Postmenopausal hormone therapy and cognition

Abstract

Prior to the publication of findings from the Women’s Health Initiative (WHI) in 2002, estrogen-containing hormone therapy (HT) was used to prevent age-related disease, especially cardiovascular disease, and to treat menopausal symptoms such as hot flushes and sleep disruptions. Some observational studies of HT in midlife and aging women suggested that HT might also benefit cognitive function, but randomized clinical trials have produced mixed findings in terms of health and cognitive outcomes. This review focuses on hormone effects on cognition and risk for dementia in naturally menopausal women as well as surgically induced menopause, and highlights findings from the large-scale WHI Memory Study (WHIMS) which, contrary to expectation, showed increased dementia risk and poorer cognitive outcomes in older postmenopausal women randomized to HT versus placebo. We consider the ‘critical window hypothesis’, which suggests that a window of opportunity may exist shortly after menopause during which estrogen treatments are most effective. In addition, we highlight emerging evidence that potential adverse effects of HT on cognition are most pronounced in women who have other health risks, such as cerebrovascular disease or diabetes. Lastly, we point towards implications for future research and clinical treatments.

Effects of menopause on cognition

Menopause refers to the time in a women’s life when her ovaries begin producing less estrogen and progesterone, and she becomes infertile. This transition period can last up to five years and in some cases longer (Harlow & Paramsothy, 2011), with the current median age of menopause at 52.5 years, defined as 12 months following cessation of menses (Gold, Crawford, Avis, Crandall, Matthews, Waetjen et al., 2013). The cessation of menstrual cycles results in a number of physiological changes that can affect emotion and cognitive processes. Importantly, women born in the USA from the 1950s onwards can expect to live more than one-third of their lives with depleted ovarian hormones (“Population and Vital Statistics”, 2007). To alleviate common menopausal symptoms such as hot flushes, night sweats, and sleep disruptions (Dennerstein, Dudley, Hopper, Guthrie, & Burger, 2000; Obermeyer, Reher, & Saliba, 2007; Ribowsky, 2011), some women opt to artificially replace hormones and boost levels through estrogen-containing postmenopausal hormone therapy (HT).

Prior research indicates that the cumulative estrogen exposure a women encounters over her lifespan influences late life cognitive ability. These accumulations stem from age at menses and at menopause, duration of breastfeeding if child-bearing, duration of HT in hormone users and time since menopause (Hesson, 2012). Additionally, experiencing menopause at younger ages has been associated with reduced cognitive performance in older adulthood (Hogervorst, 2012; Hogervorst, Kusdhany, & Rahardjo, 2011; Ryan, Scali, Carrière, Amieva, Rouaud, Berr et al., 2014), as well as greater mortality risk (Nelson, Walker, Zakher, & Mitchell, 2012). Several studies describe cognitive reductions in women during the menopausal transition (Mitchell & Woods, 2011; Schaafsma, Homewood, & Taylor, 2010; Weber, Maki, & McDermott, 2013), particularly in cognitive domains such as working memory and attention (Greendale, Huang, Wight, Seeman, Luetters, Avis et al., 2009; Keenan, Ezzat, Ginsburg, & Moore, 2001; Kimura, 2002; Wroolie, Kenna, Williams, Powers, Holcomb, Khaylis et al., 2011). Nonetheless, the literature is inconclusive with respect to the link between menopause and cognition, with some reviews indicating no substantial changes in cognitive functioning (Henderson & Sherwin, 2007), or none of clinical relevance (Henderson, Guthrie, Dudley, Burger, & Dennerstein, 2003). Other studies have reported only short-lasting cognitive decline with menopause (Greendale et al., 2009) or that only limited cognitive domains are effected such as verbal fluency (Fuh, Wang, Lee, Lu, & Juang, 2006). A 2011 review of endogenous and exogenous estrogen exposures in mid- and late-life women reported no reliable association with episodic memory or executive functions (Henderson & Popat, 2011). Potential explanations for the inconsistency in findings may include differences in the timing of menopause studied, ages of women and the effects of cognitive aging in women at retirement age (e.g. 65+ years) or other covariates not measured (Shanmugan & Epperson, 2012). Moreover, teasing apart age effects from those of menopause becomes increasingly difficult when it is considered that the menopausal transition is a period in life marked by increased intra- and inter-individual variability in physiology, and in cognitive and affective response to that variability. The Penn Ovarian Aging Study, which aimed to determine whether the menopausal transition is associated with age-independent cognitive decline, found that verbal fluency but not psychomotor speed is linked to reproductive senescence independent of age (Epperson, Sammel, & Freeman, 2013).

