Women Were Warned for Twenty Years That Hormone Therapy Raised Dementia Risk. A Study of 21,000 Points the Other Way.

This is observational. It reports an association, not a demonstrated effect. Women who take hormone therapy and continue with it differ from women who do not, in ways that are difficult to fully adjust for: they tend to be more engaged with healthcare, are more likely to have the means to pursue treatment, and are often healthier at the point of prescription. That pattern, sometimes called healthy-user bias, produces exactly this shape of result in many settings where a randomised trial later finds nothing.

The estrogen-only group is a specific population. Estrogen-only therapy is generally prescribed to women who have had a hysterectomy, because women with a uterus are given a progestogen alongside estrogen to protect the endometrium. So this is not simply a subgroup taking a simpler drug. It is a group defined by having had a major operation, and that carries its own differences in age, health history and the reasons for surgery.

Brain donors are not a random sample. People who agree to donate their brains to research, and the cohorts that recruit them, differ from the general population in education, engagement and often in health.

Does this mean the 2002 findings were wrong?

Not exactly, and the more useful framing is that they were narrower than how they were received.

The alarming dementia signal came largely from women who began hormone therapy well after menopause, frequently in their late sixties or seventies, and substantially on combined estrogen and progestogen rather than estrogen alone. What was communicated to the public was a much broader message: hormone therapy is dangerous.

The idea that timing matters, that starting near menopause may differ from starting fifteen years later, has been argued for years and remains unsettled. This study is consistent with that argument without proving it.

What should a reader actually do with this?

Nothing unilateral, and that is a genuine recommendation rather than a disclaimer.

Hormone therapy has real benefits and real risks, and the balance depends on personal and family history, the timing relative to menopause, the formulation, and the reason for considering it. A single observational study, however large and however novel its method, does not change an individual prescribing decision. It changes the weight of evidence a clinician is working from.

What it does justify is a conversation. A significant number of women were steered away from a treatment two decades ago on the basis of a message broader than the finding underneath it, and many have never revisited it. If the question is live for you, it is a reasonable one to raise, alongside the cognitive symptoms discussed in our guide to menopause, brain fog and mental health.

The research question this leaves open is the one that always follows an observational result of this size: whether a randomised trial, designed around timing and formulation, would reproduce it. That is precisely the gap the TAME trial was built to close for metformin, and it has taken years to fund. Until one does, the correct description is a strong association in a large sample, measured more directly than before, and not yet a demonstrated cause.