Gennev Director of Health, ob/gyn Dr. Rebecca Dunsmoor-Su, offers her thoughts on a recent study linking HRT to increased risk of Alzheimer’s disease.

There has been a lot of press lately about the study by Savolainen-Peltonen and associates published in the British Medical Journal that links HRT use to increased incidence of Alzheimer’s disease.

This is a complex area, and clearly we don’t have the full answer yet; however, as an ob/gyn and a bio-statician, this study does not increase my worry about prescribing HRT to my patients.

Let me explain:

The increased risk is small. First and foremost, if this study is correct (and there are good reasons to question its findings – see below) the real increase was 9-18 more women per 10,000 developed Alzheimer’s when using systemic HRT than in the general population.

The increase involved a narrow subset of women taking HRT. There was zero increase in Alzheimer’s risk in women using vaginal estrogen, and in fact, this strong of an association of Alzheimer’s and HRT appeared only in women using hormones for 10 years or longer.

And, as I said earlier, there are good reasons to question conclusions that have been drawn from this study:

Missing information

This is a large study, which is generally good; however, they used a national database and therefore do not have information on health history, family history, or why these women started hormones.

Missing this important information can lead to significant bias in the data. For example, when they started the study in 1994, we thought that HRT could be protective for mental function, so it was sometimes prescribed specifically for women who were having memory issues. It might also have been prescribed as a preventative measure for women with a family history of Alzheimer’s.



If a mental function issue was the reason for starting HRT, then those patient factors alone could be responsible for the increase in Alzheimer’s the study found in women who took HRT – not the hormones themselves.

These types of retrospective studies are not meant to tell us how to practice medicine but rather to help us design more definitive studies to answer questions.

More accurate studies; different conclusions

Several smaller, previously published studies – which rely on more credible randomized, controlled trials – do not show this association.

Studies that randomize patients tend to introduce fewer biases from factors like family history or current symptoms. We trust the results of these studies more, because they more accurately tell us the effect of the drug itself.

Of course, any study can be done in a biased way (think about the problems with the Women’s Health Initiative), but randomized, controlled trials are less biased by nature.

When we see different results from a randomized trial than we saw from retrospective trials, then we really must ask ourselves, what could be the source of bias? I think we will see larger, randomized, controlled trial results coming out in the future that will help to answer this question.

So, what is the take away?

We should continue to use HRT for the symptoms of menopause where appropriate and for the shortest time needed.

We have not used HRT to improve memory or cognition in years, and we still shouldn’t. If a woman is concerned about Alzheimer’s because of a strong family history, we should discuss if HRT is best for her and minimize the amount of time she uses it.

As we already knew from previous data, women who start HRT right at menopause, and use it for a limited time, respond differently to hormones and seem to have fewer risks from them. This has not changed.

If you're concerned about HRT or would like to know more, we encourage you to join the discussion on our community forums. 



Wendy Sloneker

March 14, 2019

Medically Reviewed By

Subscribe for our weekly newsletter for helpful articles sent straight to your inbox:

Recommended Products

No items found.
Podcast episode available on Spotify Podcasts