Reaching 90 without dementia does not close the question. In findings published on 4 August 2026 in The Lancet Healthy Longevity, Hilary L Colbeth and colleagues at UC Davis Health and Kaiser Permanente report dementia incidence in the LifeAfter90 cohort: more than 800 long-term Kaiser Permanente Northern California members, median age 92, all enrolled without dementia at age 90 or older and reassessed roughly every six months since 2018. Women in the cohort had about a two-fold higher rate of new dementia than men. Black participants had about a 75 per cent higher rate than Asian participants, and the authors report that Black and Hispanic participants had significantly higher incidence than white and Asian participants.

The genetics did not read the way they do in younger old age. Carrying APOE2, long associated with lower Alzheimer's risk, was associated with roughly 60 per cent lower dementia incidence after 90. APOE4, the strongest common genetic risk factor for late-onset Alzheimer's disease, was not associated with a meaningful increase across the cohort as a whole; in subgroup analyses it was associated with higher risk among men and with roughly double the risk among Black participants. "We saw evidence that APOE4 impacts males and females differently after age 90," Colbeth said in the university's account of the work.

The cohort was built precisely because this age band is thinly covered. Most of what is known about dementia risk comes from studies that begin at 65, in samples that are often overwhelmingly white. LifeAfter90 recruited across racial and ethnic groups from a health system whose records for some participants reach back to the 1960s, and diagnoses are made by in-person clinical evaluation, the method described in the cohort's published assessment protocol (Petrosyan and colleagues, 2023).

The counterpoint is built into the design. This is one health-system membership cohort in northern California, not a national sample, and everyone in it had already survived to 90 free of a recorded dementia diagnosis, which is a selected group by definition. It is observational, so nothing here establishes cause: the sex gap is measured, not explained, and survival differences between men and women cannot be separated from biology in these data. Dementia was ascertained clinically rather than at autopsy, and subgroup rates drawn from roughly 800 people carry wide uncertainty.