Systematic review and meta-analysis of the effects of menopause hormone therapy on risk of Alzheimer's disease and dementia

Abstract

Introduction: Despite a large preclinical literature demonstrating neuroprotective effects of estrogen, use of menopausal hormone therapy (HT) for Alzheimer's disease (AD) risk reduction has been controversial. Herein, we conducted a systematic review and meta-analysis of HT effects on AD and dementia risk.

Methods: Our systematic search yielded 6 RCT reports (21,065 treated and 20,997 placebo participants) and 45 observational reports (768,866 patient cases and 5.5 million controls). We used fixed and random effect meta-analysis to derive pooled relative risk (RR) and 95% confidence intervals (C.I.) from these studies.

Results: Randomized controlled trials conducted in postmenopausal women ages 65 and older show an increased risk of dementia with HT use compared with placebo [RR = 1.38, 95% C.I. 1.16-1.64, p < 0.001], driven by estrogen-plus-progestogen therapy (EPT) [RR = 1.64, 95% C.I. 1.20-2.25, p = 0.002] and no significant effects of estrogen-only therapy (ET) [RR = 1.19, 95% C.I. 0.92-1.54, p = 0.18]. Conversely, observational studies indicate a reduced risk of AD [RR = 0.78, 95% C.I. 0.64-0.95, p = 0.013] and all-cause dementia [RR = .81, 95% C.I. 0.70-0.94, p = 0.007] with HT use, with protective effects noted with ET [RR = 0.86, 95% C.I. 0.77-0.95, p = 0.002] but not with EPT [RR = 0.910, 95% C.I. 0.775-1.069, p = 0.251]. Stratified analysis of pooled estimates indicates a 32% reduced risk of dementia with midlife ET [RR = 0.685, 95% C.I. 0.513-0.915, p = 0.010] and non-significant reductions with midlife EPT [RR = 0.775, 95% C.I. 0.474-1.266, p = 0.309]. Late-life HT use was associated with increased risk, albeit not significant [EPT: RR = 1.323, 95% C.I. 0.979-1.789, p = 0.069; ET: RR = 1.066, 95% C.I. 0.996-1.140, p = 0.066].

Discussion: These findings support renewed research interest in evaluating midlife estrogen therapy for AD risk reduction.

Keywords: Alzheimer’s disease; dementia; menopause hormonal therapy; meta-analysis; systematic review.

Figure 1

Flow chart Studies examining HT effects on risk of AD or dementia. Reason 1: meta-analysis. Reason 2: no assessment of AD or dementia risk. Reason 3: no OR/HR/RR estimates provided and/or no data to derive estimates. Reason 4: different estimates, e.g., standardized mortality rates.

Figure 2

Meta-analysis of randomized, placebo-controlled trials of HT effects on dementia risk. Meta-analysis of randomized placebo-controlled trials investigating the risk of developing dementia with the use of systemic HT. Forest plots display individual and pooled estimates of the association between HT use (A) and dementia risk expressed as relative risk (RR) and 95% confidence intervals (C.I.). HT includes (B) estrogen-only (ET, oral conjugated equine estrogens, CEE) and (C) estrogen-plus-progestogen therapy (EPT, oral CEE and medroxyprogesterone acetate, MPA). Studies are ordered by year of publication.

Figure 3

Meta-analysis of observational studies of overall HT use on AD or dementia risk Meta-analysis of observational studies examining the risk of developing AD or dementia with the use of systemic HT. Forest plots display individual and pooled estimates of the association between overall HT use and (A) AD and (B) AD or dementia risk expressed as relative risk (RR) and 95% confidence intervals (C.I.). Overall HT includes estrogen-only and estrogen-plus-progestogen formulations, oral or transdermal. Studies are ordered by year of publication. For all studies in panel (A) the outcome is AD incidence. In panel (B), the outcome is AD plus dementia, by adding available estimates for all-cause dementia.

Figure 4

Meta-analysis of observational studies of overall HT use on AD or dementia risk by study design Meta-analysis of observational studies examining the risk of developing AD or dementia by study design. Forest plots display individual and pooled estimates of the association between use of HT in panel (A) case–control vs. (B) cohort studies and risk of AD or dementia expressed as relative risk (RR) and 95% confidence intervals (C.I.). Studies are ordered by year of publication.

Figure 5

Meta-analysis of observational studies of short vs. long duration of HT use on AD or dementia risk Meta-analysis of observational studies examining the risk of developing AD or dementia by duration of HT use. Forest plots display individual and pooled estimates of the association between (A) long vs. (B) short duration of HT use and risk of AD or dementia expressed as relative risk (RR) and 95% confidence intervals (C.I.). Studies are ordered by year of publication.

Figure 6

Meta-analysis of observational studies of estrogen-only vs. estrogen-plus-progesterone HT use on AD or dementia risk Meta-analysis of observational studies examining the risk of developing AD or dementia by HT formulation. Forest plots display individual and pooled estimates of the association between use of (A) estrogen-only vs. (B) estrogen-plus-progestogen therapy and risk of AD or dementia expressed as relative risk (RR) and 95% confidence intervals (C.I.). Studies are ordered by year of publication.

Figure 7

Meta-analysis of observational studies of midlife vs. late-life HT use on AD or dementia risk Meta-analysis of observational studies examining the risk of developing AD or dementia by timing of HT. Forest plots display individual and pooled estimates of the association between use of HT in (A) midlife vs. (B) late-life and risk of AD or dementia expressed as relative risk (RR) and 95% confidence intervals (C.I.). Studies are ordered by year of publication.

Figure 8

Summary of HT effects on AD risk Schematic overview of the main results of the study, illustrating how HT efficacy on AD risk varies based on initiation timing relative to menopause onset and HT formulation. In midlife, or more generally within 10 years of the final menstrual period, estrogen-only therapy is associated with a moderate decrease in AD risk, while estrogen-progestogen therapy is associated with a milder, non-significant risk reduction. In late-life, or more than 10 years after menopause, estrogen-only therapy presents neutral effects on AD risk, whereas estrogen-progestogen therapy is associated with a moderate risk increase, albeit non-significant.