Key takeaways
- Estradiol promotes nitric oxide release, improves arterial elasticity and suppresses inflammatory activity in the vessel wall, and the rise in women's cardiovascular events after menopause tracks its decline.
- The original WHI signal came from women starting oral conjugated equine estrogen at an average age of 63.
- Estradiol is one of the more powerful vasodilators in human physiology, and its effects on the vessel wall are not incidental to reproduction — they are direct.
- The Manson et al. 13-year follow-up of WHI participants in JAMA 2017 is the point at which the field turned.
- KEEPS and ELITE were designed specifically to answer whether timing was the explanation.
In 2002 a single trial result changed what a generation of women were told about hormone therapy, and the correction has taken more than twenty years to filter through. The correction is not that the original finding was wrong. It is that the finding applied to a specific population — older women, a decade past menopause, taking an oral formulation — and was then generalised to every woman regardless of age or preparation. Once you separate the women by when they started, the picture divides cleanly, and it divides for a reason that makes biological sense.
What estradiol does to an artery
Estradiol is one of the more powerful vasodilators in human physiology, and its effects on the vessel wall are not incidental to reproduction — they are direct.
It stimulates nitric oxide release from endothelial cells, which is the primary mechanism by which arteries relax. It improves arterial elasticity, meaning the vessel accommodates each pulse rather than transmitting it stiffly. It suppresses pro-inflammatory cytokine activity within the vessel wall, which is where atherosclerosis begins. And it supports HDL production and favourable handling of circulating lipids.
The population-level consequence is visible in every dataset on cardiovascular mortality by sex. Before menopause, women's event rates run at well under half the male rate. After menopause, the gap closes and then reverses. That inflection tracks estradiol decline closely enough that the hormone is difficult to argue away as a bystander. Estradiol and cardiovascular health covers the vascular biology in more detail.
The timing hypothesis, and why it is mechanistic rather than statistical
The Manson et al. 13-year follow-up of WHI participants in JAMA 2017 is the point at which the field turned. Women aged 50-59 who received estrogen-only therapy showed:
- Reduced all-cause mortality (risk ratio 0.69)
- Reduced coronary heart disease
- Reduced colorectal cancer
- No increase in breast cancer
Women in the same trial who started after age 70 had increased cardiovascular events. Both results are real. They are results in different populations, and the average of the two produced the 2002 headline.
The explanation that fits is not a statistical artefact but a difference in the tissue being treated. Estradiol acting on a healthy endothelium promotes vasodilation and suppresses the inflammatory processes that initiate plaque. Estradiol acting on an artery that already carries established, partially calcified plaque acts on inflamed and unstable tissue, and increasing local inflammatory and proteolytic activity in that setting can destabilise a plaque rather than prevent one. Same hormone, different substrate, opposite outcome.
That is why the operative variable is years since menopause rather than chronological age, and why the practical framing is a window rather than an age cutoff. When to start HRT covers how that window is assessed individually.
The two trials built to test it
KEEPS and ELITE were designed specifically to answer whether timing was the explanation.
- KEEPS (Kronos Early Estrogen Prevention Study, 2014) randomised newly menopausal women to four years of hormone therapy or placebo. It found no difference in carotid intima-media thickness progression, but did show improved lipid profiles and favourable movement in markers of subclinical atherosclerosis.
- ELITE (2016) directly compared women within six years of menopause against women more than ten years out. Carotid intima-media thickness progression was significantly slowed in the early-treated group and not in the late-treated group (Hodis et al., N Engl J Med 2016).
ELITE is the more informative of the two because it tested the hypothesis directly rather than by inference. It is also worth being clear about what these trials measured: arterial wall thickness is a surrogate for atherosclerosis, not a count of heart attacks. Neither trial was sized to measure hard events. What they establish is that the mechanism behaves as predicted, which is meaningful but is not the same as a demonstrated reduction in myocardial infarction.
