This pharmacovigilance assessment of seizures occurring in subjects using cannabinoids showed that cases are observed both in a context of medical and recreational use, and that their reporting was rare among cases of cannabinoid‐related adverse reactions (3%). Among the whole sample of 127 case reports of seizures as adverse effects in subjects using cannabinoids, two‐thirds (67.7%) of individuals had at least one concomitant risk factor of seizures: 31.0% among those with medical use and 78.6% among those with recreational use.

Our results showed that a majority of reported cases of seizures were among those using cannabis recreationally (77%), as compared to medical use (23%). The number of patients treated with licenced cannabinoids is low in France (200 patients/year with Epidyolex, 3200 patients in the experimental context since 2021), explaining the low number of cases reported in the context of medical use in France. More cases of seizures related to recreational cannabis use were reported in relation to the high prevalence of cannabis use in France, which has among the highest number of cannabis users in Europe [20]. Moreover, in our study, the characteristics of cannabis‐related seizures were similar to those of the general population of recreational cannabis users, in terms of sex (81% in our study compared with 82% in the French population of cannabis users) and age (mean age of 32.0 in our study compared to 54% subjects under 35 years in the population of cannabis users in France) [20]. In a survey examining the reasons for self‐medicated use associated with self‐reported cannabis use, a majority of people aged 50 and over reported medical versus recreational use of cannabis (52% vs. 18%, respectively), while the younger 18–29 age group reported a majority of recreational versus medical use (18% and 50%, respectively) [32]. In our sample, 53% of seizure cases among medical users were children, compared to 4.1% among recreational users.

In contrast to studies conducted in other pharmacovigilance systems, our study showed a lower proportion of seizures after the cannabinoid use among all cannabinoid‐related reported adverse events cases in France (3%). A pharmacovigilance study of adverse reactions to Epydiolex recorded in the FDA's FAERS (Adverse Event Reporting System) database between 2018 and 2023 showed that one third of reported ADRs were related to seizures [33]. In Europe, including French data, similar results were observed in the Eudravigilance database, where seizures accounted for 25.5% of reports with licensed cannabinoid and 19% with recreational cannabis [5, 6]. First of all, a majority of seizure reports occurred among persons with a history of epilepsy (51.2%). The low proportion of seizure cases among all cannabinoid‐related reports of adverse events, specifically in France, may be explained by a different reporter profile compared to other countries. Healthcare professionals are the main reporters in the French pharmacovigilance‐addictovigilance system [23, 34]. They may generally consider seizures as a symptom of the lack of treatment efficacy and thus not consider and report them as adverse reactions. In pharmacovigilance, the lack of therapeutic efficacy is part of the definition of ADRs. Thus, information on this suspicion should be collected and reported. Given the ambiguity for healthcare professionals of considering a seizure among patients treated for epilepsy as either a lack of treatment efficacy or as an adverse event, seizure events may be largely under‐reported, due to a lack of comprehensive knowledge of the pharmacovigilance definitions. Complicating healthcare professionals understanding of pharmacovigilance definitions is the fact that therapeutic ineffectiveness refers to the notion of efficacy (benefit) for a healthcare professional, and not to adverse reactions (risk). Another reason for the low number of reports may be that patients treated with cannabis use for medical purposes are under medical supervision, which may reduce the risk of adverse effects. We observed a relatively high number of reports in the addictovigilance data. A seizure is not perceived as the result of a cannabinoid dependence effect and is likely to be reported more frequently to the addictovigilance system.

