Six Years of Legal Cannabis in Canada: What Adverse Event Reports Reveal—and What They Can’t
Six Years of Legal Cannabis in Canada: What Adverse Event Reports Reveal—and What They Can’t
A Health Canada analysis of 698 voluntary reports submitted over six years offers a signal-generating snapshot of cannabis safety patterns in the legal market, but the inherent limitations of spontaneous reporting systems mean it cannot establish causation, calculate true incidence rates, or represent the full population of Canadian cannabis users.
Why This Matters
Canada operates one of the world’s largest federally regulated cannabis markets, yet systematic post-market safety data for legal cannabis products have been remarkably scarce. As millions of Canadians use these products for medical and recreational purposes, clinicians and regulators need structured surveillance to detect emerging safety signals. This study represents the first multi-year descriptive summary of Canada’s formal adverse event reporting system for legal cannabis, establishing a baseline dataset at a moment when product diversity, consumer demographics, and clinical applications continue to expand rapidly.
Clinical Summary
Since cannabis legalization in October 2018, Health Canada’s Canada Vigilance Program has collected voluntary and mandatory adverse reaction reports for legal cannabis products. In this cross-sectional descriptive analysis published by Health Canada investigators, the authors examined all 698 unique case reports submitted to the Canada Vigilance Database between October 2018 and December 2024. The study is a pharmacovigilance exercise rather than a controlled epidemiological investigation: it aggregates case-level data from spontaneous reports without a comparison group, incidence denominators, or longitudinal follow-up. The rationale for systematic surveillance rests on the well-established pharmacology of cannabinoids, which act on widely distributed endocannabinoid receptors and carry dose-dependent risks for neuropsychiatric, cardiovascular, and gastrointestinal effects.
The most frequently coded adverse events were hallucination, headache, nausea, dizziness, and dyspnea. Cannabis extracts were implicated in 68.8% of all cases. A substantial majority (62.3%) were classified as serious, though the primary reason cited was “other medically important condition” rather than hospitalization or death. The reporting population skewed older (mean age 56 years) and toward self-reported medical users (67.5%), with pain management as the leading indication. Causality was assessed for 668 events, and most were rated only “possibly” associated with the reported cannabis product. The authors acknowledge that without denominator data on total cannabis consumers, true adverse event rates cannot be calculated, and that the 698 reports over six years almost certainly reflect substantial underreporting relative to millions of active users.
Dr. Caplan’s Take
This study does exactly what a pharmacovigilance summary should do: it describes the signal landscape without overclaiming. What strikes me is how clearly the reporting population diverges from the typical cannabis consumer. A mean age of 56, with two-thirds reporting medical use, tells us we are looking at the patients most connected to the healthcare system and therefore most likely to file formal reports. That is not a flaw in the study; it is a structural feature of spontaneous reporting. But it means we should be cautious about treating these 698 cases as representative of what is happening across the broader legal cannabis market.
In practice, when patients ask me whether legal cannabis products are safe, I explain that serious adverse events do occur and that extracts, which can deliver high cannabinoid doses in concentrated forms, appear disproportionately in reports. I use these data the way they are meant to be used: as a reason to start low, titrate slowly, monitor systematically, and remain vigilant for neuropsychiatric symptoms like hallucinations, especially in older patients or those new to cannabinoid therapy. This dataset reinforces the importance of structured follow-up rather than assumption of safety.
Clinical Perspective
This study sits squarely at the signal-generation stage of the research arc. It confirms what smaller case series and international pharmacovigilance datasets have suggested: that neuropsychiatric and gastrointestinal events dominate cannabis adverse event profiles, and that concentrated product forms may carry disproportionate reporting burden. However, it does not and cannot distinguish genuine risk concentration from reporting bias. The absence of statistical comparisons between THC-dominant and CBD-dominant product profiles means that observed differences in event frequency across cannabinoid classes remain hypothesis-generating observations rather than evidence-based distinctions. For patient-facing recommendations, this study supports ongoing caution but does not provide the evidence base needed to quantify risk or change prescribing protocols.
Clinicians should note the prominence of hallucination as the most frequently reported event, particularly in an older population with likely polypharmacy. Cannabis products, especially THC-dominant extracts, interact with CYP3A4 and CYP2C19 metabolic pathways and may potentiate or be potentiated by commonly prescribed medications including benzodiazepines, opioids, and certain antidepressants. The most actionable takeaway from this dataset is straightforward: when initiating cannabinoid therapy in patients over 50, particularly those using extracts, clinicians should implement structured adverse event monitoring at early follow-up visits rather than relying on patients to self-report problems through passive channels.
