Risks and benefits of hormone therapy after menopause for cognitive decline and dementia: A conceptual review

Abstract

Objective: The effects on the brain of hormone therapy after the onset of menopause remain uncertain. The effects may be beneficial, neutral, or harmful. We provide a conceptual review of the evidence.

Methods: We 1) provide a brief history of the evidence, 2) discuss some of the interpretations of the evidence, 3) discuss the importance of age at menopause, type of menopause, and presence of vasomotor symptoms, and 4) provide some clinical recommendations.

Results: The evidence and the beliefs about hormone therapy and dementia have changed over the last 30 years or more. Five recent observation studies suggested that hormone therapy is associated with an increased risk of dementia, and the association appears not to change with the timing of initiation of therapy. These harmful associations may be explained by a causal effect of hormone therapy on the brain or by several confounding mechanisms. We suggest that the use of hormone therapy should be customized for different subgroups of women. It may be important to subgroup women based on age at onset of menopause, type of menopause, and presence or absence of vasomotor symptoms. In addition, the effects may vary by type, dose, route, and duration of administration of estrogens and by the concurrent use of progestogens.

Discussion: The relation of hormone therapy with the risk of dementia is complex. Hormone therapy may have beneficial, neutral, or harmful effects on the brain. Hormone therapy should be guided by the clinical characteristics of the women being treated.

Keywords: Age at menopause; Dementia; Hormone therapy; Type of menopause; Vasomotor symptoms.

Figure 1.

Cumulative incidence of dementia in the general population and age at initiation of HT in the KEEPS, ELITE, and WHIMS clinical trials. In the KEEPS and ELITE trials, the initiation of HT was adequate to test the timing hypothesis because proximate to the onset of menopause. However, the length of follow-up was too short to detect a difference in dementia risk. By contrast, in the WHIMS, the initiation of HT was inadequate to test the timing hypothesis because too remote from the onset of menopause; however, the length of follow-up was adequate to detect a difference in dementia risk. The curve for the incidence of dementia (including cognitive decline and dementia) was modified from Rocca et al., Neurology 2007 [42].

Figure 2.

Four panels illustrating alternative explanations for the association between HT and increased risk of dementia. Panel A: the association is causal. HT has a detrimental effect on the risk of dementia mediated by cardiovascular risk factors or event, neurodegenerative lesions, or both mechanisms. Panel B: the association is spurious and due to the preferential use of HT in women who are experiencing prodromal signs of cognitive dysfunction. This scenario can be named protopathic bias or cause-effect inversion. Panel C: The association is spurious and due to the confounding effect of vasomotor symptoms. Vasomotor symptoms may cause the preferential use of HT, and independently may cause an increased risk of dementia. This scenario may be called confounding by indication. Panel D: The association is spurious and due to the confounding effect of genetic variants or behavioral and environmental risk factors acting in earlier life. This scenario may be called confounding by upstream causes.

Figure 3.

Distribution of age at onset of menopause and type of menopause. The red curve represents women who underwent induced menopause, and the black curve represents women who underwent spontaneous menopause. The curves were modified from Rocca et al., Maturitas 2023 [34]. In addition, women can be grouped in premature, early, normal-age, or late menopause. The percentages by type of menopause refer to the general population that also includes a group of women who underwent hysterectomy with ovarian conservation (17.8%). This double stratification generates a total of nine subgroups of women. Women may be further characterized as having or not having vasomotor symptoms (Table 1).