For 20 years, women were warned that hormone therapy raised their dementia risk. A study of 21,000 women just looked inside their autopsied brains and found the opposite
The warning became medical consensus with unusual speed. In July 2002, the Women’s Health Initiative trial published interim results showing that combined hormone therapy, estrogen plus progestin, was associated with elevated risks of breast cancer, cardiovascular events, and what the researchers called probable dementia. Within months, hormone therapy prescriptions collapsed. Women already using it stopped. Women considering it were advised against it. The FDA added black-box warnings to hormone therapy labeling citing the dementia risk. For the next two decades, those warnings shaped how millions of women and their doctors approached one of the most common interventions in menopause care.
The scientific foundation for that guidance has since been significantly eroded. Researchers identified multiple problems with the original WHI findings: the average participant was 63 years old, more than a decade past natural menopause, an age at which estrogen’s relationship with the brain may be fundamentally different from its effects in early menopause. The cognitive harms observed in the trial were concentrated in the combined estrogen-progestin group, not the estrogen-only group, which showed a statistical trend toward fewer dementia diagnoses. And the women in the trial were generally older, already symptomatic, and not representative of the population most likely to consider hormone therapy today.
Despite this erosion, the warnings persisted. The FDA did not remove the black-box dementia warnings from hormone therapy products until February 2026, following a comprehensive scientific review and expert panel process that took more than a year.
Now a study from Stanford Medicine has provided the most direct biological evidence yet on one side of this question. Published August 12, 2026 in Neurology, the medical journal of the American Academy of Neurology, it is the first study to measure the connection between estrogen-only menopausal hormone therapy and the physical deposits that define Alzheimer’s disease in autopsied human brains, at a scale large enough to draw meaningful conclusions.
The findings run in the opposite direction from the warnings that shaped two decades of clinical practice.
What the study measured and how
The research team, led by Jennifer Bruno at Stanford Medicine, drew on two of the most comprehensive Alzheimer’s research databases available in the United States: the National Alzheimer’s Coordinating Center, which aggregates data from 37 Alzheimer’s Disease Research Centers across the country, and the Alzheimer’s Disease Neuroimaging Initiative, a large ongoing longitudinal study.
The combined dataset included 21,462 women aged 50 and older who had either used estrogen-only menopausal hormone therapy or had used no hormone therapy of any kind. Women who had used combined estrogen-progestin therapy were excluded from the analysis, allowing the team to isolate the effect of estrogen alone.
Because estrogen-only therapy is typically prescribed for women who have had a hysterectomy, most participants in the hormone therapy group had undergone that procedure. The comparison is therefore between women who had hysterectomies and used estrogen-only therapy against women who had hysterectomies and did not use hormone therapy, controlling for one of the most significant potential confounders.
For the 2,959 women in the dataset whose brains were examined at autopsy, researchers used the standard CERAD scoring system to evaluate the presence and density of amyloid plaques. They also measured neurofibrillary tangle burden using the Braak staging system, assessed the presence of neuritic plaques, and scored clinical dementia diagnoses made during participants’ lifetimes.
In the living participants from the ADNI cohort, the team measured blood and cerebrospinal fluid biomarkers of Alzheimer’s pathology, providing a window into biological changes occurring in the brain before death.
What the brain tissue showed
The autopsy findings are the study’s most significant contribution. Prior research on hormone therapy and dementia had relied almost entirely on clinical diagnoses and cognitive test scores, which are downstream of the biological processes that actually drive Alzheimer’s disease. Measuring amyloid plaques and tau tangles directly in brain tissue gets one step closer to the underlying mechanism.
Among the women whose brains were examined, those who had used estrogen-only hormone therapy had 35% lower odds of showing Alzheimer’s-related pathological deposits. They had 39% lower odds of having received a clinical dementia diagnosis during their lifetimes.
The raw proportions make the difference concrete. Of women who had used estrogen-only therapy, 18% showed no signs of Alzheimer’s disease at autopsy whatsoever, compared with 10% of non-users. At the other end of the spectrum, 40% of estrogen users showed all three hallmark pathological features of Alzheimer’s disease, compared with 51% of non-users.
The living biomarker data pointed in the same direction. Estrogen-only therapy users showed higher plasma levels of the amyloid-beta 42/40 ratio, a marker that indicates less amyloid accumulating in the brain, and higher cerebrospinal fluid levels of amyloid-beta 1-42, another indicator of reduced pathological amyloid burden.
