Cerebrum Article

A Neuroscientist Thinks About Menopause (If Her Hormones Are Willing)

The ground rules for menopausal women changed dramatically in 2002, when the U.S. government halted a study of hormone replacement therapy because increased health risks had been discovered. The controversy over the effects of hormone replacement on women’s bodies continues, but what about the effects on the brain? Can replacing the estrogen that naturally declines with age improve memory and other mental abilities? Do the benefits outweigh the risks? A neuroendocrinologist who has spent her career studying how the nervous system and hormones interact turns to the human and animal research, and a consideration of her values and priorities, in weighing her own decision about HRT. 

Published: January 1, 2004

According to the handy publication that helps you plan your IRA withdrawals, Supplement to Publication 590 of the Internal Revenue Service, in the United States today a 47-year-old woman can expect to live another 37 years. Imagine that a woman that age is already a year past menopause, having reached that stage of life only slightly earlier than the average of 51. Menopause is defined as a full year without a menstrual cycle, but that is just a symptom reflecting a decline in the production of two classes of ovarian hormones, estrogens and progestins. If this 47-year-old woman reached puberty at age 13, she will end up spending more of her adult life without estrogen and progestin than with them.

Until July 2002 (a date we will come back to later), she would have been encouraged by her doctor to replace the missing hormones. Drug companies make replacement convenient by packaging tablets that constitute a daily regimen of hormone replacement therapy (HRT). Because estrogens, if given alone, are associated with increased growth of the lining of the uterus and, in some cases, uterine cancer, women who have not had a hysterectomy are prescribed a combination of estrogen and progestin. Discussion during a visit to her doctor would have emphasized the immediate alleviation of menopausal symptoms such as hot flashes and vaginal dryness, with the added inducement of protection against osteoporosis, cardiovascular disease, and dementia.1 Studies showing that women using HRT have a better quality of life—improved mood, more restful sleep—might also have been mentioned. Of course, the risk factors for each individual patient would be weighed before a decision, but many, if not most, conscientious practitioners would have emphasized the health benefits—and the “normalcy”—of having estrogen and progestin present. Women, many with a history of successfully using hormonal contraceptives, would find little to fear in these discussions.

A neuroscientist finding herself in this situation might have carried an additional tacit agenda to her doctor’s office. She would know that animal and clinical studies leave no doubt that hormones such as estrogen specifically target the brain. Her primary concerns might have been those the doctor did not mention: Does menopause impair cognition? Can hormone replacement solve this problem? Has my IQ already dropped? The key term here, cognition, refers broadly to all of the unconscious and conscious mental work of information processing: perceiving, reasoning, remembering. For a demanding patient with a challenging, competitive professional life, losing even a tiny bit of cognitive capacity might be hard to take.

When I found myself in this situation, I was relieved when I left my doctor’s office with prescriptions for estradiol (a form of estrogen) and the most commonly prescribed progestin, medroxyprogesterone acetate. Abbreviated MPA, this is also called Provera, or, when used as an injectable contraceptive, Depo-Provera. Because I am slight of stature, my doctor was concerned about my bone density, already showing signs of thinning; estradiol would take care of this. But my primary concern was my brain. Clinical studies by respected researchers, such as Barbara Sherwin, Ph.D., of McGill University, had shown that estrogen treatment improved specific aspects of cognitive function in women, especially the recall of verbal information.2 Didn’t almost all of my professional life depend on the recall of verbal information?

Lydia DonCarlos, Ph.D., a friend and expert on hormone receptors who works at the Stritch School of Medicine of Loyola University, Chicago, titled a 2001 review article for Progress in Neurobiology “Neuroprotection by estradiol.” She and her co-authors summarized data on estrogen and the brain from more than 375 human and animal research studies.3 Their conclusion was that “Estrogen exposure decreases the risk and delays the onset of Alzheimer’s disease and schizophrenia, and may also enhance recovery from traumatic neurological injury such as stroke.” Hadn’t three of my grandparents suffered strokes? Hadn’t my father’s mother displayed symptoms of dementia before her death? Not only did I need estrogen to preserve my memory, but I also needed estrogen to protect myself against brain damage.

