Written by Neil Kurtzman | 6th September 2026
Hormone replacement therapy and dementia risk among postmenopausal women: Identifying responsive subgroups in the UK Biobank is a study published in Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association. This large observational study found that hormone replacement therapy (HRT) was associated with a modestly lower hazard of dementia, especially among women classified as having surgical menopause and those who began treatment between ages 46 and 56. It does not prove that HRT prevents dementia.
The following abbreviations are used in this discussion: hormone replacement therapy (HRT); Alzheimer’s disease (AD); standard deviation (SD); hazard ratio (HR). A P value indicates the probability of obtaining a result at least as extreme as the one observed if there were actually no association; values below 0.05 are conventionally considered statistically significant.
Investigators analyzed 183,450 postmenopausal women in the UK Biobank. Their mean age was 60.3 years, with a SD of 5.8 years. Of these women, 77,335 (42.2%) had used HRT for at least one year or were current users; 106,115 (57.8%) were nonusers or had used it for less than one year.
Mean follow-up was 13.26 years (SD, 1.96), totaling 2,433,320 person-years. There were 3,948 dementia diagnoses: 1,993 classified as Alzheimer’s disease (AD) and 1,955 as non-Alzheimer or unspecified dementia. The investigators used Cox proportional-hazards regression, adjusting for age, education, smoking, ethnicity, socioeconomic deprivation, systolic blood pressure, body mass index, cholesterol, diabetes, antihypertensive medication, and cholesterol-lowering medication.
Key findings: HRT and dementia risk
| Group or outcome | Hazard ratio (HR) | 95% confidence interval | P value | Interpretation |
|---|---|---|---|---|
| All-cause dementia | 0.90 | 0.84–0.96 | 0.001 | 10% lower hazard |
| Natural menopause | 0.94 | 0.87–1.01 | 0.091 | No significant association |
| Surgical menopause¹ | 0.74 | 0.65–0.84 | <0.001 | 26% lower hazard |
| Apolipoprotein E ε4 carriers | 0.87 | 0.80–0.95 | 0.002 | 13% lower hazard² |
| Natural estrogen exposure of less than 37 years | 0.84 | 0.77–0.93 | Not reported | 16% lower hazard³ |
| Alzheimer’s disease | 0.84 | 0.77–0.92 | <0.001 | 16% lower hazard |
| Non-Alzheimer dementia | 0.95 | 0.87–1.04 | 0.28 | No significant association |
| HRT begun at ages 46–50 | 0.87 | 0.80–0.95 | 0.002 | 13% lower hazard |
| HRT begun at ages 51–56 | 0.77 | 0.70–0.86 | <0.001 | 23% lower hazard |
| Surgical menopause; HRT begun at ages 51–56 | 0.57 | 0.44–0.74 | <0.001 | 43% lower hazard |
¹ The study’s “surgical menopause” category included hysterectomy without removal of the ovaries as well as bilateral oophorectomy, an important limitation. The interaction between menopause type and HRT was statistically significant (P = 0.001).
² Although the result among apolipoprotein E ε4 carriers was significant, the formal interaction between genotype and HRT was not. The study therefore did not establish that carriers benefit more than noncarriers.
³ The interaction between length of natural estrogen exposure and HRT was statistically significant (P = 0.018).
This study offers an intriguing possibility, not a clinical verdict. Its strongest finding-that hormone replacement therapy was associated with substantially less dementia after surgical menopause and when begun between ages 46 and 56-is biologically plausible and deserves rigorous testing. But an association derived from a prospective cohort using self-reported treatment and reproductive histories cannot establish prevention, particularly when the overall effect is modest, numerous subgroup comparisons were made, and the investigators could not distinguish estrogen alone from combined therapy or determine the effects of dose, route, and duration. Recall bias, exposure misclassification, incomplete ascertainment of dementia, and residual healthy-user bias remain possible.
Moreover, the findings must be considered alongside randomized evidence: combined oral estrogen and progestin increased dementia incidence among women aged 65 or older in the Women’s Health Initiative Memory Study, while trials involving younger women have generally shown little effect on cognitive outcomes. The present results therefore should not prompt the prescription of hormones solely to prevent dementia. Their real importance is that they identify the populations, timing, and treatment formulations that a properly designed randomized trial should examine.