“This is the first study to measure the connection between menopausal hormone therapy and reductions in Alzheimer’s-related pathological deposits in autopsied brains,” said Bruno, the study’s lead author. “Our findings provide evidence of an association between use of estrogen-only hormone therapy during later life and better outcomes on dementia and brain health.”
Why the original warnings were based on a different question
The Women’s Health Initiative findings and the new Stanford study are not directly contradictory, because they were measuring different things in different populations.
The WHI cognitive findings came from a subset of the trial called WHIMS, which enrolled women between 65 and 79 years old, most of them more than a decade past natural menopause. At that age, the brain may be past the window during which estrogen provides its strongest neuroprotective effects, a concept known in the field as the critical window hypothesis. Estrogen initiated late in the post-menopausal period may behave differently from estrogen used in the years immediately following menopause, when the brain’s estrogen receptors are still responding normally.
The WHI also used a specific oral conjugated equine estrogen formulation, and the dementia signal it found was concentrated in the combined estrogen-progestin arm, not the estrogen-only arm. When researchers later looked at the estrogen-only arm of the WHI, they found a non-significant trend toward lower dementia incidence, not higher.
The Stanford study’s population skewed toward older women who began therapy at an average age of 70, also late relative to natural menopause. That the study found protective associations at this age is notable, and it suggests that whatever biological mechanism connects estrogen to Alzheimer’s pathology may operate across a wider post-menopausal window than the critical window hypothesis would predict.
“Despite these limitations, our findings provide evidence of an association between use of estrogen-only hormone therapy during later life and better outcomes on dementia and brain health,” Bruno said.
The burden this study addresses
Two thirds of the approximately 7.4 million Americans currently living with Alzheimer’s disease are women. The disparity has long been attributed primarily to longevity, since women live longer than men on average and age is the strongest risk factor for Alzheimer’s. But a growing body of research suggests that the hormonal changes of menopause, specifically the sharp drop in estrogen production, may be an independent contributor to women’s elevated risk.
Estrogen receptors are distributed widely throughout the brain, including in the hippocampus and prefrontal cortex, regions central to memory and executive function. Estrogen promotes the clearance of amyloid protein, supports synaptic plasticity, reduces neuroinflammation, and helps regulate the metabolism of tau. When estrogen production falls sharply at menopause, these protective functions may be diminished.
The 20 years during which the WHI findings dominated clinical guidance were 20 years during which a large proportion of women who might have used hormone therapy did not. Quantifying the cognitive consequences of that period is impossible, but the Stanford study adds to a growing body of evidence suggesting the guidance that shaped those decisions was based on findings that did not represent the full picture.
What this does not establish
The researchers are direct and consistent about the limits of what observational data can show. Women who choose to use hormone therapy may differ systematically from women who do not, in ways that affect both their likelihood of using therapy and their Alzheimer’s risk. Healthier women may be more likely to be prescribed hormone therapy and to continue using it. Women with higher socioeconomic status may have both greater access to hormone therapy and greater access to the protective factors associated with lower dementia risk, including better nutrition, more exercise, and more cognitive stimulation.
The study cannot fully adjust for all these differences, and its observational design means it can document an association but cannot prove that estrogen-only therapy caused the reduction in Alzheimer’s pathology.
“While these findings help us better understand the relationship between hormone therapy use and various markers of dementia, more research needs to be done before we can make recommendations to women about their use of these therapies in relation to their brain health,” Bruno said. “This study looked back at women who were using hormone therapy decades ago with the timing and type of use differing from what is current practice for most women today, so the results are informative, but they may not apply to today’s standards.”
The study does not address combined estrogen-progestin therapy, which is used by women with an intact uterus and represents a different hormonal intervention with a different risk and benefit profile. It also does not provide data on the effect of timing, dose, or duration of estrogen-only therapy on Alzheimer’s outcomes. Those questions remain open.
What it adds, across 21,462 women and nearly 3,000 autopsied brains, is the most direct biological evidence yet that estrogen-only hormone therapy is associated with measurably less Alzheimer’s pathology in women who used it. For a question that has shaped one of the most consequential medical decisions millions of women have made over the past two decades, that evidence is worth taking seriously.
The study “Association Between Menopausal Hormone Therapy and Alzheimer Disease Neuropathology” was authored by Jennifer Bruno, Jacob S. Shaw, and S.M. Hadi Hosseini at Stanford Medicine, and published August 12, 2026 in Neurology.
Source: Stanford Medicine. DOI: 10.1212/WNL.0000000000218413
