Combined hormone therapy: 5 more breast cancers, 24 fewer fractures, per 1,000
A Cochrane review of 45,660 women finds combined hormone therapy trades fracture risk for breast cancer and stroke risk.
By The Weekend · · 5 min read
Combined hormone therapy raises breast cancer cases from about 19 to 24 per 1,000 women over roughly 5.5 years, and cuts fractures from about 111 to 87 per 1,000. Both numbers come from the same trial, the same women, the same pills. Nobody gets to choose one effect without the other.
That trade sits at the centre of a Cochrane review covering 24 studies and 45,660 participants, though almost all of the usable evidence comes from one source: the Women's Health Initiative, a US National Institutes of Health-funded trial run in the 1990s and reported in the early 2000s. The review's funders — Cochrane itself, an independent nonprofit with no drug-industry backing — didn't run new trials. They re-graded the ones that exist, and downgraded confidence in several outcomes because the evidence base is old, oral-only, and thin on the exact formulations circulating today.
A 55-year-old with a hot flush and a family history is the whole dilemma
Picture the decision as it actually lands: a woman in her early fifties, night sweats disrupting sleep for the third year running, sitting across from a doctor who has to translate population statistics into a choice for one body. If she has a uterus and takes combined oestrogen-progestogen therapy for 5.5 years, her risk of a bone fracture drops by roughly a quarter — from 111 in 1,000 to 87 in 1,000, meaningful protection against the falls and breaks that erode independence after 70. Her risk of breast cancer rises from about 19 to 24 in 1,000 — five more cases per thousand women, a number that sounds small until it's the one call that comes back abnormal. Her risk of a blood clot in a vein roughly doubles, from 10 to 20 per 1,000. Gallbladder disease requiring surgery climbs from 16 to 27 per 1,000. Heart attack risk and lung cancer risk don't move either way.
If she's had a hysterectomy and takes oestrogen alone — no progestogen needed, because there's no uterine lining left to protect — the picture shifts again. Over 7 years, fracture risk falls further, from about 141 to 103 per 1,000. Breast cancer risk doesn't budge. But stroke risk rises, from 24 to 32 per 1,000, and gallbladder surgery risk increases too. Two hormones, two different risk ledgers, and the difference between them is whether progestogen is in the mix.
The trial that scared a decade also built the numbers it now doubts
The Women's Health Initiative was designed to settle a debate that had run since the 1960s: whether hormone therapy protected postmenopausal women's hearts, as decades of observational data suggested, or whether that link was an illusion created by healthier women being more likely to take hormones in the first place. When the trial's combined-therapy arm was stopped early in 2002 — after monitors saw the breast cancer and stroke signals cross a safety threshold — hormone prescriptions in the US fell by more than half within a year. Doctors who'd spent a decade telling patients hormone therapy protected their hearts had to explain, almost overnight, that the trial said the opposite.
The reversal was so sharp it became its own cautionary tale about observational medicine, taught in methods courses as the moment a randomised trial overturned twenty years of cohort data. But the trial that caused the scare is now, more than two decades later, also the trial supplying most of the evidence for the reassurance: this Cochrane review draws its headline numbers from the same study, because almost nothing since has matched its size or duration. The field learned its most durable lesson from one experiment it has never quite repeated.
The pills tested are not the pills mostly prescribed today
Here is the catch the numbers above don't carry with them: they describe hormone therapy taken as tablets, swallowed daily, using formulations and doses common in the 1990s. Much of current practice has moved toward transdermal oestrogen — patches or gels absorbed through skin — and lower doses, changes that observational data suggest may carry a different, possibly smaller, clot and stroke risk. The review's authors flag this directly: use caution applying these findings to therapy as it's actually prescribed now. The 45,660-participant evidence base is a photograph of hormone therapy as it existed thirty years ago, not a live feed of it today. Anyone quoting "24 in 1,000" as if it applies unchanged to a patch prescribed in 2025 is repeating a number without its expiry date.
Nobody has run the numbers for women under 50
The trial populations were women already past natural menopause, mostly in their 50s and 60s. The review's authors name the open question themselves: what happens to women who go through menopause early — before 50 — or whose ovaries stop working because of surgery or cancer treatment, and who might then take hormone therapy for a decade or two rather than five years? That's a materially different exposure, and a materially different set of baseline risks, and it hasn't been tested with the same rigor. Until it is, doctors extrapolate downward from a study built for an older population, because there isn't a better dataset to extrapolate from.
Does hormone therapy protect the heart?
This review found little or no difference in heart attack risk for either combined or oestrogen-only therapy — the protective effect once claimed from observational studies didn't hold up in the randomised trial.
Is oestrogen-only therapy safer than combined therapy?
For breast cancer, yes — oestrogen-only showed no increase. For stroke, oestrogen-only actually showed a larger relative rise (24 to 32 per 1,000) than combined therapy did (13 to 18 per 1,000).