|Home | About | Journals | Submit | Contact Us | Français|
On 28–29 September 2004, the National Institute on Aging (NIA) convened scientists for a workshop on the aging female brain focused on translating into clinical practice discoveries concerning estrogens and progestogens. Workshop objectives were to examine effects of estrogen and progestogen on brain and cognitive function in relation to aging, to examine consistencies and apparent discrepancies between Women’s Health Initiative Memory Study findings and other research on cognitive function, to determine whether additional hormone interventions could be developed in this area, and to offer advice on design of clinical trials for other interventions that might ameliorate cognitive aging. Following the workshop, participants joined by other interested scientists organized into regional work groups to continue the dialogue begun in Bethesda and to propose recommendations for NIA. The resulting recommendations, referred to as the “Frontiers Proposal for Estrogen and Cognitive Aging”, acknowledge the persistence of critical gaps in our understanding of how decline in ovarian steroid secretion during reproductive aging and use of ovarian steroid hormone therapy affect normal brain function and risk for late-life neurodegenerative disorders such as Alzheimer’s disease. There is a pressing need for preclinical, human, and integrated studies on the relationship between the menopausal transition and midlife exposures to estrogens, progestogens and related compounds, and risks for age-associated cognitive disorders. Research is also needed on better predictors of adverse cognitive outcomes, valid biomarkers for risks associated with hormone therapy use, enhanced tools for monitoring brain function and disease progression, and novel forms of therapy for improving long-term cognitive outcomes.
There is considerable interest in identifying modifiable risk and protective factors for cognitive decline associated with usual aging and with common dementing disorders such as Alzheimer’s disease (AD). Among these factors are estrogens and progestogens, ovarian hormones with profound effects on many tissues and organs, including the brain. These effects include the regulation of non-reproductive processes such as perception, cognition, mood, and motor control.
On 28–29 September 2004, the National Institute on Aging (NIA) of the United States (US) National Institutes of Health convened scientists from basic science, epidemiology, and clinical trials fields for a workshop held in Bethesda, MD. Cosponsors for the workshop, entitled “Bench to Bedside: Estrogen as a Case Study” were the National Institute of Mental Health, Office of Research on Women’s Health, and the Alzheimer’s Association. The focus of the workshop was on scientific data concerning the use of estrogens and progestogens in the aging female brain: what is already known and what will be required to translate discoveries from the laboratory workbench, from epidemiology, and from clinical trials into everyday medical practice.
Defined workshop objectives were the following: (1) to examine apparent discrepancies between the findings of the Women’s Health Initiative Memory Study (WHIMS) on brain and cognitive function after estrogen and progestogen treatment, and basic research and epidemiological studies, which have often differed in terms of the formulations, doses and modes of administration, and in terms of treatment timing and duration of the hormone therapy (HT) employed; (2) to examine effects of estrogen and progestogen on brain and cognitive function in relation to aging; (3) to determine what is known and what information would be needed to establish whether additional hormone interventions could be developed in this area; and (4) to determine the lessons learned from studies on estrogen that might guide the design of clinical trials for other classes of drugs that might affect cognitive aging. The workshop agenda with abstracts and presentations is available on-line at http://www.nia.nih.gov/ResearchInformation/ExtramuralPrograms/NeuroscienceOfAging/NNA_Conferences/WorkshopAgendaAbsPres.htm; summary notes can be found at http://www.nia.nih.gov/ResearchInformation/ExtramuralPrograms/NeuroscienceOfAging/NNA_Conferences/WorkshopSummaryNotes.htm.
Following the workshop, participants joined by other interested scientists organized into regional work groups to continue the dialogue begun in Bethesda and to propose recommendations for NIA. One venue for ongoing discussion was the biennial Graylyn Conference on Women’s Cognitive Health, co-sponsored by the NIA and hosted by the Wake Forest University School of Medicine. Meetings in October 2005 and October 2007 focused on translational research and included many participants from the Bethesda Bench to Bedside workshop.
The regional work group recommendations are described below as the Frontiers Proposal for Estrogen and Cognitive Aging, implying that there remain frontiers for discoveries concerning estrogen and other sex steroids in relation to the aging brain, cognition, and cognitive disorders. The following sections provide background information, consensus recommendations by work group members, a concluding perspective, and an executive summary.
