A meta-analysis of 23,317 individuals found adolescent cannabis use associated with 37% increased odds of depression and 3.46x increased odds of suicide attempt in young adulthood, after adjusting for baseline mental health.
Parents, teens, educators, policy makers, and anyone interested in the mental health effects of adolescent cannabis use.
23,317 people: adolescent cannabis use linked to 3.46x higher suicide attempt risk
The Backstory
In February 2019, a study landed in JAMA Psychiatry that gave the adolescent cannabis debate something it had been missing: a single, clear number. Not from a brain scan, not from a rat model, but from 23,317 real people followed through the exact years that matter — the transition from teenage cannabis use to adult mental health.
Gabriella Gobbi, a psychiatrist and neuroscientist at McGill University, had spent years watching the same pattern in her clinic: young adults with depression who'd started using cannabis as teenagers. The clinical intuition was strong. The data to support it was scattered across dozens of studies, each with different methods, different populations, different conclusions. Nobody had pulled it all together.
So she did. And the number that emerged — a 37% increase in depression risk — changed policy in her own province within a year.
The Researcher Behind the Numbers
Gabriella Gobbi trained as a psychiatrist at the Catholic University of Rome before earning a PhD in neuroscience at the University of Cagliari under Gianluigi Gessa, one of Italy's pioneering psychopharmacologists. She arrived at McGill University in 2001 and built the Neurobiological Psychiatry Unit — a lab that bridges bench science and clinical psychiatry.
Her interest in cannabis and mood wasn't academic abstraction. Working in McGill's Mood Disorder Clinic, she kept encountering the same patient profile: young adults in their twenties presenting with depression who had histories of heavy cannabis use beginning in adolescence. "I started studying the role of the endocannabinoid system and THC in mood regulation in 2002," she later explained, "after observing my patients with depression who had a history of cannabis consumption."
Her animal work confirmed the pattern — adolescent THC exposure produced depression-like behavior in rats. But animal models only go so far. To move the evidence from suggestive to actionable, she needed human data at scale.
Why a Meta-Analysis Was Needed
By the late 2010s, the cannabis-psychosis link had been studied extensively. Multiple meta-analyses had established a dose-dependent association between cannabis use and psychotic disorders, particularly with early-onset and high-potency use. But the connection between adolescent cannabis and depression — a far more common condition — remained surprisingly under-synthesized.
Individual longitudinal studies pointed in the same direction, but each had limitations: small samples, short follow-up periods, incomplete adjustment for confounders. Some found significant associations; others didn't. The signal was there, but it was noisy.
How They Did It
How Gobbi Built the Evidence Base
Systematic search
Searched five major databases — Medline, Embase, CINAHL, PsycInfo, and ProQuest Dissertations — from inception through January 2017. Cast the net wide: 3,142 articles screened.
Including dissertation databases was deliberate — unpublished findings reduce publication bias, the tendency for only positive results to get published
Strict inclusion criteria
Only longitudinal or prospective studies that measured cannabis use in adolescents under 18 at baseline and then assessed depression, anxiety, or suicidality in young adulthood (18-32). Required adjustment for baseline mental health.
The age boundaries mattered: the question was specifically about adolescent exposure and adult outcomes, not concurrent use
Winnowing
269 articles selected for full-text review. 35 survived detailed scrutiny. 11 studies comprising 23,317 individuals met all criteria for the quantitative analysis.
The dropout from 3,142 to 11 reflects how few studies had both the longitudinal design and the analytic rigor needed
Quality assessment
Used the Research Triangle Institute item bank on risk of bias. Two independent reviewers at every stage. Random-effects meta-analysis to account for between-study variation.
Random-effects is the conservative choice — it assumes studies are measuring somewhat different underlying effects and produces wider confidence intervals
Gobbi et al. (2019), JAMA Psychiatry 76(4):426-434
The Numbers
1.37x
the odds of developing depression in young adulthood for people who used cannabis as adolescents, compared to those who didn't. With a 95% confidence interval of 1.16 to 1.62 and zero heterogeneity (I2 = 0%) across seven studies, this is one of the most consistent findings in the adolescent cannabis literature.
For context, the odds ratio for developing lung cancer from secondhand smoke exposure is approximately 1.20-1.30. A 37% increase in odds is modest at the individual level but massive at the population level when the exposure is as common as teenage cannabis use.
Gobbi et al. (2019)
The anxiety finding was the surprise absence: OR 1.18 (95% CI: 0.84-1.67, I2 = 42%). Not statistically significant. Given how often cannabis and anxiety are discussed together, the lack of a clear adolescent-onset signal was unexpected.