Hormone therapy

Observational Studies

Although the primary indication for HT is to alleviate uncomfortable symptoms of menopause such as nights sweats and insomnia, early observational studies reported supplementary health gains such as decreased risk for coronary heart disease and hip fracture (Grady, Rubin, Petitti, Fox, Black, Ettinger et al., 1992). Investigations conducted by our group and others found that in comparison to non-HT users, individuals undergoing estrogenic treatment performed significantly better on tests of working memory (see LeBlanc, Janowsky, Chan, & Nelson, 2001; Maki & Hogervorst, 2003; Sherwin, 2006 for a review), verbal memory (Maki, Zonderman, & Resnick, 2001) and visual memory (Resnick, Metter, & Zonderman, 1997).

Additionally, observational studies suggested that estrogen-based HT offered protection against risk for dementia (Birge & Mortel, 1997; Haines, 1998; Henderson, 1997; Hogervorst, Williams, Budge, Riedel, & Jolles, 2000; LeBlanc et al., 2001; Melton, 1999; Panidis, Matalliotakis, Rousso, Kourtis, & Koumantakis, 2001; Zandi, Carlson, Plassman, Welsh-Bohmer, Mayer, Steffens et al., 2002) and some prospective studies indicated a 50% reduction in Alzheimer’s Disease (AD) risk in those using HT compared with non-users (Kawas, Resnick, Morrison, Brookmeyer, Corrada, Zonderman et al., 1997; Paganini-Hill & Henderson, 1996; Zandi et al., 2002). However, women who opted to use HT were inclined to be healthier overall, so that probable unadjusted confounders impacted the outcomes of such studies. Thus, despite attempts to control for possible confounds known to influence cognitive abilities including age, education and socioeconomic status, the “healthy user” bias continues to be a critical confound in observational studies. Similarly, other methodological issues inherent to observational studies include the different types and doses of hormones administered, the varying durations of HT, fluctuations in temporal proximity relative to menopause, and the route of treatment delivery, including cyclic versus continuous administration. Lastly, menopausal women who have a uterus take adjuvant progesterone for protection against endometrial hyperplasia (Froom, 1991), and a number of synthetic progestins have been administered for this purpose. Thus, the varying formulations of progestins likely differ in regard to their impact on cognition and health indices, and little is known about these effects.

Intervention Studies and the Women’s Health Initiative

Randomized controlled trials are designed to overcome the limiting factors inherent in the observational studies described above. That is, by randomly assigning women to HT and placebo groups, not only is the healthy-user bias eliminated, but the dosage, hormone preparation, route of delivery and timing factors can be strictly controlled. Unlike the beneficial effects of HT reported in observational studies, randomized controlled trials in older menopausal women have produced more mixed findings. Despite some studies describing improvements in cognitive ability in HT compared with placebo groups (Haskell, Richardson, & Horwitz, 1997; Sherwin, 2006; Zec & Trivedi, 2002), especially in younger surgically menopausal women, findings from the Women’s Health Initiative (WHI) reported in 2003, did not support a protective effect of HT in older postmenopausal women.

The WHI was a randomized controlled trial designed to study the effects of HT on health outcomes in postmenopausal women. The study’s main objective was to test the popular assertion that postmenopausal HT had a protective effect against cardiovascular disease, an important public health concern. The WHI HT trial was comprised of two parallel placebo controlled trials of conjugated equine estrogen (CEE)-based HT regimens. Enrollees were 50 to 79 years of age and postmenopausal. Active therapies consisted of 0.625 mg/day CEE in women post-hysterectomy and 0.625 g/day CEE combined with 2.5 mg/day medroxyprogesterone acetate (MPA) in women with a uterus. Although study medications were terminated in 2002 (women without prior hysterectomy) and 2004 (women with prior hysterectomy), women continue to be followed. Results were unexpected, in that a protective effect against cardiovascular disease was not found (Anderson, Limacher, Assaf, Bassford, Beresford, Black et al., 2004; Rossouw, Anderson, Prentice, LaCroix, Kooperberg, Stefanick et al., 2002). Moreover, both the CEE Alone and CEE + MPA treatments were found to increase risk for stroke and blood clots in the leg and lungs, while the CEE + MPA trial was linked to an increased risk of breast cancer.

