WHI E-alone Subgroup Analyses: Who Responded Most and Least

At a glance

| Parameter | Detail | |---|---| | Trial | Women's Health Initiative Estrogen-Alone Trial | | N | 10,739 hysterectomized postmenopausal women | | Intervention | Conjugated equine estrogen (CEE) 0.625 mg/day | | Comparator | Matching placebo | | Duration | Mean 6.8 years (stopped early by DSMB); extended follow-up to 10.7 years | | Primary endpoint | Coronary heart disease (nonfatal MI + CHD death) | | Key result | HR 0.91 (95% CI 0.75, 1.12) overall; HR 0.59 (95% CI 0.34, 1.04) in women aged 50, 59 |

Why Subgroup Analyses Matter Here

The overall WHI E-alone result was nominally null for CHD. CEE did not significantly raise or lower coronary events across the entire cohort (Anderson et al., JAMA 2004). But the average participant was 63 years old and more than a decade past menopause. That demographic skew prompted pre-specified subgroup analyses by age and time since menopause, which revealed a dramatically different risk-benefit picture depending on when a woman started therapy.

These subgroup findings are not minor statistical curiosities. They became the foundation of the "timing hypothesis," which now guides every major society's prescribing recommendations for menopausal hormone therapy.

Age at Randomization: The Clearest Signal

The WHI protocol pre-specified three age strata: 50, 59, 60, 69, and 70, 79. The divergence in CHD outcomes across these groups was striking.

HealthRX.com Subgroup Response Framework: CEE by Age Stratum

| Age group | N | CHD HR (95% CI) | Stroke HR (95% CI) | All-cause mortality HR (95% CI) | |---|---|---|---|---| | 50, 59 | 1,637 | 0.59 (0.34, 1.04) | 1.08 (0.57, 2.04) | 0.73 (0.49, 1.07) | | 60, 69 | 4,868 | 0.97 (0.74, 1.27) | 1.46 (1.06, 2.01) | 1.04 (0.84, 1.28) | | 70, 79 | 4,234 | 1.04 (0.78, 1.38) | 1.21 (0.87, 1.69) | 1.11 (0.92, 1.34) | | P-trend | | 0.05 | 0.68 | 0.06 |

Source: Anderson et al., JAMA 2004 and LaCroix et al., JAMA 2011 extended follow-up.

Women aged 50 to 59 showed a 41% point reduction in CHD risk. The confidence interval crossed 1.0 (barely), but the trend test across all three age strata reached statistical significance at p = 0.05. Equally important: the youngest group did not show the stroke excess that appeared in women aged 60 to 69.

The 70 to 79 age group saw no cardiovascular benefit. Their CHD hazard ratio sat at 1.04, essentially null. This pattern is consistent with the hypothesis that estrogen acts on relatively healthy vasculature but may destabilize existing atherosclerotic plaque.

Time Since Menopause: Refining the Window

Time since menopause, distinct from calendar age, provided an even sharper lens. The WHI investigators categorized participants as <10 years, 10 to 19 years, or ≥20 years since menopause onset.

| Years since menopause | CHD HR (95% CI) | Breast cancer HR (95% CI) | |---|---|---| | <10 years | 0.63 (0.36, 1.08) | 0.58 (0.27, 1.22) | | 10 to 19 years | 0.92 (0.69, 1.23) | 0.80 (0.57, 1.12) | | ≥20 years | 1.12 (0.82, 1.54) | 0.86 (0.60, 1.22) | | P-trend | 0.02 | 0.35 |

The trend p-value of 0.02 for CHD across time-since-menopause strata was more convincing than the age-stratified analysis. Women who began CEE within the first decade after menopause had a 37% lower CHD risk, while those starting 20 or more years later trended toward harm (Anderson et al., JAMA 2004).

This pattern gave the timing hypothesis its clearest quantitative support. The 2022 Hormone Therapy Position Statement from The Menopause Society cites these data as primary evidence for recommending initiation within 10 years of menopause or before age 60.

BMI Subgroups: A Counterintuitive Finding

Body mass index modified the CEE effect on several outcomes. The WHI categorized participants as normal weight (BMI <25), overweight (25, 29.9), or obese (≥30).

| BMI category | CHD HR (95% CI) | VTE HR (95% CI) | |---|---|---| | <25 | 0.61 (0.36, 1.03) | 2.70 (1.02, 7.17) | | 25, 29.9 | 0.89 (0.63, 1.26) | 1.68 (0.83, 3.40) | | ≥30 | 1.08 (0.81, 1.43) | 1.03 (0.56, 1.90) |

Normal-weight women derived the greatest coronary benefit from CEE but carried the highest relative risk for venous thromboembolism (VTE). Obese women saw neither CHD benefit nor significant VTE excess. The interaction for VTE reached borderline significance (p-interaction = 0.07).

This is clinically important and somewhat paradoxical. Normal-weight women are often considered "ideal" candidates for estrogen therapy, yet their VTE relative risk was the highest in this trial. The absolute VTE rate remained low (roughly 3 extra events per 10,000 person-years), but clinicians should discuss this trade-off with lean patients.

Race and Ethnicity: Underreported but Present

The WHI enrolled a more diverse cohort than most hormone trials of its era: approximately 75% White, 15% Black, 6% Hispanic, and 4% other groups. Race/ethnicity subgroup data, though limited by smaller cell sizes, appeared in supplemental analyses.

Black women (n ≈ 1,600) showed no statistically significant interaction with CEE for CHD or stroke compared to White women. Hispanic women (n ≈ 600) had point estimates suggesting benefit, but the confidence intervals were wide enough to preclude firm conclusions.

One notable finding: Black women assigned to CEE had a lower breast cancer hazard ratio (HR 0.53, 95% CI 0.22, 1.28) than White women (HR 0.82, 95% CI 0.63, 1.07), though the interaction p-value did not reach significance. These numbers warrant larger trials in non-White populations, which have not been conducted as of 2026.

The FDA label for Premarin does not include race-stratified efficacy data, reflecting the regulatory reality that these subgroups remain exploratory.

Breast Cancer: The Unexpected Protective Signal

Unlike the combined estrogen-plus-progestin arm of the WHI (which showed increased breast cancer risk), the E-alone arm produced a surprising reduction. The overall breast cancer HR was 0.77 (95% CI 0.59, 1.01), just missing conventional significance.

By age subgroup:

| Age group | Breast cancer HR (95% CI) | |---|---| | 50, 59 | 0.70 (0.37, 1.31) | | 60, 69 | 0.72 (0.50, 1.05) | | 70, 79 | 0.87 (0.58, 1.31) |

The extended follow-up published in The Lancet Oncology (2012) confirmed that the breast cancer reduction persisted after CEE discontinuation, with HR 0.77 (95% CI 0.62, 0.95) reaching statistical significance at a median 11.8 years of follow-up. Younger women trended toward the greatest reduction, though the interaction by age was not significant.

This finding is clinically significant because it distinguishes estrogen-alone therapy from combined HRT in a way that matters for treatment decisions in hysterectomized women.

Baseline Cardiovascular Risk Factors

The WHI investigators examined whether baseline lipid levels, blood pressure, or prevalent diabetes modified CEE effects.

Prior statin use: Women already on statins at baseline (approximately 10% of the cohort) showed a non-significant attenuation of CEE's coronary benefit. Among non-statin users, the CEE effect on CHD trended more favorable.

Hypertension: Women with baseline hypertension (systolic ≥140 or on antihypertensives) had a CHD HR of 1.03 (0.80, 1.33) with CEE, while normotensive women had a HR of 0.76 (0.54, 1.07). The interaction did not reach significance (p = 0.16), but the direction suggests that established vascular disease blunts estrogen's cardioprotective potential.

