Dementia affects women unequally. That imbalance has kept researchers returning to a question that has resisted a clean answer for two decades: does the decline in estrogen exposure at menopause contribute to it, and if so, does replacing that estrogen help?
The biological case isn’t hard to make. Estrogen supports synaptic integrity, glucose utilization, cerebrovascular function, blood-brain barrier maintenance and the regulation of neuroinflammation. Each of those is plausibly relevant to cognitive decline. The evidence on hormone replacement therapy and dementia, however, has been conflicting enough that clinicians have had little to offer patients beyond acknowledging the uncertainty.
Two studies published in 2026 add new evidence that points in a favorable direction. Neither says hormone therapy prevents Alzheimer disease. Both suggest the relationship may depend on who is treated and when.
The studies
A 2026 cohort study in Neurology evaluated estrogen-only hormone replacement therapy against Alzheimer disease outcomes. Hormone therapy use was associated with 35% lower odds of increased Alzheimer disease pathology at autopsy, along with lower amyloid pathology measured through plasma and cerebrospinal fluid biomarkers. Users also had 39% lower odds of dementia.
A second 2026 study followed 183,450 postmenopausal women for a mean of 13.3 years. Across more than 2.43 million person-years, researchers identified 3,948 dementia cases. Hormone therapy use was associated with a 10% lower risk of all-cause dementia.
The subgroup findings in that second study are where it gets interesting. Stronger associations appeared in women with surgical menopause, in APOE ε4 carriers, and in women with shorter lifetime endogenous estrogen exposure. Initiation between ages 46 and 56 showed the greater benefit, which suggests treatment timing influences the effect.
Read together, the pattern is consistent: the women who appeared to benefit most were the ones who had lost the most estrogen exposure, or lost it earliest, or carried the highest genetic risk.
What are the implications?
Both studies indicate that the relationship between hormone replacement therapy and dementia may vary by timing of initiation, type of menopause, genetic risk and lifetime estrogen exposure. That is a meaningful refinement of a question that has usually been asked as though one answer had to apply to everyone.
The limits matter just as much. These findings are observational. They do not establish causality, and they do not support initiating hormone therapy for the sole purpose of dementia prevention.
For patients who ask
Expect questions. Coverage of studies like these tends to compress into headlines about hormone therapy preventing Alzheimer disease, which is not what either study found. The useful framing for a patient conversation is that the research is promising, that it points toward certain patients benefiting more than others, and that decisions about hormone therapy should continue to rest on the indications and risk factors they already rest on.
References
Bruno, J., Shaw, J., & Hosseini, H. (2026, August 12). Association between menopausal hormone therapy and Alzheimer disease neuropathology. Neurology. American Academy of Neurology. https://www.neurology.org/doi/10.1212/WNL.0000000000218413
Squires, S., Saleh, R. N. M., & Pilling, L. C. (2026, August 26). Hormone replacement therapy and dementia risk among postmenopausal women: Identifying responsive subgroups in the UK Biobank. Alzheimer’s and Dementia, 22(8), e71679. https://doi.org/10.1002/alz.71679
The pharmacy benefit story moves fast. Keep the whole picture in view.
The Pharmacy Benefit Brief delivers perspective on cost, policy, and member protection. One email, once a month.

The view from inside the benefit
Expert insights from pharmacists and PBM operators with an inside view of where pharmacy decisions break down, what those gaps cost, and how they affect benefit performance.
See the whole pharmacy picture.
Send us your claims data and we'll show you the avoidable cost and clinical risk hiding in it.
.jpg)