In addition to the primary and secondary outcomes tested in the main WHI trial, two inter-related ancillary intervention studies investigated the effects of HT on cognitive outcomes in postmenopausal women. The WHI Memory Study (WHIMS) (Shumaker, Reboussin, Espeland, Rapp, McBee, Dailey et al., 1998) enrolled 7,429 women 65 years of age and older to evaluate effects of HT on risk and progression of dementia, and global cognitive function. The WHI Study of Cognitive Aging (WHISCA) (Resnick, Coker, Maki, Rapp, Espeland, & Shumaker, 2004) study enrolled 2302 WHIMS participants without dementia to provide information on the effects of HT on a larger battery of specific cognitive functions, including verbal, visual, and working memory, attention, verbal fluency, and spatial ability. While WHIMS began in parallel with the main WHI trial, WHISCA was initiated on average three years after randomization to the HT.

Findings from the WHIMS and WHISCA studies were also unexpected in that the predicted cognitive advantages and decrease in risk for dementia in the HT groups were not found. Rather, the WHIMS study reported that in women aged 65 years and older an increase in risk for dementia was found in both active treatment groups. This association was significant in the combination CEE + MPA group (HR=2.05 [1.21, 3.48]; p= 0.01) (Shumaker, Legault, Rapp, Thal, Wallace, Ockene et al., 2003), but failed to reach significance in the CEE Alone trials (HR= 1.51 [0.83, 2.74;] p= 0.18) (Shumaker, Legault, Kuller, Rapp, Thal, Lane et al., 2004). Moreover, significant decreases in global cognitive function over time were reported for both active treatment arms (Espeland, Rapp, Shumaker, Brunner, Manson, Sherwin et al., 2004; Rapp, Espeland, Shumaker, Henderson, Brunner, Manson et al., 2003), the effects of which were found to persist after treatment termination (Espeland, Brunner, Hogan, Rapp, Coker, Legault et al., 2010).

Findings from the WHISCA study demonstrated that in comparison to placebo, older postmenopausal women randomized to the combination CEE + MPA group had lower verbal learning and memory scores over an average time period of 4 to 5 years (Resnick, Maki, Rapp, Espeland, Brunner, Coker et al., 2006). No other cognitive abilities measured were affected. (Note - see Singh and Su (2013) for a review of possible confounds that may have impacted findings such as the neurobiological differences between synthetic medroxyprogesterone acetate used in the combined trials compared to endogenous progesterone). For the CEE Alone trial, the WHISCA results demonstrated no significant advantageous nor detrimental effects on age-related change in specific cognitive functions (Resnick, Espeland, An, Maki, Coker, Jackson et al., 2009a). Furthermore, neither the CEE + MPA nor CEE Alone treatment significantly impacted measures of affect or depressive symptoms (Resnick et al., 2009a; Resnick et al., 2006).

In their review of estrogen effects on AD, Henderson and Brinton (2010) speculated that in addition to the healthy-user bias, results from the WHI trials also may differ from observational findings due to differences in the age at which hormone initiation occurred. While HT use is most often initiated during or shortly after the menopausal transition in women enrolled in observational studies, the initial WHIMS studies focused on older postmenopausal women, age 65 years or older.

Critical period hypothesis

The use of HT as a remedy for menopausal symptoms was reduced globally in the wake of the findings from the WHI (Ettinger, Wang, Leslie, Patel, Boulware, Mann et al., 2012). Nonetheless, HT continues to be highly effective in relieving distressing menopausal symptoms such as night sweats and hot flushes. As mentioned, HT in the WHIMS and WHISCA studies was initiated in women aged 65 years and older, and treatment initiated at an earlier time may have produced different results. This is known as the “critical period” (Resnick & Henderson, 2002) or “window of opportunity” hypothesis, whereby cognitive benefit attributable to HT may be limited to treatment close in temporal proximity to menopause (Henderson, 2013; Maki, 2013; Sherwin & Henry, 2008).

