Menopausal hormone therapy

What is given

Estrogen is the active part. Anyone who still has a uterus also receives a progestogen, because estrogen alone makes the lining proliferate. That is why there are two kinds of studies: estrogen plus progestogen with an intact uterus, estrogen alone after hysterectomy. This distinction is the key to everything that follows.

Why it works against hot flashes

The thermostat sits in the hypothalamus. When estrogen falls away, a group of nerve cells there becomes overactive, the tolerance window for body temperature narrows, and the slightest fluctuation is followed by sweating. Estrogen widens this window again.

The Cochrane analysis pools 24 studies with 3,329 women: hot flashes 75 percent less frequent than under placebo. The caveat belongs here too — placebo alone brought a decline of almost 58 percent from baseline. Without a control group, nothing can be judged here.

The WHI and what it really showed

The Women’s Health Initiative was funded by the US National Institutes of Health; the tablets came from the manufacturer. 16,608 women with a uterus were randomized to conjugated equine estrogens plus medroxyprogesterone acetate or placebo. After 5.2 years, the safety board stopped the trial: breast cancer, heart attacks, strokes and pulmonary embolisms were more frequent, hazard ratio for breast cancer 1.26.

26 percent more breast cancer was the headline. The figure that hardly anyone read out at the time is the absolute one: per 10,000 women per year, 8 more breast cancer cases, 7 cardiac events, 8 strokes, 8 pulmonary embolisms — against 6 fewer colorectal cancer cases and 5 fewer hip fractures, with overall mortality unchanged. A real risk, but not the order of magnitude that stuck.

What was misread from it

The participants were on average 63 years old, two thirds over 60. The study was not meant to treat hot flashes but to test whether hormones prevent heart disease — in mostly symptom-free women. The finding was applied to the 52-year-old with 10 sweating episodes a day.

The WHI was not a false alarm. The figures were correct, and stopping the trial was correct. What was wrong was applying a finding in 63-year-olds taking a particular preparation to every woman taking every preparation. A correction, not a reversal.

The estrogen-alone arm

10,739 women without a uterus received estrogen alone; this arm was published in 2004. Strokes were more frequent, hip fractures less frequent, and for breast cancer the hazard ratio was 0.77 — fewer, at the time narrowly not significant. Chlebowski’s 20-year analysis confirmed the picture: under estrogen alone, less breast cancer and lower mortality from it; under estrogen plus medroxyprogesterone acetate, more. The difference lies not in the estrogen but in this progestogen.

Timing: plausible, not proven

In 2017 Manson analyzed both arms: 27,000 women, 18 years. 27.1 percent died under hormones, 27.6 percent under placebo. Nothing. Only in the 50- to 59-year-olds did it look more favorable than in the 70-year-olds. From this arose the timing hypothesis: starting under 60 or within 10 years after menopause results in a different risk profile than a late start.

It was tested twice, in ELITE and in KEEPS. What was measured there was the vessel wall, not heart attacks. ELITE found the difference only in the carotid artery and only with an early start, and not at all on cardiac CT; KEEPS found none at all.

What is done differently today

Two changes, both from France. First, the route into the blood: in 2007, ESTHER found a fourfold thrombosis risk under oral estrogen, 0.9 via the skin, that is, none. The tablet passes through the liver first and stimulates clotting factors there; the patch does not.

Second, the progestogen. In the E3N cohort, the relative risk of breast cancer was 1.00 under micronized progesterone, the body’s own molecule, 1.16 under dydrogesterone and 1.69 under other synthetic progestogens.

Without hormones

For women who do not want or may not take hormones, there is fezolinetant. It blocks the neurokinin 3 receptor on precisely those overactive nerve cells in the thermostat. In SKYLIGHT 2, hot flashes decreased, less strongly than under estrogen. Liver values are monitored; the studies are manufacturer-funded.

What the studies show

WHI, combination arm (stopped July 2002)

16,608 women aged 50 to 79 with a uterus, randomized to conjugated equine estrogens plus medroxyprogesterone acetate or placebo, stopped after 5.2 years. Breast cancer 1.26; per 10,000 woman-years 8 more breast cancer cases, 7 cardiac events, 8 strokes, 8 pulmonary embolisms, 6 fewer colorectal cancer cases and 5 fewer hip fractures, overall mortality unchanged.

