The Risk/Reward Window on Hormone Therapy: What It Is, and What It Isn't
Aug 23, 2026Written by: Meredith Paci, Functional Health Coach
There is a reflex in medicine that sounds responsible, yet arguably, it's not: if it hasn't been proven in a randomized controlled trial, treat it as unproven, full stop, no further reasoning allowed. I understand where the reflex comes from. I really do. It's also not how good clinical judgment works, and it's not how the hormone therapy risk/reward window got established in the first place.
There's a second assumption riding along with that reflex, and it's the one I want to argue with here: that withholding treatment is the safe, no-cost option. It isn't. Telling a woman she doesn't qualify, or that it's too soon to know, or that the safest move is to wait, is still a decision with consequences attached to it if made not with the individual context. It just moves those consequences onto her instead of onto whoever made the call.
In July, I wrote about the "too old" myth and walked through the position from The Menopause Society (formerly NAMS, the North American Menopause Society, before its 2023 rebrand), the reanalysis of the WHI (the Women's Health Initiative, a large federally funded study that began in the early 1990s, whose original 2002 findings were reanalyzed by age and time since menopause in 2013), and the vaginal estrogen data that flatly contradicts a blanket age cutoff.
This blog is going a little deeper, because there's a question underneath all of that I didn't answer yet: why does timing matter at all? Not just what the data shows, but why the body responds the way it does, why the same molecule turns out to be protective in one woman and neutral, or riskier, in another. Answering that changes how we think about the window, because it stops being a calendar rule and starts being a tissue-state rule.
The window has a mechanism, not just a statistic
The most common critique of the timing hypothesis is that it comes from a post-hoc subgroup analysis, meaning researchers split WHI participants into groups by age and time-since-menopause after the trial had already run, rather than that split being built into the original design. That's a fair critique of how the hypothesis was first noticed. It is not a fair description of where the evidence stands now.
The Early versus Late Intervention Trial with Estradiol, ELITE, was built specifically to test this. Healthy postmenopausal women were randomized to oral estradiol or placebo, stratified by time since menopause: under 6 years versus 10 or more. After a median of 5 years, the early group on estradiol showed significantly slower progression of carotid intima-media thickness, or CIMT (in plain terms, a measurement of how much the artery wall is thickening, which reflects how much plaque is building up), compared to placebo. The late group showed no difference between estradiol and placebo at all. Same molecule, same dose class, same trial. The only variable that moved the outcome was how long the vessel had been without estrogen before treatment started.
I want to flag exactly what that trial showed and didn't show. CIMT is a surrogate marker, a proxy for arterial aging, not a count of heart attacks or strokes. ELITE showed that early estradiol slowed the arterial change. It didn't measure a reduction in cardiovascular events. This is super positive in my opinion yet staying neutral is evidence that is one step removed from the hardest outcome we'd want to see.
Why the vessel responds differently: the healthy endothelium hypothesis
This has a name and a mechanism that predates WHI entirely, developed largely out of decades of primate atherosclerosis research, with Thomas Clarkson's group producing much of the foundational data. The healthy endothelium hypothesis holds that estrogen's effect on an artery depends on the condition of that artery when estrogen shows up.In endothelium that's still intact, estrogen supports vasodilation, has anti-inflammatory effects on the vessel wall, and appears to slow the initiation of atherosclerotic changes. Once an artery already has established plaque, that same estrogen signaling doesn't get the same substrate to act on, and in some contexts can behave differently around plaque stability.
Don’t stop reading!!!
That finding is the leading biological explanation for why the window exists: time since menopause is a proxy for how likely it is that the vessel is still in the "healthy endothelium" state versus already carrying subclinical disease. This is why, if you're a mentee, or you've ever sat in mentorship with Sarah and me, you've heard me say it constantly: how healthy is the host. A 61 year old with a clean vascular profile is not the same case, mechanistically, as a 54 year old with early plaque burden, even though the calendar says the younger woman should be the "safer" candidate.
The same shape shows up in the brain
Digging into this, I came across the work of Dr. Lisa Mosconi, a neuroscientist who leads the Women's Brain Initiative at Weill Cornell and has spent years using PET imaging to track what happens to the brain across the menopause transition. Her research found that as women move through perimenopause and menopause, the brain's ability to use glucose for fuel drops in several regions, something she calls a brain energy crisis, and that estrogen receptor density in the brain increases during this same window, which suggests the brain may be upregulating receptors to compensate for a hormone it's no longer getting enough of.
Separately, researcher Roberta Brinton has proposed a mechanism for why the brain's response might depend on the state of the neuron, something she calls the healthy cell bias of estrogen action. I want to be upfront that this isn't research I've worked with directly. I came across it while digging into why the timing pattern in the brain looks the same as the timing pattern in the vessel, and I think it's worth including because the logic holds up: estrogen signaling in neurons converges on the mitochondria and increases energy metabolism, which is a benefit in a neuron that's still functioning normally. In a neuron that's already compromised, with calcium regulation already disrupted, that same convergence point becomes a liability instead of a benefit.