The Power of Hormones

My own area of research is neuroendocrinology, the study of how the nervous system and hormones interact. My professional training and research have shown me the powerful effects of hormones on animal behavior and created the strong conviction that my brain would be better off with estrogen than without it. What is it about hormones in general and estrogen in particular that is so compelling?

Experiments, conducted during the 1980s while I was a graduate student working with Donald Pfaff, Ph.D., and Joan Morrell, Ph.D., at the Rockefeller University, first showed me in the most vivid way possible the powerful effects of hormones on behavior. If a small amount of a peptide hormone called oxytocin is infused into the brain of a virgin female rat and she is then placed into a cage with young pups, she immediately begins to care for them as if they were her own litter, even crouching over them as if to give them milk. Conversely, infusion of any substance that blocks the actions of oxytocin into the brain of a recently pregnant rat supplied with an equally attractive set of pups results in exactly the opposite outcome: This second rat is supremely indifferent to the pups. My excitement about these experiments set me on a lifelong path. I began by studying rodents, but, for nearly 20 years now, I have studied hormones in insects.

“Hormone” is the general term for chemicals that produce the signals used by cells to communicate across the vast (from the perspective of a single cell) distances of the body. Hormones are found in all multicellular organisms with a circulatory system, which means not only in vertebrate and invertebrate animals but also in plants. They can be grouped according to their chemistry into several broad categories, the most important of which are the peptide (small protein) hormones and the steroid hormones.

Peptides are produced by many tissues of the body, including specialized nerve cells in the brain called neurosecretory cells. Dozens of peptide hormones such as oxytocin have been identified by scientists, but the list, even for humans, is not yet complete. These hormones dissolve readily in water, blood, or spinal fluid. They cannot cross fat-rich cell membranes, so they influence cells by attaching to special receptors embedded in the cell membranes. A peptide affects cells in much the same way as someone who arouses the occupants of a house by knocking on the door but does not come in.

When a peptide hormone attaches to its receptor, it triggers a biochemical chain reaction of events inside the cell. The result depends on the specific hormone and the type of cell involved, but it usually involves changes in the cell’s electrical excitability or the activation of enzymes, a cell’s natural chemical catalysts. Peptide hormones change cell function rapidly. Oxytocin, for example, is well known to many new parents for its ability to cause contractions of the muscles of the uterus or mammary glands shortly after it is released from its cellular storage sites or administered by an obstetrician to shorten labor.

By contrast, steroid hormones are small lipid (fatty) molecules derived from cholesterol. Cellular enzymes in organs such as the ovaries, testes, and adrenal glands convert cholesterol into a steroid hormone precursor, which can be converted by enzymes into active hormones. Which particular steroid hormone is synthesized by a specific gland depends on which enzymes are present to modify the hormone precursor. Groups of steroid hormones include the estrogens, androgens, progestins, mineralocorticoids, and glucocorticoids. Estradiol and estrone, the major products of the human ovary, are estrogens, whereas progesterone (a natural hormone) and MPA (a synthetic hormone) are progestins. Different steroid hormones share many steps in the chemical pathways used to create them, and some of these hormones are precursors for other hormones. This conversion of one hormone into another can occur far from the site where the hormone was initially synthesized. The brain, for example, can use a particular enzyme to convert testosterone to estradiol, which means a “male” hormone in the blood can act as a “female” hormone in the brain.

Because they originate in cholesterol, steroid hormones can move across cell membranes. From the cell’s perspective, steroid hormones are like the impolite visitor who enters your home without knocking. As a consequence, receptors for steroid hormones are found primarily inside the cell, often in the nucleus. There are specific receptors for each class of steroid hormone. The categories of receptors for steroid hormones roughly match the categories of steroid hormones themselves, so that all the different members of a class of steroid hormones will bind to their own receptor. Sometimes, though, more than one receptor is matched with each category of hormone. For example, the so-called estrogen receptor is actually two distinct receptors.

The existence of multiple forms of each steroid hormone, multiple receptors for each steroid, and conversions of steroids into other forms within tissues permit cellular responses to these hormonal signals to be finely tuned. But it makes interpreting the effects of specific hormone treatments much more complicated. This is no one hormone, one receptor system.