WHI, estrogen alone, and Chlebowski 2020

10,739 women without a uterus, published 2004: 12 more strokes per 10,000 woman-years, fewer hip fractures, breast cancer 0.77. After 20 years (JAMA 2020): estrogen alone breast cancer 0.78, breast cancer mortality 0.60, both significant; with medroxyprogesterone acetate 1.28.

ELITE 2016 and KEEPS

ELITE: 643 women, oral estradiol versus placebo, 5 years; carotid wall thickening half as fast with an early start, not at all more than 10 years after menopause, no effect on cardiac CT. KEEPS: 727 women, 4 years, tablet or patch versus placebo, no difference.

ESTHER 2007 and the E3N cohort

ESTHER, 271 women after a first thrombosis versus 610 controls: oral fourfold risk, transdermal 0.9. E3N, 80,000 women, 2,354 breast cancer cases: micronized progesterone 1.00, dydrogesterone 1.16, other synthetic progestogens 1.69. Neither randomized.

Where the data stop — and where things were only observed

The timing hypothesis is plausible, not proven. Both studies that were supposed to test it measured the vessel wall and not heart attacks — and even there the finding remained limited to one group and one measurement site.

The advantage of the body’s own progesterone over synthetic progestogens comes from observational data. The women in the E3N cohort were followed, not randomized. A randomized trial of progesterone versus medroxyprogesterone acetate with breast cancer as the endpoint does not exist and probably never will. The guidelines act on it nonetheless: the German S3 guideline of 2020 prefers transdermal estrogen because of thrombosis, and the North American Menopause Society states that the risk depends on preparation, route, dose and timing of initiation. Consensus based on observational data.

As prevention of chronic disease in symptom-free women, hormone therapy has been tested and has failed — that is exactly what the WHI was. No guideline recommends it for this.

Status, approval and legal

All the agents mentioned here are prescription-only; estrogens and progestogens have been approved for decades, fezolinetant in the EU since December 2023. The indication is symptoms — hot flashes, sleep, vaginal dryness — and bone protection when fracture risk is increased, not prevention. What was given in the studies, for context and not as instructions: WHI 0.625 milligrams of conjugated estrogens plus 2.5 milligrams of medroxyprogesterone acetate daily; KEEPS a patch with 50 micrograms of estradiol daily plus progesterone on 12 days a month; fezolinetant 45 milligrams daily.

Safety

Thrombosis and stroke come first, especially with oral use and a late start. Breast cancer under synthetic progestogens is in the order of 1 additional case per 1,000 women per year. For a start beyond 60, the North American guideline additionally names dementia. Anyone taking hormones should have a reason that can be felt or measured. Which preparation, which route and which age fit together belongs in a conversation with a physician that looks at medical history and findings together.

Sources

• Women’s Health Initiative, combination arm, stopped July 2002 — Rossouw et al., JAMA 2002, 16,608 women
• Women’s Health Initiative, estrogen-alone arm, 2004 — Anderson et al., JAMA 2004, 10,739 women
• Chlebowski et al., JAMA 2020 — 20-year analysis of the WHI arms
• Manson et al., JAMA 2017 — all-cause mortality in both WHI arms over 18 years
• ELITE — Hodis et al., New England Journal of Medicine 2016, 643 women
• KEEPS — Harman et al., Annals of Internal Medicine 2014, 727 women
• ESTHER — Canonico et al., Circulation 2007, oral versus transdermal route and thrombosis risk
• E3N cohort — Fournier et al., Breast Cancer Research and Treatment 2008, 80,377 women and 2,354 breast cancer cases, progestogen type and breast cancer risk
• SKYLIGHT 2 (fezolinetant, Astellas) — Johnson et al., Journal of Clinical Endocrinology and Metabolism 2023
• Cochrane analysis on hot flashes — MacLennan et al., 2004, 24 studies with 3,329 women