Several factors may explain why cannabinoids and particularly CBD may be associated with seizure events. One of those factors may be related to the origins of seizures. Shorvon et al. proposed an etiologic classification of epilepsies in four categories: idiopathic, symptomatic, provoked and cryptogenic [35]. In our analysed reports, symptomatic and provoked categories may explain most of the cases. Indeed, symptomatic epilepsy is defined as an epilepsy of an acquired or genetic cause, associated with gross anatomic or pathologic abnormalities and/or clinical features, indicative of underlying disease or condition. In this category, seizures can be due to acquired causes such as brain infection, trauma, stroke or tumour, encountered in 9% of cases in our data. Similarly, a provoked seizure is defined as an epilepsy in which a specific systemic or environmental factor is the predominant cause of the seizures, and in which there are no gross causative neuroanatomic or neuropathologic changes. Most cases of provoked seizure are caused by fever, menstrual cycle, sleep–wake cycle, metabolic and endocrine; 51% of cases in our data were related to these factors. Cannabinoids, by their excitatory activity on the brain, could thus be the cause of provoked seizures. Other provoked seizures are considered as reflex epilepsies, induced by precipitating factors such as visual, auditive or somatosensory stimuli. In our study, most individuals had associated factors that could have triggered the onset of seizures. These factors differed according to the context of cannabinoid use. However, co‐use of epileptogenic medications was the most common factor in both contexts of use. Psychotropic drugs and morphine were common in both types of use, and opiate substitution drugs (methadone and buprenorphine) in recreational use. Medication‐related factors include blood–brain barrier penetration, target pharmacological activity on brain receptors/proteins, dose level, route of administration and lipid solubility [16]. Mechanisms of medication‐induced seizures could be impacts to metabolic pathways (e.g., ibuprofen, diclofenac and indomethacin), cell death due to neurotoxicity (e.g., cisplatin, methotrexate, interferon alpha, busulfan and 5‐fluorouracil) or altered function/expression of voltage gated ion channels or neurotransmitter interferences (e.g., penicillin G, carbapenems, cephalosporins, fluoroquinolones, clozapine, opioids and lidocaine).

Among recreational cannabis users in our sample, the use of other illicit drugs and alcohol was the second most common risk factors for seizures, after the use of epileptogenic medications. These findings are not surprising, given that individuals who use cannabis recreationally are also more likely to use other illicit substances and alcohol [12, 36, 37]. Illicit drugs associated with cannabinoid use, such as cocaine, amphetamines, ecstasy and LSD, are known to cause epileptic seizures. These users may not be sufficiently informed about the risks of seizures associated with illicit drugs and multiple drug use. In addition, THC itself is thought to be a proconvulsant, although this has not been clearly explained [10]. If this effect of THC is confirmed, it could also contribute to explaining why in our data more recreational cannabis‐related cases of seizure are reported than licensed cannabinoids and medical cannabis‐related cases.

Epilepsy is a neurological disorder characterised by an imbalance between the inhibitory neurotransmitter gamma‐aminobutyric acid (GABA) and the excitatory neurotransmitter glutamate. Cannabinoids can modulate GABA transmission, since endocannabinoid binding at presynaptic type 1 cannabinoid (CB1) receptors results in the suppression of GABA‐mediated inhibition from GABAergic neurons [38, 39]. CBD is thought to act as an antagonist or even a negative allosteric modulator of the CB1 receptor, explaining its use as anticonvulsant agent [40]. However, other findings suggest an absence of direct interaction of CBD at this receptor and rather point to an indirect action mediated by different other mechanisms [41]. The observation of seizure as adverse effect related to the medical use of cannabinoids, particularly CBD, indicated for treatment‐resistant epilepsy forms, is counterintuitive and hard to explain. Previous research showed that some patients with epilepsy treated with CBD may have an increase in the number of seizures, which is a paradoxical effect [42, 43] and is currently unexplained. It is suggested that the use of non‐pure CBD products (i.e., potentially containing other cannabinoids) may be linked to an increase in seizures compared to the use of the CBD product approved by the Food and Drug Administration (FDA), which does not contain other cannabinoids [44]. For example, in a study that assessed the effect of CBD‐enriched products (i.e., containing some amount of THC) for intractable paediatric epilepsy [44], a seizure aggravation was reported in 7% of the patients [44]. However, while contamination may contribute to explaining some cases related to some preparation types of CBD, it cannot explain cases related to Epydiolex, which contains purified CBD. Other hypotheses such as the interaction of CBD with other epileptogenic medications or factors should be explored for such cases. Further hypotheses suggest that the effect of CBD could depend on whether it is combined with other antiseizure medications or not. In a preclinical study that assessed the interaction between CBD and other antiseizure medications in a rat model of recurrent severe seizures, the combination of CBD with the antiseizure medications reduced the severity and prevalence of generalised seizures. In contrast, CBD alone reduces the seizure severity, but does not decrease the expression of generalised seizures [45]. Due to the magnitude of CBD's pharmacokinetic interaction, the combination of this drug with other antiseizure medications has been suggested to be associated with a proseizure effect in certain contexts [46]. CBD may also attenuate the anticonvulsant effect of levetiracetam due to drug–drug interaction [47]. However, the majority of preclinical studies suggest a potentiation of the antiseizure medication effect while combined with CBD [47, 48, 49, 50]. Indeed, the complex pharmacology of CBD includes direct and indirect interactions at CB1 receptors, and both CBD antagonism and potentiation of THC have been suggested [41].

