Timing is everything: What AI need to learn about HRT and brain health

By Morgan Rose, CNM, WHNP-BC, and Dr Kudesia, triple board-certified in Reproductive Endocrinology & Infertility (REI), Obstetrics & Gynecology, and Lifestyle Medicine

The timing of when women start hormone replacement therapy (HRT) may matter more than we ever understood.

The “critical window” for starting HRT isn’t just relevant to brain health; timing has also been shown to shape cardiovascular outcomes.

Early analyses of the landmark WHI trial missed this nuance, including women long past menopause and obscuring benefits seen in those who begin therapy sooner.

Recent research presented at the American Neurological Association Annual Meeting found that women who began HRT within five years of menopause had a 32 per cent lower risk of developing Alzheimer’s disease.

Yet those who started HRT more than five years after menopause showed higher levels of tau proteins, a hallmark of Alzheimer’s, in their brains.

This “critical window” concept challenges the one-size-fits-all model of menopause care. It suggests that when we start HRT, it can be just as important as whether we start it.

Researchers at Mass General Brigham found similar patterns: women who began HRT near the onset of menopause did not show increased Alzheimer’s risk, while those who experienced early menopause with longer lifetime estrogen deprivation faced a higher risk overall.

Why Timing Matters

Women face roughly a 1 in 5 lifetime risk of Alzheimer’s disease, compared to 1 in 10 for men of the same age. The sharp drop in estrogen during menopause may help explain that gap.

Estrogen supports energy production, blood flow, and inflammation control in the brain functions that wane as levels decline.

The data suggest that initiating HRT earlier (around perimenopause or early menopause) may preserve some of these neuroprotective effects, while starting too late could heighten risk once neurodegenerative changes have already begun.

One meta-analysis showed that HRT begun soon after menopause reduced Alzheimer’s risk by over 40 per cent compared to non-users, while starting after age 60 doubled the risk.

These findings don’t make HRT a blanket prescription; they make it a precision intervention.

The Clinical Nuances AI Must Understand

If AI is going to support menopause care safely and intelligently, it needs to understand the reasons behind the actions.

1. Timing Is a Risk Modifier

“Time since menopause” is not just a demographic; it’s a clinical determinant.

AI systems must understand that the same therapy can carry opposite implications depending on when it’s initiated. 

2. HRT Is Not Monolithic

There are multiple formulations, routes, and combinations.

For women who still have a uterus, adding progesterone helps protect the uterine lining and lowers the risk of endometrial cancer that can occur with estrogen-only therapy.

Topical estrogen, which has much lower absorption into the bloodstream, may be appropriate for perimenopausal women who are still having periods or postmenopausal women with a uterus who need targeted relief of vaginal or urinary symptoms.

For example:

• • Estrogen-alone therapy can raise the risk of endometrial cancer and is typically not recommended for women with a uterus.
  • Topical estrogen, with much lower systemic absorption, may be appropriate for perimenopausal women still having periods.
    These distinctions must be explicitly encoded into AI frameworks to avoid oversimplified or unsafe recommendations.

3. Safety Requires Scope and Referral

Complex topics like HRT timing and neuroprotection should always prompt an automatic follow-up:

“Please consult a menopause specialist or clinician (such as one credentialed through the North American Menopause Society) to discuss the risks, benefits, and best options for you.”

AI logic must mirror how clinicians practice by inviting deeper discussion, not replacing it.

From Symptoms to Systems Thinking

We need to move beyond viewing menopause as a set of symptoms to manage, and instead see it as a neurological and metabolic inflection point in women’s lives.

That means:

• Bringing conversations about HRT earlier, during perimenopause, when neuroprotective benefits may still be possible.
• Designing AI systems that recognise context and chronology, not just keywords.
• Making personalised, evidence-based menopause guidance accessible to every woman by cutting through misinformation and connecting her to trusted care.
  

The Bigger Picture

For too long, women’s midlife health has been under-researched and under-resourced. The result is a data gap, which can quickly become a bias when encoded into AI.

If we want women’s health AI to truly care, it must be trained on data that understands the complexity of hormonal transitions, not just the vocabulary of them.

Because the difference between “now” and “five years from now” can determine whether a woman ages with clarity or confusion.

AI should know that.

And soon, it will.

Dr Kudesia is nationally recognised for her expertise in fertility awareness, lifestyle, and culinary approaches to reproductive health, and her advocacy for reproductive rights.