In this study using VigiBase^® records, nearly 10,000 safety reports related to cannabis and cannabinoids were included over five years, originating from 44 countries worldwide. The trend in cannabis- or cannabinoid-related ICSRs observed in VigiBase^® between 2020 and 2024 cannot be solely explained by overall VigiBase^® reporting patterns. Indeed, the proportion of cannabis ICSRs relative to total ICSRs increased from 2020 to 2022 before declining in 2023–2024, indicating a substance-specific pattern. The observed rise in reports may partly mirror global trends in cannabis use, which moved from an estimated 228 million users in 2002 (the highest estimate to that date) to 244 million users in 2023 [1, 23]. These observations reflect post-COVID market recovery and policy changes in many countries [1, 24]. Increasing popularity of cannabinoids and particularly CBD products, including for wellbeing purposes, may also contribute to the enhanced acceptability of recreational cannabis with impacts on global reported suspected ADRs [25].

The majority of ICSRs included in this study originated from France (47.0%) and the USA (37.8%), with smaller contributions from Germany, Greece, and the UK. France is a small country with some of the strictest cannabis regulations; its prominence as the leading source of cannabis ICSRs worldwide likely reflects its medical cannabis pilot programme and a well-established Addictovigilance system. The latter is based on the systematic recording of suspected ADRs related to substances with an abuse potential and may thus provide insights not captured through conventional pharmacovigilance [26–29]. In the USA, high reporting reflects long-standing pharmacovigilance practices as this country accounted for 36.1% of the total ICSRs recorded between 2020 and 2024, well above the UK, the leading contributor in Europe, which accounted for 5%. These findings underscore that geographic differences are influenced by national safety monitoring and pharmacovigilance systems, legal frameworks, and healthcare practices. In Europe, reporting is strongly encouraged for healthcare professionals and consumers, or even mandatory for all suspected ADRs in certain Member States.

Differences across countries are also influenced by local regulatory and market changes, along with evolving consumption patterns. In the USA, there was a marked decline in Cannabis sativa reports and a concomitant increase in CBD-related reports between 2020 and 2024 compared with 2015–2019. Concurrently, the use of CBD for self-reported health conditions has been increasing in North America [30]. In 2024, the global consumer health CBD market was valued at USD 20.9 billion, with North America leading the market at 61.1%, followed by Europe at 17.5% [31]. In Europe, France showed a relative increase in CBD or dronabinol reports; however, reports associated with Cannabis sativa remained predominant. In Greece, all included ICSRs were recorded in 2024, the vast majority of which were associated with THC:CBD. In this country where cannabis is illegal for recreational use, the use of cannabis-based medicines for a defined list of indications has been regulated from March 2018; however, these drugs were scarcely available until 2024 [32]. In Germany, the regulatory status of cannabis has markedly evolved, with the authorisation of cannabis for medical use in 2017, followed by that for recreational use in 2024 [33]. The proportion of ICSRs related to THC:CBD has decreased in Germany compared with 2015–2019, while the proportion of ICSRs related to CBD has increased. An even more pronounced trend was observed in Italy, with a sharp rise in CBD-related reports and a corresponding decline in THC:CBD-related reports. In Italy, non-prison penalties regarding cannabis have initially been increased before being decreased: both strategies were followed by a similar rise in the prevalence of cannabis use among adults aged 15–34 years [4, 34]. Based on these findings, reports of suspected ADRs may serve as indirect indicators of access to and use of cannabis and cannabinoids.

Cannabis sativa remained the most frequently reported substance, followed by CBD and THC:CBD combinations. Importantly, cannabis extracts reported during the French pilot programme were recorded as Cannabis sativa in VigiBase^®, representing a total of 1477 ICSRs between March 2021 and March 2024 (i.e., 14.8% of total ICSRs and 31.5% of ICSRs from France) [29]. Overall, median age was 29 years, with predominance of males, consistent with known sex and age distributions among cannabis users [1]. Cannabidiol reports were associated with younger ages, and THC:CBD with older ages, likely reflecting the medical use of Epidiolex^®/Epidyolex^® for refractory epilepsy, and Sativex^® for spasticity related to multiple sclerosis [35, 36].