Central nervous system consequences of menopausal estrogen and progesterone loss is an area of intense study, because many actions of these sex steroids are potentially relevant to cognitive aging and dementia. Both in vitro and in vivo analyses indicate that estrogen protects against a wide spectrum of neurodegenerative insults, a substantial number of which are linked to AD. A key finding relevant to the development of AD is the ability of estrogen to reduce formation of β-amyloid (Petanceska et al. 2000; Yue et al. 2005), a hallmark biochemical marker of this illness. Moreover, estrogen promotes both morphological and electrophysiological correlates of learning and memory in vitro and in vivo (Brinton et al. 1997; Hao et al. 2006; Murphy and Segal 1996; Toran-Allerand et al. 1980; Woolley and McEwen 1992). Myriad neurotransmitter systems relevant to cognition (acetylcholine, serotonin, noradrenalin, dopamine, glutamate) are regulated by estrogen. Studies in adult and aged non-human primates have revealed both behavioral and neurobiological effects of estrogen, some of which counter the effects of normal aging on cognition (Hao et al. 2006; Rapp et al. 2003a, b; Tinkler et al. 2004; Voytko 2002). Overall, basic science analyses using both in vitro and in vivo model systems indicated that estrogen—typically 17β-estradiol but also conjugated equine estrogens—protect neurons against insults associated with AD (Brinton 2005). Moreover, these same estrogens in the same model systems can activate biochemical, genomic, cellular and behavioral mechanisms of memory (Frye et al. 2007; Singh et al. 1994; Toran-Allerand 2000). An important aspect of these studies, and of virtually all of the basic science in vitro and in vivo analyses, is that neurons were healthy prior to estrogen exposure and prior to exposure to neurodegenerative insults or lesions (Brinton 2005). Typically, basic science analyses have used a preventive model, i.e., neurons are pretreated with an estrogen or progestogen prior to exposure to the neurological insult. Studies on aged rodents have reinforced this concept, i.e., that healthy neurons respond differently than compromised neurons, in that hippocampal synapses are less responsive to estrogen in aged rats (Adams et al. 2001), and the behavioral impact of estrogen is less impressive in aged rats than in young animals (Daniel et al. 2006; Markowska and Savonenko 2002; Savonenko and Markowska 2003). These data have led to the notion that there may be correlates with age or with duration of time after menopause that establish a “window of opportunity” for estrogen to be effective in protecting circuits mediating mood and cognition, with decreased responsiveness once this time frame passes (Gibbs 2000; Resnick and Henderson 2002). Indeed, increasing basic science evidence indicates the importance of the hormonal state, the duration of estrogen or progesterone privation, and the impact of aging on the neural response to estrogen (Brinton 2005; Gibbs 2000; Johnson and Sohrabji 2005).
The issue of progestogen action in the brain is less well explored, but the existing data indicate that the type of progestogen is critical to functional outcome in neurons. In some model systems, progestogens such as medroxyprogesterone acetate (MPA) are associated with outcomes very different from those associated with progesterone (Nilsen and Brinton 2003). At the same time, there is increasing evidence that progesterone has neuroprotective effects that may be very useful clinically (Stein 2007).
The Women’s Health Initiative, the parent study for WHIMS (Shumaker et al. 2003, 2004), was a multicenter, randomized, double-blind, placebo-controlled trial designed to assess effects of a commonly prescribed oral HT on cardiovascular disease, breast cancer, and other health outcomes (Anderson et al. 2003; Rossouw et al. 2002). Active treatment was with combined HT for women with a uterus [conjugated equine estrogens (CEE) 0.625 mg/day with MPA 2.5 mg/day as a continuous combined preparation; CEE+MPA trial) or with estrogen alone (CEE 0.625 mg/day; CEE-alone trial) for women without a uterus. WHIMS assessed the incidence of dementia and other cognitive outcomes among community-dwelling women who at baseline were between the ages of 65 and 79 years inclusive. After a mean follow-up of 4–5 years, 61 women in the CEE+MPA trial developed dementia, as did 47 women in the CEE-alone trial. About one-half of cases were attributed to AD, but separate analyses were not reported for specific dementia subtypes. In both trials, dementia incidence was increased for women assigned to active treatment compared to placebo [CEE+MPA hazard ratio (95% confidence interval) 2.05 (1.21–3.48); CEE-alone, 1.49 (0.83–2.66); Shumaker et al. 2003, 2004]. Mild cognitive impairment (MCI), thought to indicate increased risk of dementia, was a secondary WHIMS outcome; an insignificant increase in MCI risk was attributed to HT in the CEE+MPA trial [hazard ratio 1.07 (0.74–1.55)] and in the CEE-alone trial [hazard ratio 1.34 (0.95–1.89); Shumaker et al. 2003, 2004]. WHIMS findings on global cognition are considered below.
