Estrogen-Only vs Combined

The Women's Health Initiative was not one trial; it was two, and they pointed in different directions. In the combined arm, breast cancer went up; in the estrogen-only arm, it went down. Stroke and blood clots, meanwhile, rose in both. Lumping the arms together — as the 2002 headlines did — hides the single most instructive comparison in the whole saga. This page lays out the two risk profiles side by side, explains the biology behind the difference, and explains what neither arm lets you conclude.

Why the Trial Had Two Arms

The split was anatomical. Estrogen taken without opposition thickens the uterine lining, and over years that thickening raises the risk of endometrial cancer — a fact established by the epidemiology of the 1970s, when unopposed estrogen prescriptions were followed by a surge of uterine cancers (Ziel & Finkle, NEJM, 1975). The fix is a progestogen, which keeps the lining in check. So the WHI ran two parallel trials: women with an intact uterus got conjugated equine estrogens plus medroxyprogesterone acetate (E+P), and women who had had a hysterectomy — and therefore had no lining to protect — got estrogen alone. The two arms asked two different questions, and they produced two different answers. Everything that follows is about the gap between them.

The Combined Arm, in Its Own Column

The E+P arm is the one the world heard about: breast cancer up, stroke up, blood clots up, coronary events up — the full accounting, with the absolute numbers behind each headline, is the subject of the forensic reporting page. For this page's purpose, the E+P column serves as the reference frame: a real but modest increase in breast cancer (8 extra per 10,000 women per year), a similar stroke increase, a larger and front-loaded clotting increase (18 extra per 10,000), and a small coronary increase concentrated in the first year. The arm also delivered real benefits — about a third fewer hip fractures and colorectal cancers. Keep that column in mind, because the other arm's column is where the surprises begin.

The Estrogen-Only Arm: The Breast Cancer Reversal

The estrogen-only trial enrolled 10,739 women and ran until February 2004, when stroke crossed its boundary and the trial was stopped after 6.8 years. The breast cancer result, published that year in JAMA, was the opposite of the combined arm's: 26 cases per 10,000 women per year on estrogen versus 33 on placebo, a hazard ratio of 0.77 with a confidence interval that just touched 1.0 (0.59–1.01). In other words, in this specific trial, conjugated equine estrogen alone was associated with fewer breast cancers than placebo — about 7 fewer per 10,000 women per year.

What Did Not Differ: Stroke and Clots

The reversal was real but partial, and the "estrogen-only is safer" reading collapses as soon as you look at the vascular columns. Stroke was increased in the estrogen-only arm — 44 versus 32 per 10,000 women per year, a hazard ratio of 1.39, which is if anything a larger absolute excess than the combined arm's 8 extra strokes. Venous thromboembolism also rose: 28 versus 21 per 10,000, a hazard ratio of 1.33, versus the combined arm's sharper 2.06. Coronary events, by contrast, were neutral in the estrogen-only arm (hazard ratio 0.91, not statistically significant). The honest summary: estrogen alone looked better on breast cancer and coronary disease, but both arms carried a stroke and clotting penalty. There is no arm of this trial that a woman with a history of stroke or blood clots should read as reassuring.

Side by Side: The Two Profiles

Two patterns jump out. Breast cancer and coronary events moved in opposite directions between the arms; stroke, clots, and hip fractures moved the same way in both. Whatever the progestin is doing, it is doing it selectively — and the selectivity is the whole story of the formulations page.

The 13-Year Follow-Up: What Lasted, What Faded

Stopping a trial does not stop its data. The 2013 cumulative analysis covering the intervention plus post-stopping years found that the E+P arm's breast cancer excess persisted after stopping, while the estrogen-only arm's breast cancer deficit also persisted — the curves stayed apart for over a decade (Manson et al., JAMA, 2013). A later follow-up with over 18 years of median tracking went further: breast cancer mortality was lower in the estrogen-only arm (hazard ratio around 0.60) and higher in the E+P arm (around 1.44) (Chlebowski et al., JAMA, 2020). Those mortality findings are real, but they carry the caveat the whole topic runs on: the trials were not powered for breast cancer death, and a mortality signal in a subgroup of a subgroup deserves respect, not certainty. The arm difference itself, however, is no longer in dispute.

Why the Arms Differ: The Progestin Question

The standard explanation for the gap runs through the progestogen. Breast tissue responds to progesterone-receptor stimulation with cell proliferation, and medroxyprogesterone acetate — the synthetic progestin used in WHI — is a potent stimulator of that pathway. When the world's pooled observational data were analyzed in the 2019 Collaborative Group review of more than 100,000 breast cancer cases, the excess risk concentrated in current users of combined therapy and grew with duration beyond about five years, while estrogen-only use carried a much smaller or absent excess (Collaborative Group, Lancet, 2019). The natural experiment of the two WHI arms fits that pattern precisely. Whether a different progestogen — micronized progesterone in particular — carries the same breast signal is one of the open questions the formulations page examines, and the honest answer there is that comparative trial data are thinner than anyone would like.

The Uterus Complication

The reason most postmenopausal women cannot simply take estrogen alone is the reason the trial had two arms in the first place. Unopposed estrogen raises endometrial cancer risk markedly — the 1970s experience, before progestogens were added, produced a sharp rise in uterine cancers among estrogen users (Ziel & Finkle, NEJM, 1975). The E+P arm of WHI confirmed the protection: endometrial cancer was not increased with combined therapy (hazard ratio 0.81, not significant). So the estrogen-only option exists only for women without a uterus — a biological constraint, not a preference. Anyone weighing "estrogen alone looks better" against "I still have my uterus" is weighing the wrong things; the lining rules that question out.

Reading the Two Arms Honestly

• 🎯 Ask "which arm?" before you believe any WHI number. "Hormone therapy raises breast cancer risk" is true of E+P and false of E-alone in this trial — the arm is not a detail.
• 🩸 No arm is vascularly neutral. Both raised stroke and clots; a history of either weighs heavily against either regimen.
• 🏛️ The uterus decides the options. Estrogen alone is for women without a uterus, full stop — unopposed estrogen and an intact lining are a known cancer risk.
• 🧭 Arm-level data is group-level data. It narrows the conversation; it does not end it — the individual weighing lives on the decision framework page, with a clinician in the room.

Related Topics