How Hormones Can Shape Behavior

The most intensely studied effect of steroid hormones on behavior is how sexual receptivity can be produced in female rodents by treating them with estradiol followed by progesterone.4 In female rats, natural sexual receptivity is restricted to a short part of their four-day reproductive cycle and is signaled—to both male rats and neuroendocrinologists—by a posture called lordosis in which the female conspicuously elevates her head and rump. This behavior is typically not seen until hours after estrogen treatment begins because, in contrast to the rapid actions of the peptide oxytocin, the effects of estrogen depend on the synthesis of new proteins by the body, which takes time.

A cell that has a receptor for a particular hormone is often referred to as that hormone’s target. In the examples of both oxytocin (maternal behavior) and estrogen (sexual behavior), the target cells for hormone action are nerve cells in the brain. We know this because application of the hormones to the brain can induce changes in behavior. Estrogen alters the brain by changing its chemistry and its structure. For example, Victoria Luine, Ph.D., of Hunter College in New York City has shown that female rats who have had their ovaries surgically removed have lower levels of a brain enzyme called choline acetyltransferase if they are not treated with estrogen. Choline acetyltransferase synthesizes the neurotransmitter acetylcholine, and a reduction in brain acetylcholine is associated with memory deficits not only in laboratory rats but also in humans with Alzheimer’s disease. In another example, research in the laboratory of Bruce McEwen, Ph.D., at the Rockefeller University showed that ovariectomy alters the fine structure of nerve cells in the hippocampus of the brain, reducing the number of spines, which are sites of synaptic communication between nerve cells. Treatment of ovariectomized rats with estrogen returned spine numbers to normal levels.5

Oxytocin and estrogen are examples of how a hormone acts in the brain to ensure that behavior is coordinated with what is going on in the body. The actions of oxytocin in the brain link birth and lactation to maternal behavior; the actions of estrogen in the brain promote fertility by linking ovulation (also regulated by estrogen) to sexual receptivity. Research in the 1980s and 1990s, however, enlarged our view of these hormones (and many others) by showing a broader range of effects on the brain. Oxytocin is now known to regulate many social and affiliative behaviors. In small rodents called prairie voles, for example, oxytocin underlies the formation of monogamous pair bonds.6

Similarly, estrogens are now known to exert effects on the brain that have no obvious direct link to reproduction, and some of these effects are on cognition.7 The research reports describing these effects are both fascinating and controversial, with the controversy stirred up, in part, because hormone action on the brain, especially early in life, appears to result in differences between the sexes in performance on various cognitive tasks. Adding to this uncertainty is that the effects of estrogen on cognition are typically rather modest, so they are not always detected, and the importance in daily life of any effects seen in the laboratory is debatable.

Doing Menopause Right

The widespread use of HRT to alleviate the symptoms of menopause and to prevent osteoporosis and heart disease allows us to study the effects of estrogen on cognition in one specific large group of humans: postmenopausal women. These studies have important implications from a public health perspective, as well as for individual women trying to make good choices. After all, the chance to “do menopause right” comes only once.

Let me describe what happens (I like to think) after I take my daily tablet of estradiol. After being absorbed into my bloodstream, some estradiol is carried to my brain. Subpopulations of nerve cells in specific areas of my brain, particularly the preoptic area, hypothalamus, hippocampus, and forebrain, are targets for estrogens. Therefore, the estradiol I have taken binds to the estrogen receptors in the nucleus of these cells; the hormone/receptor combination then binds to responsive elements in my DNA, encouraging the synthesis of the proteins encoded by these specific genes.

Some of these proteins affect the processing of neurotransmitters, chemical signals used for communication between nerve cells. Other proteins alter the electrical activity of nerve cells; still others are cell building blocks, structural proteins that change the length of nerve cell processes or the size of synapses, the points of contact through which nerve cells communicate. Some cells might even be stimulated by the estradiol I have taken to produce a few new nerve cells.8

I then put a positive spin on this schema by assuming that the net outcome of all of these actions is improved cognitive function. After taking my estradiol, I believe I recall what I have read more easily, I am articulate as a lecturer, I am clever in discussions with my colleagues. I never misplace my car keys or forget the name of a student I have just met. Given these effects, who wouldn’t want to take estrogen?

Estrogen, Memory, and Cognition

The preceding scenario is wholly imaginary, in part, because I have never taken an estrogen without also taking a progestin and, in part, because the effects of hormones on cognition are subtler than the effects of hormones on behaviors such as maternal care and sexual receptivity. Still, granted this element of wishful thinking, my scenario has a basis in animal studies of hormones and cognition.

My specialty, neuroendocrine studies, often uses animal models, because animals can be randomly assigned to different treatment groups, compliance with treatment is ensured, and surgery (removal of the ovaries, referred to as ovariectomy) can be used to produce a coordinated onset of the equivalent of menopause. This kind of research also permits the effects of low estrogen per se to be studied disentangled from the effects of aging. Yet another advantage is that chemical and structural changes in the brain can be examined in detail when the brains are dissected at the end of the experiments. So what do we observe in these studies?

When challenged with specific problems to solve, such as remembering which arm of an eight-arm, sunburst-shaped maze has just yielded a food reward and will now be empty, rats with their ovaries removed and no estrogen treatment make significantly more wrong turns than comparable rats treated with estrogen (referred to as “estrogenreplaced”). In another task called a water maze, ovariectomized rats made to swim in a child’s wading pool take longer than estrogen-replaced rats to re-find a submerged platform now obscured by dye in the water. The untreated ovariectomized rats act as if they cannot remember the location of the platform they have just visited, despite the presence of abundant landmarks helpfully provided by the experimenter. Another task asks rats to associate a signal, such as a sound or a flash of light, with a subsequent mildly painful electrical shock to the feet. If, on detecting the signal, the rats move to a designated safe location in the cage, they avoid the shock entirely. In a 1994 study by Meharvan Singh, Ph.D., and colleagues, ovariectomy decreased the number of avoidance responses displayed (so the rats got more foot shocks); estrogen replacement restored avoidance behavior to normal levels (fewer foot shocks).

If I were a rat, would I want to be in the group that forgets the location of a Cheerio or cannot learn to move to the other side of the cage to avoid an electric shock? Give me the estrogen!

Postmenopausal women who are not taking estrogen also show mild deficits in some laboratory tests of cognitive function. Many of these tasks measure some aspect of verbal fluency or memory, such as the abilities to recognize words from a recently read paragraph, learn to associate pairs of words, and name colors. Although the problems human subjects are asked to solve are different from those used in studies of rodents, researchers believe they tap the same cognitive processes. For example, both human and animal memory can be categorized on the basis of duration. Working memory is a form of short-term memory used to store information that will be immediately used to solve a problem. In the eight-arm maze task, rats must use working memory to stop themselves from re-entering arms they have just visited. In humans, verbal recall of words recently read requires a similar, temporary form of information storage. Evidence from many studies suggests that both rats and women of any age perform better on tests that require working memory if estrogen is present.

Evidence for a beneficial effect of estrogen on mental functions other than working memory is harder to come by. A recent review of hormones and cognition by Gary Dohanich, Ph.D., of Tulane University summarized 20 different studies published between 1994 and 2001 that investigated the effects of estrogen or estrogen plus a progestin on what some call reference memory. Reference memory is the ability to learn a rule and use it later to solve a problem: It is the type of memory displayed when a rat in a radial maze learns that it is a waste of time to enter certain arms of the maze, because they never contain any food. Estrogen improved performance in 3 out of the 20 reference memory studies, had no effect in 6 studies, and actually impaired performance in 11 studies.

Another category of tasks that has been studied requires mental manipulation of objects, sometimes referred to as spatial learning. Estrogen often causes a slight impairment on the performance of these tasks, which in daily life might translate into whether you can read a map correctly or give someone accurate driving directions.

To Continue or Stop HRT?