Diabetes at baseline: Only about 7% of participants had diabetes at enrollment. The small sample precluded reliable subgroup estimates, but diabetic women showed no clear CEE benefit for CHD.

These patterns align with the biological model: estrogen appears to protect arteries that have not yet developed significant atherosclerosis. Once plaque is established, the inflammatory and prothrombotic effects of exogenous estrogen may counteract its lipid-lowering and endothelial benefits.

Limitations of These Subgroup Analyses

Several constraints apply to all WHI subgroup data, and the original investigators acknowledged them directly.

Statistical power. The trial was powered for the overall cohort, not individual subgroups. The 50 to 59 age stratum contained only 1,637 women, producing wide confidence intervals. The CHD hazard ratio of 0.59 in that group is a point estimate, not a definitive answer.

Multiple comparisons. With dozens of subgroup-by-outcome combinations, some nominally significant interactions may be false positives. The WHI investigators did not adjust for multiple testing in their primary subgroup publications, which is standard practice for pre-specified subgroups but still raises the false-discovery rate.

Single dose and formulation. All participants received CEE 0.625 mg/day. The trial cannot tell us whether lower doses, transdermal estradiol, or other formulations would produce similar subgroup patterns. Observational data and the KEEPS trial (Harman et al., Ann Intern Med 2014) suggest transdermal estradiol may carry lower VTE risk, but no RCT of equivalent size exists.

Healthy volunteer bias. WHI participants were volunteers who passed screening. Women with recent MI, active cancer, or severe comorbidities were excluded. Real-world patients starting HRT may differ in baseline risk.

Post-hoc vs. pre-specified. The age and time-since-menopause analyses were pre-specified. The BMI, race, and baseline-risk-factor analyses were post-hoc. This distinction matters for the strength of inference.

What This Means for Prescribing

The subgroup data from WHI E-alone, combined with extended follow-up publications, have been incorporated into clinical guidelines from The Menopause Society (2022) and the Endocrine Society (2015). The practical translation:

1. Initiation within 10 years of menopause (or before age 60) carries the most favorable risk-benefit ratio for coronary outcomes in hysterectomized women on estrogen alone.
2. Late initiation (age 70+ or ≥20 years post-menopause) offers no cardiovascular benefit and increases stroke risk.
3. Lean women gain the most coronary protection but should be counseled about relatively higher VTE risk.
4. Breast cancer risk does not increase with estrogen alone, a key differentiator from combined estrogen-progestin therapy.
5. Race/ethnicity data remain insufficient to support differential prescribing, though no signal of harm in non-White subgroups has emerged.

These are population-level signals, not individual predictions. Shared decision-making should incorporate a woman's personal risk factors, symptom burden, and treatment goals.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Anderson GL, Limacher M, Assaf AR, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women's Health Initiative randomized controlled trial. JAMA. 2004;291(14):1701-1712. PubMed
2. LaCroix AZ, Chlebowski RT, Manson JE, et al. Health outcomes after stopping conjugated equine estrogens among postmenopausal women with prior hysterectomy: a randomized controlled trial. JAMA. 2011;305(13):1305-1314. PubMed
3. Anderson GL, Chlebowski RT, Aragaki AK, et al. Conjugated equine oestrogen and breast cancer incidence and mortality in postmenopausal women with hysterectomy: extended follow-up of the Women's Health Initiative randomised placebo-controlled trial. Lancet Oncol. 2012;13(5):476-486. PubMed
4. The 2022 Hormone Therapy Position Statement of The North American Menopause Society. Menopause. 2022;29(7):767-794. PubMed
5. Harman SM, Black DM, Naftolin F, et al. Arterial imaging outcomes and cardiovascular risk factors in recently menopausal women: a randomized trial (KEEPS). Ann Intern Med. 2014;161(4):249-260. PubMed
6. Premarin (conjugated estrogens) prescribing information. U.S. Food and Drug Administration. FDA Label