Dementia is one of the most significant causes of disability and mortality worldwide, with over 55 million individuals currently estimated to suffer from the condition.¹ The medical community has long recognized that women are at substantially higher risk of developing dementia than men – a discrepancy that persists even after accounting for longer average lifespan among women relative to men in developed countries. This observation suggests that estrogen or other sex hormones may play an important role in cognitive decline, an idea which led researchers Pourhadi et al. to conduct a recent study on whether hormone replacement therapy (HRT) – i.e., therapy designed to offset the drop in estrogen levels that occurs during menopause – might be associated with increased risk of dementia later in life.²
What they did
In light of hints from previous observational data suggesting a possible link between hormone therapy and dementia, Pourhadi et al. conducted a case-control study comparing HRT use among women with incident dementia (cases, n=5,589) and age-matched dementia-free women (controls, n=55,890). Cases and controls were identified from across the female Danish population between 2000-2018 using Danish national registries. All participants were between the ages of 50 and 60 as of January 1, 2000, and all were free of any history of dementia, breast or gynecological cancers, blood clots or clotting disorders, liver disease, bilateral oophorectomy (surgical removal of the ovaries), or hysterectomy.
The primary exposure of interest was the use of combined estrogen and progestin (E+P) for menopausal hormone replacement therapy. Using data from the Danish National Prescription Registry, the investigators assessed participants’ HRT use from January 1, 1995 until two years prior to the index date (for cases, this was the date of either dementia diagnosis or start of dementia treatment, while for controls, this was the date at which their age-matched case received a dementia diagnosis or initiated dementia treatment).
What they found
Of the 5,589 dementia cases, 4,436 (79.4%) were late-onset all-cause dementia, while 1,458 (26.1%) cases were registered as Alzheimer’s disease (AD).
The results showed that a total of 1,782 (31.9%) cases and 16,154 (28.9%) controls had received combined E+P menopausal hormone therapy before the index date. Most of these individuals (66.2%, or 11,879) had ceased E+P use >8 years before the index date, while 1,555 (8.7%) were still users at the time of the index date.
Compared with never-users of menopausal E+P therapy and estrogen-only or progestin-only users, ever-users of menopausal E+P treatment were reportedly 24% more likely to develop all-cause dementia after adjusting for various confounders (adjusted HR: 1.24; 95% CI: 1.17-1.33). In looking at subtypes of dementia, the authors reported that ever-users of menopausal E+P treatment were also 21% more likely than controls to develop late-onset dementia (HR: 1.21; 95% CI: 1.12-1.30) and 22% more likely to develop Alzheimer’s disease (AD) specifically (HR: 1.22; 95% CI: 1.07-1.39).
Moreover, longer durations of hormone use were associated with further increases in dementia risk. While use of E+P for less than one year was associated with a 21% increase in risk of all-cause dementia relative to never-users (HR: 1.21; 95% CI: 1.09-1.35), women who used these hormones for 8-12 years were at a 39% higher risk than never-users (HR: 1.39; 95% CI: 1.21 to 1.58). Risk elevation among women on E+P for over 12 years was stronger still – an increase of 74% over never-users (HR: 1.74; 95% CI: 1.45-2.10). A similar pattern held for late-onset dementia, though not for AD-related dementia.
Collectively, these results indicate that women who use menopausal hormone therapy are at higher risk of dementia than those who do not use HRT, and that longer periods of use are associated with further increases in risk.
Reasons for doubt
So, how do these findings impact my generally positive views on menopausal HRT? Not at all, as this retrospective study suffers from a number of flaws (some of which are pointed out by the authors themselves) which ultimately prevent drawing any clinically meaningful conclusions from these results.
As is common in observational studies, confounding variables may certainly have played a hand in these results. Cases were significantly more likely than controls to live alone, have lower education levels and household income, and have hypertension, diabetes, and thyroid disease, all of which are known risk factors for dementia. The authors adjusted for these factors in their statistical analyses, but they are likely mere examples of many distinctions between the two groups. Based on these reported differences alone, we can infer that controls likely had better overall health and were engaging in more health-conscious behaviors which would independently reduce their risk of dementia, such as exercising or maintaining lower lipid levels. These possible confounds likely explain why even women with hardly any HRT exposure (i.e., those on HRT for less than one year) were still found to have significantly elevated dementia risk. (Indeed, risk among women taking E+P for <1 year was higher than among women taking E+P for 1-4 years or 4-8 years.)
Further, we have reason to doubt the accuracy of certain elements of Poudhari et al.’s source data. For instance, data on hormone therapy use, duration, formulation, and dosage was based on filled prescriptions, but prescriptions filled are not necessarily prescriptions taken, potentially introducing error into the subsequent analyses. Additionally, we cannot trust the data on AD diagnoses since AD cases were almost certainly vastly underreported in this cohort. While AD is estimated to account for 60-80% of all dementia cases worldwide, in this study, it nominally accounted for only 26% – a discrepancy which the authors themselves point out as reflecting incomplete information on dementia diagnoses in health registries. In effect, this shortcoming renders any reported associations between HRT and Alzheimer’s disease completely meaningless.
But perhaps the greatest limitation of this study with regard to its clinical relevance is that the hormone formulations used by women in this study do not reflect the formulations most commonly used for HRT in the present day. The authors report that the primary progestin used by women in this study was norethisterone (83.5% of cases on E+P and 80.6% of controls), followed by medroxyprogesterone and levonorgestrel, with ~90% of women in both groups using oral formulations. As I’ve discussed previously, use of these synthetic progestins has fallen dramatically over the years, and bioidentical micronized progesterone is now overwhelmingly preferred as a source of systemic progesterone in HRT. Indeed, systemic progesterone in general is less common with HRT than it was two decades ago, as progesterone-coated intrauterine devices are gaining increasing favor. In other words, even if we assume Poudhari et al.’s results are accurate and haven’t been influenced significantly by confounding or source data inaccuracies, they have little bearing on the question of cognitive health implications of HRT in a modern setting.
The bottom line
Through this case-control study, authors Pourhadi et al. attempted to determine how combined estrogen and progestin therapy in menopausal women impacted risk of dementia. Though the investigators reported higher risk of all-cause, late-onset, and Alzheimer’s-associated dementia with hormone use, these results were likely heavily confounded, preventing us from making any real conclusions that hormone therapy underlies the apparent association. Even if such an association were to exist, the formulations used by most women in this study are not representative of formulations more commonly used today, so these results tell us nothing about a potential link between HRT and dementia in the present day.
As I’ve discussed at length on the podcast, far too many women have been denied the benefits of HRT over the years as a result of heavily flawed reports of associations with disease, and I sincerely hope that the results by Poudhari et al. do not add further fuel to the anti-HRT fire. As exemplified by this study, fears over HRT use are largely founded in misleading correlations with dubious relevance to modern clinical practice, whereas the many benefits of HRT for menopausal women remain clear and robust.
For a list of all previous weekly emails, click here.
References
- World Health Organization. Dementia. Published March 2023. Accessed June 26, 2024. https://www.who.int/news-room/fact-sheets/detail/dementia
- Pourhadi N, Mørch LS, Holm EA, Torp-Pedersen C, Meaidi A. Menopausal hormone therapy and dementia: nationwide, nested case-control study. BMJ. 2023;381:e072770. doi:10.1136/bmj-2022-072770
