Hormone Therapy and Memory: What Three Trials Measured, Two Decades Apart
The Reading Room, Vol. 19 · 10 min read · What the trials tested, why age at starting changes the answer, and the gap between what was measured and what is being asked
Based on the published research of Carey E. Gleason, PhD — Associate Professor, Division of Geriatrics and Gerontology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, and co-lead investigator on the KEEPS-Cognitive and Affective Study. Faculty profile · KEEPS-Cog paper
Does HRT help brain fog? Three randomised trials measured cognition in more than 5,000 women. None of them found a cognitive benefit, and none of them measured the symptom women actually describe.
The question is usually asked in an appointment lasting ten minutes, after the word that gets used is “fog”, and it deserves a better answer than the two that circulate. One says hormone therapy protects the brain. The other says it causes dementia. Both are drawn from the same body of trials, read badly, in opposite directions.
What follows is what those trials actually did, in whom, and at what age.
The short answer:
- KEEPS-Cog randomised 693 women, mean age 52.6 and 1.4 years past their final period, to two hormone regimens or placebo for up to 4 years. No treatment-related cognitive benefit was found.¹
- The same women were reassessed roughly 10 years later. 275 had usable data from both points, and there were no long-term cognitive effects of the earlier short-term exposure.²
- WHIMS randomised 4,381 women aged 65 and over, mean age 71, and found the risk of probable dementia doubled on combined therapy: HR 2.05 (95% CI 1.21-3.48).³
- That was 1 case per 222 women on treatment against 1 per 455 on placebo, and the trial’s own authors noted applicability to women under 65 was unknown.³
- In KEEPS-Cog, oral estrogen improved symptoms of depression and anxiety. Transdermal did not.¹
The source, and who else has measured this
Carey E. Gleason, PhD is Associate Professor in the Division of Geriatrics and Gerontology, Department of Medicine, at the University of Wisconsin School of Medicine and Public Health, and was co-lead investigator on the KEEPS-Cognitive and Affective Study. Of this line of research, she has said: “Recent findings have added to our understanding of the complex effects of hormones on the brain.” Faculty profile · KEEPS-Cog paper
Who else has measured this: the contrasting finding above WHIMS came from a team led by Sally A. Shumaker, PhD (Faculty profile), Professor of Public Health Sciences at Wake Forest University School of Medicine, who followed more than 4,000 women aged 65 and over. Read the WHIMS trial. Both teams published in peer-reviewed journals with named authors and open data; neither has reviewed, approved or endorsed this article, and Blue Leaf Journal has no commercial or other relationship with either institution.
Scope and method of this review
This piece covers what randomised trials found when they gave menopausal hormone therapy and then measured cognition. It draws on three: the cognitive arm of the Kronos Early Estrogen Prevention Study and its ten-year continuation, and the Women’s Health Initiative Memory Study.¹²³
Sources were selected on one criterion: randomised, placebo-controlled trials with cognition as a prespecified outcome. Observational studies were excluded, because the question of who chooses hormone therapy is precisely the confound that made this literature confusing for twenty years.
What this review does not cover: whether hormone therapy relieves hot flashes, night sweats, or genitourinary symptoms, which is a separate and much stronger evidence base; cardiovascular or breast outcomes; and any recommendation about whether an individual woman should take it, which is a clinical decision that depends on her history.
One limitation deserves stating at the top rather than the bottom. These trials measured performance on cognitive tests. Brain fog, as women describe it, is a subjective experience of slowed thinking, lost words and effortful recall. The two are related and they are not the same outcome, and no trial in this review measured the second one directly.