WHIMS results on dementia were unanticipated based on the basic science research described above, which provided a strong rationale why estrogen might be neuroprotective, and based on observational studies that had considered the association between HT use and AD risk. Most observational studies have shown a protective association when ever-users of HT were compared to women who had never used HT, with meta-analyses suggesting overall risk reductions of about one-third (Hogervorst et al. 2000; Yaffe et al. 1998). In the Cache County prospective cohort, the protective association of HT was modified by timing, and was seen with past, but not current, HT use (Zandi et al. 2002); in the Multi-Institutional Research in Alzheimer Genetic Epidemiology case-control study, the protective association of HT was modified by age and was seen among younger, but not older, postmenopausal women (Henderson et al. 2005).
Complaints of poor memory are common around the time of the menopausal transition (Mitchell and Woods 2001), but forgetfulness is common at other ages as well (Neugarten and Kraines 1965) and memory symptoms may be attributed to a variety of cognitive and noncognitive factors (Mitchell and Woods 2001). After ovarian estrogen production is halted abruptly by oophorectomy or by use of a gonadotropin-releasing hormone analog, there is clinical trial evidence that short-term use of estrogen has a beneficial effect on cognitive performance, at least for memory for verbally encoded information (Phillips and Sherwin 1992; Sherwin and Tulandi 1996); whether other younger women may similarly benefit from HT is less apparent (Henderson and Sherwin 2007). Importantly, long-term cognitive effects of HT used in middle age are unknown. Available evidence suggests that natural menopause per se does not have an important effect on verbal memory or most other cognitive skills (Fuh et al. 2006; Henderson et al. 2003; Kok et al. 2006), at least in the short run, and serum estrogen concentration at midlife is unrelated to cognitive performance (Henderson et al. 2003; Herlitz et al. 2007). The relationship between endogenous estradiol levels and dementia risk is also controversial (Geerlings et al. 2003; Schupf et al. 2006). However, one follow-up study of women who had participated in randomized clinical trials of HT during middle age found that those randomized to active treatment had a lower risk of cognitive impairment 5–15 years later than those randomized to placebo (Bagger et al. 2005).
Among older postmenopausal women, observational studies on HT and cognition provide conflicting evidence (Carlson et al. 2001; Kang et al. 2004), but randomized clinical trials are more consistent in showing no cognitive benefit of HT in this older age group (Grady et al. 2002; Viscoli et al. 2005; Yaffe et al. 2006; reviewed by Henderson and Sherwin 2007). These clinical trial results are similar to those reported from WHIMS, where over a mean follow-up of about 5 years participants were assessed annually with a test of global cognition (Espeland et al. 2004; Rapp et al. 2003a, b). In HT and placebo groups, global cognitive scores tended to increase over time, presumably representing practice effects. The magnitude of increase was somewhat less among CEE-MPA and CEE-alone users than placebo-users, but mean between-group differences were small and not clinically important (Espeland et al. 2004; Rapp et al. 2003a, b). In a subgroup of WHIMS participants assessed more extensively (the Women’s Health Initiative Study of Cognitive Aging), combined estrogen-progestin HT had contradictory effects, impairing verbal memory but showing a trend toward improving nonverbal memory (Resnick et al. 2006). Other comparisons in this analysis were not significant, and mean magnitudes of any between-group differences were small.
Differences between results of observational studies and WHIMS are most evident with respect to dementia outcomes. Reasons for apparent discrepancies might include (1) differences in outcomes (AD vs all-cause dementia), (2) differences in HT formulations, doses and routes of administration (although estrogen exposure in most observational studies was primarily oral CEE and progestogen exposure was primarily MPA), (3) differences in timing of HT exposures (often used at a younger age near the time of menopause in observational studies vs used only at an older age remote from the time of menopause in WHIMS), and (4) systematic bias in observational studies (Henderson 2006). Concerning the second possibility, it can be noted that women in the WHIMS trial with a uterus received HT as a continuous combined CEE+MPA preparation, putatively a less physiological approach than sequentially administered hormones and a formulation that was not yet available at the time most observational results were conducted. The third and fourth possibilities may be most crucial: are observational studies fatally flawed by biases such as recall bias or the healthy user bias (Hemminki et al. 1993; Matthews et al. 1996), or do WHIMS findings of enhanced dementia risk—using a specific HT formulation in older postmenopausal women—fail to generalize to younger women more likely to use HT for moderate to severe vasomotor symptoms?
