| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Departments of Clinical Neurosciences, Brown University, Providence, Rhode Island
b Departments of Community Health, Brown University, Providence, Rhode Island
c Brown University Center for Gerontology and Health Care Research, Providence, Rhode Island
d College of Pharmacy, University of Rhode Island, Kingston
Brian R. Ott, Neurology Department, Memorial Hospital of Rhode Island, 111 Brewster Street, Pawtucket, RI 02860 E-mail: Brian_Ott{at}mhri.org.
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. This retrospective cohort study compared 191 women estrogen users with 663 women nonestrogen users, matched according to age, nursing home facility, year of assessment, and baseline level of cognitive function. The outcome was decline in cognition, measured by the Cognitive Performance Scale, over a minimum follow-up period of 6 months.
Results. No significant difference was found in the rate of cognitive decline among estrogen users and nonusers.
Conclusions. Estrogen therapy administered to nursing home residents is not associated with a reduction in cognitive decline. This study lends further support to recent controlled clinical trials that found no benefit for estrogen treatment on cognition in outpatients with dementia.
DEMENTIA is a major precipitating factor leading to placement in long-term care institutions (1)(2). Nearly 60% of nursing home residents are diagnosed with some form of dementia, with Alzheimer's disease (AD) being the most common etiology (2)(3). In 1991, the estimated annual cost of AD was $80 billion in the United States, making it the third most expensive illness after heart disease and cancer (4). Medicaid costs alone for long-term care of people with AD were $5.7 billion (4). Consequently, there is a socioeconomic as well as medical imperative to find ways to delay the disease as well as reduce disease severity through pharmacologic intervention.
While the etiology of AD is unknown, there have been many theories regarding its pathogenesis. The incidence and prevalence of AD are reported to be greater in women than in men and rise with increasing age (4)(5)(6)(7). Consequently, one proposed mechanism that may partially account for deterioration of cognitive function in women is declining estrogen levels associated with menopause.
Estrogen exerts numerous positive physiologic actions relevant to cognitive function, which could counteract the effects of neurodegeneration. Estrogen reduces monoamine oxidase activity, allowing for an increased level of neurotransmitters at the synapse, potentially resulting in improved cognition and behavior (8). Estrogen enhances neurite growth and promotes the expression of mRNA for a brain-derived neurotrophic factor (9)(10). Estrogen possesses antioxidant activity that hinders the development of neuronal plaques and may also reduce deposition of beta-amyloid by accelerating its metabolism (11). Another secondary effect of estrogen is to increase cerebral blood flow (9)(12).
Despite these numerous plausible biologic mechanisms, some cohort and case-control studies have reported that estrogen has no effect on cognition or the incidence of dementia in women (13)(14). Brenner and coworkers (15) found that oral estrogen use was associated with a 50% reduction in AD risk; however, their sample size was too small to achieve statistical significance. Other epidemiological studies indicate a statistically significant decreased risk of dementia and cognitive decline among estrogen users (16)(17)(18)(19).
The potential beneficial effects of estrogen therapy on cognition reported in these and other studies have been extensively reviewed (20)(21)(22). To date, no studies have examined the effect of estrogen therapy on cognitive decline among residents in nursing homes. This population is important to study, because nursing home residents have such a high prevalence of established and incipient dementia. Therefore, the objective of this study was to quantify the beneficial effect of estrogen use on the rate of cognitive decline among women residing in U.S. nursing homes.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Embedded within the MDS are several scales to define cognition, physical function, and behaviors. Cognitive performance is measured using a six-item scale, the Cognitive Performance Scale (CPS). The CPS is scored on a seven-point ordinal scale (ranging from a score of 0 to 6) in which higher scores characterize worse cognitive performance (25). Seven separate direct measures of cognitive performance were included in the MDS: short- and long-term memory, orientation items (recall of season, location of room, identity of staff, and that the resident is in a nursing home), and a single item assessing the ability of the resident to make decisions about activities that are a part of daily life. The reliability and validity of the CPS scale have been shown and also have been compared to other research instruments, including the Mini-Mental State Examination (MMSE), a widely used, reliable, and valid test of cognitive status (26). A CPS score of 0 to 1, which is minimal impairment, is equivalent to an MMSE of 23 to 24. A CPS score of 2 to 3, which is of moderate cognitive impairment, constitutes an MMSE score of about 6 to 17. Lastly, a CPS score of 4 to 6 indicates severe impairment and is equivalent to an MMSE of 2 to 4 (27)(28)(29).