Even if they were not worried about postmenopausal declines in cognitive function, menopausal women choosing hormone replacement before July 2002 probably felt good about their choice because the other benefits of HRT seemed so clear. This situation changed dramatically with the announcement that July that the Women’s Health Initiative (WHI) would terminate one of its studies of HRT in year five, rather than after a planned eight and a half years, because of risks that had been identified.9

The WHI, funded by the National Heart, Lung, and Blood Institute, is a comprehensive, multiple-medical-center study of women’s health that included a study of the most commonly prescribed HRT regimen for postmenopausal women: daily oral treatment with Prempro (Wyeth-Ayerst). Prempro (Premarin/Provera) combines 0.625 milligrams of horse (equine) estrogens and the synthetic progestin, MPA (2.5 milligrams). The study commands attention because of its scale and design: Between 1993 and 1998, more than 26,000 women between the ages of 50 and 79 years were randomly assigned to HRT or a placebo group. The study was designed so that neither the women nor their doctors knew which treatment they were receiving. Women who had not had a hysterectomy (16,608) received either Prempro or placebo; the remaining women were treated with equine estrogens (Premarin) but no progestin.

The scientists responsible for monitoring progress of the study chose to end the Prempro study three and a half years ahead of schedule because their analyses indicated that women taking Prempro had a higher risk of breast cancer, stroke, and cardiovascular disease than did the group of women who did not take the hormones. At the time the Prempro study was terminated, no such effect was seen in the estrogen-only study, and participants were instructed to continue taking their pills. But the estrogen-only study was also terminated earlier than originally planned (in February 2004, instead of 2005) because of a slightly increased risk of stroke in the treatment group. Overall, though, fewer negative effects were detected from estrogen alone, compared with estrogen plus progestin.

News coverage of the decision to terminate the Prempro study was intense. A treatment that had seemed as unremarkable as taking a daily multivitamin suddenly elicited fear. Stories in newspapers and on television said many women immediately stopped HRT, and this was borne out by analyses of prescriptions for HRT. Prescription records are easily obtainable in countries such as Canada and New Zealand where the government is the insurer, and these records revealed a drop in the number of new prescriptions for HRT after July 2002. At the annual editorial board meeting of the journal Hormones and Behavior in Orlando in November 2002, one of my colleagues asked, only half-joking, whether the journal should publish a special issue documenting the changes in behavior and cognitive ability of postmenopausal women in the United States after so many abruptly stopped HRT.

I read the original article on the WHI study and then the commentaries that followed. I soon felt confident that I did not immediately need to stop HRT. The increase in cancers associated with HRT was a slight effect, detectable by using a rather sophisticated procedure for monitoring data. In terms of absolute numbers, the overall risk for an individual woman is small. In making my own decision, I emphasized for myself the positive effects of HRT that WHI demonstrated. These positive effects were on prevention of osteoporosis and colon cancer. I also focused on the substantial differences between the women in the study group and myself: they were older (average age of 63 years on enrollment) and had been included in the study because they had not taken HRT continuously since the onset of menopause, whereas I had gotten an early start. They had been prescribed equine estrogens, whereas I had chosen the form of estrogen produced naturally by humans. And I eagerly awaited publication of the results of companion studies that used subgroups of the WHI participants to explore the effects of HRT on mental function and quality of life. Of particular interest was the Women’s Health Initiative Memory Study (WHIMS). Participants in WHIMS would take a standard screening test for dementia (the Modified Mini-Mental State Examination) before and after starting HRT.

But, when I was honest with myself, I could not deny the defensive and possibly irrational nature of my response. I was choosing a possibly risky treatment for a problem, a decline in cognitive abilities, that I had no evidence even existed. Taking on even a slight health risk with no benefit makes no sense. To make a decision based on all the available data, I needed to take a break from my research on the insect nervous system and re-immerse myself in mammalian neuroendocrinology. On the one hand, what I found in the animal literature made me less confident in my choice to pursue HRT. On the other hand, the literature also gave me reason to be encouraged, as I believe women should be generally, because my search revealed results from a new generation of animal studies suggesting that the “absence of a hormone leads to a deficit” model is overly pessimistic.

Animal vs. Human Studies

My comfort level with HRT was reduced because of mismatches between the human and animal studies. Rats do not take Prempro. Animal studies focus on the beneficial effects of estrogen, yet, like the women in the terminated arm of the WHI study, I am also taking a progestin to protect against uterine cancer. Also, animal studies focus on a surgically induced equivalent of menopause in young adult rats, not on changes that occur with age.