And the clinical data tracks with that, though not equally on both sides of the WHIMS data. WHIMS, which studied women 65 and older, found that CEE plus MPA significantly raised dementia risk, HR 1.76. The estrogen-alone arm trended the same direction, HR 1.49, but didn't reach statistical significance, so I'm not leaning on it the same way. Both arms are a late-initiation, older-brain population, which is where the mechanism predicts a worse response, but the combined-hormone finding is the one carrying real statistical weight. On the other side, the KEEPS Continuation study followed women who'd started transdermal or oral estradiol within three years of menopause, and found no long-term cognitive harm a decade later. I want to be precise here: that's a reassurance of safety, not proof of cognitive benefit. KEEPS didn't show hormone therapy protects cognition. It showed early initiation didn't hurt it. Those are two different specific claims, and I'm not going to blur them into one because when you are creating a study you need to keep the goal the goal. ‘This is what we saw, this is what we didn’t” Yet! Me being me, I can’t help but see the benefit…
What it means when two independent mechanisms agree
This is what I find genuinely important here, and it's the direct answer to the "nothing's proven" objection, though I want to be precise about how strong each piece is, because it isn't all carrying the same weight.
The vascular side is the tightest evidence in this piece: a mechanism proposed from primate research before WHI existed, and a trial, ELITE, built specifically to test that mechanism's prediction, which confirmed it. Mechanism and purpose-built trial, matched to each other. That's pretty dang solid evidence.
The brain side is a looser parallel, and I'd rather say that plainly than let it sound equally airtight. Mosconi's and Brinton's mechanisms were derived independently of each other, which is very important. But WHIMS and KEEPS Continuation are two different trials, different populations, different endpoints, not an early-versus-late arm of the same study the way ELITE was designed. Lining them up gives you a pattern that matches what the mechanisms predict. It is not the same as a trial built specifically to test timing in the brain the way ELITE tested it in the vessel.
So what you have is one tight, matched confirmation in the vasculature, and one arguably super suggestive parallel in the brain. Both point the same direction, and I'm not going to flatten the difference between them for a cleaner argument. A dedicated late-initiator brain trial doesn't exist, and yes that absence is a gap, a bigger one on the neurological side than the vascular one.
What this doesn't give you permission to do
None of this means mechanism-based reasoning is a blank check. The precise cutoffs, 6 years, 10 years, age 60, come out of real, important research, the trial design choices and subgroup boundaries that got us this far, and those studies are important. But those cutoffs aren't a line that exists in nature at that exact point, and, from what I can tell, they were never meant to function as one. All of this should run on individual risk exposure: lipid history, inflammatory load, blood pressure history, blood glucose regulation, genetics. This is why Sarah and I are so adamant about starting now, starting today, and walking forward with your health, because two women at the same number of years since menopause can have very different amounts of "healthy endothelium" left to work with.
Which is really the point I want to land on: this isn't a case for swapping one blanket rule for another, starting everyone or withholding everyone past a certain point. It's a case of looking at the individual woman sitting in front of you, her labs, her vascular and metabolic history, her risk profile, and weighing that against the population-level pattern, instead of letting either the pattern or the fear override her individual case.
Where I, Meredith Paci, land, today, August 2026:
Here's what we're working with: a named mechanism showing up in two different tissue systems, a trial that was built specifically to test one of them, and real-world data in both systems lining up with what that mechanism predicts. That's a well-supported pattern, and coaches can feel steady saying so.
What isn't settled: the exact cutoffs, whether individualized vascular and metabolic risk markers can meaningfully widen or narrow that window on a person-by-person basis, and what a late initiator's real risk looks like on modern hormonal regimens that weren't the ones studied in WHI or ELITE.
I'll also say plainly that where I land on this in five months, or next year, could look different than where I'm landing right now. That's not hedging. That's what happens when you stay committed to growth in research and stay in the room with real clients long enough: experience updates you, and so does science. Both are still moving. Hold the confidence where the evidence earns it, hold the uncertainty where it belongs, and don't round either one up.
Where this lands for me in practice: it was never going to be settled by a rule that works for every woman regardless of who she is. It gets settled by sitting with her specific labs, her specific history, her specific risk profile, and weighing that against what the population data shows, instead of letting either the population data or the fear override her individual case. And I'll say the uncomfortable part directly: telling her no because the individualized picture is inconvenient to gather isn't the cautious choice, it is a reckless one that carries its own risk.
If you're a coach and you want in on a group mentorship built to train you at this level, please reach out. Women need more of you and you need to be ready.If you want something more direct, we also do 1:1 strategy sessions.
Contact us about group mentorship · Inquire about 1:1 strategy sessions
If you have a question of your own, the anonymous form that started this series is still open.
Meredith Paci is the co-founder of Fortify Health Coaching. I am not a licensed medical provider. The content in this post is educational and does not constitute medical advice or a patient-provider relationship. If you are considering hormone therapy, work with a qualified and well experienced provider who will assess your individual history.
Sources referenced:
