Understanding the impact of sex and gender in Alzheimer's disease: A call to action

2.1.1. Cardiometabolic risk factors

Several studies have identified modifiable cardiometabolic diseases including type 2 diabetes, metabolic syndrome, obesity, and other cardiovascular risk factors for the development and progression of AD [17]. Despite the well-known fact that the development, symptoms, and treatment of cardiometabolic diseases differ by sex [18], few studies have examined sex differences in cardiovascular risk factors for AD. Most studies to date have simply adjusted for sex in regression models. This is unfortunate because work on sex differences in the vasculature has shown that microvascular disease is a greater contributor to cardiovascular disease in women than in men, whereas obstructive coronary artery disease is a greater contributor in men than in women [19]. Moreover, women have a higher risk of diabetic complications than men, including myocardial infarction (MI), depression, and coronary heart disease, all of which are risk factors for AD [18]. Thus, while a diagnosis of hypertension, high cholesterol, and diabetes in midlife has been associated with a greater risk of developing AD for both women and men, the risk of AD for women with these factors may be greater than for men [20, 21]. New research is needed to understand whether the relationship between cardiometabolic factors and risk of AD differs by sex, and how this difference varies by age. Given that men typically develop cardiometabolic diseases at earlier ages than women (perhaps due, in part, to the protective effects of estradiol before menopause), additional research is needed to understand the interrelationship between sex differences in the timing of the risk factors and development of AD, and how this difference contributes to the age-specific distribution of AD.

2.1.2. Depression

While depression can be a symptom of AD in older adults, depression at midlife is believed to increase the risk of AD by as much as 70% [22]. Women have twice the risk of depression compared to men, a difference that emerges at puberty [23, 24], and worsens during the menopausal transition [25, 26]. Depression has implications for cognition across the life span given shared brain regions between mood and memory and shared etiologic pathways such as immune and stress hormone dysregulation. Some cohort studies found that elevated depressive symptoms were associated with AD only in men [27, 28]. Conversely, in the Women’s Health Initiative Memory Study (WHIMS), clinically significant depressive symptoms were associated with a nearly twofold increased risk of MCI and dementia over a 5.4-year average follow-up in a sample of 6000 women [29]. A past history of depression was also associated with a near doubling of the risk of dementia in WHIMS [29]. Another study found that depression was associated with hippocampal volume loss only in women [30, 31], while in WHIMS depressive symptoms were associated with prefrontal volume loss [32].

2.1.3. Sleep

Sleep and circadian rhythms disturbances are common among patients with and may impact the development of AD pathology as well [33]. The production and clearance of Aβ peptides is associated with the sleep-wake cycle, with wakefulness associated with Aβ production and sleep associated with Aβ clearance [34]. Subjective reports of poor sleep quality and short sleep duration as well as objective measures of high sleep fragmentation are associated with increased Aβ accumulation, poorer cognition, and increased risk of AD in older adults [35, 36]. Similar results are seen during midlife, where objective reports of short sleep duration, long sleep duration, and poor sleep quality are associated with poorer cognitive function [37]. Recent evidence suggests that disruption of sleep stage 3 (slow-wave sleep) in particular increases Aβ levels [38]. Interestingly, slow wave activity is greater in women than men across all ages [39, 40], although the effects of this difference on Aβ clearance is unknown. There is also a general decline in slow wave sleep with age [39–41]. In addition, the prevalence of sleep disturbances and sleep disorders can increase with age, with menopause being a particularly vulnerable time for women. For example, while sleep apnea is more prevalent in men, its incidence in women greatly increases after menopause [42]. Individuals with sleep apnea show cognitive decline at an earlier age than those without sleep apnea [43]. Taken together, current findings suggest that more research is needed to understand the relationship between sleep and AD risk, and whether that risk differs between women and men.

2.2.1. Education

Low socioeconomic status, education, and occupational attainment pose similar risk for AD in both women and men. However, in the past century, women have had fewer opportunities for higher education and occupational attainment. As a result, many more women than men are affected by this risk factor [6]. More recently, educational attainment for women has been higher than men in the United States [44]. The improvement in education and occupational attainment in women over the last few decades may be one explanation as to why the incidence of dementia may be declining more for women [45, 46].