Study at a Glance
What Kind of Evidence Is This
This is a cross-sectional descriptive analysis of spontaneous pharmacovigilance reports, a form of regulatory surveillance that sits near the base of the evidence hierarchy. It has no comparison group, no denominator population, no randomization, and no longitudinal follow-up. The single most important inference constraint this design imposes is that it is structurally incapable of supporting causal conclusions or estimating incidence rates. Its value lies entirely in hypothesis generation and early signal detection, not in quantifying risk or establishing drug-event causality.
How This Fits With the Broader Literature
The event profile described here aligns with findings from other national pharmacovigilance systems, including the FDA’s FAERS database and the WHO’s VigiBase, both of which have identified neuropsychiatric symptoms, gastrointestinal complaints, and dizziness as common cannabis-associated reports. The overrepresentation of cannabis extracts is consistent with emerging concerns about concentrated product forms documented in U.S. poison control data and the Colorado Department of Public Health surveillance reports. What this study adds is a Canadian-specific regulatory baseline covering the entire first era of federal legalization, which can be tracked longitudinally as the market matures and product categories evolve.
Notably, the demographic skew toward older medical users echoes patterns seen in Canadian provincial surveys but diverges sharply from recreational use demographics captured in the Canadian Cannabis Survey, where younger adults predominate. This divergence reinforces the conclusion that the reporting population is not representative of the total consumer base.
Common Misreadings
The most likely overinterpretation is to treat these 698 reports as evidence that legal cannabis causes the adverse events described at meaningful population-level rates. In reality, the study cannot establish causation for any individual event, as most were rated only “possibly” associated with cannabis. Equally importantly, the absence of denominator data means that 698 reports over six years, drawn from a market serving millions of consumers, could represent a vanishingly small fraction of actual adverse events or a nearly complete capture of serious ones. The data simply cannot distinguish between these scenarios. A related misreading is to interpret the 62.3% serious case rate as evidence that most cannabis adverse events are serious; this figure almost certainly reflects a well-documented reporting bias in which serious events are far more likely to be submitted to voluntary surveillance systems than mild ones.
Bottom Line
This Health Canada pharmacovigilance summary provides the first comprehensive descriptive profile of adverse event reports for legal cannabis products in Canada over six years. It identifies plausible safety signals, particularly around neuropsychiatric events and cannabis extracts in older medical users, but it cannot establish causation, calculate incidence, or represent the broader consumer population. Its value is as a regulatory baseline and signal-generation tool. Clinical practice should incorporate structured adverse event monitoring, especially for extract users, while awaiting controlled studies that can quantify actual risk.
Frequently Asked Questions
Does this study prove that legal cannabis causes hallucinations or other serious side effects?
No. The study catalogues what people reported to Health Canada’s voluntary surveillance system, but it cannot prove that cannabis caused any specific event. Most events were assessed as only “possibly” associated with the cannabis product involved. Proving causation requires controlled studies with comparison groups, which this analysis does not include.
Why were only 698 reports collected when millions of Canadians use cannabis?
Spontaneous adverse event reporting systems consistently capture only a small fraction of actual events. Reporting is voluntary for consumers and healthcare professionals (though mandatory for licence holders), and most people who experience mild or moderate side effects never file a report. The low number confirms substantial underreporting and means these data cannot be used to estimate how common adverse events actually are.
Should I be concerned about cannabis extracts specifically?
Cannabis extracts accounted for 68.8% of reports in this dataset, which is notable but requires careful interpretation. Extracts can deliver higher cannabinoid doses per serving than dried flower, which could contribute to more pronounced effects. However, the disproportionate reporting may also reflect the fact that extract users in this dataset were predominantly medical consumers who are more engaged with the healthcare system and more likely to file formal reports. If you use cannabis extracts, starting with low doses and increasing gradually remains the most prudent approach.
Does this study apply to younger recreational cannabis users?
The reporting population had a mean age of 56 years and was predominantly composed of self-reported medical users. This demographic profile does not represent younger recreational consumers, who constitute a large segment of the Canadian cannabis market. The safety patterns observed in this dataset may not generalize to populations with different ages, product preferences, consumption patterns, and baseline health conditions.
What should I tell my doctor about this study?
Have thoughts on this? Share it:
