The Study That Scared a Generation Off Hormone Therapy (HRT), Revisited

If you have ever floated the idea of hormone therapy with a doctor and watched their expression shift, you and that doctor may have been flinching at the same thing: a single set of headlines from the summer of 2002. That July, a large American trial called the Women’s Health Initiative, or WHI, was halted before its planned end date, and within days the message reaching most women was blunt and frightening. Hormones cause breast cancer. Hormones cause heart attacks. Put them down.

More than twenty years later, that message is still doing damage in exam rooms across Canada, including to women who would clearly benefit from treatment. So this is a longer, more careful piece than most on this site, because the story earns it. We are going to walk through what the trial set out to do, what it found, how a piece of careful science turned into one of the most consequential miscommunications in modern women’s medicine, and where the evidence has landed in the years since. The sources are all linked, so you can read them for yourself. If you would rather have the short version than the full breakdown, we have a plain-English summary here.

The warning was written from women in their 60s and 70s, then handed to women in their 40s.

The first thing almost nobody was told in 2002 is that the WHI was never designed to study the women most likely to be asking about hormones.

It was a chronic disease prevention trial. The question it asked was whether starting hormones could prevent heart disease, fractures, and bowel cancer in older women, the group most at risk of those conditions. To answer that, it enrolled a population that was, on average, well past menopause. The combined trial, published in JAMA in 2002, recruited 16,608 postmenopausal women aged 50 to 79 with an intact uterus, and the average age at enrollment was 63. Only about 30 percent were under 60, and just 12 percent were aged 50 to 54. The typical woman in the trial was more than a decade past her last period.

That detail matters more than any single statistic that followed, because the women who take hormones for symptoms are usually in their late 40s and 50s, in the thick of the transition, not 63 and symptom-free. The trial enrolled very few of them. By one account, only around 4 percent of participants had moderate to severe hot flashes, as a 2005 commentary in the US-based AMA Journal of Ethics later pointed out.

There were also two separate trials, not one, a distinction that got flattened almost immediately. Both tested oral conjugated equine estrogen, sold as Premarin. One arm added a synthetic progestin, medroxyprogesterone acetate or MPA, for women with a uterus. The other gave estrogen alone to women who had had a hysterectomy. They produced very different results, which is the part of the story that got lost.

Here is where the gap between the science and the headlines becomes hard to ignore.

The combined trial was stopped early because a breast cancer monitoring statistic crossed a pre-set boundary and the overall risk-benefit tally tipped the wrong way. The reported hazard ratios looked alarming at a glance: 1.26 for invasive breast cancer, 1.29 for heart disease. To a reader without a statistics background, 1.26 sounds like a 26 percent chance of developing breast cancer. It does not mean that. It means a 26 percent relative increase over a baseline risk that was already small, which is why the absolute numbers in the box above are so modest. In plain terms, that worked out to roughly eight extra cases of breast cancer per 10,000 women each year. The 26 percent was a jump on a number that was already low, not 26 women in 100.

Even the headline breast cancer finding was not statistically significant. On the nominal confidence interval, the result brushed right up against 1.0, the line that separates a real effect from something that could be chance. Once the analysis was corrected for the fact that the data had been examined many times over the years, as WHI investigator Robert Langer later detailed in Climacteric, the confidence interval for breast cancer ran from 0.83 to 1.92, comfortably crossing 1.0. The same was true for heart disease. By the trial’s own adjusted statistics, neither headline harm reached significance. In plain terms, a confidence interval is the range the real answer probably falls in, and once that range includes 1.0, the value that means no difference at all, the result can no longer be told apart from chance. After the honest correction, both the breast cancer and the heart findings landed there.

The authors’ own conclusion was narrow and specific. This regimen, they wrote, should not be started or continued to prevent chronic disease, describing it as “not consistent with the requirements for a viable intervention” for that purpose. That is a statement about using hormones as a preventive medicine in older women. It is not a statement about treating hot flashes, night sweats, or low mood in a 49-year-old. The trial never tested that.

The clearest account of what went wrong comes from inside the trial itself.

A 2017 reappraisal in the journal Climacteric by Robert Langer, who was a principal investigator at one of the WHI’s clinical centres, describes a publication process he calls highly irregular. The first results paper was written by a small group at the coordinating centre and submitted to JAMA without the clinical site investigators seeing it. By the time the broader investigator group was shown the paper, at a meeting in Chicago, the journal had already been printed and shipped. Edits the investigators proposed, to soften the tone and correct the interpretation, arrived too late to matter.

Then came the press release. The announcement led with the breast cancer finding, and the study’s lead investigator told the press conference the results applied to all women regardless of age. The hardcopy journal was still in the mail, so for several days the trial was being discussed through a press release rather than the data. By the time scientists could read the paper closely, the story was already set.

Langer’s summary is direct: the reporting implied the trial had been stopped because hormones caused breast cancer and heart attacks, when in fact there was no statistically significant harm for either on the properly adjusted analysis. A 2017 article in the Canadian Medical Association Journal relaying his account put it plainly, describing how the unusual circumstances of the early termination produced misinformation and, in his words, hysteria, and noting that hormone use fell by as much as 80 percent in the years that followed.

There is a further detail, rarely told, that suggests the breast cancer harm may not have been real at all. Langer’s reappraisal points out that among the women who had never used hormones before the trial, breast cancer turned up at the same rate whether they took the combined therapy or the placebo. No increase at all. So where did the frightening number come from? From a quirk in the comparison group. Among the placebo women who had used hormones in the past, breast cancer appeared unusually rarely, lower than anyone expected, for reasons still unexplained. A hazard ratio works by measuring the treated group against the placebo group, so when that placebo number came in abnormally low, the treated women looked high beside it, the way you would look tall standing next to an unusually short crowd, even though your own height never changed. The figure that frightened the world leaned on that one low comparison, not on a real rise among the women taking hormones.

Even the reassuring numbers got distorted on the way out the door. A newsletter the coordinating centre mailed to participants charted the event rates on a vertical scale that stopped at 60, though the rates themselves were counted per 10,000 women. Drawn that way, a true difference of a few cases per 10,000 filled a big stretch of the chart and looked like a wall. On a scale that ran all the way to 10,000, the same gap would have been a barely visible sliver. Same numbers, a frightening picture.

Two years after the combined trial made headlines, the second WHI trial reported. It pointed in nearly the opposite direction, and it received a fraction of the attention.

The estrogen-alone trial, published in JAMA in 2004, which gave conjugated equine estrogen without a progestin to women who had had a hysterectomy, found no increase in breast cancer. The result trended the other way, toward less breast cancer, with a hazard ratio of 0.77. In extended follow-up reported in Lancet Oncology, that reduction became statistically significant: roughly a 23 percent drop in invasive breast cancer, and a sizeable reduction in breast cancer death.

Sit with that, because it reverses the takeaway most women absorbed. Estrogen by itself did not raise breast cancer risk in this trial. It lowered it. The signal in the combined trial appears to track with the added progestin, MPA, not the estrogen. The popular shorthand, that estrogen causes breast cancer, had the chemistry close to backwards.

The reanalyses that followed pointed to one variable the original coverage ignored entirely: when a woman starts.

This is the timing hypothesis, sometimes called the window of opportunity. A pooled WHI analysis published in JAMA in 2007 found that the risk-benefit picture for heart disease was more favourable in women who began hormones near menopause than in women who started a decade or more later, even though stroke risk stayed elevated across all age groups. The proposed mechanism is that estrogen helps keep relatively healthy arteries healthy, but can destabilize plaque that has already built up in older ones.

The WHI’s own integrated analysis of both trials, published in JAMA in 2013, put numbers on that age effect. For women aged 50 to 59 taking estrogen alone, the overall balance of benefits and harms came out in favour of the treated group, about 19 fewer adverse events per 10,000 women each year, while women in their 70s on the same therapy landed on the wrong side of the same measure.

