Abstract
A recent study by Rocca et al. examined the link between premenopausal oophorectomy and dementia [1]. The benefit of hormone replacement therapy to menopausal women is a hotly debated topic especially since the publication of the Women's Health Initiative (WHI) studies. Estrogen therapy or estrogen plus progestin (hormone) therapy was initially prescribed for the management of menopause-associated symptoms and, subsequently, other health benefits were also ascribed to this treatment, such as cardiovascular benefits, bone health and preserving cognitive function. Many of these conclusions were based on observational studies and, in the case of brain function, a large body of experimental evidence also indicated that estrogen was beneficial for this organ.
Our recent understanding of the role of estrogen in dementia, however, is heavily influenced by the WHI study and the associated Memory Study (WHIMS). This multiyear, multicenter, placebo-controlled, double-blind study indicated that hormone replacement (estrogens plus progestins) increased the risk of cardiovascular disease [2], while estrogen therapy increased the risk of stroke but did not increase cardiovascular disease [3]. The findings of the WHIMS indicated that hormone or estrogen therapy (considered together) increased (all-cause) dementia [4]. However, this landmark study was widely divergent from other epidemiological studies as well as the experimental literature, which showed that estrogen therapy was beneficial to various aspects of brain function.
In the experimental literature, the effects of estrogen therapy are typically tested on a bilaterally ovariectomized female animals, which is the equivalent of the surgical menopause. This model system has been instrumental to our understanding of the behavioral, cellular and molecular actions of estrogen and other ovarian steroids on the brain. A significant majority of such studies have shown that estrogen replacement to ovariectomized animals is neuroprotective [5] and improves performance on learning and memory tasks [6]. The few studies that have examined estrogen's effects in a mouse model of Alzheimer's disease show that estrogen delays the onset of plaques and improves the behavioral deficits that are associated with the transgenic mice [7,8], although not all studies agree [9]. In addition, studies have also shown that estrogen replacement to ovariectomized females modulates risk factors associated with dementia; such as maintaining the integrity of the blood–brain barrier [10], reducing infarct size following stroke [11] and reducing neural inflammation [12,13]. Recently, several animal studies have shown that the effect of estrogen is not always beneficial and may depend on the reproductive age of the animal [14]. Hence, research on essential questions regarding the role of estrogen and the duration and timing of estrogen treatment is very urgently needed.
Results
The study by Rocca et al. takes a novel approach to understanding the neuroprotective role of ovarian hormones, using a retrospective analysis to determine the association between an oophorectomy performed before menopause and the incidence of dementia. Women enrolled in this study were part of the Mayo Clinic Cohort, which includes patients who underwent a unilateral or bilateral oophorectomy as well as a reference group who did not undergo this surgery. The surgical group was restricted to those who were premenopausal at the time of surgery. In addition, women who underwent surgery for malignancies were not included. Controls were case-matched to each surgery case by year of birth. Follow up was performed by phone by trained assistants who were blind to the participant's status (oophorectomy or referent). All living women who could be interviewed were administered a modified Telephone Interview for Cognitive Status (TICS-m). A proxy was identified for women who could not be administered the TICS-m, owing to death or because they were incapacitated, and the proxy was queried on a brief dementia questionnaire. A low score on the TICS-m, a previous diagnosis of dementia or a proxy report of impaired daily living activites were used to classify women with dementia. Age at onset of dementia was therefore the day of the test administration or the date of report by a proxy, the actual date of diagnosis or the age at death if the proxy was unable to report the age of onset. Analysis was conducted for overall oophorectomy surgery and separately for the unilateral and bilateral sub-cohorts. All groups were stratified by age of surgery and in the case of the bilateral oophorectomy group, who received estrogen treatment after the oophorectomy, analysis was also stratified by the age at estrogen deficiency (at the end of estrogen treatment).
The results reveal that women who underwent oophorectomy had a higher incidence of cognitive impairment or dementia. Furthermore, the risk for dementia was greater if the surgery was performed at a younger age. In women with a unilateral oophorectomy, there was a significantly greater risk for dementia compared with women in the reference group, and a linear increase in this risk when the surgery was performed at a younger age. On the other hand, the risk of dementia was not increased in the bilateral oophorectomy group as compared with the reference group. When the age of oophorectomy was considered, there was a greater incidence of dementia in the younger two tertiles considered together compared with the oldest tertile. Furthermore, in this population there was a significant increase in the incidence of dementia among women who did not receive estrogen treatment up to age 50 years, while those who did receive estrogen treatment up to age 50 years were no different from the reference group.
Significance
This study supports a role for ovarian hormones in reducing the risk for dementia and also underscores the importance of timing of hormone loss as a critical variable in determining the risk for dementia. Women who experienced any type of ovarian loss (unilateral or bilateral) were more likely to develop dementia than the reference group. Interestingly, this association was much stronger in the unilateral oophorectomy group and less so in the bilateral oophorectomy group. The authors indicate that the women with a bilateral oophorectomy were typically prescribed estrogen therapy, while this was not typically the case where women had the possibility of hormone synthesis owing to the remaining ovary. The authors also speculate that since the majority of these unilateral oophorectomies were concurrent with a hysterectomy, the remaining ovary may have been less functional owing to a surgery-induced loss of blood supply. Women in this group also sought earlier intervention for menopause-related symptoms, supporting the idea that these women did experience a premature menopause. Hence, the unilateral oophorectomy group may actually be more hormone-deprived than the bilateral oophorectomy group; however, the present experimental design does not provide independent confirmation of this.