At least two clinical trials will consider differential effects of early and late HT, the Kronos Early Estrogen Prevention Study (KEEPS) (Harman et al. 2005), and the Early versus Late Intervention Trial with Estrogen (ELITE) (http://www.clinicaltrials.gov/ct/show/NCT00114517). In both, cognition will be assessed as secondary outcomes, although the trials are relatively small and further studies will almost certainly be required. Results will be very important in guiding future research initiatives.
There is considerable evidence from both pre-clinical and human studies that ovarian steroids have widespread neuroregulatory potential throughout the nervous system. These hormones mediate effects within the brain relevant to the pathophysiology, course, and treatment response characteristics of several brain disorders. However, critical gaps exist in our knowledge of both the effects on brain function of declining ovarian steroid secretion during reproductive aging, and the role of ovarian steroid HT in the prevention or treatment of brain diseases. Ovarian steroids exert a wide range of physiologic effects within the body. Their signaling pathways are extremely complex; involve many different organ systems, and interact with a variety of other hormonal systems. Thus the manifold nature of their regulatory capacity is accompanied by a wide range of potential actions—some beneficial under certain conditions while detrimental under others. Indeed, one observation emerging from the literature is that the effects of ovarian steroids on brain function are not uniform, and several potential sources of variability have been identified that may explain these otherwise discordant results. First, both the chronological and reproductive age of the individual, as well as their health status, influence the observed response to administered ovarian steroids. Thus, the timing of administering HT is an important determinant of both the health benefits and risks to the individual. Second, the type of steroid preparation could mediate a range of diverse effects within the same organ system, resulting in benefits accrued with one hormonal preparation and either no benefits or adverse health events caused by another preparation. Finally, the schedules of administration of estrogen and progestogen (i.e., cyclic or continuous combined) are important to consider with respect to effects on brain. As a caveat, the work group membership was concerned about two trends in prior studies that have examined the role of ovarian steroids (both endogenous and exogenous forms) in human and animal brain function: (1) the majority of clinical studies were not designed to measure brain function specifically as the primary endpoint, and (2) basic science studies were focused narrowly on healthy models and did not incorporate the spectrum of aging and disease found in the human condition.
Future efforts should focus on developing the following: (1) improved understanding of the variables that predict those women in whom the menopause transition adversely alters the course of specific brain diseases; (2) biomarkers for those women who will benefit from the introduction of ovarian HT and those for whom HT is contraindicated; (3) novel therapeutics, such as selective ovarian steroid (i.e., estrogen, progesterone, and androgen) modulators that enhance the potential therapeutic benefits of these steroids in the brain but have more acceptable long-term safety profiles; and (4) clinically relevant assessment tools of both brain function and disease progression.
The following is a Frontiers Proposal for the development of research initiatives to investigate effects of ovarian steroid hormones on cognitive aging and brain function. This proposal for the NIA of the US National Institutes of Health has four goals:
It is recommended that the NIA develop an integrated program to coordinate NIA-sponsored basic and clinical research studies in the areas described above. Such a program would provide a firm scientific foundation for the endocrine aspects of neural aging and increase the knowledge available to women before they enter menopausal transition, particularly with respect to confusing discrepancies between observational and clinical trial data on HT and cognition. This investigator-based initiative identified several issues that require the attention of clinicians and scientists who work in this arena:
Finally, it is recommended that NIA provide financial support for a future workshop to develop these questions into specific investigations that would warrant new requests for applications or could be integrated into the existing portfolio of ongoing NIA-sponsored studies.
This report was prepared by the National and Regional Chairs, with input from the membership of regional Work Groups.
National Chair: Bruce S. McEwen, Rockefeller University, New York, NY
East Coast Work Group:
West Coast Work Group:
Mid-West Work Group:
Frontiers Proposal for Estrogen and Cognitive Aging Writing Group on behalf of the Frontiers Proposal for Estrogen and Cognitive Aging Work Groups. Writing Group members are Sanjay Asthana, Roberta Diaz Brinton, Victor W. Henderson, Bruce S. McEwen, John H. Morrison, and Peter J. Schmidt. Work Group members and their affiliations are given in the Appendix.
Sanjay Asthana, Email: ude.csiw.enicidem@as.
Roberta Diaz Brinton, Email: ude.csu@notnirbr.
Victor W. Henderson, Email: ude.drofnats@nosrednehv.
Bruce S. McEwen, Email: ude.rellefekcor@newecm.
John H. Morrison, Email: firstname.lastname@example.org.
Peter J. Schmidt, Email: vog.hin.liam@tdimhcsretep.