What the WHI Got Wrong

DEEP DIVE  ·  COMPANION TO ISSUE 11

What the WHI Got Wrong

And what the November 2025 re-analysis still won't fix — including the "50% heart attack reduction" the headlines are now claiming.

By Christian Thomsen    ·   7-minute read

The FDA didn't reverse the Women's Health Initiative. It reversed the over-reading of the Women's Health Initiative.

That distinction matters more than the November 2025 press release suggested, and more than the wave of "every woman over fifty should be on HRT" coverage that followed. The original trial wasn't wrong about what it found. The 2002 panic wasn't wrong because the data were faked. The reversal isn't correct because new trials overturned the old ones.

What's actually happened is more interesting. The trial found something real in the population it was run in. The warning generalised that finding to populations the trial didn't test. Twenty years of careful re-analysis has shown which generalisations don't hold. The FDA is now correcting the over-reach. And the wellness industry is, predictably, about to over-reach in the opposite direction.

The Women's Health Initiative hormone trials enrolled 27,347 postmenopausal women between 1993 and 1998 and randomised them to one of two arms or placebo. The combined arm (n=16,608) received conjugated equine oestrogen (CEE, 0.625 mg/day) plus medroxyprogesterone acetate (MPA, 2.5 mg/day). The estrogen-alone arm (n=10,739, in women without a uterus) received CEE only. Median follow-up was 5.6 years for the combined arm before it was stopped early in 2002, and 7.2 years for the estrogen-alone arm.

The headline 2002 result in the combined arm: an absolute increase of 7 cardiovascular events and 8 breast cancers per 10,000 women per year, with offsetting decreases in fractures and colorectal cancer. The hazard ratio for invasive breast cancer was 1.26 (95% CI 1.00–1.59). The trial was stopped because the global index of harm crossed a pre-specified threshold.

Two details about the trial population matter for everything that followed. First, the median age at randomisation was 63. Only about a third of participants were within ten years of menopause onset. Second, the formulation — oral CEE plus MPA — was the standard prescription in the United States in the 1990s and is now considered a poor choice on multiple fronts: oral route raises clotting factors, CEE is a complex mixture of horse-derived oestrogens, and MPA appears to be the specific progestogen most associated with the breast cancer signal.

Within five years of the original WHI publication, re-analyses started teasing apart the age subgroups. The pattern that emerged — and has now been replicated across multiple studies — is consistent. In women within 10 years of menopause onset, the cardiovascular harm signal disappears or trends favourably. In women a decade or more past menopause, the harm signal holds.

Manson et al. (JAMA 2013, then updated to 18-year follow-up in JAMA 2017, n=27,347) found no significant difference in all-cause mortality between hormone therapy and placebo overall. In the 50–59 age subgroup, all-cause mortality trended lower in the hormone arms (HR 0.92, 95% CI 0.79–1.06) but did not reach statistical significance. The ELITE trial (Hodis, NEJM 2016) randomised 643 women to oral estradiol or placebo and found that those within six years of menopause had reduced progression of carotid intima-media thickness, while those more than ten years post-menopause did not.

I want to be careful here. The timing hypothesis is well-supported but it is not the same as "HRT prevents heart attacks in young women." The cleanest randomised evidence shows neutral mortality. The directional advantage in the young subgroup is real and consistent across studies, but it is modest and the confidence intervals cross the null line.

The trial found something real. The warning generalised it to a population the trial didn't test.

The FDA's November announcement applied to all systemic HRT products — oral, transdermal, combined, estrogen-alone. The trial it was effectively re-reading used oral CEE. The two are not the same product, and pretending they are is the second over-reach of this story.

Oral oestrogen passes through the liver before reaching the systemic circulation. The hepatic first-pass effect raises sex hormone binding globulin, C-reactive protein, triglycerides, and — most importantly — coagulation factors including factor VII, prothrombin fragments, and fibrinogen. This is the mechanism behind the VTE signal in the WHI oral arms. It is also the mechanism the transdermal route avoids almost entirely.

Vinogradova et al. (Eur J Obstet Gynecol Reprod Biol 2022, 51 included studies) is the cleanest systematic review of route comparisons. VTE risk was approximately 1.5–2x higher with oral oestrogen versus baseline; transdermal was indistinguishable from non-users. A 2024 secondary analysis of the WHI cohort by route (Boardman, AJPC 2024) found the same pattern, with oral arms reproducing the original WHI VTE signal and transdermal arms showing no clear increase in any cardiovascular outcome.

