Surprise UK Study Finds No Link Between Cannabis Use and Cognitive Decline in Older Adults

Challenging Assumptions on Cannabis and Brain Health

A groundbreaking study from researchers at the University of Oxford has upended conventional wisdom about cannabis use in older adults. Published in BMJ Mental Health on February 25, 2026, the research reveals no association between lifetime cannabis consumption and accelerated cognitive decline or increased dementia risk. This finding comes from one of the largest analyses to date, drawing on data from the UK Biobank and the US Million Veteran Program.

The surprise element? Cannabis users in the study cohort actually performed modestly better on certain cognitive tests at baseline compared to non-users. Lead author Saba Ishrat, a DPhil candidate in Oxford's Department of Psychiatry, emphasized that this should not be misconstrued as a protective effect of cannabis. Instead, differences are largely attributable to users' higher socioeconomic status, education levels, and other demographic factors.

As cannabis use rises among those over 50—driven by medical prescriptions and shifting attitudes—this UK-led research provides crucial reassurance for policymakers, clinicians, and aging populations concerned about long-term brain health.

Unpacking the Study Design and Methodology

The Oxford team's work leveraged the UK Biobank (UKB), a vast repository of health data from over 500,000 participants recruited between 2006 and 2010, aged 40-69 at baseline. Nearly 19,000 ever-users of cannabis were compared to over 60,000 non-users across cross-sectional and longitudinal analyses. Cognitive function was rigorously assessed using five touchscreen-based tests covering numeric memory (attention and working memory), fluid intelligence (problem-solving and logic), trail making (visual search, processing speed, executive functions), symbol digit substitution (visual memory), and pairs matching (episodic visual memory, executive functions, processing speed).

Longitudinally, up to 15,841 users and 49,006 non-users were tracked over time with linear mixed-effects models adjusting for age, sex, income, deprivation, education, occupation, alcohol, smoking, body mass index (BMI), blood pressure, and assessment center. Complementing UKB data, the US Million Veteran Program (MVP)—with 12,222 cases of cannabis use disorder (CanUD)—examined dementia incidence via Cox proportional hazards models, controlling for similar confounders plus head injury, depression, and post-traumatic stress disorder (PTSD).

Mendelian randomization (MR) analyses used genome-wide association study (GWAS) SNPs for CanUD and lifetime cannabis use (CanLU) to probe causality, employing inverse-variance weighted (IVW), Egger regression, and weighted median/mode methods. This genetic approach helps disentangle correlation from causation by leveraging inherited variants as proxies for exposure.

Headline Results: Better Baseline Performance, No Decline

Cross-sectionally, UKB cannabis users outperformed non-users on numeric memory (beta=0.07, 95% CI 0.06-0.09, p<0.001) and fluid intelligence (beta=0.12, 95% CI 0.10-0.13, p<0.001), with trends in other domains fading after false discovery rate (FDR) correction. No dose-response emerged—all frequency levels (1-10 times vs. 11+ times lifetime) showed similar patterns. Males exhibited stronger fluid intelligence associations, but age interactions were absent.

Critically, longitudinally, no associations with cognitive change appeared (e.g., fluid intelligence beta=0.00155, 95% CI -0.00082 to 0.00393, p=0.20). In MVP, CanUD conferred no elevated dementia risk (European ancestry HR=1.11, 95% CI 0.97-1.26, p=0.12; African ancestry HR=0.97, 95% CI 0.79-1.20, p=0.80). MR yielded null results—no causal links from cannabis to cognition or dementia (e.g., CanUD IVW beta=0.004 for fluid intelligence, p=0.88), nor reverse causation.

Senior author Dr. Anya Topiwala, from Oxford Population Health, noted: "Across large UK and US cohorts... we found no evidence that cannabis use was associated with accelerated cognitive decline or increased dementia risk in older adults."

Explore the full BMJ Mental Health paper for detailed statistics and figures.

Context from Prior UK Biobank Research

This Oxford study builds on earlier UK Biobank analyses. A November 2025 paper in Age and Ageing by University of Haifa researchers (using UKB data from 67,713 adults aged 60+) found lifetime cannabis users (17%) outperformed non-users across five domains: attention (B=0.071), executive function (B=0.047), processing speed (B=0.363), visual memory (B=0.062), working memory (B=0.181)—all p<0.001. Former users showed slower executive decline longitudinally (B=0.008, p=0.030), though longer duration predicted faster processing speed loss.

Another 2026 UKB study (26,000 middle-aged/older adults) linked lifetime use to larger brain volumes in hippocampus, orbitofrontal cortex, and amygdala—regions dense in cannabinoid receptors—and better memory/attention, hinting at neuroprotective potential, though causality unproven.

These converge on no harm—and possible inverse associations—in older adults, contrasting adolescent-focused studies showing impairments from heavy use.

Photo by Iulia Topan on Unsplash

Why the Surprise? Reconciling with Earlier Concerns

Prior research often highlighted risks from adolescent or heavy chronic use: IQ drops (up to 8 points in Dunedin study), memory deficits, white matter changes. However, adult-onset or moderate use shows minimal persistent effects post-abstinence. Oxford's genetic MR addresses reverse causation (e.g., low cognition driving use) absent here.

Confounders like education explain baseline advantages: UKB users skew healthier, wealthier. Potency unmeasured—modern strains (20-30% THC vs. 4% historically) may differ, but recreational users dominate samples.

Cannabis Compounds and Brain Mechanisms

Cannabis contains over 100 cannabinoids; delta-9-tetrahydrocannabinol (THC, psychoactive) binds CB1 receptors in hippocampus/prefrontal cortex, modulating memory/executive function. Cannabidiol (CBD, non-psychoactive) may counter THC risks, promoting neuroprotection via anti-inflammation, BDNF upregulation.

Older adults metabolize THC slower (50% reduction in clearance), prolonging effects but risking falls/psychosis. Yet, low-dose medical cannabis aids chronic pain (prevalent in 50% over-65s), sleep, without evident cognitive toll per Oxford data.

View the Oxford press release for researcher insights.

Rising Prevalence Among UK Older Adults

Cannabis use in UK adults 16-59: 6.8% past-year (Crime Survey England & Wales, 2024). Over-65 lifetime use ~2-3%, but doubling since 2012 amid medical legalization (2018). Medical prescriptions surged 400% post-2018 for pain/nausea; 10,000+ patients by 2026, many seniors.

• Medical: NICE guidelines for chemotherapy nausea, MS spasticity; off-label for arthritis, anxiety.
• Recreational: 1.7% over-65 lifetime (older data), rising with destigmatization.

Implications for Policy and Clinical Practice

UK policy: Class B drug, medical specialist-prescribed. Oxford findings support expanded trials for geriatric conditions without cognitive fears. Clinicians: Screen use routinely, advise low-THC/CBD ratios, monitor interactions (e.g., warfarin).

Stakeholders: NORML hailed "contradicts stereotypes"; caution from authors on cardiovascular risks, use disorder (30% regular users). No Alzheimer's Society direct response, but prior notes no proven benefit/harm in dementia.

Actionable: Older patients—start low (2.5mg THC), titrate slow; vaporize over smoke to minimize lungs.

Photo by Oliver Sjoberg on Unsplash

Caveats, Limitations, and Future Directions

UKB bias: Healthier volunteers, self-report recall/potency gaps. MVP: Severe cases only. No heavy/recent use granularity. MR power limited by SNP effects.

Future: Dose-response trials, high-potency longitudinals, THC/CBD ratios. Oxford calls for detailed exposure studies in diverse cohorts.

Balanced view: No green light for unchecked use—psychosis, addiction risks persist.

Broader Impacts on Higher Education and Research

This Oxford-led work underscores UK universities' role in evidence-based policy amid cannabis debates. Funds like UKRI support Biobank analyses, training neuroscientists. Implications for psych/pharma courses: update curricula on adult vs. youth effects.

Explore UK Biobank for similar projects.