Incidence of dementia after age 90 years and association with APOE genotype, race, and sex in the USA: the LifeAfter90 prospective cohort study.
Level 3 - non-randomized controlled study
Prospective observational cohort study
PubMed 42551464 · doi:10.1016/j.lanhl.2026.100882
What was done
The LifeAfter90 prospective cohort study evaluated dementia incidence and risk factors among Kaiser Permanente Northern California members aged 90 years or older between July 2018 and November 2024. Participants underwent in-person or remote clinical evaluations every 6 months using physician assessment, the Clinical Dementia Rating, and a Functional Activities Questionnaire. Salivary DNA was collected for APOE genotyping. Of 1,120 enrolled individuals, 96 with prevalent dementia and 219 with only one visit were excluded, leaving 805 participants (413 with APOE data). Fine-Gray subdistribution hazard ratio (sHR) models adjusting for age were used to evaluate associations with sex, race and ethnicity, and APOE status while treating death as a competing risk.
What was found
Over a mean follow-up of 2.0 years (SD 1.7), 138 participants (17%) developed dementia and 295 (37%) died. The age-standardized incidence rate was 116.82 cases per 1,000 person-years (95% CI 93.69–139.96). In Fine-Gray competing-risk models, dementia risk was higher in female than male participants (sHR 1.89, 95% CI 1.30–2.76) and higher in Black than Asian participants (sHR 1.75, 95% CI 1.07–2.88). Risk was lower in APOE ε2 carriers than non-carriers (sHR 0.39, 95% CI 0.17–0.88), but was not significantly higher in APOE ε4 carriers (sHR 1.51, 95% CI 0.92–2.47). Education showed no significant association with dementia risk.
Why it matters
This study provides prospective data on dementia risk in a diverse cohort of individuals aged 90 and older. It demonstrates that ethnoracial and sex disparities persist into extreme old age, while the protective association of APOE ε2 remains evident and the APOE ε4 risk effect is attenuated.
Limits
Follow-up was short (mean 2.0 years) with high mortality (37%). APOE genotyping was available for only 413 of 805 participants (51%), and 219 enrolled individuals were excluded due to having only one evaluation, introducing potential attrition or selection bias. The sample was drawn exclusively from an insured population in Northern California, limiting generalizability.
Cited by
- supports The likelihood of developing dementia increases the longer a person lives.