France’s Medical Cannabis Trial: 3% Serious Adverse Events in 3,164 Patients Over Three Years
By Dr. Benjamin Caplan, MD | Board-Certified Family Physician, CMO at CED Clinic | Evidence Watch
France’s three-year government-regulated medical cannabis experiment tracked safety in 3,164 patients and found that 3% experienced a serious adverse drug reaction, most commonly involving neurological, gastrointestinal, or psychiatric symptoms. While rare cardiovascular and suicidality signals emerged, this observational study lacks a comparator group, meaning these rates cannot be attributed solely to cannabis rather than to underlying disease burden.
France’s Medical Cannabis Trial: 3% Serious Adverse Events in 3,164 Patients Over Three Years
A national registry-based safety study finds neurological, gastrointestinal, and psychiatric reactions are the most common adverse drug reactions in France’s regulated medical cannabis experimentation, with rare but notable cardiovascular and suicidality signals warranting further investigation in comparative trials.
#72
Strong Clinical Relevance
One of the largest government-mandated pharmacovigilance datasets for medical cannabis, directly informing safety counseling and monitoring protocols in clinical practice.
Medical Cannabis Safety
Adverse Drug Reactions
European Regulation
Real-World Evidence
Clinicians who recommend or manage medical cannabis have long been working with an evidence base dominated by small trials and voluntary adverse event reporting. Government-mandated pharmacovigilance at national scale, with obligatory reporting from all prescribers and dispensers, provides a fundamentally different quality of safety data. The French RECANN registry represents one of the first European programs to generate this kind of structured, large-cohort real-world safety intelligence. For practitioners counseling patients on risks, these data offer a concrete reference point for safety discussions, even as their interpretation requires caution given the absence of a control group.
France’s medical cannabis experimentation, launched in March 2021, enrolled patients with treatment-refractory conditions across five approved indications: refractory neuropathic pain, certain forms of epilepsy, symptoms related to cancer or its treatment, palliative situations, and spasticity in multiple sclerosis. The RECANN registry, combined with the French Pharmacovigilance and Addictovigilance Database, served as the mandatory safety reporting infrastructure. Between March 2021 and March 2024, 3,164 patients were enrolled and received standardized pharmaceutical-grade cannabis products (oils and dried flower) under supervised titration. Mandatory reporting from all participating healthcare professionals distinguishes this dataset structurally from voluntary reporting systems, though underreporting remains possible even in mandated frameworks.
Over three years, 1,186 patients (37.5%) experienced at least one adverse drug reaction, and 81 patients (3%) experienced at least one serious adverse drug reaction. The most common categories were nervous system disorders (37.2% of all ADRs), gastrointestinal disorders (16.9%), and psychiatric disorders (15.2%), a profile broadly consistent with known cannabinoid pharmacology. Among serious events, six cases of acute coronary syndrome were notable because ACS was a prespecified enrollment contraindication, raising questions about screening adequacy. Eight cases of suicidal ideation or attempt were reported, with half occurring in patients without documented prior psychiatric history. Addiction-related signals were strikingly rare, with only 12 addictovigilance reports analyzed and one confirmed case of medical cannabis use disorder, though the authors acknowledge that systematic addiction history collection was legally precluded during the study. The authors emphasize that this descriptive, non-comparative design cannot establish whether the observed ADR rates represent excess risk attributable to cannabis or reflect the underlying disease burden of this complex, treatment-refractory population.
This is exactly the kind of systematic safety surveillance that the medical cannabis field needs more of. A mandatory reporting framework at national scale, covering more than three thousand patients over three years, is a real contribution. The 3% serious ADR rate feels clinically plausible for this population, but without a comparator group, we simply cannot say how much of that risk belongs to cannabis versus the underlying conditions that made these patients eligible in the first place. That is the fundamental gap. The cardiovascular and suicidality signals deserve focused follow-up, and the rarity of addiction-related events is noteworthy, though I would caution against reading too much into that given the legal barriers to systematic addiction screening in this cohort.
In my own practice, these data reinforce what I already do: careful cardiovascular screening before initiation, routine mental health check-ins during titration, and honest conversations about the range of possible side effects. I pay particular attention to patients without psychiatric history who are starting cannabis for pain or spasticity, because the suicidality data from this study suggest that assumed low-risk patients may still warrant monitoring. What I take from this report is not a change in my protocols but a validation of their necessity, and better language for discussing safety with patients who ask me directly about risk.
This study sits at an early but critical position in the research arc for medical cannabis safety. We have moved past the era where safety data were limited to recreational-use epidemiology or small clinical trials with narrow populations. What the RECANN registry offers is a regulatory-grade safety signal catalog from a structured, government-supervised program. However, it remains descriptive: it can generate hypotheses about cardiovascular, psychiatric, and neurological risk but cannot test them. The natural next step is a comparative pharmacovigilance analysis, ideally against matched cohorts receiving standard-of-care treatments for the same refractory conditions.
From a pharmacological standpoint, the predominance of nervous system and gastrointestinal ADRs is entirely consistent with CB1 and CB2 receptor pharmacology and does not represent a novel signal. The cardiovascular findings demand closer attention. Six acute coronary syndrome events in a population where ACS was an exclusion criterion suggest either incomplete screening or a genuine cannabis-attributable cardiovascular risk that screening alone cannot eliminate. For clinicians managing
Have thoughts on this? Share it:
