The Hormone Therapy Puzzle: Why Decades of Research Still Can't Give Women a Clear Answer on Dementia Risk

For the 1.1 million American women who enter menopause each year, the decision of whether to take hormone therapy carries stakes that extend far beyond managing hot flashes. Alzheimer's disease disproportionately devastates women — nearly two-thirds of the 7.2 million Americans living with the disease are female [1] — and the question of whether the hormones that decline during menopause could protect or harm the aging brain has become one of the most consequential and confounding puzzles in modern medicine.

Now, a cluster of major studies published between late 2025 and early 2026 has simultaneously advanced and complicated that puzzle, offering new tools for assessing individual risk while underscoring just how far the field remains from a definitive answer.

The Million-Woman Meta-Analysis: No Clear Signal

The most comprehensive synthesis of evidence to date arrived in December 2025, when an international team led by Melissa Melville at University College London published a systematic review and meta-analysis in The Lancet Healthy Longevity [2]. Commissioned by the World Health Organization to inform forthcoming dementia-prevention guidelines, the review encompassed one randomized controlled trial and nine observational studies totaling 1,016,055 participants.

The conclusion was striking in its ambiguity: there was "no significant association between menopause hormone therapy and the risk of dementia or mild cognitive impairment" [2]. Subgroup analyses examining different timing windows, durations of use, and formulation types also returned null results. The researchers found no evidence that hormone therapy protects against dementia — but no consistent evidence that it increases risk either.

"The findings reinforce current clinical guidance that menopause hormone therapy should be guided by perceived benefits and risks and not for dementia prevention," said Professor Aimee Spector, the study's senior author at UCL [3]. The review will directly inform the WHO's updated guidelines on reducing the risk of cognitive decline and dementia, expected later in 2026.

Yet the Lancet review's reassuring neutrality tells only part of the story. Other studies published in the same period have pointed to biological mechanisms through which hormones could accelerate Alzheimer's pathology in certain women — mechanisms invisible to the epidemiological lens of meta-analysis.

The Tau Problem: Brain Imaging Reveals a Troubling Pattern

In March 2025, a team led by Gillian Coughlan and Rachel Buckley at Massachusetts General Hospital published findings in Science Advances that moved the conversation from population-level statistics to the molecular level [4]. Using PET brain imaging, the researchers tracked amyloid-beta and tau protein accumulation in 146 women aged 51 to 89 — half of whom had previously used hormone therapy, half of whom had not.

The results revealed a critical age divide. Among women over 70, those who had previously used hormone therapy showed significantly faster accumulation of tau protein in brain regions most vulnerable to Alzheimer's disease — the entorhinal cortex, inferior temporal gyrus, and fusiform gyrus [4]. This accelerated tau buildup was not observed in younger women. Nor was there any significant difference in amyloid-beta accumulation between the two groups.

The distinction matters because tau pathology tracks more closely with cognitive decline than amyloid plaques do. "Hormone therapy use in older women affects the pathological progression of tauopathy, with implications for cognitive decline," the researchers concluded [4].

Rachel Buckley, an associate professor of neurology at Mass General, has emphasized estrogen's fundamental role in the brain: "It's found in the hippocampus, which is an area that we know is very closely associated with memory and learning" [5]. The implication is that the same hormone that supports brain function before menopause may, under certain conditions and timing, interact with Alzheimer's pathology in harmful ways.

A Biomarker Breakthrough: Predicting Who Is Vulnerable

The most recent contribution arrived on March 10, 2026, when researchers published a study in JAMA Network Open drawing on data from 2,766 women enrolled in the Women's Health Initiative Memory Study between 1996 and 1999, with follow-up extending to 2021 [6][7].

The study examined plasma phosphorylated tau 217 (p-tau217), a blood-based biomarker for Alzheimer's disease that can now be measured with a simple blood test. The headline finding: higher baseline p-tau217 levels were associated with approximately three times the risk of developing dementia over the following 25 years [6].

But the study's most provocative finding concerned how that risk interacted with hormone therapy. Among women randomized to combined hormone therapy — estrogen plus progestin — higher p-tau217 levels were linked to roughly four times the dementia risk [7]. That amplified risk pattern was not observed among women taking estrogen-only therapy.

Professor Eef Hogervorst of Loughborough University, who analyzed the findings, noted that the association was strongest in women over 70, white women, and carriers of the APOE4 genetic variant — a well-established Alzheimer's risk factor [7]. The implication is that a blood test could, in theory, identify which women face heightened vulnerability before they begin hormone therapy — a prospect that would fundamentally change how treatment decisions are made.

The Progesterone Question

Across multiple studies, a consistent pattern has emerged: combined hormone therapy — estrogen paired with a progestogen — appears to carry greater cognitive risks than estrogen alone.

This divergence traces back to the Women's Health Initiative Memory Study (WHIMS), which enrolled women aged 65 and older beginning in the late 1990s [8]. In that trial, 40 of the 61 women diagnosed with probable dementia were in the estrogen-plus-progestin group, compared with just 21 in the placebo group — a hazard ratio of 2.05. The companion WHI Study of Cognitive Aging found that the combined regimen worsened verbal memory, while estrogen alone had no measurable effect on cognition [8].