The progestin issue is important. Many researchers who study the complex effects of hormones on the brain have raised the possibility that treatment with progestin might neutralize the beneficial effects of estrogen. Receptors for progestins are found in many of the same parts of the brain as receptors for estrogens, but their role in cognition has been studied much less. This could be, in part, because my generation of researchers was taught in their basic classes that progestins decrease the electrical activity of the brain, sometimes acting just like anesthetics. Women who take progesterone for HRT or other medical reasons are told to take it at night because of its sleep-inducing effects. Differences in the actions of estrogens and progestins on the brain extend to brain chemistry, with estrogens increasing and progesterone decreasing the amount of certain neurotransmitters. Studies of dendritic spine densities in the hippocampus of rats sampled on different days of the estrous cycle show that progesterone reduces, rather than increases, spine number. Furthermore, a recent study at the University of Illinois by my colleague Janice Juraska, Ph.D., and her student Elissa Chesler showed that treating ovariectomized rats with a combination of estrogen and progesterone shortly before testing impaired their performance in a water maze task, but that estrogen by itself had no negative effect.

Even when a progestin is included, animal studies have primarily used progesterone, but relatively few women on HRT take this natural hormone. The synthetic, MPA, is cheaper than progesterone, which requires special processing to be absorbed if taken orally. In my case, my doctor said she feels more comfortable prescribing MPA than progesterone because the supporting literature on its use in HRT is much larger. Unfortunately, studies in progress suggest that, in cultured nerve cells, MPA in particular nullifies the actions of estrogens.10 Among the small number of studies of hormones and cognition that included progestins, I could find no studies that compared the effects of MPA and progesterone on learning and memory.

The question of age is my second concern. We do not know if it is valid to generalize results from studies of surgically induced menopause (in rats or humans) to the conditions of low estrogen associated with aging. Changes associated with aging, but independent of hormones, might interact with HRT to produce results different from those seen in younger ovariectomized rats and women. After all, both men and women show changes in cognition with aging, independent of hormones. Another important question is whether it matters when HRT begins. Estrogen might support the function only of healthy nerve cells and might not be able to reverse age-related damage that has already occurred, at least to an extent detectable in tests of cognition. There is evidence that women already diagnosed with dementia do not improve with estrogen treatment, but I could not find an animal study that compared aging rats and ovariectomized young rats receiving the same treatments and tackling the same tasks.

My concerns were magnified by the publication in 2003 of the WHIMS results that showed that the WHI Prempro regimen was associated, in a sample of more than 4,000 women age 65 and older, with a doubled risk of dementia and no improvement in test scores on the Modified Mini-Mental State Examination. In fact, the performance of the women in the hormone group was inferior to that of women receiving the placebo.11 A related study demonstrated that, for the women in the WHI study, Prempro did not improve overall quality of life.

New Ideas

So, obviously, animal studies have their limitations, their downside. Nonetheless, they do offer some intriguing suggestions for women considering the effects of hormones on the brain. One implication of recent animal studies is that, rather than assuming that the hormonal changes of menopause mean an overall decrease in brainpower, women might profitably think in terms of slight shifts in their cognitive strengths.

One such discovery that should provoke broader discussion is the observation that the effect of estrogen on some learning tasks is not an improvement in performance per se, but rather a shift in the strategy used to solve the problem. Donna Korol, Ph.D., and her colleagues at the University of Illinois studied rats in a simple maze shaped like the letter T. Hungry rats repeatedly placed in the maze learn that only one arm of the maze will ever contain food. To remember which way to turn to get this food, the rats can use one of two strategies: They can remember that they should always make the same response (for example, turn right) or they can remember that they should always walk in the same direction in relation to landmarks (for example, walk to the side of the room where the window is). These two methods are referred to as response and place strategies; they both get the rat to the food.

The point here is to find out which rats choose which strategy (which the researchers discover by manipulating the variables) and compare how well they learn. The study demonstrated a striking influence of hormones on choice of strategy. It was conducted in young adult rats experiencing the natural hormone surges and declines of the estrous cycle, and place strategists were strongly overrepresented in the high-estrogen group, wherease response strategists dominated the low-estrogen group. Indeed, the only difference between the two groups was in strategy, as both took the same amount of time to learn which arm was baited. Although these were young rats, I think the study offers a clue that the changes in cognitive performance associated with menopause should be characterized as differences, rather than decrements.