Psychiatric and nervous system disorders were the most frequent suspected ADRs. From a mechanistic perspective, THC acts as a partial agonist at central CB1 receptors, which are widely distributed in brain regions involved in cognition, emotion, and reward processing. This interaction modulates neurotransmitter systems and potentially contributes to psychiatric ADRs [37]. In contrast, CBD has indirect effects on the endocannabinoid system and other neurological targets, which may influence neuronal excitability and help explain the neurological adverse reactions observed [38]. Cannabis sativa was primarily associated with psychiatric disorders, including substance use disorders, suicidal ideation and completed suicide. This aligns with the well-known addictive potential of cannabis: indeed, approximately 42% of substance use disorders worldwide have been attributed to cannabis use [1, 37]. The relationship between cannabis use and suicidal behaviours has been recently confirmed in the meta-analysis by Maffre Maviel et al., despite the known implication of depression, established to be a partial confounder [39]. Nervous system disorders were mainly related to Cannabis sativa and CBD. Cannabis sativa-related reports included coma, amnesia, and altered consciousness, which correspond to the known effects of THC [37]. Seizures and epilepsy were most frequently associated with CBD, which was also frequently associated with the “drug ineffective” PT, inconsistent with previous results suggesting the effectiveness of highly purified CBD extracts across a broad range of epilepsy disorders [40]. Nonetheless, a detailed analysis of the French pharmacovigilance database identified the ineffectiveness of pharmaceutical CBD as the main risk factor for cannabinoid-related seizures [41]. Convulsions have been reported in preclinical studies following the administration of high dose of CBD (150–300 mg/kg/day) [38, 42]. Regarding vascular disorders, the use of cannabis has been associated with MACE, with a dose-dependent relationship [43]. These substance-specific mechanistic considerations complement broader questions regarding the role of the endocannabinoid system in the pathophysiology of neuropsychiatric, vascular, and other disorders, and the corresponding risks among individuals exposed to cannabis and cannabinoids [44].

Substance-specific patterns observed highlight the need for careful risk-benefit assessment prior to cannabis or cannabinoid use. Suspected ADRs reflect both medical and recreational contexts, with effect variability influenced by dose, formulation, and route of administration. In particular, the multifactorial complexity of cannabis use encompassing its various forms and contexts, which themselves depend on legal status and the availability or lack of cannabis-based medications or medical cannabis, as well as the interplay of many individual factors, makes it difficult to accurately assess its relationship with mental health [45]. The high proportion of serious cases, with a significant proportion of the “death” criterion, emphasises the importance of continued clinician awareness and pharmacovigilance follow-up. Global trends in CBD use, often for self-medication, may drive new patterns of suspected ADRs, particularly neurological disorders.

One of the main strengths of this study is that it provides a comprehensive global overview of suspected ADRs related to cannabis and cannabinoids, using data from VigiBase^®, the largest international pharmacovigilance database [46]. This large-scale approach allowed the identification of substance-specific patterns and temporal trends across diverse regions and regulatory contexts. However, several limitations inherent to spontaneous reporting systems must be acknowledged. First, despite the extensive geographic coverage of the WHO PIDM, data from VigiBase^® are not representative of the true incidence or distribution of ADRs at the population level. Differences in national pharmacovigilance practices, healthcare systems, and reporting behaviours, including underreporting, notoriety bias related to increased attention to cannabis and cannabinoids, and time-related reporting biases following changes in product availability or regulatory status, contribute to heterogeneity in the data’s quality, completeness, and consistency. The completeness score of ICSRs included in the present analysis ranged from 0.2 to 1. Furthermore, at the time of data extraction, no cases originating from Canada were identified in VigiBase^® in 2020–2024 (whereas it was the third-largest contributing country in 2015–2019), likely reflecting reporting delays rather than an absence of suspected adverse reactions. This may result in an underrepresentation of Canadian cases in the analysed sample. Second, in countries where both medical and recreational cannabis uses coexist, it is not possible to reliably distinguish the context of use in ICSRs [47]. Third, VigiBase^® does not capture patient-level information such as sociodemographic characteristics, medical history, comorbidities, concomitant drug use, or contextual details including history of use and the chronological sequence of events. This lack of detailed information limits the possibility of conducting robust pharmacological assessments. Consequently, no causality assessment was systematically performed by the researchers to determine whether cannabis or cannabinoids were responsible for the suspected adverse drug reactions reported, including serious outcomes such as death, and causality cannot be inferred from these data. However, causality is evaluated in certain national pharmacovigilance systems and is therefore incorporated into the corresponding VigiBase^® records. Also, we did not specifically analyse additional suspect and/or interacting drugs that may be co-reported with cannabis or cannabinoids in the included ICSRs. The sensitivity analysis of ICSRs involving cannabis or cannabinoids without co-exposures suggested different use patterns and a likely selection bias (toward therapeutic use). Due to this limitation, results from this complementary analysis are not detailed, but they informed the overall interpretation. Fourth, a single ICSR may include multiple mentions of the same cannabinoid, potentially leading to an overestimation of substance frequency. To address this, we used ICSRs as the unit of analysis to minimise the risk of inflated frequency estimates. Fifth, some products of interest in this study, such as synthetic cannabinoids, were hardly visible in the database. More specifically, the potential presence of unlabelled cannabinoids in cannabinoid products, including in CBD products, remain out of reach. Lastly, while children under 2 years of age were excluded, it is possible that cases of accidental paediatric ingestions in older children were included in the analysed sample, even though these cases were not the primary focus of the study. Notwithstanding these limitations, and particularly the variability in the level of causality assessment, VigiBase^® centralises suspected ADRs associated with the use of cannabis-based products from monitoring systems [46]. Taken together, these limitations highlight the need for cautious interpretation of the findings and underline the exploratory nature of pharmacovigilance-based analyses.