2.2.2. Exercise

Physical activity at midlife is associated with a decreased risk of AD [47]. Women exercise less than men [48], and gender differences in parenting roles accounts for only some of this difference. The effect of exercise might also vary depending on estradiol level and menopausal stage, with greater benefits observed when estrogen levels are high [49].

2.2.3. Marital status

Compared with women, men who have never married or are widowed, have a greater risk of developing AD [20, 50, 51]. A potential reason for this consistent observation is that women have historically been responsible for the health-care of their family (e.g., getting offspring and partners/husbands to healthcare providers for regular checkups, assuring everyone has a healthy diet, and so forth), sometimes at the expense of their own health. Single women, compared with single men, are also more likely to see a health-care provider and to engage in social activities, which are beneficial for cognition.

2.2.4. Caregiving for AD

Family caregivers for AD patients are usually either spouses or adult-children [52]. Women make up on average 60% of all family caregivers [52] and these rates are especially high for Hispanic and African American caregivers [53]. Women caregivers report a twofold higher level of caregiver burden compared to male caregivers [54]. Compared to sons, daughters of AD patients provided more routine assistance, experienced more guilt, were less likely to receive support from their spouses, and were more likely to leave the labor market to care for their parent [55, 56]. In contrast, sons were more likely to ask for help from other family members. When sons identified themselves as caregivers, it was relatively common that it was the son’s wife who provided the care [56, 57]. Caregiving is associated with elevated levels of cortisol and impaired attention and executive function [58]. It has been hypothesized that spousal caregivers may be at higher risk of cognitive impairment or dementia than noncaregiver spouses in response to several psychosocial (e.g., depression, social isolation, and sleep problems), behavioral (e.g., exercise and diet), and physiological (e.g., metabolic syndrome and inflammation) variables [59]. More research is needed to better determine how gender differences in caregiver responsibilities may impact risk of AD.

2.3.1. Hypertensive pregnancy disorders

Hypertensive pregnancy disorders (HPDs) are a major cause of maternal and fetal morbidity and mortality. These disorders affect approximately 12% of all pregnancies and include gestational hypertension, preeclampsia, eclampsia, chronic hypertension, and preeclampsia or eclampsia superimposed on chronic hypertension. Previous studies reported associations between HPD and subjective cognitive complaints [60, 61] or brain white matter hyperintensities on magnetic resonance imaginge [62–64]. However, these studies had small sample sizes and only assessed brain structure and cognitive function less than 10 years after the HPD. Recently, the long-term association between HPD (i.e., >10 years after pregnancy) and brain structure and cognitive function was assessed in a multiethnic study of 1279 women (mean age=61) enrolled in the Family Blood Pressure Project Genetic Epidemiology Network of Arteriopathy study [64]. Women with a history of HPD had greater brain atrophy decades after their pregnancies than women who had normotensive pregnancies. There was also a trend for more white matter hyperintensities in those with HPD. Given the increasing prevalence of HPD with increasing obesity and later maternal age, further research examining HPD and risk of AD is needed. There is also a need to determine whether there is a common underlying factor that increases the risk of both HPD and AD or whether HPD is, itself, an independent risk factor.

2.3.2. Menopause, hormone therapy, and cognition

A key determinant of sex differences in cognition and brain function is sex steroid hormones. Menopause is a universal event experienced by all women who live to midlife and beyond. Given that women living today will on average spend one-third of their life in the postmenopausal stage, it is critical to consider menopause in brain function and AD.

2.3.3. Natural menopause and memory decline

Women demonstrate a decrease in verbal memory during the menopausal transition [65], a change that has been linked to alterations in hippocampal function associated with loss of estradiol [66, 67]. In some women, this decline appears to be temporary, as two longitudinal studies show a decline in verbal memory during perimenopause, but not during postmenopause [65, 68]. Recent cross-sectional studies, however, found impairment in verbal memory among postmenopausal women but not among perimenopausal women [66, 67, 69].