What the three trials found
| Trial | n and population | Finding | Limitation |
|---|---|---|---|
| KEEPS-Cog (2015), randomised, placebo-controlled, up to 4 years | 693 women, mean age 52.6, 1.4 years post final period | No treatment-related benefit on global cognition or any of four domains | Mean cognitive follow-up 2.85 years; healthy volunteers |
| KEEPS-Cog, mood outcomes | Same cohort | Oral estrogen improved depression and anxiety symptoms; transdermal did not | Secondary outcome; effect sizes small to medium |
| KEEPS Continuation (2024), observational follow-up | 275 women with data at both points, 8 to 14 years later | No long-term cognitive effects of the earlier short-term exposure | Under half the original cohort; those who return differ from those who do not |
| WHIMS (2003), randomised, placebo-controlled | 4,381 women aged 65 and over, mean age 71 | Probable dementia risk doubled on combined therapy, HR 2.05 (1.21-3.48) | Started an average of two decades after menopause; one formulation only |
| WHIMS, absolute numbers | Same cohort | 1 case per 222 treated against 1 per 455 on placebo | Authors stated applicability to women under 65 was unknown |
Sources: Gleason et al. (2015), PLOS Medicine 12(6):e1001833; Gleason et al. (2024), PLOS Medicine, 10.1371/journal.pmed.1004435; Shumaker et al. (2003), JAMA.¹²³
Read together, the three produce one coherent picture and two headlines that should never have been written.
The picture is that hormone therapy started close to menopause, in healthy women around 52, neither helped nor harmed cognition over four years, and had no detectable effect a decade later.¹² The same class of treatment, started at 71 in women two decades past menopause, raised the risk of dementia.³
The headline “hormone therapy protects the brain” has no randomised support in this literature. The headline “hormone therapy causes dementia” is a finding about women aged 65 and over, on one specific formulation, which the trial’s own authors declined to generalise downward.³
If the fog is arriving on days that follow particular nights or particular weeks, that pattern is worth having in front of you before an appointment. The Quiet Audit is a private ten-minute pass through what the week is carrying. Start here.
Key terms
Cognitive domain means a grouped set of mental abilities measured together, such as verbal learning, working memory or processing speed. KEEPS-Cog assessed four of them plus a global score.¹
Hazard ratio compares the rate of an event between two groups. A hazard ratio of 2.05 means the event occurred at roughly twice the rate in the treated group. It says nothing on its own about how common the event was.
Conjugated equine estrogens are the oral estrogen formulation used in both WHIMS and one KEEPS arm. Transdermal estradiol is estrogen delivered through the skin by patch, the other KEEPS arm. They behave differently and the trials treated them separately, which is why “HRT” as a single category is not a useful unit of analysis.
Probable dementia is a trial endpoint reached through a defined diagnostic protocol, not a clinical diagnosis made at a single visit.
How strong is this evidence?
Moderate, with an important qualification about what was asked. The evidence that hormone therapy does not improve cognitive test performance in recently menopausal women is good: a randomised placebo-controlled trial and a ten-year follow-up agree.¹² The evidence of increased dementia risk when treatment begins after 65 comes from a large randomised trial and is also good, within that population.³
What the grade does not cover, and this is the whole difficulty: none of these trials measured brain fog. They measured performance on cognitive tests in healthy volunteers. A woman losing words in meetings is not reporting a test score, and the literature is close to silent on whether treatment changes her experience. The most plausible route by which it might, indirectly, is through sleep and symptom relief, and that has not been tested as a cognitive question.
Where the argument goes wrong in both directions
Two readings dominate the search results and both fail on the same detail, which is age at initiation.
The protective reading generally cites observational data showing that women who took hormone therapy had better cognitive outcomes. Those women also differed systematically in education, health and access to care, which is exactly what randomisation exists to remove. When the question was randomised in women close to menopause, the benefit did not appear.¹
The alarming reading cites WHIMS without its age. The trial enrolled women 65 and over, average age 71, beginning treatment roughly two decades after their final period.³ Applying that result to a woman of 52 considering treatment for hot flashes is applying a finding outside the population it was measured in, and the trial’s own authors said the applicability to women under 65 was unknown.³
What survives both errors is narrow and useful: for cognition specifically, in women near menopause, the honest answer from randomised evidence is no measurable effect either way. The broader question of what the transition does to the brain is a different one with a more encouraging answer, and the question worth taking to the appointment is not “will this fix the fog”.
