Cannabis and Psychosis Link Gets Stronger, Not Weaker, With Age

For decades, researchers have warned that cannabis use is tied to psychosis. A sweeping new analysis of more than 70,000 people reveals something far more unsettling: the link doesn’t weaken with age. It grows stronger. People who use cannabis and develop psychosis later in life are losing far more years to early onset than younger users, upending a widely held assumption about who is most at risk.

Psychosis is a condition in which a person loses touch with reality, experiencing hallucinations, delusions, or severely disorganized thinking. It is a hallmark of disorders like schizophrenia. Researchers have long suspected that cannabis could push forward the clock on when these symptoms first appear, but the picture of exactly who is affected, and by how much, has remained blurry. This new analysis sharpens that picture considerably, and what it shows should prompt a serious conversation about public health warnings on cannabis.

Published in the journal Biological Psychiatry, the meta-analysis, a study that statistically combines results from many individual studies into a single, more powerful analysis, pooled data from 149 studies covering 18,272 cannabis users and 52,801 non-users. On average, cannabis users experienced the onset of psychosis 2.50 years earlier than non-users. That figure reaffirms prior research. The deeper result is the one that stands out: the older the age group studied, the bigger the gap between cannabis users and everyone else.

Cannabis and Psychosis: The Age Twist That Surprised Researchers

When the research team, based at the University of New South Wales in Sydney, Australia, sorted the study samples by their average age of psychosis onset, a clear pattern emerged. In samples where onset typically occurred before age 20, cannabis use showed no significant connection to earlier onset. That was unexpected, since much of the existing science has focused on the teenage brain as the primary target of cannabis-related risk.

In adult samples, the story flipped dramatically. In studies where the average onset age fell between 20 and 25, cannabis users developed psychosis about 1.60 years earlier. In samples with an average onset age of 25 to 30, that gap widened to 4.43 years. In studies where onset averaged 30 or older, cannabis users developed psychosis 6.30 years earlier than non-users.

Researchers weigh cumulative exposure as one possible explanation, though the analysis itself does not prove it. Because cannabis use typically begins in adolescence, someone who develops psychosis at 35 has, on average, likely been using it longer than someone who develops it at 19. Under that hypothesis, longer exposure could gradually lower the brain’s resistance to psychosis, which would fit the widening gap in age groups where onset naturally happens later in life. Notably, the study did not find a statistically significant difference between heavier and lighter users, so this remains an idea the authors discuss rather than something the data settle.

Why Teenagers Didn’t Show the Cannabis-Psychosis Link

Perhaps the most counterintuitive result is the absence of a significant link in adolescent-onset cases. Many scientists have focused on teenagers because the brain is still developing during those years, and cannabis is thought to potentially interfere with that process. So why didn’t the data show cannabis pushing psychosis onset earlier in that group?

A few possibilities are on the table. Psychosis before age 20 may be driven mostly by factors like genetic predisposition or lower cognitive ability before symptoms ever appear, leaving less room for cannabis to make a measurable difference. Another is a floor effect: psychosis before the mid-teens is so rare that there may not be much statistical room for cannabis to accelerate it further.

What the researchers do not claim is that cannabis is therefore safe for teenagers. Absence of evidence is not evidence of absence, and the study was not designed to measure all the ways cannabis might harm young developing brains.

How the Psychosis Study Was Conducted, and Why the Results Hold Up

Researchers searched three major medical databases for studies published through April 2025. From nearly 9,600 records initially identified, the researchers included 181 separate data samples drawn from 149 peer-reviewed studies. To qualify, a study had to report the average age at psychosis onset for both cannabis-using and non-using groups.

Steps were taken to avoid double-counting. When multiple studies drew from overlapping patient groups, they kept only the largest dataset. Removing five outlier samples barely moved the overall finding, dropping it by just 1%, a sign of how consistent the results were across the broader pool of data.

Researchers found no statistical evidence of publication bias, the tendency for journals to publish positive results more than negative ones. By one measure, it would take more than 5,700 additional studies with zero effect to erase the overall finding, which suggests the pattern is not simply a product of selective reporting.

Heavier cannabis users showed an earlier onset of about 2.88 years compared to non-users, while lighter users showed a gap of about 1.90 years. That difference between the two groups did not reach statistical significance, but the direction is consistent with the idea that more exposure may mean greater risk.

On gender, women-only samples showed a mean earlier onset of about 4.64 years among cannabis users, while men-only samples showed a gap of about 3.32 years. A separate analysis found that samples with a greater proportion of men in the cannabis-using group also tended to show a larger overall difference in onset age. The authors interpret this cautiously, suggesting it may reflect a known pattern in which men generally develop psychosis earlier than women, a pattern that could skew results when men are overrepresented among cannabis users. They flag it as a point that warrants further study rather than a settled conclusion.

Contrary to what some might expect, more recently published studies did not show larger gaps between cannabis users and non-users, despite widespread increases in cannabis potency in recent decades. According to the researchers, this null result may reflect limitations in how studies are designed and published, rather than a genuine absence of a potency effect.

Why the Cannabis-Psychosis Findings Matter

If cannabis use genuinely causes earlier psychosis onset rather than simply being correlated with it, the practical stakes are high. Developing psychosis years earlier can mean losing critical windows for education, career development, and building the social connections that help buffer against mental health decline. Earlier illness also tends to mean more time spent in the health care system. Psychotic disorders already account for a significant share of health care costs in developed countries, according to the paper.

Declaring a definitive cause-and-effect relationship remains out of reach, the authors note, because observational research cannot fully untangle correlation from causation. A person genetically predisposed to both cannabis use and psychosis might develop both for reasons unrelated to one causing the other. That possibility remains open.

Still, the size of this analysis and the overall consistency of the association across vastly different studies and countries point in the same direction. For people who carry some vulnerability to psychosis, the evidence increasingly indicates that cannabis use is not a neutral choice, and that the risk may follow them well past their teenage years.

Disclaimer: This article summarizes findings from a single peer-reviewed meta-analysis and is intended for general information, not medical advice. The research shows an association between cannabis use and earlier psychosis onset; it does not prove that cannabis causes psychosis, and results reflect group-level study data rather than individual risk. Anyone with concerns about cannabis use or mental health should consult a qualified healthcare professional.

---