2.3.4. Risk of initiating hormone therapy late in life

There is considerable confusion about whether estrogen containing hormone therapy (HT) is protective or harmful in women with respect to the development of AD. Findings from WHIMS showed that combination estrogen plus progestin therapy in women aged 65 years and older doubled the risk of dementia [70, 71]. Analysis of the data from WHIMS showed that the most serious adverse events, including those on brain and cognition, occurred in women who began with low cognitive function at baseline [70]. These findings suggested that initiating HT at an older age and among women whose cognitive performance is atypically low may have adverse consequences. The relevance of these findings to clinical practice was brought into question because the very large majority of women initiate HT earlier in life, during the menopausal transition. Furthermore, these findings were at odds with findings from a recent Women’s Health Initiative (WHI) publication on the effects of HT on mortality over 18 years of cumulative follow-up [72]. Women randomized to estrogen therapy had a significantly lower risk of dying from AD or dementia than women randomized to placebo. This effect was not observed in women randomized to estrogen plus progestin therapy in the WHI. The number of cases was too small to determine whether the effect was dependent on timing of the initiation of HT in relation to menopause. The mortality findings need to be interpreted with caution because the cause of death was reported by National Death Index and in some cases next of kin, whereas dementia cases were prospectively adjudicated in WHIMS. On the other hand, the smaller number of dementia cases in WHIMS (n=237 dementia cases) compared with WHI mortality study (n=758 cases) leads to less statistical power in WHIMS, and a sizeable proportion of referrals for dementia determination in WHIMS was not adjudicated. Currently, there are insufficient data to determine the effects of menopausal HT on late-life dementia.

Observational data show that women who initiate HT early in the menopausal transition or at a younger age have a lower risk of AD than women who initiate HT later [73–75]. This pattern is consistent with the critical window hypothesis of HT, which suggests that early initiation of HT in relation to the menopausal transition confers greater cognitive benefit than initiation later in life [76, 77]. In contrast, a recent population-based study of 489,105 Finnish women showed that the risk of death from AD was reduced by 15%−19% in women who used HT for at least 5 years. Risk of death from vascular dementia was reduced from 37% to 39% regardless of the length of exposure or timing [78]. One benefit of the Finnish data is that compared to US data, Finnish data appear to be less influenced by the healthy user bias, the tendency of women who go on using HT to be healthier and better educated than other women.

2.3.5. Effects of HT on cognitive function in the early postmenopause

High-quality clinical trials provide reliable evidence that HT early in the postmenopausal period does not confer cognitive benefit but is safe for cognitive function. Specifically, three large trials demonstrated no immediate or enduring adverse effects of HT on cognition when HT was initiated within 5 years of the final menstrual period [79–81]. One of those trials focused on women from the WHI who were aged 50–59 years (mean age=53) at randomization and found no negative cognitive effects when women were tested 14–15 years later. Another included a site-specific neuroimaging study which indicated that Aβ deposition was lower in APOE ε4- positive women randomized to receive estradiol compared with those randomized to receive placebo [82]. It is possible that the effect was observed only in APOE ε4-positive women, because APOE ɛ4 carriers have greater Aβ deposition at an earlier age than APOE ɛ4 noncarriers. Unfortunately, although the primary indication for HT is hot flashes, none of those studies selectively enrolled women with bothersome vasomotor symptoms. Some evidence demonstrates that vasomotor symptoms are associated with memory deficits [83, 84]. Overall, there is no support for a “critical window” of HT in the early postmenopausal period but that hypothesis has not been tested in symptomatic women or women earlier in the menopausal transition. Studies of oral contraceptives or HT in perimenopause are warranted given that the memory decline that occurs during the menopausal transition emerge in early perimenopause and appear to resolve in the early postmenopausal period [65, 68].

2.3.6. Gynecological surgeries

Most studies [85–93], but not all [94, 95], have shown that earlier age at menopause (natural or surgery-induced) is associated with an increased risk of cognitive decline and dementia. Notably, the increased risk of cognitive impairment and dementia did not vary by indication for the oophorectomy and was eliminated by estrogen therapy that was initiated after the surgery and continued up to 50 years of age or longer [85]. Collectively, these studies suggest that bilateral oophorectomy may be a risk factor for cognitive decline and dementia. Such a conclusion is supported by findings that an earlier age at surgical menopause was associated with a greater burden of neuritic plaques at autopsy [93]. However, the women who underwent these surgeries can differ from other women in ways that put them at higher risk of cognitive impairment and dementia independent of the surgery [13]. For example, women with adverse childhood experiences or adult abuse were found to be at increased risk of undergoing bilateral oophorectomy before menopause [13], and early childhood trauma has been associated with an increased risk of late-life dementia [96]. It may be that early oophorectomy interacts with such risk factors to influence risk of dementia. More generally, well-controlled studies measuring AD pathology in vivo (cerebrospinal fluid [CSF] positron-emission tomography [PET]) in women with early surgical menopause are warranted to better interpret the epidemiological findings.