The message, I think, is that studies of the effects of hormones during aging, whether in rodents or humans, need to focus on specific tasks, rather than global measures, and to look for shifts in strategy, as well as gains and impairments. Thinking along these lines requires that we change our attitude, including the idea that working out strategies to maintain function in the face of age-related changes in memory (such as writing lists and buying a Palm Pilot) is somehow inferior to the strategies used earlier in life. I like to keep in mind that, in Korol’s study, both high-estrogen and low-estrogen rats were equally good at getting what they needed from their environment.

What Should I Do?

Animal studies to date do not tell us all we want to know about how HRT as it is currently practiced (or at least was before July 2002) affects cognition. But my own experience as a researcher makes me confident that neuroscientists working on animal models will be energized by the WHI findings and by the public’s interest in connections between estrogen and cognition. Discrepancies between expected and actual findings lead to further research that eventually will explain the discrepancies and thus advance our understanding.

There are many intriguing leads. One is a report from Elizabeth Gould, Ph.D., and colleagues at Princeton University that, in addition to making new dendritic spines, estrogen adds new nerve cells to the hippocampus—and new hippocampal neurons enhance the performance of rats on certain tasks. The importance of this brain region for cognition is such that, were this result to be confirmed in either humans or a nonhuman primate, I believe that many women would consider HRT for the cognitive benefits alone. But first they would need to see proof that new nerve cells are produced under conditions of HRT with a progestin. Documenting beneficial effects of estrogen by itself would not be persuasive.

Each woman must make the decision that—given what we know now—appears best for her particular situation. It seems unlikely that the questions I want to ask my doctor now can be answered in time to permit me to choose the optimal strategy of HRT for myself. Should I find an old-fashioned compounding pharmacy that can provide me with progesterone so I can switch from MPA? Should I just stop HRT now and wait to see what new treatments are developed in the future? Should I take estrogen and a progestin intermittently, rather than chronically, to mimic the hormone patterns of a menstrual cycle? Should I take testosterone and let my brain convert it to estradiol where it wants to? Should I have a medically unnecessary hysterectomy so I can take estrogen without a progestin?

If my doctor tells me that currently these questions have no answers, this neuroscientist will be forced to agree. I am still taking HRT for my bones, but I no longer believe that I am helping my brain. And I review this decision annually.


  1. Davison, S, and Davis, SR. “Hormone replacement therapy: current controversies.” Clinical Endocrinology 2003; 58: 249-261.
  2. Sherwin, BB. “Estrogen and cognitive functioning in women.” Endocrine Reviews 2003; 24: 133-151.
  3. Garcia-Segura, LM, Azcoitia, I, and DonCarlos, LL. “Neuroprotection by estradiol.” Progress in Neurobiology 2001; 63: 29-60.
  4. Pfaff, DW. Estrogens and Brain Function. New York. Springer-Verlag, 1980.
  5. Gould, E, Woolley, CS, Frankfurt, M, et al. “Gondal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood.” Journal of Neuroscience 1990; 10: 1286-1291.
  6. Carter, CS, and Keverne, EB. “The neurobiology of social affiliation and pair bonding.” In: Pfaff, DW, Arnold, AP, Etgen, AM, et al, eds. Hormones, Brain, and Behavior, Volume 1. New York. Academic Press, 2002: 299-337.
  7. Dohanich, G. “Gonadal steroids, learning, and memory.” In: Pfaff, DW, Arnold, AP, Etgen, AM, et al, eds. Hormones, Brain, and Behavior, Volume 2. New York. Academic Press, 2002: 265-327.
  8. Tanapat, P, Hastings, NB, Reeves, AJ, et al. “Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat.” Journal of Neuroscience 1999; 19: 5792-5801.
  9. Writing Group for the Women’s Health Initiative Investigators. “Risks and benefits of estrogen plus progesterone in healthy post-menopausal women.” Journal of the American Medical Association 2002; 288: 321-333.
  10. Wickelgren, I. “Brain researchers try to salvage estrogen treatments.” Science 2003; 202: 11381139.
  11. Rapp, SR, Epeland, MA, Shumaber, SA, et al. “Women’s Health Initiative memory study: A randomized controlled trial.” Journal of the American Medical Association 2003; 289: 2663-2672.