Two trials were then built specifically to test this. The ELITE trial, published in the US-based New England Journal of Medicine in 2016, gave oral estradiol to women sorted by how long they had been postmenopausal, and found it slowed the progression of artery thickening only in the women who started within six years of menopause. The KEEPS trial in recently menopausal women found no serious cardiovascular harm and benefits on some symptoms, with reassuring long-term safety data.

A Danish randomized trial pointed the same way, with harder endpoints than artery scans. The Danish Osteoporosis Prevention Study, published in the BMJ in 2012, gave hormones to just over a thousand recently menopausal women aged 45 to 58. After a decade, the treated group had roughly half as many deaths, heart attacks, and heart-failure admissions as the untreated group, with no increase in cancer, clots, or stroke. It was a small trial, but it was a randomized one, in exactly the women the WHI barely enrolled.

Perhaps the most reassuring number of all came from the longest look. An 18-year follow-up of both WHI trials, published in JAMA in 2017, found no increase in all-cause mortality, cardiovascular mortality, or total cancer mortality with either regimen. Across nearly two decades, women who had been on hormones were no more likely to have died than women who had been on placebo.

A practical takeaway sits inside all of this. For a healthy woman in her late 40s or 50s, near menopause and without specific contraindications, the evidence that frightened a generation was drawn from women who were neither her age nor in her situation. The timing that applies to her is the one the original trial was least equipped to study.

One piece of this remains unsettled, and it deserves to be presented as the live debate it is.

Two camps of credible researchers disagree about how much the progestin in combined therapy raises breast cancer risk. On one side, a 2023 critical review in Menopause by Avrum Bluming, Howard Hodis, and Robert Langer argues the alarm was overstated from the start. They note that the combined-therapy increase amounts to roughly one extra case per 1,000 women per year, was not statistically significant on the properly adjusted analysis, and came with no increase in breast cancer death. They also point out that the WHI’s own later publications credit estrogen alone with reducing breast cancer by about 23 percent.

On the other side, the WHI investigators, led by Rowan Chlebowski in a 2020 analysis in the US-based journal JAMA, maintain that combined estrogen-plus-progestin therapy does increase breast cancer incidence, and that this is a real and durable finding. Their more recent position narrows the disagreement in a useful way. In that same 2020 analysis, breast cancer fell in the estrogen-alone arm and rose in the combined arm, which points to the added progestin, MPA, rather than the estrogen itself, as the part that raises risk.

The part both sides agree on is the reassuring one: estrogen on its own lowers breast cancer risk.

So the honest state of the evidence is this. That estrogen alone reduces breast cancer is well supported and agreed on by both sides. Whether, and by how much, the progestin in combined therapy raises it is still disputed between serious people reading the same data. Both camps agree the absolute risk, if it exists, is small, and that it attaches to one older formulation, not to hormones as a category.

The major medical bodies have converged, and the direction is clear.

In Canada, the 2021 SOGC and Canadian Menopause Society guideline states that hormone therapy is the most effective option for vasomotor symptoms, the hot flashes and night sweats, and can be safely started in women without contraindications who are younger than 60 or less than 10 years past menopause. That is graded as high-quality evidence. It is worth saying plainly, since the old framing still lingers: for hormonally driven symptoms in this window, hormone therapy is a legitimate first-line treatment, not a last resort to be reached only after everything else has failed.

The Menopause Society, formerly NAMS, in its 2022 US position statement reaches the same conclusion: for women younger than 60 or within 10 years of menopause onset and with no contraindications, the benefit-risk balance favours treatment for bothersome symptoms and protection against bone loss.

In November 2025, the picture shifted again. The US Food and Drug Administration announced it would remove the “black box” warning for cardiovascular disease, breast cancer, and probable dementia from estrogen-containing menopausal hormone products, a process it has since carried into approved label changes for several products. The boxed warning for endometrial cancer on systemic estrogen-alone products remains in place. The SOGC publicly welcomed the change, noting the science had been reanalyzed, and reaffirmed its 2021 guidance. The European Menopause and Andropause Society called the original WHI interpretation overly broad and, in several respects, misleading.

The collapse in hormone use after 2002 was not abstract. It changed prescribing in Canada within two years, and the consequences cut in more than one direction.

A Canadian study in the Journal of the National Cancer Institute found that combined hormone use among women aged 50 to 69 fell from 12.7 percent in 2002 to 4.9 percent in 2004. The same study linked that drop to a decline of close to 10 percent in breast cancer incidence in that age group. That is the honest counterweight in this story: for combined therapy specifically, less use plausibly did prevent some breast cancers, and a fair account has to say so. To be precise about what this does and does not mean, the risk here attaches to the combined estrogen-plus-progestin therapy, the one with the synthetic progestin, not to estrogen on its own, which the same body of evidence shows lowering breast cancer risk. The asterisk belongs on one particular combination, not on hormone therapy as a whole.

What that figure leaves out is the cost on the other side of the ledger, because stopping hormones was never the risk-free default it was treated as. Langer’s reappraisal catalogues what followed the exodus. A large American health maintenance organization found that women who stopped lost bone density and then began breaking hips at a rising rate, with the risk climbing from no measurable increase in the first year to nearly double five years after stopping. A Finnish registry covering close to two million woman-years found that women who stopped hormones had roughly twice the risk of cardiac death and stroke in the first year compared with women who stayed on them.

The most contested estimate of that cost is also the most arresting, and it should be read as disputed rather than settled. A 2013 analysis in the US-based American Journal of Public Health by Philip Sarrel and colleagues calculated that estrogen avoidance among hysterectomized American women in their 50s, the group the estrogen-alone trial suggested benefited most, was associated with somewhere between roughly 18,600 and 91,600 premature deaths over the decade after 2002. The WHI investigators rejected that estimate, arguing that the underlying mortality difference was not solid enough to carry it and that a true effect of that size would have shown up in national death rates. Both things can hold at once: the precise number is unknowable and probably overstated at the top end, and the direction of harm from blanket avoidance in this particular group was real.

Stopping hormones was treated as the safe choice. For a lot of women, it was not the safe one.

The harder cost does not show up cleanly in any chart. A generation of Canadian women mid-transition went undertreated, told their symptoms were either trivial or too dangerous to address. A 2023 Canadian primary care review notes that hormone use among women aged 45 to 74 fell from around 43 percent to about 11 percent after 2002. A SOGC history of menopause medicine records that antidepressant prescriptions rose roughly in step with the fall, while an unregulated market in alternative remedies grew to fill the space hormones had left.

Some of that cost is measurable in dollars, and the figure is not small. When symptoms go untreated, women cut their hours, pass on promotions, or leave work altogether, often at the peak of their earning years.

The encouraging part is that this is turning around. Canadian insurer data show hormone therapy claims up about 21 percent over three years, as the science finally reaches the women who needed it all along.

The WHI was good science aimed at the wrong question for the women who most wanted answers. It correctly showed that starting hormones in your 60s and 70s to prevent chronic disease is not a good trade. It was never built to tell a 50-year-old whether treating her night sweats was safe, and the people who turned it into a blanket warning against all hormones, for all women, at all ages, went well past what the data could support.

Two decades of follow-up have been broadly reassuring. For healthy women under 60 or within 10 years of menopause, the major Canadian and international bodies now agree that the benefits of hormone therapy generally outweigh the risks for treating symptoms and protecting bone. The breast cancer question around combined therapy remains open, the absolute risks are small, and the route and formulation can change the math. What is no longer in serious doubt is that the wholesale retreat from hormones carried real costs of its own, in fractures, in cardiac risk, and in years of needless suffering, even if the exact toll will always be argued over.

None of this means hormone therapy is right for everyone. It means the decision belongs to you and a clinician who knows the current evidence, not to a headline from 2002. If a doctor still quotes that summer at you as though nothing has been learned since, you are allowed to bring the last twenty years into the room.

