Menopause Hormone Therapy: What the WHI Got Wrong and What We Know in 2026

On July 9, 2002, the estrogen-plus-progestin arm of the Women’s Health Initiative was stopped early, and the resulting press conference did something almost no clinical trial has ever done: it changed a nation’s prescribing behavior in a matter of weeks. Use of menopausal hormone therapy fell off a cliff and never fully recovered. A generation of women went through menopause without treatment that would have worked.

In November 2025, the FDA began the opposite move, stripping the boxed warningsboxed warning The FDA’s strongest drug warning, printed inside a black border at the top of the label. Reserved for risks the agency judges serious enough that prescribers must see them first. about cardiovascular disease, breast cancer, and dementia off hormone therapy labels — calling them a product of outdated interpretation.

Both moments deserve skepticism. The first over-read a trial; the second risks over-correcting it. Here is what sits in between.

---

What the WHI Actually Studied

The trial itself was well conducted. The problem was who was in it and what it was asking.

The estrogen-plus-progestin arm randomized 16,608 postmenopausal women with an intact uterus to conjugated equine estrogensConjugated equine estrogens (CEE) A mixture of estrogens originally derived from pregnant mares’ urine, sold as Premarin. This was the estrogen used in the Women’s Health Initiative — not the estradiol most commonly prescribed today. 0.625 mg plus medroxyprogesterone acetateMedroxyprogesterone acetate (MPA) A synthetic progestin used in the Women’s Health Initiative. It appears to carry a larger breast cancer signal than micronized (body-identical) progesterone. 2.5 mg daily, or placebo. It was stopped after a mean 5.2 years.

Three design facts explain most of the subsequent confusion:

The participants were old. Mean age at enrollment was 63, and the majority were more than a decade past their final period. The trial was designed to test whether hormone therapy prevents chronic disease in older women — not whether it safely treats symptoms in a 51-year-old.

One regimen was tested. Oral conjugated equine estrogens with a synthetic progestin. Not transdermaltransdermal Delivered through the skin, as a patch or gel. Because the drug enters the bloodstream directly, it bypasses the first pass through the liver that oral pills undergo — which is why patches carry less clot risk than estrogen pills. estradiolestradiol The main and most potent form of estrogen during the reproductive years. It is also the form used in most modern hormone therapy patches and gels., not micronized progesteronemicronized progesterone Progesterone chemically identical to what the body makes, ground into fine particles so it can be absorbed as a capsule. Often preferred over synthetic progestins in modern hormone therapy. — the formulations most commonly prescribed today.

The endpoint was prevention, not symptoms. WHI was never designed to answer the question most women are actually asking.

---

The Numbers Behind the Headlines

The reported relative risks — a 26% increase in invasive breast cancer, a 29% increase in coronary events — are accurate and also nearly useless without their denominators. Here is what the trial reported in absolute terms, per 10,000 women per year:

Eight extra breast cancers per 10,000 women per year is a real harm. It is also a very different fact from the one most women absorbed in 2002, which was closer to “hormones cause breast cancer.”

And when the WHI investigators followed both arms out to 18 years and looked at death from any cause, they found nothing. All-cause mortality was 27.1% with hormone therapy versus 27.6% with placebo — a hazard ratiohazard ratio A number comparing how fast an outcome happens in two groups. Above 1 means more risk in the treated group, below 1 means less, and exactly 1 means no difference. of 0.99. No increase in cardiovascular death. No increase in cancer death.

---

The Timing Hypothesis: Supported, Not Proven

The leading explanation for why WHI looked worse than decades of observational data is that estrogen’s vascular effects depend on the state of the arteries receiving it. In relatively healthy vessels early in menopause, estrogen appears to slow plaque development; in vessels with established plaque, it may destabilize what is already there.

The cleanest test was ELITE, published in the New England Journal of Medicine in 2016. It randomized healthy postmenopausal women — stratified into those less than 6 years past menopause and those 10 or more years past — to oral estradiol or placebo, and tracked carotid artery wall thickness over a median of five years. Progression slowed significantly in the early group and not at all in the late group.

That is genuine support — but it is a surrogate endpointsurrogate endpoint A stand-in measurement — like artery wall thickness or bone density — used because it is faster to measure than the outcome people actually care about, such as heart attacks or fractures. Surrogates do not always predict the real outcome.. ELITE measured artery wall thickness, not heart attacks, and the two do not always track. The WHI mortality data points the same direction — the ratio of hazard ratios comparing women aged 50–59 to those aged 70–79 was 0.87 (95% CI 0.76–1.00) over 18 years — but that sits right at the edge of significance and comes from subgroup analysis.

So: timing almost certainly matters, the biological rationale is sound, and no trial has ever proven that starting early prevents cardiovascular events. Nobody should take it for that reason.