Nursing home staff recorded as many as 18 different drugs received by each resident in the 7 days preceding the assessment. Information on the resident's drug therapy included brand/generic name, dosage, route and frequency of administration, and whether the order was standing or as needed (30). We translated the drugs initially coded using National Drug Codes into therapeutic classes and subclasses using the Master Drug Data Base (Medispan Inc, Indianapolis, IN) (30)(31).
Previous studies have supported the association between the use of unopposed estrogen (rather than in combination with progestins) and higher cognitive performance (32). Brenner and colleagues (15) reported on different formulations and different routes of estrogen, but the oral conjugated form had the lowest relative risk for cognitive decline. Actual medical history regarding hysterectomy is not available in the SAGE database, yet, about one third of this population is likely to have had hysterectomies and qualify for treatment with unopposed estrogen (33).
In the present study, we defined estrogen use as any oral estrogen product (whether it is opposed or unopposed) being used more than three times weekly. The Medispan code range for estrogens that were used in this study were 24-00-00 to 24-99-40, omitting estrogens in combinations with antianxiety agents (Medispan codes 24-99-30, 24-99-40, and 24-99-20, respectively). Because there were only two women using estrogens in combination with progestins, we were unable to evaluate this group separately.
Subjects
The study design was a retrospective cohort study. We included 255,819 women in nursing homes, of whom 184,557 were matched to the Medicare denominator file, and of whom 5742 were younger than 65 years of age. From the 178,815 women who were 65 years and older, 135,506 were without contraindications to estrogen use (such as vegetative state, history of pulmonary embolism, deep vein thrombosis, stroke, Parkinson's disease, hydrocephalus, brain tumor, and Huntington's disease). Of 135,506 women, 63,726 had at least two assessments recorded (roughly equal to 6 months of follow-up assessments). We required persons in this study to have at least two assessments so that the study sample would not include those who where critically ill or nonpermanent nursing home residents.
From this group of 63,726 participants, 31,726 had a CPS score of 3 or less. From this group of women with mild to moderate cognitive impairment, we identified 226 women who had used unopposed estrogen. For each estrogen user, we identified women not using estrogen and matched according to age (within 5 years), nursing home facility, year of assessment, and baseline CPS score (0–1 for mild and 2–3 for moderate). We were unable to identify matched nonusers for 35 estrogen users. For the 191 users for whom we could identify at least one matching nonuser, we found 663 nonestrogen users (median = 4, range 1–10, nonusers to user).
Analysis
The measure of effect estimated for this study was the incidence rate ratio. We determined the number of women experiencing a decline in cognitive function as indicated by at least a one unit increase in the CPS score from baseline. We estimated the denominators of the incidence rate of the users and nonusers using the same methodology. Person-time was estimated by calculating the number of days between the baseline and the date of a change of at least 1 point on the CPS. We censored outcomes on date of death or hospitalization when appropriate.
An increase of at least 1 point on the CPS is equivalent to more than a 3-point change on the MMSE and, therefore, is clinically relevant (34). We also considered as a separate endpoint the rate of at least a 2-unit increase in the CPS score.
We considered a number of demographic variables, ADL scales, and concomitant diagnosed diseases as indicated by the MDS or the Medicare claims via ICD-9 codes as potential confounders. Demographic variables included age, race/ethnicity, and marital status. Measure of physical function was captured through the ADLs. The ADL score, a five-item, six-level score, measures performance in physical function. The scale is based on the resident's dependency in the areas of dressing, eating, toileting, bathing, locomotion, transferring, and incontinence (35). We classified residents as mildly impaired (ADL score of 0 to 1), moderately impaired (ADL score 2 to 3), or severely impaired (ADL score 4 to 5).
People with concomitant diagnosed diseases including cerebral vascular accidents, Parkinson's disease, depression, schizophrenia, hypothyroidism, anemia, Alzheimer's disease, and dementia (non-Alzheimer type) can exhibit cognitive impairment. We also evaluated the prevalence of cancer, osteoporosis, and cardiovascular or peripheral vascular diseases because of their indication of the use of estrogen.