When this belongs with a doctor
Nothing in this review is a reason to start, stop or change hormone therapy. That decision depends on symptoms, personal and family history, age, time since menopause and formulation, and it belongs with a clinician who knows all of them. What is worth raising promptly rather than at the next routine appointment: memory change that others have noticed before you did, getting lost on familiar routes, difficulty with words that is worsening month on month, or any cognitive change that arrived suddenly. Those are assessed differently from the gradual, effortful, word-finding version most women mean by fog. The treatment itself is set out separately here.
The one move
The one move
Before the appointment, write down which symptom you most want treated, in one sentence, and keep the fog separate from it. The trials support hormone therapy for vasomotor symptoms and do not support it for cognition, so a conversation that names one target is a conversation that can actually be answered.
If the fog turns out to track the weeks rather than the hormones, the Quiet Audit is a private pass through what those weeks are carrying. Start here.
Questions this review is asked
Does HRT help brain fog?
No randomised trial has shown a cognitive benefit. KEEPS-Cog found none in 693 recently menopausal women over up to four years, and a ten-year follow-up found none either.¹² No trial measured subjective brain fog directly, so the question as women ask it remains formally unanswered.
Does it cause dementia, then?
In women aged 65 and over starting combined therapy, WHIMS found the risk of probable dementia doubled, from 1 case per 455 to 1 per 222.³ The trial’s authors stated that applicability to women under 65 was unknown, and later trials in younger women found no cognitive harm.¹²³
Why do the two sets of results look so different?
Age at initiation is the main candidate. WHIMS treated women an average of two decades past menopause; KEEPS treated women 1.4 years past it.¹³ That difference in timing is the leading explanation, and it is a hypothesis supported by the pattern rather than a proven mechanism.
Did anything in these trials improve?
Yes. In KEEPS-Cog, oral estrogen improved symptoms of depression and anxiety against placebo, with small to medium effect sizes. Transdermal estradiol did not show the same mood benefit.¹ That is a mood finding, not a cognitive one.
If the fog is not hormonal, what is it?
That framing is too binary. Fragmented sleep, sustained cognitive load and low mood all degrade the kind of thinking women describe as fog, and all three are common in this decade for reasons that have nothing to do with a prescription. Ruling those in or out is usually more productive than treating fog as a hormone question first.
Should this stop me taking hormone therapy for hot flashes?
Nothing here speaks to that. Vasomotor symptoms are the indication with the strongest evidence base, and this review deliberately did not assess it. The point is only that cognition should not be the reason on the form.
Informational, not medical advice. Decisions about hormone therapy, and any memory change that worries you or that others have noticed, belong with your clinician. Blue Leaf Journal is not affiliated with, and has no commercial relationship with, Dr. Gleason, Dr. Shumaker, or their institutions; neither has reviewed, approved or endorsed this article. Checked against the primary sources linked above on 6 September 2026. Corrections: norawhitfield@blueleafjournal.com.
Two headlines that should never have been written.
References
1. Gleason, C. E., Dowling, N. M., Wharton, W., et al. (2015). Effects of Hormone Therapy on Cognition and Mood in Recently Postmenopausal Women: Findings from the Randomized, Controlled KEEPS-Cognitive and Affective Study. PLOS Medicine, 12(6), e1001833. Read the trial
2. Gleason, C. E., Dowling, N. M., et al. (2024). Long-term cognitive effects of menopausal hormone therapy: Findings from the KEEPS Continuation Study. PLOS Medicine. Read the follow-up
3. Shumaker, S. A., Legault, C., Rapp, S. R., et al. (2003). Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study. JAMA, 289(20), 2651-2662. Read the trial
Miriam Alderton is Research Editor at Blue Leaf Journal. She reads the methods section first and the abstract last, and every figure in this piece is linked to its source above.
Written by Miriam Alderton, Research Editor, for Blue Leaf Journal. Updated: 31 August 2026.