---

Breast Cancer: The Risk That Is Real

This is where enthusiasm most often outruns the data. The largest analysis is a 2019 Lancet individual-participant meta-analysis pooling 58 studies — roughly 143,000 women with breast cancer. Every type of systemic hormone therapy except vaginal estrogen was associated with excess risk, and risk rose with duration of use.

For a woman of average weight starting at age 50 and using it for 5 years, the estimated additional breast cancers diagnosed between ages 50 and 69 were roughly:

• 1 extra case per 50 women — daily combined estrogen-progestagen
• 1 extra case per 70 women — intermittent combined regimens
• 1 extra case per 200 women — estrogen alone

Ten years of use roughly doubles those figures, and some excess risk persists for more than a decade after stopping.

One real tension: within WHI itself, the estrogen-alone arm did not show an increase in breast cancer, and long-term follow-up suggested a reduction. The randomized and observational data genuinely disagree here, and anyone who tells you the question is closed is overselling. What both agree on is that combined therapy carries more risk than estrogen alone, and that duration is the main lever.

---

Route and Formulation: The Underrated Decisions

If there is one practical thing this article can change, it is this. Oral estrogen passes through the liver before reaching circulation, where it increases production of clotting factors. Transdermal estrogen largely skips that step.

A 2019 BMJ study of 80,396 women with venous thromboembolismVenous thromboembolism (VTE) A blood clot that forms in a vein — usually in the leg (deep vein thrombosis) — and can break loose and travel to the lungs. matched to 391,494 controls quantified the difference: oral preparations were associated with a 58% increase in clot risk (adjusted OR 1.58), while transdermal patches and gels showed no increased risk at all (OR 0.93, 95% CI 0.87–1.01).

Transdermal isn’t automatically right for everyone, but for a woman with obesity, a clotting history, migraine with aura, or simply a preference for the lower-risk option, it is the more defensible default — and it remains under-prescribed.

Two related points. If you have a uterus, you need a progestogenprogestogen An umbrella term for progesterone and its synthetic lookalikes. Added to estrogen therapy in women who still have a uterus, to keep estrogen from overgrowing the uterine lining. to protect the uterine lining from estrogen-driven overgrowth, and micronized progesterone appears to carry a smaller breast cancer signal than synthetic progestins like medroxyprogesterone acetate. And if your only symptoms are vaginal dryness, irritation, or painful sex, local vaginal estrogen is a different conversation entirely — systemic absorption is minimal, it was the one exception in the Lancet analysis, and it is dramatically under-used.

---

What It Doesn’t Do

Two claims circulate widely and are not supported. Dementia prevention: the KEEPS Continuation study re-evaluated women roughly 10 years after hormone therapy started within three years of their final period, and found neither cognitive benefit nor harm — a reassuring null, but not evidence of protection. Longevity: the 18-year WHI mortality hazard ratio was 0.99. Hormone therapy does not extend life.

If cognition is your concern, the evidence-based options in midlife are unglamorous: sleep, resistance training, cardiovascular risk management, and possibly creatine, where the menopause-specific data is early but real.

---

The 2025 Label Change — and How to Read It

On November 10, 2025, following an expert panel and public comment period, the FDA began removing the boxed warnings covering cardiovascular disease, breast cancer, and probable dementia. The boxed warning about endometrial cancer for systemic estrogen used alone was retained, and revised labels carry age-specific framing emphasizing initiation within 10 years of menopause.

Two things are true at once. Warnings derived from a trial in 63-year-olds were being applied to women in their early fifties, and that deterred appropriate treatment for two decades. And: removing a warning does not remove a risk. The Lancet breast cancer estimates did not change on November 10. Neither did the clot data.

The practical effect should be a more individualized conversation — age, years since menopause, symptom burden, personal and family history of breast cancer and clotting — not a new default. The Menopause Society’s 2022 position statement still frames it correctly: favorable benefit-risk for women under 60 or within 10 years of menopause onset who have bothersome symptoms or bone-loss risk, less favorable outside that window.

For women who can’t or don’t want to use hormones, the options are better than they were: fezolinetant (2023) and elinzanetant (October 2025) both target the brain pathway driving hot flashes.

---

Where This Leaves Us

The WHI was not wrong. It was over-generalized — a prevention trial in older women read as a safety verdict on symptom treatment in younger ones, with relative risks reported in the form most likely to frighten. Two decades later the answer is clear enough to act on, and narrower than either camp wants: an excellent treatment for the symptoms of menopause and for bone loss, with small absolute risks when started early and a real breast cancer signal that grows with duration. It is not a longevity drug. The moment it is marketed as one, we will have swung from one distortion to its mirror image.

---

A note on what’s still open: No randomized trial has ever tested whether hormone therapy started in early menopause prevents cardiovascular events. That trial would be large, long, and expensive, and it has never been funded — so everything written about the timing hypothesis, here included, rests on surrogate endpoints and subgroup analyses.