Initially, we compared estrogen users to nonusers on sociodemographic characteristics, comorbid conditions, behavioral problems, and concomitant medication use. The effect of estrogen use was estimated using Cox proportional hazards models, considering nonusers as the reference group and simultaneously controlling for all the potential confounding variables. The hazard rate ratio derived from these analyses was interpreted as an estimate of the incidence rate ratio. As the goal of our analysis was to provide an estimate of the effect of estrogen on the outcomes of interest adjusted for potential confounders, we did not use statistical significance to guide our decisions about which variables would remain in the models. To evaluate whether the effect of estrogen use on rate of cognitive decline differed by baseline level of cognitive function, we performed a stratified analysis. We censored residents at their time of death or time of first hospitalization. We examined the log-log survival function to evaluate (and rule out) departures from the proportionality assumption for each model. From these models, we derived estimates of effect and corresponding 95% confidence intervals (CIs). We performed all analyses using SAS (v.8.1, Cary, NC).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
12%), users (23%) were less likely to have dementia of the non-Alzheimer's type than nonusers (30%). Yet, distributions of behavioral problems including wandering, verbal abuse, and inappropriate behavior were similar in users and nonusers.
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Although there is now good evidence that estrogen therapy should not be prescribed specifically for treating AD, clinical trials examining the potential of estrogen for dementia prevention or delay in asymptomatic women are still in progress. The largest of these clinical trials is the Women's Health Initiative, which may not report data for at least 3 to 4 years (39).
Of interest, there was a trend toward less cognitive decline in estrogen users in our study, particularly among those with minimal cognitive impairment at study entry, in whom there was a 30% reduction in cognitive decline. Our sample size may have been too small to detect a significant difference in this subgroup. Furthermore, we cannot rule out the possibility that our outcome measure, the CPS, was too insensitive to detect a clinically significant effect of estrogen on cognition. Although the CPS has good correlation with other measures of cognitive function, its use as an outcome variable for research studies has not been firmly established (40). Reliability of the CPS may be limited by the fact that nonresearch personnel completed the assessment. Also, the range of scores may be too restricted to detect relatively small or subtle changes in cognitive performance produced by drugs such as estrogen.
Confounding factors further limit our ability to draw firm conclusions from these data. Nonestrogen users were somewhat older and more functionally and cognitively impaired than estrogen users. One would expect that the nonusers would have declined at a faster rate based upon these factors. Although we controlled for these factors in the analysis and through the matched study design, the potential for residual confounding cannot be ruled out. Estrogen users received more concomitant medications than estrogen nonusers, including tricyclic antidepressants, antihypertensives, and narcotics. One would expect that the estrogen users would have declined at a faster rate based on these factors. Yet, adjustment for these agents in our study did not significantly alter the estimates of effect derived from our study.
After adjusting for multiple confounding variables, the hazard ratio of our estrogen model approached one. Therefore, we are unable to conclude that there was a significant relationship between estrogen use and cognitive decline in this sample of nursing home residents. The impact of estrogen therapy on dementia in the nursing home is likely to be very low in light of these negative findings; however, we cannot exclude a clinically relevant effect of estrogen on cognitive decline in nondemented women or in those with only mild cognitive impairment.
|
Acknowledgments |
|---|
Received January 2, 2002
Accepted April 2, 2002
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. E. Morley The Need for a Men's Health Initiative J. Gerontol. A Biol. Sci. Med. Sci., July 1, 2003; 58(7): M614 - 617. [Full Text] [PDF]
|
|
|
|
|
|
J. E. Morley and H. M. Perry III Androgens and Women at the Menopause and Beyond J. Gerontol. A Biol. Sci. Med. Sci., May 1, 2003; 58(5): M409 - 416. [Full Text] [PDF]
|
|
|
|
|
|
W. A. Banks and J. E. Morley Memories Are Made of This: Recent Advances in Understanding Cognitive Impairments and Dementia J. Gerontol. A Biol. Sci. Med. Sci., April 1, 2003; 58(4): M314 - 321. [Full Text] [PDF]
|
|
|
|
|
|
S. Asthana Estrogen and Cognition: The Story So Far J. Gerontol. A Biol. Sci. Med. Sci., April 1, 2003; 58(4): M322 - 323. [Full Text] [PDF]
|
|
|
|
|
|
G. T. Grossberg and A. K. Desai Management of Alzheimer's Disease J. Gerontol. A Biol. Sci. Med. Sci., April 1, 2003; 58(4): M331 - 353. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
