Why Hormone Therapy Is Not Recommended to Prevent Heart Disease or Dementia

Two questions about menopausal hormone therapy sound almost identical but have entirely different answers. Should a person with no symptoms take hormones to lower their future risk of heart disease or dementia? And should a person with disruptive hot flashes take hormones to feel better? The US Preventive Services Task Force answers the first question with a clear no. In its 2022 statement in JAMA, it assigns a Grade D recommendation to hormone therapy for the primary prevention of chronic conditions, meaning it advises against starting estrogen plus progestin, or estrogen alone after hysterectomy, purely to reduce long-term disease risk. The Task Force deliberately does not answer the second question at all. Keeping those two questions apart is the whole point.

Key points#

• The Grade D verdict applies only to using hormones to prevent chronic disease in people without symptoms.
• It does not apply to treating hot flashes, night sweats, or vaginal dryness, which are a separate clinical question.
• The evidence comes largely from the Women's Health Initiative, which tested prevention directly and found net harm.
• A headline that says "hormones do not prevent disease" is accurate; one that says "no one should take hormones" is not.

What a Grade D actually claims#

A USPSTF grade is not a verdict on a drug. It is a verdict on one carefully defined question, built from three parts: a specific population, a specific action, and a specific goal. Change any of the three and the grade may no longer apply.

Here the population is postmenopausal people who have no menopausal symptoms. The action is taking hormones. The goal is lowering future risk of chronic illness. Within that box, the Task Force issued two parallel conclusions. Do not use combined estrogen and progestin to prevent chronic conditions in people who still have a uterus. Do not use estrogen alone for the same purpose in people who have had a hysterectomy. A D grade carries a defined meaning in this system: there is at least moderate certainty that the intervention produces no net benefit, or that its harms outweigh its benefits.

The boundaries are stated plainly in the recommendation itself. It does not cover people taking hormones for hot flashes, night sweats, or genitourinary symptoms such as vaginal dryness. It does not cover people with premature or surgically induced menopause. It does not cover anyone who already has the diseases in question. Prevention in a symptom-free person is the entire scope.

The evidence behind the no#

The recommendation leans heavily on the Women's Health Initiative, a set of large randomized trials that tested the preventive idea head on. That idea had real momentum before the trials. Observational studies had suggested that women on hormones seemed to have healthier hearts, and it was reasonable to hope a trial would confirm it. It did not.

As the National Heart, Lung, and Blood Institute summarizes, estrogen plus progestin after menopause raised the risk of heart disease, stroke, blood clots, breast cancer, and dementia. Estrogen alone raised the risk of stroke and blood clots. The 2022 JAMA statement describes the same picture: no meaningful drop in coronary heart disease, more strokes and venous clots, and, in the memory substudy of older women, more probable dementia with combined therapy.

Estrogen alone, given to people who had a hysterectomy, showed a somewhat different balance. It still raised stroke and clot risk, but it did not carry the same breast cancer signal, and in longer follow-up breast cancer risk was actually lower. Both regimens reduced fractures, which is a genuine benefit. Weighing that fracture reduction against the documented harms, the Task Force concluded that for the narrow goal of preventing chronic disease in people without symptoms, the trade was not worth it. No net benefit was the operative finding.

One drug, two report cards#

It helps to remember that a medication can earn a failing grade for one job and a passing grade for another. Aspirin makes the point. Once recommended broadly to prevent first heart attacks, its role has been steadily narrowed as its bleeding risk was weighed against a shrinking benefit, yet it remains standard after a heart attack has already happened. The useful question is almost never whether a drug is simply good or bad. It is good or bad for whom, for what purpose, and at what stage of life. Hormone therapy studied as a preventive across a broad, older, symptom-free population answers one version of that question. It says little about the version a 51-year-old with drenching night sweats is actually asking.

Why symptom relief is a different calculation#

Treating menopausal symptoms runs on a different set of numbers. The people involved tend to be younger and closer to the menopausal transition. The benefit is relief that is often quick and easy to feel. The typical duration of use is shorter. Because the population, the benefit, and the time frame all differ, the risk and benefit math differs too.

Professional societies that focus on menopause evaluate symptom treatment on its own terms, weighing an individual's symptom burden, age, time since menopause, and personal risk factors for clots, stroke, and breast cancer. Their conclusions can look quite different from a blanket D grade without contradicting it, because they are grading a different question. That is not the guidelines disagreeing with each other. It is two questions receiving two answers.

The most common reporting error is to collapse the two. "Hormones do not prevent disease" is a fair summary of the Task Force verdict. "Hormones are dangerous and no one should take them" is not something the evidence supports. The D grade rejects a preventive strategy. It does not condemn a class of medicines.

How to read a recommendation like this one#

Two habits make guidelines far easier to interpret. First, find the population and the endpoint before you react to the grade. A recommendation aimed at symptom-free prevention tells you little about symptomatic treatment, and the reverse is equally true. Second, separate what a guideline or drug label has actually demonstrated from what people assume it means. A recommendation describes what was tested and found. It is not a personal instruction, and it does not automatically transfer to every situation a patient brings into the room.

The strength of the USPSTF approach is precisely its refusal to stretch. By declining to answer beyond the question the trials tested, it produces a conclusion you can rely on for that question and a clear flag that other questions need their own analysis. For anyone weighing menopausal hormone therapy, the practical lesson is that "should I take this to prevent dementia?" and "should I take this to live better through menopause?" are two separate conversations, and the answer to the first should never be borrowed to settle the second.