News · Brain & Mental Health
Women who used estrogen-only hormone therapy had less Alzheimer's damage in 2,959 autopsied brains
Hormone therapy and dementia have been argued about for 25 years, mostly from questionnaires. A Neurology team went to the tissue instead, and found lower odds of Alzheimer's pathology at autopsy among women who had used estrogen.
- Women who had used estrogen-only hormone therapy had less Alzheimer's damage in their brain tissue.
- The odds of significant Alzheimer's pathology at autopsy were about a third lower.
- They were also less likely to have been diagnosed with dementia during life.
- Autopsy is the point: plaques counted in tissue cannot be swayed by who the patient was.
- Hormone use was self-reported and looked back on, so this cannot show cause.
The question of whether hormone therapy protects the brain has been argued for a quarter of a century, almost entirely from questionnaires and diagnosis codes.
That is a weak place to argue from. Whether a woman gets a dementia diagnosis depends on whether she has a doctor, whether that doctor is looking, and whether she is the kind of patient whose complaints get investigated. All three track wealth and education, and so does taking hormone therapy in the first place.
Writing in Neurology, researchers went somewhere harder to argue with. They tested the association between estrogen-only MHT use, or menopausal hormone therapy, and what was actually found in the brain after death.
The odds of increased Alzheimer pathology on autopsy were significantly decreased in hormone therapy users relative to nonusers, by about a third.
Why tissue is a better witness
A brain at autopsy does not know how much its owner earned.
Alzheimer’s disease is the most common form of dementia among older people, and people with Alzheimer’s have plaques and tangles in their brain, abnormal buildups of different proteins. Counting them is a measurement, not an assessment. It does not depend on whether anyone thought to order a scan.
That matters because the alternative explanation for every previous version of this finding has been diagnostic. If hormone therapy users are simply more likely to be well-off, well-connected and articulate, they will also be more likely to be diagnosed with something treatable and less likely to be written off as old. Tissue is immune to that particular story.
The two cohorts, and what each was for
Neuropathologic data were collected from NACC for 258 hormone therapy users and 2,701 non-users, with a mean age of death around 82 in both groups. That is the autopsy series, and the primary result.
The second data set did something different. In the ADNI cohort, 110 users and 1,948 non-users were followed with living-brain measures rather than tissue, and hormone therapy use was associated with decreased amyloid load assessed through plasma and spinal fluid.
Two independent samples, two different kinds of measurement, one direction. Hormone therapy use was also associated with significantly lower odds of clinical dementia diagnoses.
The agreement is the argument. Any single one of those results has an obvious objection; the objections are not the same for all three.
What this does not settle
Nobody was randomized. Nobody was even asked in advance.
Hormone use here was self-reported, often about decisions taken decades earlier, and the women who made those decisions were not a random sample of women. In the 1990s, hormone therapy in the United States was disproportionately taken by women with private insurance, more education and better access to a gynecologist, and each of those independently predicts a lower risk of dementia.
The authors are direct about the limits. Their results do not address causality and have limited generalizability due to the retrospective nature of the study, though they suggest a protective effect of MHT use in the dementia course. They describe them as small but significant associations.
There is also a definitional narrowness worth holding onto. This is estrogen-only therapy, which in practice means women who have had a hysterectomy. Most women who take hormone therapy take a combined preparation, and this study says nothing about it.
The timing problem underneath all of this
The reason the field is a mess is that “hormone therapy” describes several different exposures that people keep treating as one.
Menopause is the time in your life when you stop having your period and can no longer get pregnant, reached when you have not had a period for 12 months, and the most common age for reaching it is between 45 and 55. Starting estrogen at 51 to manage symptoms is a different biological event from starting it at 68 to prevent disease, and the trial that shaped public opinion in the early 2000s largely tested the second.
The idea this paper leans on is that estrogen decline during menopause may contribute to increased risk of Alzheimer disease in women, which is a hypothesis about a specific window rather than about hormones in general.
What a woman should actually do with this
Nothing yet, and the reason is not caution for its own sake.
Aging is the biggest risk factor for dementia, and there is no cure for most types of dementia, including Alzheimer’s disease. That combination makes any hint of prevention attractive enough to act on prematurely, which is exactly how the field got into trouble the first time.
Hormone therapy is currently prescribed to treat menopausal symptoms, and that decision already involves a real conversation about clots, breast cancer and timing. This study does not change any input to that conversation. What it does is make the case that the question deserves a trial designed to answer it, in the window where the hypothesis says it should work.
People also ask
What did the study find?
The odds of increased AD pathology on autopsy (primary outcome) were significantly decreased in MHT users relative to nonusers (odds ratio 0.65, 95% CI 0.48-0.88, p = 0.005). MHT use was also associated with significantly lower odds of clinical dementia diagnoses (OR 0.61, 95% CI 0.55-0.67) and with decreased amyloid load in plasma and cerebrospinal fluid.
Why does autopsy evidence matter here?
Because a dementia diagnosis is a judgement made by a clinician, and judgements can be influenced by who the patient is. Plaques and tangles counted in tissue after death are not. Finding the same direction in both is harder to explain away as diagnostic bias.
What is estrogen-only hormone therapy?
Hormone therapy given without a progestogen, usually to women who have had a hysterectomy. It is a different exposure from the combined therapy used in most women with a uterus, and this study looked only at the estrogen-only form.
Does this contradict the Women's Health Initiative?
Not exactly. The trial that alarmed everyone in the early 2000s tested combined therapy started well after menopause, and this is estrogen-only in an observational sample. Both can be true; they are not measuring the same thing.
How big is the effect?
The authors call it small but significant. A roughly one-third reduction in the odds of an outcome is not nothing, but it comes from an observational comparison in which the two groups differed in many ways beyond hormone use.
What is the main weakness?
Hormone use was self-reported, the design is retrospective, and women who took hormone therapy in the 1990s were, on average, wealthier and better educated than women who did not. Every one of those things independently predicts less dementia.
Should this change anyone's prescription?
No. Hormone therapy is prescribed for menopausal symptoms, and its risks and benefits are decided on that basis with a doctor. An observational association with brain tissue is a reason for a trial, not a reason to start a drug. This is general information rather than medical advice.