WHI E+P Extension Data and What Happened After the Trial Ended

At a glance

| Detail | Value | |---|---| | Trial | Women's Health Initiative Estrogen-plus-Progestin (WHI E+P) | | N | 16,608 postmenopausal women aged 50, 79 | | Intervention | CEE 0.625 mg/d + MPA 2.5 mg/d | | Comparator | Matching placebo | | Original duration | Median 5.6 years (stopped early July 2002) | | Extension follow-up | Up to 18+ years post-randomization | | Primary endpoints (original) | CHD (primary benefit), invasive breast cancer (primary safety) | | Key extension result | Breast cancer HR remained elevated at 1.28 (95% CI 1.11, 1.48) through 18-year follow-up; cardiovascular hazard attenuated toward null |

Why the Trial Was Stopped and What Came Next

The original WHI E+P publication in July 2002 reported that combined CEE+MPA increased coronary heart disease events (HR 1.29 to 95% CI 1.02, 1.63), invasive breast cancer (HR 1.26 to 95% CI 1.00, 1.59), stroke, and pulmonary embolism. The Data Safety Monitoring Board stopped the trial three years ahead of schedule because the global index of risks exceeded benefits.

That decision launched two decades of reduced HRT prescribing worldwide. But it also created an unusual scientific opportunity: a large, placebo-controlled cohort that could be followed for years after the intervention ended. The WHI investigators designed three sequential extension phases to capture exactly that data.

Extension Study Design

The post-intervention follow-up operated in distinct phases, each with different consent structures and data collection methods.

Extension 1 (2005 to 2010) re-consented 12,788 of the original 16,608 participants for continued annual health outcome tracking. Women were no longer taking study medication. Roughly 4% of those originally assigned to CEE+MPA and 6% of placebo-group women initiated open-label hormone therapy on their own during follow-up, a crossover rate the investigators monitored but could not prevent.

Extension 2 (2010 to 2020) continued health event surveillance through National Death Index linkage and periodic questionnaires. This phase allowed the 18-year cumulative analyses published in JAMA 2020.

The critical methodological point: these extensions preserved the original randomization. Women were analyzed in their original assigned groups regardless of post-trial hormone use. This intention-to-treat approach is the backbone of the follow-up data's validity, though it also means estimates are conservative since any crossover would dilute between-group differences.

The HealthRX Duration-Signal Framework for Reading WHI Extension Data

Not all risks from the original trial behaved the same way after stopping. We categorize the post-trial signal trajectories into three patterns that help clinicians interpret the extension data:

| Signal Pattern | Outcome | What Happened Post-Cessation | |---|---|---| | Persistent and growing | Invasive breast cancer | HR remained elevated and became more statistically significant over time | | Attenuating toward null | CHD, stroke, PE | Excess risk diminished within 3 to 5 years of stopping | | Benefit that faded | Hip fracture, colorectal cancer | Protective effects during treatment disappeared after cessation |

This framework matters because the original 2002 paper presented a single "global index" that combined these divergent trajectories into one number. The extension data shows that collapsing them was misleading for clinical decision-making.

Breast Cancer: The Signal That Did Not Fade

The breast cancer finding is the most clinically consequential extension result. During the active trial, the hazard ratio for invasive breast cancer was 1.26 (95% CI 1.00, 1.59), barely crossing statistical significance.

By the 2010 post-stopping analysis (median 11.0 years of cumulative follow-up), the picture sharpened considerably:

| Timepoint | HR for Invasive Breast Cancer | 95% CI | Excess Cases per 10,000 Woman-Years | |---|---|---|---| | Active trial (5.6 yr) | 1.26 | 1.00, 1.59 | 8 | | Post-stopping (2.4 yr off drug) | 1.27 | 1.06, 1.51 | 9 | | Cumulative 13-year | 1.28 | 1.11, 1.48 |, | | Cumulative 18-year | 1.28 | 1.13, 1.45 |, |

Several things stand out. The hazard ratio barely changed across nearly two decades of follow-up, but the confidence interval narrowed as more events accrued. The 2010 analysis by Chlebowski et al. in JAMA further showed that breast cancers diagnosed in the CEE+MPA group were more likely to be node-positive and that breast cancer mortality was higher in the hormone group (HR 1.96 to 95% CI 1.00, 4.04). By the 18-year follow-up, breast cancer mortality in the CEE+MPA arm remained numerically elevated.

The biological explanation most cited: MPA promotes proliferation of existing subclinical breast lesions. Once those lesions progress past a point of no return (probably sometime during the 5.6 years of active therapy), stopping the hormone does not reverse the trajectory. This is consistent with the observation that the estrogen-alone arm of WHI (in hysterectomized women, without MPA) showed no breast cancer increase and possibly a decrease.

Cardiovascular Outcomes: The Attenuation Story

The cardiovascular data followed the opposite pattern. The CHD signal, which drove the early stopping, weakened substantially once women stopped taking study medication.

| Outcome | Active Trial HR | Post-Stopping HR | Cumulative 18-Year HR | |---|---|---|---| | CHD | 1.29 (1.02, 1.63) | 1.04 (0.84, 1.28) | 1.09 (0.96, 1.24) | | Stroke | 1.41 (1.07, 1.85) | 1.16 (0.90, 1.49) | 1.16 (1.00, 1.35) | | Pulmonary embolism | 2.13 (1.39, 3.25) | 1.29 (0.81, 2.04) | 1.56 (1.17, 2.08) |

The PE risk remained statistically elevated cumulatively, but the CHD excess essentially vanished within a few years of cessation. This argues against CEE+MPA causing permanent vascular damage and aligns with the "thrombotic and inflammatory" mechanism that operates primarily during active exposure.