Observational Studies

Many observational studies lend support for the critical period hypothesis. For example, an Australian study of 428 women > 60 years old found that those who had initiated HT early postmenopause performed significantly better in tests of psychomotor speed, verbal fluency and global cognition (MacLennan, Henderson, Paine, Mathias, Ramsay, Ryan et al., 2006). Another large scale study of 343 postmenopausal Danish women found similar advantages of early HT on cognitive outcomes (Bagger, Tankó, Alexandersen, Qin, Christiansen, & Group, 2005). Lastly, in a large epidemiological study in Cache County, investigators reported that in comparison to current users of HT, former users (that is, more likely to have initiated HT at menopause onset), had a reduction in risk for AD (Zandi et al., 2002). This reduction was also associated with duration of HT use.

Intervention Studies

To test the “critical period” hypothesis directly, two recent intervention studies were conducted examining the effects of HT initiation soon after menopause on cognitive function. Firstly, the WHIMS-Young (WHIMS-Y) study investigated 1326 women who had taken part in the WHI CEE-based randomized controlled trials when aged 50–55 years. An average of 14.2 years after randomization to treatment and 7.2 years after treatment discontinuation, when the women were approximately 67.2 years of age, a battery of cognitive tests were administered via a telephone interview (Espeland, Shumaker, Leng, Manson, Brown, LeBlanc et al., 2013). This study addressed the critically important question of whether cognition is impacted years later when women undergo HT during early menopause. Contrary to the initial WHI results, these data indicated neither harm nor benefit to cognitive ability in women initiating HT early in menopause.

A second large scale randomized controlled clinical trial examined the relationship between estrogenic treatment during early menopause and cognition and mood (Wharton, Gleason, Miller, & Asthana, 2013). In the Kronos Early Estrogen Prevention Study (KEEPS), 728 women, with a mean baseline age of 53 years and with a 4 year follow up period, were randomized during early menopause to one of three groups: lower dose CEE (0.45 mg/day); transdermal estradiol (50 μg/day); or placebo (Harman, Brinton, Cedars, Lobo, Manson, Merriam et al., 2005). In addressing a methodological shortcoming of prior studies, participants in the active estrogen arms were administered oral progesterone (200mg), as opposed to the synthetic progestin used in WHI, for 12 days each month. The parent KEEPS study’s primary endpoints were cardiac health measures, however, an ancillary study (KEEPS-Cog) was designed to examine several cognitive domains such as memory, attention and mood in 571 of the KEEPS women (Wharton, Gleason, Dowling, Carlsson, Brinton, Santoro et al., 2014). As with the WHIMS-Y results, there were neither advantageous nor harmful effects of HT on measures of memory or other cognitive functions. Interestingly, the KEEPS investigators did find an improvement in symptoms of depression and anxiety in women randomized to oral CEE. Taken together, women who are considering the health benefits and risks of initiating HT for the relief of menopausal symptoms can take some comfort in these findings, and they are in keeping with prior findings in older WHIMS participants that healthier women are less susceptible to any deleterious effects of HT (Resnick, Espeland, Jaramillo, Hirsch, Stefanick, Murray et al., 2009b).

Estrogens and surgical menopause

Surgical menopause is a type of medically induced menopause where both ovaries are surgically removed by bilateral oophorectomy. The most common indications for this type of surgery in premenopausal women are cervical, endometrial and ovarian cancer (Novetsky, Boyd, & Curtin, 2011; Shuster, Gostout, Grossardt, & Rocca, 2008). As the main producers of estrogen, surgical removal of the ovaries results in a sharp withdrawal of these steroid hormones, whereas naturally menopausal women encounter a more gradual reduction in ovarian hormone production over a number of years.

Observational Studies

From a cognitive standpoint, younger age at surgical menopause has been linked to an increased risk of cognitive dysfunction and dementia (Bove, Secor, Chibnik, Barnes, Schneider, Bennett et al., 2014; Nappi, Sinforiani, Mauri, Bono, Polatti, & Nappi, 1999; Phung, Waltoft, Laursen, Settnes, Kessing, Mortensen et al., 2010; Rocca, Bower, Maraganore, Ahlskog, Grossardt, De Andrade et al., 2007; Rocca, Grossardt, Shuster, & Stewart, 2012). Observational results from longitudinal studies of surgical menopause indicate that both young and old women who have undergone bilateral oophorectomy carry an increased risk of cognitive impairment and dementia (Rocca et al., 2007; Rocca, Grossardt, & Maraganore, 2008), as well as reductions in global cognition and memory. Additionally, such cognitive changes have been observed as soon as 3–6 months following surgery. On the other hand, some studies have described no or negligible effects of surgical menopause on cognitive abilities in middle-aged (Kok, Kuh, Cooper, van der Schouw, Grobbee, Michael et al., 2006; Vearncombe & Pachana, 2009) and older women (Bove et al., 2014; Kritz-Silverstein & Barrett-Connor, 2002). In general, observational studies examining the effects of HT use following surgical menopause have described cognitive benefits (Bove et al., 2014; Rocca et al., 2007). One study reporting the combined results from two different longitudinal cohorts of older women (average age 78 years) found that HT undergone within 5 years of surgery, and for a minimum of 10 years, was associated with a reduced decline in global cognitive function (Bove et al., 2014).

