from UCB and Eisai as well as from government entities (Canadian Institutes of Health Research, Fonds de Recherche Québec-Santé), academic institutions (Centre Hospitalier de l'Université de Montréal), and foundations (TD Bank, TSC Alliance, Savoy Foundation, Quebec Bio-Imaging Network). MK's salary was supported by the Fonds de Recherche Québec-Santé. DN reports unrestricted educational grants from UCB, Eisai, Liva Nova, Pendopharm, Paladin Labs, and Sunovion; research grants for investigator-initiated studies from UCB and Eisai as well as from government entities (Canadian Institutes of Health Research, Fonds de Recherche Québec-Santé, Natural Sciences and Engineering Research Council), academic institutions (Centre Hospitalier de l'Université de Montréal), and foundations (TD Bank, Savoy Foundation, Quebec Bio-Imaging Network). DN's salary was supported by Canada Research Chair Program.
The remaining authors declare that the research was conducted in the absence of any commercial or financial .
relationships that could be construed as a potential conflict of interest.

Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.