But the suicidality findings were the ones that stopped people:
- Depression moderate
Population attributable risk: ~7.2%, translating to approximately 413,000 young adult depression cases in the US potentially attributable to adolescent cannabis exposure
OR 1.37 (95% CI: 1.16-1.62). Heterogeneity: I2 = 0%. Seven studies, 23,317 individuals. The most robust finding — consistent across every included study with no statistical heterogeneity.
- Suicidal Ideation high
Clinically significant: suicidal ideation is already a leading concern among young adults, and a 50% increase in odds from a highly prevalent exposure demands attention
OR 1.50 (95% CI: 1.11-2.03). Heterogeneity: I2 = 0%. Three studies. A 50% increase in odds of thinking about suicide — consistent across studies.
- Suicide Attempt high
The wide confidence interval (1.53-7.84) and high heterogeneity mean the true effect size is uncertain, but even the lower bound (53% increase) is concerning
OR 3.46 (95% CI: 1.53-7.84). Heterogeneity: I2 = 61.3%. Three studies. The strongest association but also the most variable across studies.
- Anxiety low
The null finding for anxiety was unexpected given the common clinical association between cannabis and anxiety disorders
OR 1.18 (95% CI: 0.84-1.67). Not statistically significant. Three studies. Moderate heterogeneity (I2 = 42%).
Gobbi et al. (2019), JAMA Psychiatry
The Population Math That Changed Policy
Gobbi's team didn't stop at effect sizes. They calculated the population attributable risk — the proportion of depression cases in young adults that could theoretically be prevented if adolescent cannabis exposure were eliminated.
~413,000
estimated young adult depression cases in the United States potentially attributable to adolescent cannabis exposure, based on a population attributable risk of 7.2%. In Canada: approximately 25,000 cases. In the United Kingdom: approximately 60,000 cases.
This is the distinction between individual risk and public health burden. An individual teenager who uses cannabis faces a modest 37% increase in depression risk. But because so many teenagers use cannabis, the absolute number of attributable cases is enormous. This is the same math that makes even small increases in common-disease risk matter more than large increases in rare-disease risk.
Gobbi et al. (2019)
This number had real-world consequences. In 2019, Gobbi testified before Quebec's parliamentary commission studying Bill No. 2, which proposed raising the province's minimum legal age for cannabis from 18 to 21. Her evidence was part of the case that carried the legislation. Quebec became — and remains — the only Canadian province with a legal cannabis age of 21 instead of 18 or 19.
A subsequent analysis confirmed the policy worked: the increase in cannabis use among 18-to-20-year-olds was 51% lower in Quebec than in other provinces after the age change.
What the Critics Said
The study's reception was overwhelmingly positive — it was published in JAMA Psychiatry, one of the top journals in the field, and has been cited hundreds of times. But legitimate criticisms sharpened the picture.
The data is outdated
In an Addiction commentary, McDonald et al. (2019) pointed out that the weighted average year of cannabis use measurement across the 11 studies was 1997. THC potency has risen from roughly 4% in the mid-1990s to 17% or higher by 2017. The studies captured a cannabis landscape that no longer exists.
This criticism cuts both ways. If the association exists with low-potency cannabis, higher-potency products might produce stronger effects — meaning the study may underestimate current risk. But it also means the specific odds ratios may not directly apply to today's cannabis products.
Confounding remains a problem
Not all included studies adjusted for other substance use (alcohol, tobacco), socioeconomic status, or adverse childhood experiences. Teenagers who use cannabis differ from those who don't in many ways that independently predict depression.
The included studies were required to adjust for baseline mental health — the most important confounder — and the sensitivity analysis excluding potentially overlapping cohorts produced a nearly identical OR of 1.40 (95% CI: 1.17-1.67). But perfect confounding control is impossible in observational research.
Reverse causation
A 2020 Quebec longitudinal study of 1,606 adolescents found that depression at baseline predicted subsequent cannabis use, raising the possibility that depressed teens self-medicate with cannabis rather than cannabis causing depression.
Gobbi's design partially addresses this by requiring adjustment for baseline depression. But the reverse pathway — that teenagers use cannabis because they're already developing depression — remains plausible and is not mutually exclusive with a causal effect of cannabis on mood.
The suicide attempt heterogeneity
The most alarming finding — the 3.46x increase in suicide attempt risk — had the highest heterogeneity (I2 = 61.3%), meaning the included studies disagreed substantially about the magnitude of the effect.
High heterogeneity doesn't invalidate the finding, but it means the precise number should be interpreted cautiously. The confidence interval (1.53-7.84) tells us the true effect is likely somewhere between a 53% and 684% increase — a very wide range.
McDonald et al. (2019) Addiction; Quebec Longitudinal Study (2020); Gobbi et al. sensitivity analyses
How This Fits With Other Adolescent Brain Research
Gobbi's meta-analysis is one piece of a converging body of evidence about the adolescent brain's vulnerability to cannabis.