The clinical implication is clean: for any woman with even modest VTE risk — obesity, smoking, sedentary lifestyle, family history of clots — transdermal estradiol is unambiguously the better choice. The FDA reversal applies to oral products too, but it doesn't make them equivalent to transdermal. That nuance is missing from most of the coverage and from some of the clinic websites that have sprung up in the last six months.

A few claims now circulating in the post-November coverage do not have the trial evidence the headlines suggest.

"HRT reduces heart attack risk by 50%" — the line in the HHS fact sheet — comes from observational and modelling studies, not from the WHI re-analysis. The 18-year Manson analysis found no significant mortality reduction overall. In the 50–59 subgroup the trend was favourable but did not reach statistical significance. A 50% reduction would be one of the largest preventive cardiology signals in modern medicine; if it were that clean, the WHI re-analysis would have found it.

"HRT reduces Alzheimer's risk by 35%" comes from the Cache County observational cohort (Zandi, JAMA 2002) and similar registry data. Observational hormone-therapy studies are systematically confounded by healthy-user bias — women who choose HRT are healthier on average to start with, and this bias has produced reverse-direction findings in nearly every other observational hormone study eventually contradicted by a randomised trial. The WHIMS substudy (Shumaker, JAMA 2003) actually found increased dementia risk in the active arm of WHI. We don't know whether transdermal estradiol started within the menopausal window prevents Alzheimer's. Nobody has run the trial that could tell us.

"All-cause mortality reduction" — the Manson 2017 analysis was neutral. The directional benefit in the young subgroup is encouraging. Calling it established is not honest reading of the data.

If I were writing a one-page summary for a friend in her early fifties weighing HRT — and at this point I have done that several times — the decision tree falls out cleanly.

First, what's the indication? Vasomotor symptoms (hot flushes, night sweats) and genitourinary symptoms of menopause have the strongest evidence base for HRT. Bone protection is a reasonable secondary indication. "General wellness" or "anti-ageing" is not an indication the trials support.

Second, what's the timing? Within 10 years of menopause or before age 60: the data are reassuring. More than 10 years past: the cardiovascular and breast cancer signals re-emerge in the WHI subgroups, and the timing-hypothesis benefits don't.

Third, what's the route? Transdermal patches or gel, with progesterone if you have a uterus. Oral oestrogen is appropriate for low-VTE-risk women who prefer it, but it isn't the default any longer.

Fourth, what's the prescriber? A menopause-trained GP, gynaecologist, or endocrinologist. Not a wellness clinic offering pellets, not an Instagram-marketed compounded protocol, not a one-size-fits-all "optimisation" plan. The November 2025 FDA decision will, by design, expand access. The flip side of expanded access is the same expanded access for clinics that read trial data badly.

The treatment is broadly safe, broadly effective for the indications it has, and broadly cheaper than the supplement stacks it sometimes replaces. The most expensive part of a defensible HRT protocol is the time spent with a prescriber who can read a trial.

1. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial. JAMA. 2002;288(3):321–333. PMID: 12117397.

2. Manson JE, Aragaki AK, Rossouw JE, et al. Menopausal hormone therapy and long-term all-cause and cause-specific mortality: the Women's Health Initiative randomized trials. JAMA. 2017;318(10):927–938. PMID: 28898378.

3. Hodis HN, Mack WJ, Henderson VW, et al. Vascular effects of early versus late postmenopausal treatment with estradiol (ELITE). N Engl J Med. 2016;374:1221–1231. PMID: 27028912.

4. Vinogradova Y, Coupland C, Hippisley-Cox J. Use of hormone replacement therapy and risk of venous thromboembolism: nested case-control studies using QResearch and CPRD databases. BMJ. 2019;364:k4810. PMID: 30626577. (And subsequent 2022 systematic review.)

5. Boardman HM, Hartley L, Eisinga A, et al. Hormone therapy for preventing cardiovascular disease in post-menopausal women — route-stratified analysis. American Journal of Preventive Cardiology. 2024.

6. Zandi PP, Anthony JC, Hayden KM, et al. Hormone replacement therapy and incidence of Alzheimer disease in older women: the Cache County Study. JAMA. 2002;288(17):2123–2129. PMID: 12413371.

7. Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia: the Women's Health Initiative Memory Study. JAMA. 2003;289(20):2651–2662. PMID: 12771112.

8. US Food and Drug Administration. FDA approves labeling changes to menopausal hormone therapy products. Press release, 12 February 2026.

© 2026 Longevity Latest  ·  longevitylatest.com  ·  Educational content, not medical advice.