The individuals in the study were also exposed to anti‐epileptic drugs, which may also lead to worsening of seizures, with an increase of the frequency or severity of existing seizures, emergence of new types of seizure or the occurrence of status epilepticus. Discrimination of drug‐induced worsening of seizures from spontaneous fluctuation or epilepsy evolution is not always easy. For example, epilepsies may be severely aggravated by inappropriate treatment due to an adverse interaction between the mode of action of the drug and the pathogenetic mechanisms underlying specific seizure types or syndromes. For example, seizure aggravation has been reported in Lennox–Gastaut syndrome with gabapentin, carbamazepine and intravenous clonazepam [51, 52]. Finally, other circumstances such as drug–drug interactions, under‐dosing and lack of treatment adherence may explain the recurrence of seizures.

The strengths of this pharmacovigilance assessment lie in the availability and the analyses of the narrative reports of cases. These narrative reports provide a more accurate clinical description of the circumstances of the occurrence of seizures after the cannabinoid use. The French pharmacovigilance system is one of the largest ADR reporters in the world, and the addictovigilance system is unique in Europe [23, 34]. Moreover, as in France, most reporters are healthcare professionals, the reported cases of seizures have a high degree of accuracy (i.e., validated seizure diagnosis).

Limitations of the study include the spontaneous nature of the reporting in the pharmacovigilance and addictovigilance system leading to possible under‐reporting. A likely important contributing factor to the under‐reporting of cannabis‐related seizures in these French data is their interpretation as resulting from a lack of efficacy rather than as a proper ADR. The level of investigation into the causes of seizures and the quality of the data collected may exclude some risk factors associated with cannabinoid‐related seizures. In particular, co‐exposure medication data, therapeutic indication or context of medication use are likely lacking, or heterogeneously recorded in these adverse reaction reporting databases. Nonetheless, even if the used data are incomplete, they are reported by healthcare professionals and investigated by pharmacologists in the pharmacovigilance and addictovigilance centres before being recorded into the database. Moreover, paradoxical adverse reactions, such as seizures after seizure treatment with CBD, have been shown to be difficult to detect in pharmacovigilance databases [53]. However, even if this possible under reporting of paradoxical adverse effects might be the case for CBD prescribed to treat epilepsy (i.e., medical use of CBD), there is no reason to expect such an underreporting bias for seizure cases with recreational cannabis. Finally, it is not possible to conclude if cannabinoids are the cause of seizures, as other risk factors of seizure were also present among the studied cases. The objective of the study was not to assess a causal relationship between seizure and cannabinoids, but to characterize the profile of cannabinoid users who have seizures, so that preventive measures can be taken.