The Rocca et al. paper is subject to the shortcomings common to retrospective/observational studies, especially when the population pool is drawn over a 40-year period with changing medical practice, surgical techniques and lifestyle factors. Owing to cost issues, dementia was classified by a phone interview rather than a face-to-face neurologic interview. The specificity of the phone test was high, although the sensitivity was 73.3% for Alzheimer's disease and much lower (44.6%) for all types of dementia. Therefore, it is possible that the extent of dementia was underdetected in these groups, although it may be assumed that such underestimation would be random across groups. An issue that was not specified in the paper was the type of hormone therapy used, which would be a relevant variable in view of the WHIMS findings that the estrogen-only preparation was less deleterious than the estrogen–progestin preparation [4]. This may be especially pertinent since the unilateral oophorectomy group would predominantly receive estrogen therapy since most of them had a prior hysterectomy, while the bilateral group might receive the estrogen–progestin combination. Finally, an additional analysis that would be useful is a direct comparison between the unilateral and bilateral groups themselves, although such analysis may be underpowered with the present sample size.
The Rocca et al. study is consistent with both human and animal studies using a similar model. Specifically, this study provides a long-term perspective of oophorectomy and cognition that fits well with the short-term consequences of oophorectomy and cognition in a series of studies published by Barbara Sherwin and colleagues almost 20 years ago. In a prospective crossover design, estrogen alone, androgen alone or combined hormone treatment given postoophorectomy did not affect scores on a short- or long-term task or a task of logical reasoning, while oophorectomized females who received placebo had lower scores postsurgery on all measures of cognitive functioning [15]. This and other studies [16,17] were short-term studies but suggest that the cumulative effect of the oophorectomy over several years, especially without estrogen supplementation, is likely to cause serious cognitive deterioration, consistent with the findings of the Rocca study.
Although the ovariectomy model is widely used in the experimental literature, there are no exact parallels to the design used in the Rocca study. The few studies that have compared ovariectomized animals with intact females have shown inconsistent findings. Intact females outperformed ovariectomized animals on a radial arm maze; however, that advantage is lost when intact animals become acyclic as a result of food deprivation [18]. Ovariectomized females have poor performance on object recognition and object placement tasks as compared with intact controls [19], although ovariectomy improved performance on a maze-learning task [20]. Ovariectomy reduces the expression of key molecules such as the growth factor receptor trkA and the cholinergic enzyme choline acetyltransferase (ChAT) in the basal forebrain as compared with intact females [21]. Estrogen replacement to ovariectomized females rats improves performance on cognitive tasks and the timing of such replacement is critical, since hormone replacement soon after the surgery is more effective in improving performance than when it is given at a delayed period after ovariectomy [22]. This finding dovetails well with the observation that the unilateral oophorectomy patients were more susceptible to cognitive impairment than the bilateral oophorectomy patients who received estrogen therapy immediately.
In recent clinical and preclinical studies, the issue of early versus late hormone therapy has been suggested as an explanation to resolve the inconsistencies in the effects of estrogen (protective in some cases, deleterious in others) [23]. In animal studies, estrogen replacement to ovariectomized young adult females is generally neuroprotective and anti-inflammatory; however, delaying estrogen treatment to ovariectomized females or to older acyclic females is not neuroprotective [24] and paradoxically proinflammatory [25]. This may underlie the differences seen in the unilateral oophorectomy group (who likely had an extended period of hypoestrogencity) as compared with the bilateral oophorectomy group who received immediate hormone replacement. Some animal studies have demonstrated that estrogen treatment to females after a prolonged period of hypoestrogenicity is either ineffective as a neuroprotectant or actually detrimental to the brain. In the context of the Rocca study, it would therefore be useful to know if the delayed estrogen therapy to the unilateral oophorectomy group was a detrimental factor, perhaps by examining the incidence of dementia in this group stratified for estrogen therapy use at menopause versus no estrogen therapy at menopause.
An important issue that this study raises is whether oophorectomy is specifically deleterious to the brain or whether there is an overall decline in other ovarian hormone-sensitive organ systems. In a related study, the authors have shown that unilateral or bilateral oophorectomy increases the risk for Parkinson's disease [26], suggesting that there may be a general decline in neural or neuronal function as a result of ovarian-hormone loss. In a previous study of this population, the authors also found an increased mortality in the bilateral oophorectomy group, especially if the surgery occurred before 45 years of age [27]. Thus, one possibility is that the increased risk of dementia as a result of oophorectomy may be a consequence of overall health changes in these individuals.
Future perspective
The Rocca study is an important addition to the continuing debate on the role of ovarian hormones and the risk for dementia. Relevant to both patients and physicians is the observation that a gynecological surgery may have profound impact on overall health and may specifically impair cognitive function. The results of this study also emphasize that women with unilateral oophorectomies should be more closely monitored for premature menopausal symptoms postsurgery and postsurgical hormone therapy should be considered for this group as it is for bilateral oophorectomy patients.
Executive summary
Hormone therapy increases the risk for dementia in postmenopausal women, according to the Women's Health Institute studies, although observational studies suggest that estrogen therapy may be neuroprotective.
The present study found that unilateral and bilateral oophorectomy is associated with an increased risk for dementia.
The risk for dementia was worse in women with unilateral oophorectomy compared with those with bilateral oophorectomy, possibly owing to the fact that women with unilateral oophorectomy are less likely to receive hormone therapy immediately after surgery.
Short-term hormone therapy should be considered for both unilateral and bilateral oophorectomy patients.
Removal of the ovaries and the consequent loss of endogenous ovarian hormones may contribute to a general decline in health and a specific decline in neurologic function.
Footnotes
The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties. No writing assistance was utilized in the production of this manuscript.