Research Timeline
The Adolescent Cannabis Evidence Base
Meier et al. (Dunedin Study)
Silins et al. (Three Cohorts)
Scott et al. (Cognition Recovery)
Gobbi et al. (This Study)
Di Forti et al. (EU-GEI)
Albaugh et al. (ABCD/IMAGEN)
Multiple sources; see individual study pages
The pattern across these studies is remarkably consistent: the adolescent brain responds differently to cannabis than the adult brain, and the consequences manifest across cognitive, structural, and psychiatric dimensions. Gobbi's contribution was demonstrating that the psychiatric consequences — specifically depression and suicidality — follow the same developmental-vulnerability pattern.
What People Get Wrong
Myth vs. Reality
The Evidence
Evidence-based analysis
The Clinical Implications
For parents trying to navigate conversations about cannabis with their teenagers, Gobbi's findings offer something specific to say: the risk isn't hypothetical, it's quantified. Not as a scare tactic, but as information. A 37% increase in depression risk and a potential 3.5-fold increase in suicide attempt risk are numbers worth knowing, even with the caveats.
For teenagers and young adults who started using early and are now dealing with depression or suicidal thoughts, the study suggests a connection worth exploring with a mental health professional — while also offering the reassurance that the adult brain has more resilience than the adolescent brain.
For policy makers, the population-attributable-risk calculation was the game-changer. Individual risk is modest. Population burden is substantial. That's the math that moved Quebec's legal age from 18 to 21.
Key Takeaways
Association of Cannabis Use in Adolescence and Risk of Depression, Anxiety, and Suicidality in Young Adulthood
Gobbi G, Atkin T, Zytynski T, Wang S, Askari S, Boruff J, Ware M, Marmorstein N, Cipriani A, Dendukuri N, Mayo N (2019) · JAMA Psychiatry
Related Research
Key studies in this area
The Dunedin IQ Study
Meier et al. (2012)
The landmark longitudinal study linking adolescent cannabis to cognitive decline — from the same era of evidence about developmental vulnerability
Cannabis and educational attainment across three cohorts
Silins et al. (2014)
Three-cohort analysis showing dose-response relationship between teen cannabis use and educational, occupational, and suicide outcomes
The ABCD Study: Cortical thickness in adolescents
Albaugh et al. (2021)
Prospective neuroimaging evidence of accelerated prefrontal thinning in adolescent cannabis users — the structural mechanism that may underlie Gobbi's psychiatric findings
THC and CBD have opposite effects on the brain
Bhattacharyya et al. (2010)
THC's opposing effects to CBD on emotion-processing brain regions help explain why cannabis composition matters for mood outcomes
High-potency cannabis and psychosis (EU-GEI)
Di Forti et al. (2019)
The companion psychiatric risk study — while Gobbi focused on depression and suicidality, Di Forti documented the psychosis risk from early, heavy, high-potency use
Memory fully recovers after quitting
Scott et al. (2018)
Evidence for resilience: cognitive deficits from cannabis are largely reversible, suggesting the adult brain can recover from adolescent exposure
Does teenage cannabis use cause depression?
This meta-analysis found adolescent cannabis use associated with 37% higher odds of depression in young adulthood, after adjusting for pre-existing mental health. The longitudinal design strengthens the causal argument, but observational studies cannot definitively prove causation. Reverse causation (depressed teens self-medicating) and residual confounding remain possible explanations. The most honest answer: cannabis likely contributes to depression risk in vulnerable adolescents, but it is probably not the sole cause in most cases.
How alarming is the suicide attempt finding?
The 3.46-fold increase in suicide attempt risk is the strongest association found, but it comes with important caveats. Only three studies contributed to this estimate, and the heterogeneity was high (I2 = 61.3%), meaning the true effect size is uncertain. The confidence interval (1.53-7.84) means the actual risk increase could be anywhere from 53% to nearly 8-fold. Even the conservative lower bound is clinically concerning, but the headline number of 3.46x should be cited with appropriate uncertainty.
Does this apply to all forms of cannabis or just smoking?
The included studies primarily captured smoked cannabis from the 1990s and 2000s, at potencies far below today's products. The studies did not distinguish between different cannabinoid profiles, consumption methods, or potency levels. Whether CBD-rich products carry similar risks, or whether high-potency concentrates carry greater risks, remains unknown.
If I used cannabis as a teenager and now have depression, is it because of the cannabis?
It's impossible to know for any individual case. The study identifies a population-level association — among groups of people, those who used cannabis as teenagers had higher rates of depression. But depression has many causes, and many teenage cannabis users never develop it. If you're concerned about a connection in your own life, this is worth discussing with a mental health professional. The encouraging news from other research is that the adult brain shows significant capacity for recovery.
Why didn't the study find a significant link with anxiety?
