Cannabis use in middle-aged and older adults does not appear to accelerate cognitive decline or raise the risk of dementia, according to one of the largest studies yet published on the question — drawing on UK Biobank data alongside records from more than 220,000 US veterans.

The findings, published in BMJ Mental Health in February 2026, come as cannabis use rises among older people, driven by expanding medical access and changing social attitudes. Clinicians have long lacked robust evidence on what that means for memory and brain health in later life.

What the researchers did

Led by researchers at the University of Oxford, the team analysed two major cohorts. In UK Biobank, they compared up to 18,975 lifetime cannabis users with up to 60,598 non-users across five cognitive tests, including numeric memory and fluid intelligence, both at a single time point and again roughly seven years later.

In the US Million Veteran Program, they examined whether cannabis use disorder — identified through hospital diagnostic codes rather than self-report — was linked to new dementia diagnoses in more than 12,000 affected veterans, compared with controls across European and African ancestry groups.

To probe whether any associations might be causal rather than coincidental, the researchers also ran Mendelian randomisation analyses, using genetic variants linked to cannabis use as natural experiments.

No sign of faster decline or higher dementia risk

At baseline, cannabis users in UK Biobank scored slightly higher on two cognitive tests — numeric memory and fluid intelligence — than people who had never used the drug. But over follow-up, there was no significant difference in how scores changed over time.

In the veteran cohort, cannabis use disorder was not significantly associated with incident all-cause dementia. The hazard ratio in the European ancestry group was 1.11, with a confidence interval that crossed 1.0, meaning the result could plausibly reflect no real effect.

Genetic analyses likewise found no evidence that cannabis use or cannabis use disorder causally affects cognitive performance or dementia risk.

"Within the limits of these cohorts, we found no evidence that cannabis use contributes substantially to cognitive ageing or dementia risk in older adults," the authors concluded.

Why users scored higher at baseline

The researchers were clear that the cross-sectional cognitive advantage among cannabis users is unlikely to mean the drug improves brain function.

In UK Biobank, cannabis use is socially patterned: users tended to be younger, less deprived and more likely to hold a degree than non-users. Those factors are themselves strongly linked to better cognitive performance, and may not be fully captured even after statistical adjustment.

When longitudinal and genetic data failed to support a causal benefit — and no dose–response relationship emerged between frequency of use and test scores — the team concluded residual confounding was the more plausible explanation.

That distinction matters for public interpretation. A snapshot in which cannabis users appear sharper does not necessarily mean cannabis protected their brains; it may reflect who chooses to use, and who ends up in the study.

What clinicians are being told

In their clinical implications, the authors said occasional or prior cannabis use may not be a major driver of cognitive ageing in older adults, based on current observational and genetic evidence.

They stressed that this is not an endorsement of cannabis, particularly at higher doses or with prolonged use, where safety remains uncertain. Doctors should still ask about cannabis history and consider cognitive screening as part of routine care.

"Further longitudinal and mechanistic studies are needed to guide evidence-based advice for older adults considering cannabis for medical or recreational purposes," they wrote.

Important caveats

The study has limits that temper how far the null findings should travel. UK Biobank participants are healthier and more advantaged than the general population, and cannabis exposure was based on self-reported lifetime use rather than detailed dosing or THC potency.

The veteran analysis relied on diagnosed cannabis use disorder, which likely captures more severe use patterns than occasional recreational consumption. Dementia was identified from electronic health records, so timing of true disease onset remains uncertain.

Since publication, researchers writing to the journal have questioned whether survivor bias and practice effects from repeat cognitive testing could partly explain the null longitudinal results. The authors replied that Mendelian randomisation, which uses genetic variation fixed at conception, is less vulnerable to those biases — though they acknowledged UK Biobank's selected sample means conclusions apply most directly to older adults who remain engaged with follow-up.

Earlier work linking heavy adolescent cannabis use to midlife cognitive decline may also reflect different exposure patterns. This study primarily captures lifetime or later-life use in population cohorts, not persistent use from teenage years.

Relevance for UK readers

With medical cannabis prescriptions continuing to grow in Britain and debate over wider reform never far from the headlines, the study offers a nuanced data point rather than a green light.

For older adults already using cannabis — whether via prescription or otherwise — the research suggests no clear signal that past or moderate use is accelerating dementia risk. But it does not settle questions about high-potency products, daily long-term use, or use alongside other health conditions common in later life.

Reporting based on findings published in BMJ Mental Health, February 2026.