Route of delivery is the second variable
Oral estrogen is absorbed from the gut and passes through the liver before reaching the systemic circulation. The liver, seeing a large estrogen signal arrive at high concentration, increases production of clotting factors, C-reactive protein and other acute-phase proteins, and triglycerides. This is a pharmacokinetic consequence of the route, not a property of the hormone.
Transdermal estradiol enters the circulation directly through the skin and bypasses that first pass. The clotting factor response does not occur, which is why observational data consistently finds a venous thromboembolism signal with oral preparations and not with transdermal ones (Canonico et al., ESTHER study).
The formulation used in the original WHI arm was oral conjugated equine estrogen — a different molecule from the estradiol the body produces, delivered by the route that maximises hepatic effect. Modern practice defaults to transdermal bioidentical estradiol for cardiovascular reasons specifically. When the 2002 result is quoted today, it is being quoted about a preparation and a route that most clinicians would no longer choose.
What actually moves on a lab panel
Typical year-one changes in cardiovascular markers on transdermal therapy:
| Marker | Direction | Magnitude |
|---|---|---|
| HDL | Up | +5-10% |
| LDL | Down | −5-10% |
| ApoB | Down | −10-15% |
| Triglycerides (transdermal) | Slight down or unchanged | 0 to −10% |
| hs-CRP (transdermal) | Down or unchanged | 0 to −20% |
| Endothelial function (FMD) | Up | +15-25% |
Two notes on reading that table. The triglyceride and hs-CRP rows are annotated transdermal for a reason — oral preparations move both in the opposite direction, which is the first-pass effect showing up in the numbers. And ApoB is the row worth watching most closely, because it counts atherogenic particles rather than the cholesterol they happen to be carrying; ApoB, the real cardiovascular number explains why that distinction changes risk estimates.
The clinical pearl: the cardiovascular effects of hormone therapy show up in subclinical markers — arterial elasticity, lipid profile, inflammation — within the first year. Hard outcomes take five years or more to separate. A woman who checks her labs at six months is looking at the mechanism working, not at the outcome, and the two should not be confused in either direction.
The progestogen partner matters too
Any woman with an intact uterus needs endometrial protection alongside estrogen, and the choice of agent is not neutral. Bioidentical micronised progesterone has broadly neutral or slightly favourable cardiovascular effects. The synthetic progestin used in the original WHI combined arm partially offset estrogen's benefit, and the combined arm's results differ from the estrogen-only arm's for that reason.
Modern protocols default to micronised progesterone, which also has effects on sleep and mood that most women notice long before any lab changes — see progesterone for sleep and mood.
How the decision is actually made
None of the above tells an individual woman what to do, and it is not meant to. What it does is identify which facts the decision depends on: how many years since the final period, whether there is existing cardiovascular disease or established plaque, personal and family history of breast cancer and of clotting, whether the uterus is intact, and what symptoms are actually being treated. Cardiovascular effect is a consideration in that evaluation, not an indication on its own — hormone therapy is prescribed for menopausal symptoms and for bone, and the vascular picture informs how and whether that is done.
It is worth saying plainly that the risks are real rather than imaginary; they are simply smaller, more route-dependent and more timing-dependent than the 2002 coverage implied. HRT side effects, honestly sets those out without the softening.
Bottom line
The claim that hormone therapy raises cardiovascular risk is not false so much as unqualified. It described older women, starting oral conjugated equine estrogen a decade or more past menopause, treating arteries that already carried established disease. For a woman starting near menopause with transdermal bioidentical estradiol and micronised progesterone, the measurable picture is favourable: better endothelial function, better lipids including ApoB, lower inflammation, and reduced all-cause mortality in the 50-59 subgroup of the trial that started the panic. Timing and route dominate the story. The evidence for surrogate markers is strong; the evidence for hard event reduction is suggestive rather than settled, and it should be described that way. Whether it is right for any individual is a physician's judgement after a full history — the 60-second assessment is where that evaluation begins.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