The mechanism remains debated. Some researchers theorize that synthetic progestogens — particularly medroxyprogesterone acetate (MPA), the type used in the WHI — may antagonize estrogen's neuroprotective effects or even amplify neurotoxicity [9]. Whether bioidentical progesterone behaves differently from synthetic progestins is a question that current evidence cannot definitively answer, though it is one of the field's most active areas of inquiry.

The Timing Hypothesis: A Window That Opens and Closes

If the research landscape has a unifying thread, it is that when a woman begins hormone therapy may matter as much as whether she takes it at all.

A 2023 analysis led by Lisa Mosconi at Weill Cornell Medicine, examining more than 50 studies, found that hormone replacement therapy initiated within 10 years of menopause significantly reduced dementia risk [5]. A separate analysis found that women starting HRT within five years of menopause had up to 32% lower Alzheimer's risk compared to controls [10]. Estrogen-only therapy in midlife was associated with a 32% risk reduction, while combined estrogen-progestogen therapy showed a 23% reduction [10].

But these protective effects appear to reverse if therapy begins too late. The WHIMS trial demonstrated increased dementia risk in women who started hormones at 65 or older [8]. The Mass General imaging study showed accelerated tau accumulation specifically in women over 70 [4]. And the new JAMA Network Open biomarker study found its strongest risk signals in women beyond that same age threshold [6].

This pattern has given rise to what researchers call the "critical window" or "timing hypothesis" — the idea that estrogen is neuroprotective during a perimenopause window but potentially harmful when introduced to a brain that has already undergone significant aging or begun accumulating Alzheimer's pathology.

Source: Compiled from multiple studies including WHI/WHIMS, CNN/Mosconi 2023 analysis

Data as of Mar 15, 2026CSV

Why the Research Is So Difficult to Resolve

The persistent contradictions in this literature are not merely academic — they reflect fundamental challenges in studying hormones and the brain.

The landmark WHI trials, which shaped clinical practice for two decades, enrolled only women aged 65 and older — well past the typical age of menopause [8]. Their findings cannot be straightforwardly applied to younger women considering therapy for menopausal symptoms. The formulations studied (conjugated equine estrogens and medroxyprogesterone acetate) differ from the transdermal estradiol and micronized progesterone increasingly prescribed today. And observational studies, while often showing protective effects, are plagued by "healthy user bias" — women who choose hormone therapy may be healthier to begin with.

"We do not use hormone therapy for Alzheimer's prevention right now," Mosconi has cautioned. "Clinical guidelines currently do not endorse using hormone therapy only for Alzheimer's prevention" [5].

The $50 Million Bet on Answers

The scope of uncertainty has prompted a major new research investment. The CARE initiative (Cutting women's Alzheimer's Risk through Endocrinology), a $50 million global research program led by Mosconi at Weill Cornell Medicine, aims to analyze biomarkers from nearly 100 million women to understand why females face elevated Alzheimer's risk and how hormonal transitions contribute [5].

The program reflects a broader recognition that menopause — experienced by half the world's population — has been dramatically under-studied relative to its health consequences. The U.S. menopause treatment market was valued at $5.93 billion in 2024 and is projected to reach $9.12 billion by 2033, driven in part by growing awareness and the destigmatization of menopause [11]. Utilization of menopausal hormone therapy among women with symptoms has risen from 12.9% in 2022 to 15.4% in 2024 [11].

Source: JAMA Health Forum

Data as of Mar 15, 2026CSV

What Women and Their Doctors Should Know Now

For the millions of women weighing hormone therapy decisions today, the current evidence supports several practical conclusions, even amid the broader uncertainty:

Menopause symptom management remains the primary indication. The Lancet meta-analysis confirms that hormone therapy should not be taken solely to prevent dementia, nor should dementia fears alone drive women away from treatment they need for debilitating symptoms like hot flashes, sleep disruption, and mood changes [2].

Timing matters. The weight of evidence favors initiating therapy close to menopause onset — ideally within five to ten years — rather than later in life. Current clinical guidelines recommend against starting hormone therapy after age 65 or more than 10 years post-menopause [4][10].

Type may matter. Combined estrogen-progestogen therapy has consistently shown more concerning cognitive signals than estrogen-only therapy, though the latter is only appropriate for women who have had a hysterectomy [7][9].

Individual risk assessment is advancing. The p-tau217 blood test, while not yet recommended for routine clinical screening, represents a potential future tool for identifying women who may be particularly vulnerable to adverse cognitive effects from certain therapies [6].

The APOE4 genotype is relevant. Women carrying this genetic variant may face amplified risk from combined hormone therapy and could benefit from genetic testing to inform their decisions [7].

The WHO's forthcoming guidelines, expected in 2026, may provide updated clinical frameworks that incorporate these nuances. Until then, the most honest message the research can offer is one of informed individualism: the right answer depends on who you are, when you start, and what you take — and science is still working out exactly how those variables interact.