Adults with attention-deficit/hyperactivity disorder (ADHD) were markedly more likely to develop mild cognitive impairment or dementia after age 50 than adults without the diagnosis, according to a study of 187,341 U.S. adults published October 7, 2026, in JAMA Psychiatry. People with a higher genetic predisposition to ADHD also faced a higher risk, even without a diagnosis.
The peer-reviewed retrospective cohort study, led by researchers at the University of Pittsburgh, drew on electronic health records and genomic data from the National Institutes of Health’s All of Us Research Program. Participants were aged 50 or older at entry (mean age 56.4 years; 57% female) and were followed for a mean of 10.3 years. In all, 3,026 participants (1.6%) had an ADHD diagnosis, identified with a validated algorithm requiring multiple diagnosis or medication records on separate occasions.
Over follow-up, 6,462 participants (3.4%) developed cognitive impairment. Among those with ADHD, 6.7% did, compared with 3.4% of those without, or 8.96 versus 3.29 cases per 1,000 person-years. After adjustment for sex, race and ethnicity, APOE ε4 status and cohort entry year, ADHD was associated with a hazard ratio of 4.60 (95% CI, 3.99–5.29). The estimate fell to 3.08 (95% CI, 2.66–3.58) in a model that also accounted for physical and psychiatric comorbidities, education, smoking and health care use, and to 2.26 (95% CI, 1.75–2.92) in a matched comparison.
The genetic analysis pointed the same way. Participants in the top fifth of an ADHD polygenic risk score had a 35% higher risk of cognitive impairment than those in the bottom fifth (hazard ratio 1.35; 95% CI, 1.25–1.46). To test whether early dementia might be driving ADHD diagnoses rather than the reverse, the authors required a gap between the two diagnoses; the association held with gaps of up to eight years but was no longer significant at nine, by which point 74% of cases had been reclassified, reducing statistical power.
Among people who went on to develop cognitive impairment, those with ADHD reported poorer mental health, lower quality of life and lower social satisfaction, and had 34% more outpatient visits in the following year. The All of Us program is supported by the NIH Office of the Director, and the study authors reported grants from the National Institute on Aging and the Alzheimer’s Association; the funders had no role in the analysis. One author disclosed a range of pharmaceutical industry relationships.
Why it matters
The findings replicate a pattern reported in other countries, this time in a large and diverse U.S. sample, and add genetic evidence. “This doesn’t prove that ADHD causes dementia, but seeing the same pattern show up independently in medical records gives us greater confidence that the association is not simply an artifact of how ADHD gets diagnosed,” senior author Tharick A. Pascoal, MD, PhD, associate professor of psychiatry and neurology at the University of Pittsburgh, said in a university statement carried by News-Medical.
The authors suggest the results support cognitive monitoring and integrated care for older adults with ADHD. Co-author Douglas Teixeira Leffa, MD, PhD, who recently completed his psychiatry residency at UPMC, told Healio that most people with ADHD in the study did not develop cognitive impairment and that the findings do not show ADHD causes dementia, but that clinicians should be alert to meaningful changes in memory or thinking “rather than attributing new difficulties to longstanding ADHD.”
The caveats
This is an observational study, so it cannot establish cause, and the authors list several limitations. Residual confounding and reverse causation from the early phase of dementia cannot be excluded. People with ADHD may have more contact with health care, raising the chance that cognitive problems are detected. Diagnoses taken from electronic records may be under-ascertained or misclassified, and record-keeping for most participants began only in midlife, so the age of first recorded ADHD diagnosis may not reflect true onset.
The polygenic score was derived mainly from people of European ancestry, limiting its applicability to others, although associations were also significant among participants of African ancestry. The study could not adequately evaluate whether ADHD treatment changes the risk, which the authors flag as a question for long-term studies. All of Us participants are volunteers, and the ADHD group was mostly female and more highly educated than in the broader literature, which may limit how far the results generalize.
This article is for general information and does not replace advice from a qualified clinician. Corrections: email the editor using the address on the Contact page.
