Women who took estrogen-only menopausal hormone therapy showed fewer signs of Alzheimer’s in their brains and were less likely to be diagnosed with dementia — though researchers stress this shows an association, not cause and effect.
Conducted at Stanford Medicine and published August 12, 2026 in Neurology, the study analysed data from 21,462 women, including thousands with brain scans, biomarker tests or autopsies (average autopsy age 82).
What was found
Compared with non-users, hormone therapy users had 39% lower odds of a dementia diagnosis and 35% lower odds of Alzheimer’s-related brain pathology at autopsy.
18% of therapy users showed no Alzheimer’s signs at autopsy versus 10% of non-users, and their brains had less amyloid buildup. The analysis examined amyloid plaques, tau tangles and neuritic plaques.
Why autopsy data changes the strength of the finding
Most studies of this question rely on clinical dementia diagnoses, which are imperfect. A meaningful proportion of people diagnosed with Alzheimer’s turn out to have different or mixed pathology, and some with substantial pathology are never diagnosed.
Autopsy examines the brain directly, and finding a difference in pathology rather than only in diagnosis makes a diagnostic artefact far less likely as an explanation. If hormone therapy users were simply less likely to be diagnosed — through better healthcare access, higher education or diagnostic bias — their brains at autopsy would look the same. They did not.
Why estrogen might plausibly matter
The biological rationale is not thin. Estrogen receptors are abundant in brain regions affected early in Alzheimer’s, including the hippocampus, and estrogen influences synaptic function, glucose metabolism in neurons, inflammation and cerebral blood flow.
The epidemiology also invites the question. Roughly two-thirds of Alzheimer’s patients are women, and while greater longevity explains part of that, it does not obviously explain all of it — and the abrupt loss of estrogen at menopause is an obvious candidate for the remainder.
The detail that constrains everything
The benefit was seen specifically with estrogen-only therapy, not estrogen combined with progestin — and that distinction is clinically decisive.
Estrogen-only therapy is prescribed to women who have had a hysterectomy. Women with an intact uterus require progestin alongside estrogen, because unopposed estrogen causes endometrial thickening and raises uterine cancer risk substantially.
So a finding applying only to estrogen-only therapy applies only to women who have had their uterus removed. It cannot be extended to the larger population of women considering hormone therapy at menopause.
The timing problem
The most important limitation is generational. This historical group started therapy around age 70, while typical current practice begins in the late 40s to early 50s.
That gap matters because of the “timing hypothesis” in hormone therapy research, which holds that effects on the brain and cardiovascular system depend heavily on when treatment starts relative to menopause. Starting near menopause and starting two decades later may produce genuinely different results — and the direction is not predictable from these data.
A finding in late starters therefore may not transfer to early starters, and vice versa.
The confounding that remains
“The results show an association and do not demonstrate that hormone therapy can prevent dementia,” cautioned study author Jennifer Bruno.
Women who received hormone therapy in that era differed systematically from those who did not. They were more likely to have regular medical care, to be healthier at baseline, and — because hormone therapy was withdrawn from some women over cardiovascular concerns — to be at lower cardiovascular risk to begin with.
Cardiovascular health is itself among the stronger predictors of dementia risk, so a group selected partly on cardiovascular grounds would be expected to have less dementia regardless of hormones.
The wider history
This question has a fraught past. Early observational work suggested hormone therapy protected against dementia and cardiovascular disease; a large randomised trial then reported the opposite for dementia in older women, contributing to a collapse in prescribing.
Subsequent reanalysis complicated both, and the timing hypothesis emerged partly as an attempt to reconcile them. Observational data here suggesting benefit is therefore not new evidence in an empty field — it is another data point in a literature where observational and randomised findings have repeatedly disagreed.
Why the sex disparity in Alzheimer’s is contested
The background question motivating this line of research — why roughly two-thirds of Alzheimer’s patients are women — is less settled than it appears, and the competing explanations matter for how the hormone finding should be read.
The simplest explanation is survival. Women live longer on average, age is the dominant risk factor, and more women therefore reach the ages where the disease is common. That accounts for a substantial share of the disparity and, most researchers think, not all of it.
Beyond that, candidates include the loss of estrogen at menopause, sex differences in how the APOE4 risk gene operates — carrying it appears to confer greater risk in women — differences in tau accumulation, and historical differences in education and occupational complexity that affect cognitive reserve.
These are not mutually exclusive, and disentangling them is difficult because they correlate. A finding that hormone therapy associates with less pathology fits the estrogen explanation without excluding the others — and the women who received hormone therapy also differed in education and healthcare access, which are themselves candidate explanations for the disparity.
Hormone therapy has genuine risks and benefits, and remains a decision for each woman and her doctor rather than something to start based on one study. This summarises an observational study and is not medical advice. Do not start or stop hormone therapy without consulting your healthcare provider.