Estrogen and Alzheimer Disease: Title and subTitle BreakPlausible Theory, Negative Clinical Trial

Whether to use estrogen replacement therapy in postmenopausal women continues to represent a major concern for women and their physicians. In the last few years, it has become clear that estrogen exerts a wide range of effects, including (in addition to its effects on reproduction) effects on the cardiovascular system,¹ skeletal system² and, of particular interest, neuronal systems serving cognitive function, especially memory.^{3 - 7}

There are good biological reasons to suspect that estrogen may influence cognition in postmenopausal women. Preclinical evidence suggests that estrogen exerts neuronal effects through mechanisms involving both genomic and cell surface receptors. Two estrogen receptors, α and β, are expressed in brain tissue and appear to be involved in the genomic effects.⁸ Nongenomic cell surface actions are known as well, with mechanisms that involve second messenger systems and effects on neuronal excitability and ion channels. Effects of estrogen may be observed both on brain morphology and cognitive performance. For example, in oophorectomized rats, estrogen can increase neurite outgrowth and dendritic spine formation in the hippocampus; in other studies, administration of estrogen to oophorectomized rats improved performance in cognitive tasks involving maze learning.⁸ Furthermore, in animal models, pretreatment with estrogen reduces the extent of injury due to stroke, suggesting a neuroprotective role for estrogen and providing some evidence that estrogen might be useful in ameliorating the decline in function in women with Alzheimer disease (AD).⁹ The way estrogen is administered also may have relevance; animal studies suggest that short-term (but not long-term) administration of estrogen in older female rats produces significant effects on hippocampal neuronal systems involved in memory.¹⁰

The influence of estrogen on cognitive function in women has been much more difficult to establish. For instance, it is unclear whether the effects of estrogen on cognitive function refer to effects on cognition in the immediate postmenopausal period, effects on cognition later in life, the ability of estrogen to prevent cognitive decline—or more specifically, to prevent the onset of the most common dementia, AD—or to the effects of estrogen on cognitive processes in women with established AD. The study by Mulnard and colleagues¹¹ in this issue of THE JOURNAL addresses this last question.

This well-designed multicenter clinical trial addressed the specific question of whether exogenous estrogen alters the inexorable downhill course of AD in women who had had hysterectomy. Subjects were women with mild to moderate AD and an average age of 75 years who were enrolled in a randomized, double-blind study, and received either estrogen (low or high dosage) or placebo. Adherence was carefully monitored by both caregiver reports and estrogen levels at each visit. The results are clear and unequivocal: estrogen does not halt the decline in cognitive function in AD. However, the conclusions are valid only for a very specific population, women of advanced age (approximately 75 years) with AD of mild to moderate severity.

Still unanswered are questions concerning, for example, whether estrogen given in the early postmenopausal period can prevent or delay the onset of AD or diminish its severity. Two general lines of investigation have been used to examine these questions: observational studies and clinical trials. In studies of aging women who did not have symptoms of AD, the data are far from consistent, with some studies suggesting benefit from estrogen and others showing no effect. Such inconsistency may reflect differences in the ages of the women studied; thus, studies involving younger women are more likely to show a positive effect of estrogen on cognitive function.^{12 - 13} Results of studies in older populations are far more varied, with some studies indicating a positive influence of estrogen on cognitive function^{14 - 16} and others failing to show an effect.^{17 - 18} Another possible factor in explaining the variability in results is whether estrogen was administered intermittently or continuously. The effects of estrogen on cognitive function may be observed most strongly when the agent is first used. In this regard, the study by Mulnard et al found that after 8 weeks of treatment, estrogen produced significantly better scores than placebo on the Mini-Mental State Examination. It is intriguing to speculate that, as suggested in animal studies, it is the effects of relatively short-term administration of estrogen that might lead to alterations in those neuronal systems critical for learning.

Similar research strategies have been used to address the question of whether estrogen administration might reduce the risk for the development of AD in women at risk for the disorder, but studies to date have produced inconsistent results.⁷ Such studies are quite difficult to perform and interpret and are subject to a number of confounding factors. For example, women who choose to use estrogen are often better educated and generally healthier and less depressed than nonusers. Furthermore, estrogen use might appear to be protective if estrogen is discontinued in women who develop cognitive symptoms or AD; in this case, if women with AD were surveyed, it would be found that fewer women with AD are taking estrogen. Moreover, women with AD may not remember whether they ever took estrogen in the past and surrogates may not know either; even verified pharmacy records indicating that estrogen was prescribed provide no assurance that the medication was actually consumed.

Recent advances in technology, particularly functional brain imaging, may provide new opportunities to better address questions about the effects of estrogen on cognitive function. For example, functional brain imaging may detect brain metabolic abnormalities in patients with AD before behavioral deficits are evident, thus presenting an opportunity to intervene very early in the course of the disease process.¹⁹ In addition, functional brain imaging may represent a more sensitive tool¹² to examine issues, such as the effects of different dosages of estrogen, routes of administration, length of treatment, and even the effects of newer, more specific selective estrogen receptor modulators.

Several important conclusions can be drawn from the investigation by Mulnard et al in the context of the several decades of studies that have examined the effects of estrogen. First, it is apparent that the questions that now need to be addressed are much more refined than simply asking whether estrogen improves cognition. Rather, future studies need to investigate, for example, the effects on particular cognitive functions (eg, memory, reading, attention) of specific types of estrogen, when prescribed under particular conditions (dosage, frequency) to well-defined groups of women (eg, middle aged, without any evidence of AD). Mulnard et al examined a very specific question and provided evidence that, under the conditions described, estrogen does not prevent decline in cognitive function in older women with mild to moderate AD.

Second, the study by Mulnard et al sends a cautionary message. The findings are a reminder that there is not a simple translation from laboratory experiments to clinical application, and illustrate that hypotheses—no matter how compelling the results on which they are based—remain provisional suppositions that must be proven before they can inform clinical practice. Given the consequences of declining cognitive function and the number of women who will spend an increased proportion of their lives with decreased levels of circulating estrogen, surely the questions are important enough to mobilize the kinds of multisite, randomized, double-blind trials reported by Mulnard et al. Clearly, such studies will not be easy or inexpensive, but what they offer is an opportunity to carefully examine, in a prospective fashion and with careful assessment of adherence, the effects of estrogen on cognitive functions and to minimize the confounding factors that have plagued previous investigations. Several large-scale studies (eg, Women's Health Initiative Memory Study,²⁰ Women's International Study of Long Duration Oestrogen After Menopause²¹ ) are in progress, and their results will be of considerable interest to women and their physicians. The study by Mulnard et al indicates that estrogen does not slow the progressive loss of cognitive function in elderly women who already have AD; it remains for future studies to determine the role of estrogen, if any, in preventing or delaying the onset of AD.