The key sources behind this piece, for anyone who wants to read further:

• Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women. JAMA, 2002.
• Women’s Health Initiative Steering Committee. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy. JAMA, 2004.
• Anderson GL, et al. Conjugated equine oestrogen and breast cancer incidence and mortality in postmenopausal women with hysterectomy: extended follow-up. Lancet Oncology, 2012.
• Langer RD. The evidence base for HRT: what can we believe?. Climacteric, 2017.
• Eggertson L. Trial overstated HRT risk for younger women. Canadian Medical Association Journal, 2017.
• Rossouw JE, et al. Postmenopausal hormone therapy and risk of cardiovascular disease by age and years since menopause. JAMA, 2007.
• Manson JE, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women’s Health Initiative randomized trials. JAMA, 2013.
• Manson JE, et al. Menopausal hormone therapy and long-term all-cause and cause-specific mortality. JAMA, 2017.
• Hodis HN, et al. Vascular effects of early versus late postmenopausal treatment with estradiol (ELITE). New England Journal of Medicine, 2016.
• Miller VM, et al. The Kronos Early Estrogen Prevention Study (KEEPS): what have we learned?. Menopause, 2019.
• Schierbeck LL, et al. Effect of hormone replacement therapy on cardiovascular events in recently postmenopausal women: randomised trial (Danish Osteoporosis Prevention Study). BMJ, 2012.
• Bluming AZ, Hodis HN, Langer RD. ‘Tis but a scratch: a critical review of the WHI evidence associating menopausal hormone therapy with breast cancer risk. Menopause, 2023.
• Chlebowski RT, et al. Association of menopausal hormone therapy with breast cancer incidence and mortality during long-term follow-up of the WHI randomized clinical trials. JAMA, 2020.
• Sarrel PM, Njike VY, Vinante V, Katz DL. The mortality toll of estrogen avoidance: an analysis of excess deaths among hysterectomized women aged 50 to 59 years. American Journal of Public Health, 2013.
• Prentice RL, et al., for the WHI Investigators. Women’s Health Initiative view of estrogen avoidance and all-cause mortality. American Journal of Public Health, 2013.
• Yuksel N, et al. Guideline No. 422a: Menopause: vasomotor symptoms, prescription therapeutic agents, complementary and alternative medicine, nutrition, and lifestyle. Journal of Obstetrics and Gynaecology Canada, 2021.
• The 2022 Hormone Therapy Position Statement Advisory Panel. The 2022 hormone therapy position statement of The North American Menopause Society. Menopause, 2022.
• FDA / HHS. HHS advances women’s health, removes misleading FDA warnings on hormone replacement therapy. US Food and Drug Administration, 2025.
• SOGC. SOGC statement regarding the use of menopausal hormone therapy (MHT). Society of Obstetricians and Gynaecologists of Canada, 2025.
• EMAS. EMAS statement on the FDA decision to remove black-box warnings from menopausal hormone therapy. European Menopause and Andropause Society, 2025.
• De P, Neutel CI, Olivotto I, Morrison H. Breast cancer incidence and hormone replacement therapy in Canada. Journal of the National Cancer Institute, 2010.
• Black D. Menopause hormone therapy: 2023 update. Canadian Primary Care Today, 2023.
• Reid R, et al. Menopause medicine: past, present, and future. Journal of Obstetrics and Gynaecology Canada, 2019.
• Manulife. Hormone replacement therapy for menopause. Manulife Group Benefits News, 2024.
• Menopause Foundation of Canada and Deloitte. Menopause and Work in Canada. Menopause Foundation of Canada, 2023.
• After the Women’s Health Initiative: menopausal women and hormone therapy. AMA Journal of Ethics, 2005.

Related reading: When you are ready to take this into an appointment, our companion guide, How to Talk to Your Doctor About Hormone Therapy, walks through the timing window to mention, the questions to ask about dose and form, and what to do if you get brushed off.

This article is for informational purposes only and does not constitute medical advice. Always consult your healthcare provider for personal medical guidance.