Intervention Studies

Barbara Sherwin and her colleagues performed a series of studies in the late 1980s and early 1990s examining the effects of HT in younger surgically menopausal women. The first intervention study tracked 50 pre-operative patients who were randomly assigned to hormone (estrogen-androgen, estrogen or androgen) treatment or placebo control groups. Following one month of baseline observation where four tests of cognition were administered, the patients underwent surgery. Results showed that in comparison to the active hormone group, significantly lower scores for the placebo group post-surgery were found for logical reasoning and short- and long-term memory (Sherwin, 1988). A later intervention study followed 19 women before and after bilateral oophorectomy and hysterectomy surgery, tested on a variety of memory measures (Phillips & Sherwin, 1992). Again, the women were randomly assigned to HT or placebo groups. There was a significant group by time interaction in that HT prevented the decline in immediate and delayed verbal memory associated with surgical menopause. However, the beneficial effect of HT was domain specific as immediate and delayed recall of visual material and digit span scores were not subject to hormonal effects in this small trial. These studies highlight that hormone treatments in younger women appear to protect against post-surgery cognitive decline, and may be especially beneficial to verbal memory performance.

Overall these studies combined suggest that with surgically induced menopause, the temporal proximity and duration of HT play a critical role in its neuroprotective efficacy. Clinical studies also lend support for the idea that cognitive function is negatively affected by surgical menopause and that women undergoing this procedure may experience greater benefit from HT (Farrag, Khedr, Abdel-Aleem, & Rageh, 2002; Henderson & Sherwin, 2007). Nonetheless, findings are still mixed and the literature is generally inconsistent in its conclusions (Maki & Hogervorst, 2003). Further, there is a lack of both observational and intervention studies where surgical menopause is clearly defined and objective cognitive deficits, in contrast to cognitive symptoms, are assessed.

Estrogenic treatment and the vulnerable brain

As mentioned, there is reason to believe that from a cognitive standpoint, HT in older women may have variable effects in naturally menopausal women, and that this variation may be linked to health status. WHIMS-MRI was conducted to better understand the mechanism contributing to the deleterious effects of HT on cognitive outcomes as well as dementia risk (Coker, Espeland, Hogan, Resnick, Bryan, Robinson et al., 2014; Coker, Hogan, Bryan, Kuller, Margolis, Bettermann et al., 2009; Resnick et al., 2009b). As HT increases risk for stroke and thromboembolic events, it was hypothesized that increases in ischemic burden might explain the adverse effect of HT on cognition. In addition, WHIMS-MRI examined potential effects of HT on variations in regional brain volumes, including hippocampal and frontal regions. Ischemic lesion burden and regional brain volumes were investigated in 1403 women from the WHIMS study randomly assigned to CEE + MPA, CEE Alone, (on average 3.0 and 1.4 years after cessation of treatment, respectively) or placebo groups. Contrary to prediction, HT versus placebo groups did not differ significantly with respect to ischemic lesion burden (Coker et al., 2009). In contrast, results showed a decline in total brain, frontal and hippocampal volumes in women formerly assigned to HT versus placebo. Furthermore, the original WHIMS study found that baseline mental status at enrollment, as measured by a modified Mini-Mental State Exam (3MSE), was a significant factor moderating the deleterious cognitive effects of HT (Espeland et al., 2004). That is, the women with lower 3MSE scores at baseline had significantly greater HT-related cognitive reductions than the women with higher 3MSE scores at baseline. Analogous to these findings, WHIMS-MRI also demonstrated that the deleterious cognitive effects of HT on brain volumes were most marked in women with the lowest 3MSE scores at study commencement in the WHI, as well as those with the highest burden of white matter lesion volumes (Resnick et al., 2009b). These data suggest that poorer cognitive health at HT initiation may increase vulnerability to adverse effects of estrogens on the postmenopausal brain. Consistent with these findings, recent results from the Three Cities study indicated an interaction between diabetes and endogenous estradiol levels on risk for dementia. In women with diabetes, higher estradiol levels were associated with a marked increase in risk for dementia in older women (HR=14.2 [1.6,123]), with a more modest increase in women without diabetes (HR=3.4 [0.1,147]) (Carcaillon, Brailly-Tabard, Ancelin, Rouaud, Dartigues, Guiochon-Mantel et al., 2014).