2.3.7. Neurobiological model of aging and menopause

Both clinical and basic science studies provide a theoretical framework for understanding menopause as a female-specific risk factor for AD. Specifically, it has been proposed that perimenopause is a bioenergetic transition state characterized by a decline in mitochondrial function and a change in fuel source from glucose to lipids which in turn could lead to remodeling of the brain, loss of synaptic spines (especially in the hippocampus), and ultimately neurodegeneration [97–103]. In support of this view are neuroimaging data demonstrating that compared with premenopausal women, perimenopausal and postmenopausal women show AD-like reductions in glucose metabolism, and these reductions are related to platelet mitochondrial activity [104].

3.3.1. The APOE ε4-sex interaction on AD

The ε4 allele of the APOE gene is a major genetic risk factor of late-onset AD dementia [121] and is consistently linked with abnormal accumulation of the Aβ protein [122]. The apoE protein is involved in the transport of cholesterol and other lipids in the periphery of the brain. The apoE4 protein has been shown to be less effective at Aβ clearance [123], and contributes to a diminished response to neuronal injury compared to the apoE2 and apoE3 proteins [124]. An analysis of nearly 58,000 participants showed sex differences in the risk of AD dementia and MCI by APOE genotype [125]. Among persons aged 65–75 years with the APOE ε3/ε4 genotype, the risk of AD dementia is fourfold higher in women than that in to men. Similarly, among those aged 55 to 70 years with the APOE ε3/ε4 genotype, the risk of MCI is 43% higher in women than that in men. Although the mechanisms underlying the interaction between sex and the APOE genotype remain unclear, results across research groups have suggested that APOE ε4 females may show greater levels of AD pathology, more compromised brain network integrity, and/or accelerated longitudinal decline for a given level of AD pathology [126–128].

3.3.2. Women APOE ε4 carriers have worse longitudinal performance than men

In an analysis of over 5400 clinically normal participants, women who were APOE ε4 carriers had an elevated risk of progression on the Clinical Dementia Rating scale compared with both APOE ε4 men and APOE ε4-negative women; this effect was also observed in analyses restricted to APOE ε3/ε4 carriers. Although an analysis of MCI patients (N>2500) found no sex by APOE ε4 (homozygote and heterozygotes combined) interaction in progression to AD dementia, an analysis restricted to APOE ε4 heterozygotes revealed a similar sex by APOE ε4 interaction [126]. Several studies have shown that APOE ε4 women had a faster decline of cognitive function compared to either women noncarriers or men of any genotype [129, 130].

Because women with an APOE ε4 allele are at increased risk of developing AD compared with women without ε4, more work is needed to clarify whether APOE status modifies the effects of female-specific risk factors for AD. This approach has been undertaken in studies of HT and AD risk, but without clear consensus. Some studies suggest that the effects of HT on cognition in women is influenced by APOE genotype, with benefits of HT observed in APOE ε4 noncarriers only, whereas other studies suggest that HT reduces AD risk in APOE ε4 carriers [131] or that HT reduces AD in both carriers and noncarriers [132, 133].

Overall, these data suggest that the APOE ε4 allele interacts with sex to influence risk of AD dementia. The increased risk among women may be due to multiple underlying mechanisms, including direct effects of apoE on AD pathology and on individual trajectories of decline for a given level of pathological burden. An important limitation regarding the observed APOE-associated sex differences in AD, is evidence of earlier mortality of APOE ε4 men that results in a selective APOE ε4 survival bias for women [134]. Future research is also needed to examine the sex-specific differences in the influence of the APOE ε4 over the life span. For instance, it is possible that the interaction between sex and APOE ε4 occurs during menopause, at a time when the risk of Aβ accumulation first begins to increase among APOE ε4 individuals [135]. However, most studies investigating dementia risk have examined older ages (>65), making it difficult to isolate the potential impact of sex during midlife.