A study has raised a question that many women would understandably want answered: could hormone replacement therapy (HRT) protect against dementia? The research found that women who had used HRT were less likely to receive a dementia diagnosis during follow-up, particularly if they had undergone surgery involving their reproductive organs. But the results do not establish that HRT caused the difference.
The researchers analysed data from 183,450 postmenopausal women in the UK Biobank, followed for an average of 13.3 years. During that period, 3,948 received a dementia diagnosis, including 1,993 diagnosed with Alzheimer’s disease.
Women who reported using HRT for at least a year, or who were still taking it when they joined the study, had a 10% lower relative risk of dementia than non-users after researchers accounted for factors including age, education, smoking, deprivation, blood pressure and diabetes.
The difference was larger among women the researchers classified as having had “surgical menopause”, meaning surgery involving removal of the womb or ovaries. HRT use was associated with a 26% lower relative risk of dementia in this group. There was no statistically clear association among women who had undergone natural menopause.
However, the surgical group included both women who had undergone a hysterectomy, which removes the womb, and women whose ovaries had been removed. Removing both ovaries causes an abrupt fall in oestrogen, while removing the womb alone does not necessarily affect hormone production in the same way. Grouping these women together makes it harder to determine whether HRT protects against dementia after a sudden loss of oestrogen.
Why the findings need care
Women who take HRT may differ from those who do not in ways that also affect dementia risk. Education, income, general health and access to care can influence both the decision to seek treatment and later health. Researchers can adjust for measurable differences, but observational studies cannot account for everything.
An early study found that women who later chose to take HRT were already healthier and at lower risk of dementia eight years before starting treatment. They tended to have more education, lower blood pressure and diabetes risk, were more active and were less likely to be overweight. These factors are themselves associated with dementia risk.
The new study also could not distinguish between women using oestrogen alone and those using oestrogen combined with a progestogen, or compare doses and ways of taking HRT. Oestrogen-only HRT is generally prescribed after hysterectomy, while a progestogen is usually added for women with a womb to protect its lining against cancer.
In a review of national registry studies that I worked on, longer use of combined hormone therapy was associated with a small increase in Alzheimer’s risk. The same pattern was not seen with oestrogen-only treatment. Those studies were observational too, so they cannot prove treatment caused the increase.
Results from the same UK Biobank data have also varied. An earlier analysis reported an association between lifetime HRT use and higher dementia risk in women whose natural menopause occurred after 50. The studies asked different questions and analysed different groups, so they cannot be compared directly. Their differing results do, however, show why one new association cannot settle the debate.
The new paper also suggested that timing may be important. Lower dementia risk appeared among women who started HRT between their late forties and mid-fifties, but the pattern was less convincing among those who started younger. This makes it difficult to argue that replacing hormones as early as possible protects women whose menopause occurs particularly young.
Researchers also examined women with the APOE ε4 genetic variant, which increases Alzheimer’s risk. Although the association appeared stronger among carriers, statistical testing did not establish that HRT affected them differently from non-carriers. Previous studies have also produced conflicting results.
What does this mean for someone taking HRT?
Trials in women closer to menopause have not shown consistent improvements in cognition after several years of HRT. Some short-term trials have reported benefits, including among women with substantial menopausal symptoms, but these findings have not established that HRT prevents dementia.
A trial involving women aged 65 and over found an increased incidence of dementia among those taking a form of combined HRT commonly used at the time. That does not answer every question about treatments used today or about women who undergo early surgical menopause. However, trial evidence does not currently support using HRT to prevent dementia in women who undergo menopause at the usual age of around 50 or 51.
That is reflected in Nice guidance, which says HRT should not be offered specifically to prevent dementia or cardiovascular disease. It does recommend HRT for menopausal hot flushes and night sweats. Those benefits should remain central to treatment decisions, alongside medical history and the risks of the particular treatment.
Women who have their ovaries removed at a young age deserve particular attention because early loss of hormones may have different health effects from menopause at the usual age. This study gives researchers a reason to investigate that group more closely. For now, it provides no sound reason to start, stop or change HRT solely in the hope of preventing dementia.
Eef Hogervorst is affiliated with Loughborough University. She acted as a non-paid dementia and MHT/HRT use expert for NICE 2024 Guidelines and ESHRE 2016 Guidelines for Early and Premature Menopause. In the past she also performed analyses and reviews for Organon and acted as expert to discuss WHI data having been invited by Wyett in 2002. She has done Cochrane analyses showing at best only time limited benefits (2-3 months) of HRT after RCT in highly symptomatic women, but also in women with AD negating the window of time theory. Negative effects were also seen in older women with longer (>1 year) treatment durations.