Part 6 — Hormones And The Brain: Timing, Form, And How We Know

The hormones-and-cognition conversation is one of the oldest and most relitigated debates in women’s health. A big study in 2003 (the Women’s Health Initiative Memory Study) scared a generation of doctors off hormone replacement therapy. Two later studies (KEEPS and ELITE) partially rehabilitated it by showing that two things matter: when you start, and which form you take. In 2025, the FDA began the process of removing broad boxed-warning language from menopausal hormone therapy, and in 2026 it approved labeling changes that removed boxed-warning language about cardiovascular disease, breast cancer, and probable dementia from several products. Early 2026 closed the loop with brain imaging and mechanism, and the clinical takeaway is now as clear as a precision-medicine recommendation gets.

The FDA acts

Menopause hormone therapy has been one of the most misunderstood interventions in modern medicine. For over two decades, the legacy of the Women’s Health Initiative (WHI) cast a long shadow, leading to black box warnings and a culture of fear that discouraged both clinicians and patients from considering hormone replacement, even when the potential benefits were substantial. The key problem was generalization. WHI and WHIMS asked a late-start, population-level question. Many women making decisions today are asking an earlier, symptom-driven, formulation-specific question.

That distinction is the whole story: WHIMS found increased dementia risk when estrogen plus progestin was started in women 65 and older (Shumaker 2003). That is not the same clinical question as transdermal estradiol, started near menopause, in a symptomatic woman with brain-risk factors. The 2024 JAMA review led by Manson made the same distinction: hormone therapy is not a chronic-disease-prevention drug, but the risk-benefit profile is more favorable for symptomatic women younger than 60 or within 10 years of menopause (Manson 2024).

Under Dr. Marty Makary’s leadership at the FDA, this evolving evidence base has begun to reshape national policy. The agency is rolling back broad black box warnings on menopausal hormone therapy and reframing risk to emphasize timing, dose, route, and individual risk factors rather than one size fits all fear (FDA 2025; FDA 2026). The hormone question is no longer “safe or unsafe?” It is “which woman, which hormone, which route, and which decade?”

What brain MRIs across many studies now show

He and colleagues pulled together the brain imaging evidence on menopausal hormone therapy from many studies across many years (He 2026). The consistent picture across those studies:

• Early initiation matters. Starting hormone therapy during the menopause transition or within the first few years after menopause is associated with preserved gray matter, healthier hippocampus (the memory structure), and better white matter. Starting more than 5 to 10 years after menopause is neutral at best, and in some studies, unfavorable.
• Form matters. Transdermal estrogen — the kind delivered through a patch or gel absorbed through the skin — consistently does better on these brain imaging outcomes than oral estrogen (a pill). Transdermal skips the first pass through the liver and avoids some of the downstream effects on clotting and hormone-binding proteins.
• APOE ε4 makes this conversation bigger. For women carrying ε4, the interaction between estrogen loss and Alzheimer-related biology appears larger than for non-carriers. Timing and form matter more, not less.

The mechanism study

Salathe and the University of Kansas Medical Center group did careful proteomics — measuring thousands of proteins in the brain — in a model of surgical ovarian loss and estrogen replacement (Salathe 2026). Losing ovarian function disrupted specific protein networks in the brain. Estrogen replacement partially reversed the disruption. This is not vague “neuroprotection,” but specific, measurable, reproducible protein changes of the kind any pharmacologist would demand before accepting that a drug has a real mechanism of action.

“Partially reversed” is an important qualifier. Some protein networks recovered with replacement, while others did not, which means the timing of replacement matters biologically, not just clinically: a brain that’s been estrogen-deprived for ten years may have accumulated changes that a brain replaced earlier has not. This is the molecular story underneath the clinical timing window.

What this means for you

Four developments now point in the same direction: regulatory recognition by the FDA, behavioral outcomes from earlier KEEPS and ELITE data, brain structure from the 2026 MRI review, and molecular mechanism from the 2026 proteomics study. That is not a guarantee, but it is a much clearer signal than women were given for the last twenty years.

If you are a woman approaching menopause, in perimenopause, or in the first few postmenopausal years:

• The right time to have the hormone-therapy conversation with a clinician is now, not ten years after your last period. The window of benefit for brain outcomes closes.
• If you and your clinician decide to proceed, transdermal estradiol (patch or gel) is preferred over oral estrogen for brain-relevant outcomes.
• If you still have a uterus, progesterone is added. Micronized progesterone is preferred over synthetic progestins, again for the brain-relevant safety profile.
• Dosing should aim for symptom control and normal physiologic estradiol levels, not supraphysiologic levels.
• Follow-up includes tracking cognition along with the usual breast and uterine monitoring.
• If you carry APOE ε4, the case for an active conversation is stronger. This is not automatic prescribing. It is earlier and more deliberate decision-making.

This is a specific recommendation for a specific patient situation, not a universal recommendation, and it is grounded in evidence that finally lines up.

What this means in our clinic

In practice, we open the hormone conversation at perimenopause, not at the first memory complaint a decade later. Route is transdermal estradiol. Progesterone is added when needed, preferring micronized. Dosing targets symptom control plus estradiol levels in the physiologic range. And the decision to continue beyond the first five postmenopausal years is a deliberate conversation, not a default, informed by APOE status, family history, and cognitive trajectory.

References — Part 6

1. He Z, Wang Y, Tang W, Li J. Timing and route of menopausal hormone therapy critically shape structural brain outcomes: A systematic review of MRI evidence. Neuroscience & Biobehavioral Reviews. 2026:106684. doi: 10.1016/j.neubiorev.2026.106684
2. Salathe SF, Franczak E, Busick Z, Boakye FB, Allen J, Lutkewitte A, Wilkins H, Thyfault J, Kugler BA. Loss of ovarian function and estrogen therapy remodel the brain’s synaptic and metabolic proteome. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology. 2026. doi: 10.1152/ajpregu.00031.2026
3. Shumaker SA, Legault C, Rapp SR, Thal L, Wallace RB, Ockene JK, Hendrix SL, Jones BN, Assaf AR, Jackson RD, Kotchen JM, Wassertheil- Smoller S, Wactawski-Wende J. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study. JAMA. 2003;289(20):2651-2662. doi: 10.1001/jama.289.20.2651
4. Manson JE, Crandall CJ, Rossouw JE, Chlebowski RT, Anderson GL, Stefanick ML, et al. The Women’s Health Initiative Randomized Trials and Clinical Practice: A Review. JAMA. 2024;331(20):1748-1760. doi: 10.1001/jama.2024.6542
5. U.S. Food and Drug Administration. HHS Advances Women’s Health, Removes Misleading FDA Warnings on Hormone Replacement Therapy 2025. URL: https://www.fda.gov/news-events/press-announcements/hhs-advances-womens-health-removes-misleading-fda-warnings-hormone-replacement-therapy
6. U.S. Food and Drug Administration. FDA Approves Labeling Changes to Menopausal Hormone Therapy Products. 2026. URL: https://www.fda.gov/news-events/press-announcements/fda-approves-labeling-changes-menopausal-hormone-therapy-products