HRT is linked to less dementia — but protection is not yet proven
Among 183,450 women, hormone therapy use was associated with a 10% lower dementia risk; the observational design cannot establish that treatment prevented the disease.
Observation — The study followed 183,450 postmenopausal women in the UK Biobank for an average of 13.3 years and recorded 3,948 dementia diagnoses. Survival models compared participants who reported hormone replacement therapy, or HRT, use for at least one year with those who never used it or used it for less than a year, adjusting for age, blood pressure, body mass index, cholesterol, education, socioeconomic deprivation, smoking, diabetes, and medications.
Observation — Across the full sample, HRT use was associated with a 10% lower risk of all-cause dementia, with a hazard ratio of 0.90 and a 95% confidence interval from 0.84 to 0.96. The association was stronger after surgical menopause, with a ratio of 0.74, and among women with shorter natural estrogen exposure, with a ratio of 0.84. The estimated relative reduction for Alzheimer's disease was 16%; no comparable reduction appeared for non-Alzheimer dementias.
Inference — The findings point to heterogeneity: combining every formulation, age, and hormonal history into one question may conceal meaningful differences. They do not identify a protective mechanism. Although the association looked stronger in APOE ε4 carriers, the overall interaction between genotype and HRT was not statistically significant, so the study does not show that genetic testing can select who will benefit.
Hypothesis — The lower risk among women who started HRT from ages 46 to 56 is consistent with the critical-window hypothesis, under which neural tissue may respond differently near the menopausal transition than it does decades later. Consistency is not confirmation: starting age is also related to clinical indication, access to care, prior health, and other factors that could produce the same observed pattern.
Limitation — This was an observational cohort, not a randomized trial. HRT users may have more education, income, medical follow-up, or better overall health, creating healthy-user bias even after statistical adjustment. The database also lacked adequate detail on therapy composition, dose, and route; estrogen alone, estrogen-progestogen combinations, pills, and patches should not be treated as one intervention.
Responsible speculation — Trials and finer causal analyses may test whether timing, formulation, and reproductive history truly modify the effect and might eventually support more personalized decisions. For now, these data do not make HRT a dementia-prevention treatment and do not justify starting, stopping, or changing therapy without individual clinical assessment of symptoms, risks, and benefits.
Key points
- The overall finding was a 10% lower associated risk, not demonstrated causal prevention.
- Surgical menopause, estrogen exposure, and starting age showed differences that still require confirmation.
- This study alone is not a basis for changing hormone therapy without individual medical review.