The age-stratified re-analyses published by Manson et al. in NEJM 2013 provided further context. Women who started CEE+MPA at ages 50, 59 had no statistically significant increase in CHD during the trial or follow-up. The cardiovascular excess was concentrated in women who started therapy a decade or more after menopause. This "timing hypothesis" data, visible only through extension follow-up, fundamentally changed how clinicians interpreted the original trial.

Mortality: The Bottom Line

All-cause mortality is the outcome that cannot be gamed by competing risks. Across 18 years of cumulative follow-up in the JAMA 2017 analysis, the results were:

• All-cause mortality HR: 1.02 (95% CI 0.96, 1.08)
• Cardiovascular mortality HR: 1.01 (95% CI 0.90, 1.12)
• Cancer mortality HR: 1.03 (95% CI 0.92, 1.15)

No statistically significant difference in death from any major cause category. This is a striking finding given the alarm of 2002: after nearly two decades of follow-up with over 7,000 deaths, the overall mortality curves for the CEE+MPA and placebo groups were essentially superimposed. The risks (breast cancer) and benefits (fewer fractures, possibly fewer colorectal cancers during therapy) appear to have roughly balanced each other in aggregate.

The age-stratified mortality data showed a trend toward lower all-cause mortality in women who started at 50, 59 (HR 0.87 to 95% CI 0.76, 1.00), though this did not achieve conventional significance.

The Fracture and Colorectal Cancer Reversals

During active therapy, CEE+MPA reduced hip fractures (HR 0.66 to 95% CI 0.45, 0.98) and colorectal cancer (HR 0.56 to 95% CI 0.38, 0.81). Both protective effects, per the original WHI publication, were among the clearest benefit signals.

Post-cessation data showed these benefits disappeared rapidly. Hip fracture rates equalized within two to three years of stopping. The colorectal cancer protection also attenuated, with the cumulative HR moving toward 1.0 by the 13-year analysis. For clinicians, this means the fracture and colorectal benefits of combined HRT require continued treatment to maintain, unlike the breast cancer risk which persists after stopping.

Limitations the Authors Acknowledged

The WHI investigators have been transparent about several design features that complicate interpretation of the extension data.

Age at enrollment. The mean age was 63. Only 33% of participants were aged 50, 59. The extension data cannot fully resolve the timing hypothesis because the younger subgroup was underpowered for many endpoints. The Premarin/MPA FDA label reflects this by recommending the lowest effective dose for the shortest necessary duration, without age-specific risk tiers.

Adherence during the trial. By year 6, only 54% of women in the active arm were still taking their assigned medication. Nonadherence dilutes both risk and benefit estimates. The as-treated analyses (which censor women at discontinuation) generally show larger effect sizes, suggesting the intention-to-treat estimates understate the biological effects of CEE+MPA.

Post-trial crossover. Roughly 4 to 6% crossover in each direction during extension follow-up adds noise. The investigators adjusted for this in sensitivity analyses without materially changing the conclusions.

Single regimen tested. WHI tested one dose of CEE and one dose of MPA. It cannot speak to lower doses, transdermal estradiol, micronized progesterone, or other commonly used regimens. The Endocrine Society clinical practice guideline notes this limitation explicitly when citing WHI data.

Self-selection in extension. Women who consented to extension follow-up were healthier on average than those who did not. This survival bias could attenuate long-term risk signals.

What Changed in Clinical Practice

The extension data, accumulated over nearly two decades, gradually softened the absolutist "HRT is dangerous" message that followed the 2002 publication. Several professional society positions shifted:

The 2022 Hormone Therapy Position Statement from the North American Menopause Society now endorses initiation of HRT in symptomatic women under 60 or within 10 years of menopause, citing the age-stratified WHI extension data as key evidence that the benefit-risk ratio is favorable in this population.

The persistent breast cancer signal, however, remains the primary guardrail. Clinicians now distinguish clearly between estrogen-alone therapy (no breast cancer signal in WHI) and combined estrogen-progestogen therapy (persistent signal), a distinction the original 2002 media coverage failed to make.

The Unanswered Question

The WHI extension data tells us what happens after stopping one specific combined HRT regimen in predominantly older postmenopausal women. It does not tell us what would happen with modern formulations (transdermal estradiol plus micronized progesterone), earlier initiation, or different durations. The KEEPS and ELITE trials addressed some timing questions but were too small and short to capture hard endpoints like breast cancer or mortality. A definitive trial of early-initiation, modern-formulation HRT remains unfunded and, given the sample sizes required, may never be conducted.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Writing Group for the Women's Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women. JAMA. 2002;288(3):321-333. PubMed
2. Chlebowski RT, Anderson GL, Gass M, et al. Estrogen plus progestin and breast cancer incidence and mortality in postmenopausal women. JAMA. 2010;304(15):1684-1692. PubMed
3. Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women's Health Initiative randomized trials. JAMA. 2013;310(13):1353-1368. PubMed
4. Manson JE, Aragaki AK, Rossouw JE, et al. Menopausal hormone therapy and long-term all-cause and cause-specific mortality. JAMA. 2017;318(10):927-938. PubMed
5. Chlebowski RT, Anderson GL, Aragaki AK, et al. Association of menopausal hormone therapy with breast cancer incidence and mortality during long-term follow-up of the Women's Health Initiative randomized clinical trials. JAMA. 2020;324(4):369-380. PubMed
6. FDA. Premarin/medroxyprogesterone acetate prescribing information. AccessData
7. The 2022 Hormone Therapy Position Statement of The North American Menopause Society. Menopause. 2022;29(7):767-794. PubMed