Research with more basic science models has provided insights into mechanisms that may explain disparate findings between the generally positive impacts of estrogenic action in the brain in animal models (Luine, 2014) and the adverse consequences of therapeutic human interventions. For example, Brinton (2005, 2008) has proposed the ‘healthy cell bias of estrogen action’, based on the premise that exposure of healthy neurons to estrogens invokes a neuroprotective response that promotes both neurological function and cognition. On the other hand, in compromised or diseased cells, estrogenic treatment over time will exacerbate neurodegeneration, pointing to a vulnerability in certain brains for HT. These findings are consistent with in vitro analyses of estrogen treatment of hippocampal neurons prior to, or after amyloid-β insult, which show neuroprotective outcomes versus no benefit or worsened degeneration, respectively (Morrison, Brinton, Schmidt, & Gore, 2006). Taken together, these findings suggest that women who are less neurologically healthy, including older women, may be most vulnerable to any deleterious cognitive effects of HT, and in contrast, that HT may confer neuroprotective benefits on healthy brains.

In addition, studies in animal models suggest that specific regimens of HT may influence cognitive outcomes. In contrast to the WHI findings in older women where continuous hormone supplementation was administered, research in aged surgically menopausal rhesus monkeys showed that a cyclic regimen of HT (21 days via injection) compared to vehicle-injection is sufficient to preserve prefrontal cortex-dependent cognition, measured by spatial working memory (Rapp, Morrison, & Roberts, 2003). The large effect found on delayed response performance implicates the dorsolateral prefrontal cortex as the potential target of estradiol treatment (Morrison & Baxter, 2012). Thus, optimal HT regimens to maximize brain health and cognitive outcomes remain a subject of ongoing research (Baxter, Roberts, Gee, Lasley, Morrison, & Rapp, 2013).

Concluding remarks

This article provides an overview of the current literature concerning the effects of postmenopausal hormone treatments on cognitive function. We review the relationships between sex steroid hormones and cognitive functioning, highlighting variation in potential HT effects in relation to age, hormone regimen, timing of treatment relative to menopause, and baseline cognitive and brain integrity. Although observational studies suggested that HT would benefit cognitive function and reduce the incidence of cognitive impairment, including dementia, results of the WHIMS randomized clinical trials in older postmenopausal women did not support the observational findings and showed poorer cognitive outcomes in women randomized to HT. The “critical period” hypothesis (Resnick & Henderson, 2002) posits that early treatment of younger postmenopausal women, closer to the menopausal transition, might be more beneficial to cognition. However, two recent randomized trials (WHIMS-Y and KEEP-Cog) testing this hypothesis showed neither harm nor benefit of HT interventions closer to menopause. Although the lack of cognitive benefit in these trials is disappointing, the absence of short and long-term harm to cognitive function should reassure women who choose to use HT for treatment of menopausal symptoms. Lastly, MRI (Resnick et al., 2009b) and cellular studies (Brinton, 2005, 2008) point towards a greater adverse effect of postmenopausal estrogens on a vulnerable brain, whereby women who are less healthy, either cognitively or neurologically, seem to be at greater risk of having an adverse response.

In summary, investigations of hormone treatments in menopausal women have produced many inconsistent findings. The literature is vast and diverse with long-term cognitive outcomes still being realized. However, moving forward, results from more recent clinical intervention trials can help inform clinical practice with regards cognitive function as well as other important late-life outcomes such as risk for dementia. Ongoing research can build on the wealth of information from both the basic sciences and human studies to inform the types and timing of treatments that best serve menopausal women in the future.

Highlights.

• Effects of menopause on cognition

• Hormone therapy

• Intervention studies and the women’s health initiative

• ‘Critical period’ hypothesis

• Estrogens and surgical menopause

• Estrogenic treatment and the vulnerable brain