A systematic review and meta-analysis covering more than a million women found no clear overall association between menopause hormone therapy and the risk of dementia or mild cognitive impairment. Published in Lancet Healthy Longevity, the study pooled data from both observational research and randomized trials, arriving at a conclusion that conflicts with at least one major national registry study. The result lands as millions of women weigh hormone therapy decisions against a backdrop of contradictory evidence that has persisted for more than a decade.
Why the dementia question divides researchers and patients
The tension behind this finding is straightforward: women approaching or experiencing menopause are frequently told that hormone therapy carries cognitive risks, yet the largest pooled analysis to date does not support that warning as a blanket rule. The Lancet meta-analysis synthesized evidence across study designs and hormone formulations, and its top-line message is that the data do not point to a consistent increase or decrease in dementia risk.
That conclusion, however, sits uneasily alongside a Danish nationwide nested case-control study that reported positive associations between menopausal hormone therapy and dementia, including among women treated at age 55 or younger. The Danish study drew on national registries covering the full population, giving it statistical power that smaller trials cannot match. The disagreement between these two bodies of evidence is not a minor academic footnote. It shapes prescribing decisions, patient anxiety, and clinical guidelines across multiple countries.
One explanation for the conflicting results is that existing datasets have not stratified participants by genetic variants that affect estrogen metabolism. If certain women process estrogen differently at the cellular level, the same hormone regimen could carry different cognitive consequences depending on genotype. Neither the large registry studies nor the clinical trials published so far have tested this hypothesis with individual-level genomic data, which means the observed heterogeneity across studies could reflect real biological differences that current research designs are not equipped to detect.
Another source of tension is how risk is communicated. Many women hear about dementia findings from older hormone therapy trials that enrolled participants in their mid-60s or beyond, long after the typical age of natural menopause. Applying those results to women in their early 50s may be misleading. Yet public messaging often compresses nuanced age and timing effects into a single warning label, amplifying fear while obscuring the underlying uncertainties.
Trials and registries that shaped the pooled estimate
The meta-analysis drew on several landmark studies. Among the most influential is the WHI Memory Study-Younger cohort, known as WHIMSY, which examined women who started hormone therapy at ages 50 to 55. That trial found no detectable long-term cognitive harm or benefit in this younger group, according to a National Institute on Aging summary. The WHIMSY result is significant because earlier WHI data had raised alarms about dementia risk in women who began therapy at age 65 or older. By isolating women closer to typical menopause age, WHIMSY helped establish that the cognitive signal seen in older participants did not carry over to midlife initiators.
Two additional randomized trials reinforced that pattern. The KEEPS-Cog trial enrolled women in their early 50s and compared oral conjugated equine estrogens plus progesterone, transdermal estradiol plus progesterone, and placebo over up to four years. It reported no meaningful cognitive improvement and no major cognitive decline. The ELITE-Cog trial tested the so-called timing hypothesis directly, comparing estradiol’s cognitive effects in women closer to menopause against those much further past it. Neither trial produced evidence of a strong cognitive benefit window for estradiol, but neither suggested a clear dementia hazard for typical midlife use either.
On the observational side, a large UK primary-care database analysis using QResearch and CPRD examined dementia risk by hormone therapy exposure, type, and duration. That nested case-control analysis offered granular stratification and is frequently interpreted as finding no overall association with dementia risk, although some subgroups showed small differences. Because it was based on routine clinical records, the study captured real-world prescribing patterns, including varied formulations and treatment lengths. Together, these trials and cohort analyses form the evidential backbone of the new pooled estimate, supporting the view that menopausal hormone therapy does not universally raise or lower dementia risk.
The outlier is the Danish registry study, also published in The BMJ, which reported elevated dementia risk even among women treated before age 55. Because Denmark’s national registries capture virtually every prescription and diagnosis, the study’s statistical reach is difficult to dismiss. Its positive association with dementia complicates the “no clear link” narrative and has not been fully reconciled with the trial data. Some clinicians read the Danish findings as a signal to avoid long-term systemic hormone therapy when possible, while others argue that the randomized evidence should carry more weight for causal inference.
Gaps in the data that affect real treatment decisions
Several questions remain open. The full individual-level data and exact study weights from the Lancet Healthy Longevity meta-analysis have not been released beyond the bibliographic record. Without those details, outside researchers cannot fully assess how much each contributing study influenced the pooled result, or whether certain subgroups were underrepresented. That lack of transparency makes it harder to test whether factors such as age at initiation, duration of therapy, or specific estrogen–progestogen combinations drive subtle differences in risk.
The Danish registry authors have not published a direct response explaining why their positive association diverges from the new pooled estimate. Possible explanations include differences in hormone formulations prescribed in Denmark versus other countries, longer follow-up periods captured by the registry, or confounding by indication, where women with early cognitive symptoms may have been more likely to seek hormone therapy. None of these explanations has been tested in a head-to-head reanalysis that applies the same inclusion criteria and statistical models across datasets, leaving room for interpretation on both sides.
Another gap involves outcomes short of dementia. Many women are more worried about day-to-day memory lapses, concentration problems, or “brain fog” than about a formal dementia diagnosis decades later. Yet most large trials and registry studies focus on clinically coded dementia or standardized neuropsychological test scores, not on subjective cognitive complaints. The absence of consistent measures for these more subtle symptoms means that current evidence cannot fully answer whether hormone therapy meaningfully affects how women feel and function cognitively in midlife.
There are also important equity concerns. Much of the randomized evidence comes from predominantly white, higher-income participants in North America and Europe. Registry data, while broader, still reflect health systems with relatively robust access to care. Women from underrepresented racial and ethnic groups, or those with limited access to healthcare, may have different baseline dementia risks and different patterns of hormone use. Without more diverse cohorts and stratified analyses, it is unclear whether the “no clear association” conclusion applies uniformly across populations.
What this means for women considering hormone therapy
For patients and clinicians, the current evidence supports a cautious but not alarmist approach. The Lancet meta-analysis and several large trials suggest that, for many women who start hormone therapy around the time of menopause, there is no strong signal of increased or decreased dementia risk. The Danish registry findings, however, remind prescribers that long-term systemic use may carry uncertainties that have not been fully resolved, especially when therapy continues well beyond the menopausal transition.
In practice, that means decisions about hormone therapy should still be driven primarily by menopausal symptom burden, cardiovascular and breast cancer risk profiles, and personal preferences, rather than by fear of dementia alone. Women with severe hot flashes, night sweats, or sleep disruption may reasonably choose hormone therapy for quality-of-life benefits, with the understanding that current data do not show a consistent cognitive penalty or protection. At the same time, it is prudent to use the lowest effective dose for the shortest duration that meets treatment goals, revisiting the decision regularly as health status and evidence evolve.
For researchers, the conflicting dementia signals underscore the need for more transparent data sharing, harmonized outcome definitions, and studies that integrate genomics, imaging, and long-term clinical follow-up. Until those gaps are addressed, women and their clinicians will continue to navigate hormone therapy decisions in the gray zone between reassuring trial results and cautionary registry data-balancing real symptom relief today against still-uncertain cognitive risks decades down the line.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