The anxiety analysis included only three studies, limiting statistical power. The confidence interval (0.84-1.67) was wide enough to include both no effect and a substantial effect. Additionally, the relationship between cannabis and anxiety is known to be bidirectional and dose-dependent — low doses may reduce anxiety while high doses increase it — which could wash out the signal in a meta-analysis that doesn't distinguish dose levels.
What the researchers found
Adolescent cannabis use was associated with depression in young adulthood (OR 1.37, 95% CI: 1.16-1.62), suicidal ideation (OR 1.50, 95% CI: 1.11-2.03), and suicide attempt (OR 3.46, 95% CI: 1.53-7.84). The association with anxiety was not statistically significant (OR 1.18, 95% CI: 0.84-1.67).
Why it matters
Published in JAMA Psychiatry, this is one of the most rigorous analyses of the adolescent cannabis-depression link. The elevated suicide attempt risk is particularly concerning given the high prevalence of adolescent cannabis use worldwide.
The numbers in context
23,317 individuals across 11 studies. Depression OR: 1.37 (I2=0%). Suicidal ideation OR: 1.50 (I2=0%). Suicide attempt OR: 3.46 (I2=61.3%). Anxiety OR: 1.18 (not significant). All adjusted for baseline depression/anxiety/suicidality.
How the study worked
Systematic review and meta-analysis searching five databases (Medline, Embase, CINAHL, PsycInfo, Proquest) for longitudinal and prospective studies of cannabis use in adolescents under 18 with depression outcomes in young adults 18-32. Eleven studies comprising 23,317 individuals were included. Random-effects meta-analysis with quality assessment.
Who was studied
23,317 individuals from 11 longitudinal/prospective studies across multiple countries. Cannabis use measured in adolescence (under 18), mental health outcomes assessed in young adulthood (18-32).
What this study cannot tell us
Observational studies cannot prove causation. Residual confounding possible despite adjustment for baseline mental health. Suicide attempt association had high heterogeneity (I2=61.3%). Dose-response relationships not fully characterized across all studies.
How to read the evidence
Strong: meta-analysis of longitudinal studies published in JAMA Psychiatry with low heterogeneity for depression (I2=0%) and suicidal ideation (I2=0%). Funded by Canadian Institutes of Health Research.
When this study was published
Published in 2019 in JAMA Psychiatry. Included studies primarily measured cannabis use in the 1990s-2000s, when average THC potency was substantially lower than today.
The bigger picture
Even with individual-level risk being moderate, the high prevalence of adolescent cannabis use means a large absolute number of depression and suicide cases may be attributable to cannabis at the population level. This is a public health math problem, not just an individual risk assessment. The study directly contributed to Quebec raising its legal cannabis age from 18 to 21 — the only Canadian province to do so.
Questions still open
- Is there a threshold of adolescent cannabis use below which depression risk is not elevated? What biological mechanisms might explain the association with suicidality? Would these associations hold for CBD-only products? Do today's higher-potency products carry greater or different risks than the ~4% THC cannabis captured in these 1990s-era studies?
Common questions
Does teenage cannabis use cause depression?
How concerning is the suicide attempt finding?
Why didn't the study find a significant anxiety link?
Are today's higher-potency products more dangerous?
Read the original research
Association of Cannabis Use in Adolescence and Risk of Depression, Anxiety, and Suicidality in Young Adulthood: A Systematic Review and Meta-analysis.
JAMA psychiatry, 76(4), 426-434
Citation
Gobbi, Gabriella; Atkin, Tobias; Zytynski, Tomasz; Wang, Shouao; Askari, Sorayya; Boruff, Jill; Ware, Mark; Marmorstein, Naomi; Cipriani, Andrea; Dendukuri, Nandini; Mayo, Nancy. (2019). Association of Cannabis Use in Adolescence and Risk of Depression, Anxiety, and Suicidality in Young Adulthood: A Systematic Review and Meta-analysis.. JAMA psychiatry, 76(4), 426-434. https://doi.org/10.1001/jamapsychiatry.2018.4500
Explore the wider topic
- How to Talk to Your Teenager About Weed (Without Losing Them)
- My Kid's Friends Smoke Weed: A Parent's Guide
- I Smoke Weed But Don't Want My Kids To: Navigating the Hypocrisy
- Quitting Weed as a Parent: Why It Is the Hardest and Most Important Quit
- Quitting Weed as a Teen or Young Adult
- My Kid Is Smoking Weed: A Parent's Guide
- Using Weed Under 18: What It Does to Your Developing Brain
- Weed and Anxiety: The Paradox That Traps People
- Quitting Weed with Depression: What You Need to Know
- Weed and Suicidal Thoughts: What You Need to Know
- Cannabis and the Developing Brain: What Every Teenager (and Parent) Should Know