Skip Navigation


American Journal of Epidemiology Advance Access originally published online on January 23, 2008
American Journal of Epidemiology 2008 167(6):692-700; doi:10.1093/aje/kwm362
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
167/6/692    most recent
kwm362v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Petitti, D. B.
Right arrow Articles by Chui, H. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Petitti, D. B.
Right arrow Articles by Chui, H. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology © The Author 2008. Published by the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.

ORIGINAL CONTRIBUTIONS

Incidence of Dementia in Long-term Hormone Users

Diana B. Petitti1,2, Valerie C. Crooks3, Vicki Chiu3, J. Galen Buckwalter3,4 and Helena C. Chui5

1 Department of Clinical Analysis, Kaiser Permanente Southern California, Pasadena, CA
2 Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA
3 Department of Research and Evaluation, Kaiser Permanente Southern California, Pasadena, CA
4 Department of Research and Development, eHarmony, Pasadena, CA
5 Department of Neurology, Keck School of Medicine, University of Southern California, Los Angeles, CA

Correspondence to Dr. Valerie Crooks, Department of Research and Evaluation, Kaiser Permanente Southern California, 100 S. Los Robles, 2nd Floor, Pasadena, CA 91101 (e-mail: Valerie.c.crooks{at}kp.org).

Received for publication June 20, 2007. Accepted for publication November 9, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Results from epidemiologic studies of postmenopausal hormone use and dementia have been conflicting. Investigators from the Women's Health Initiative Memory Study reported that the incidence of dementia in women aged ≥65 years assigned to hormone use was increased. Here the authors report results from a prospective cohort study of 2,906 dementia-free women (1,519 hormone users and 1,387 hormone nonusers) aged ≥75 years who were recruited from a Southern California health plan in 1999 and followed through 2003. Cognitive status was assessed annually using the Telephone Interview of Cognitive Status–modified, supplemented by the Telephone Dementia Questionnaire and medical record review. The mean self-reported age at initiation of hormone use was 48.3 years for users of estrogen alone (n = 1,072) and 54.9 years for users of estrogen plus progestin (n = 447); self-reported mean durations of hormone use were 30.5 years and 23.2 years, respectively. There were 283 incident dementia cases identified during follow-up. After adjustment for age, education, and medical history, hazard ratios for incident dementia were 1.34 (95% confidence interval: 0.95, 1.89) in estrogen/progestin users and 1.23 (95% confidence interval: 0.94, 1.59) in estrogen users. These findings do not provide support for an effect of estrogen or estrogen/progestin use in preventing dementia.

Alzheimer disease; dementia; hormone replacement therapy

Abbreviations: CI, confidence interval; SD, standard deviation; TDQ, Telephone Dementia Questionnaire; TICSm, Telephone Interview of Cognitive Status–modified


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Several lines of biologic data suggest that estrogens protect neural function (1, 2). Results from epidemiologic studies of the relation between hormone therapy and Alzheimer's disease conflict (3–20). Alzheimer's disease is the end result of pathologic processes that begin in middle age and progress. In the Cache County Study, a prospective epidemiologic study of hormone therapy and dementia incidence, Zandi et al. (20) reported a protective association of discontinued and long-term hormone therapy with dementia. Although they were based on a small number of Alzheimer's disease cases (n = 26) among hormone therapy users, these findings were interpreted as evidence for a limited time window around menopause during which hormone therapy reduces dementia risk (21).

In clinical trials, administration of estrogen alone to women with clinically diagnosed Alzheimer's disease has not prevented cognitive decline (22–24). Investigators from the Women's Health Initiative Memory Study, an ancillary study of the Women's Health Initiative (a randomized trial of hormone use), reported higher incidences of dementia and Alzheimer's disease in women aged ≥65 years who were randomized to receipt of combined estrogen/progestin hormone therapy (25) or estrogen alone (26). In the Women's Health Initiative Memory Study, use of hormone therapy was short-term.

Here we provide 5-year follow-up from a prospective cohort study that addressed the limitations of prior epidemiologic studies—small sample size and short follow-up—and report on the incidence of dementia in a large number of long-term users of hormone therapy defined by prescription and self-reported information.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
As previously reported (19), subjects were recruited from a defined population of female members of a Southern California health plan aged ≥75 years on July 1, 1998, who had been continuously enrolled in the health plan from 1992 to 1998. Computer-stored prescription data (available since 1992) were used to identify current hormone therapy users and nonusers in 1998. Hormone therapy users (n = 3,058) were defined as women who had filled at least one prescription for oral estrogen at a health plan pharmacy in every calendar year from 1992 to 1998. Nonusers (n = 24,276) were defined as women without any estrogen prescriptions from 1992 to 1998. Women who had intermittent prescriptions from 1992 to 1998 were not eligible. To assure an equal number of women in the hormone therapy comparison group, 3,394 women from the non-hormone-therapy group were chosen, frequency-matched by age and zip code with hormone therapy users.

Of the 6,542 women eligible for the study, 3,681 were interviewed at baseline in 1999. Data were gathered on personal habits, medical history, and self-reported use of hormone therapy. The Telephone Interview of Cognitive Status–modified (TICSm) (27) was administered. Women who were eligible but not interviewed at baseline were significantly older and more likely to be nonusers of hormone therapy (both p's < 0.05). Medical record review for a random sample of baseline nonrespondents (n = 565) showed a prevalence of physician-diagnosed dementia of 11 percent in baseline nonrespondent hormone therapy users (n = 165) and 10 percent in nonrespondent nonusers (n = 400) (19).

Despite up to 10 contact attempts, 636 of the 3,681 women interviewed at baseline could not be contacted subsequently and were excluded from this analysis. The 139 women classified with dementia at baseline (68 hormone therapy users and 71 nonusers) were also excluded.

The 2,906 dementia-free women with at least one contact after baseline comprised the prime cohort for this analysis (figure 1). Compared with women in the prime cohort, the 636 women who dropped out after baseline were significantly older, less educated, and more likely to have a history of diabetes and Parkinson's disease (all p's < 0.05). Dropouts were more likely to be hormone therapy nonusers than to be hormone therapy users (20 percent vs. 15 percent; p < 0.001); nonuser dropouts died at higher rates than user dropouts (7 percent vs. 5 percent). After excluding the 217 dropouts who died, the percentages of living women not contacted at all after baseline were 10 in hormone therapy users and 15 in nonusers (p < 0.001).


Figure 1
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 1. Flow of participants from recruitment to analysis in a study of dementia incidence and long-term hormone use, Southern California, 1999–2003. TICSm, Telephone Interview of Cognitive Status–modified.

 
Annual telephone interviews were attempted for all women in the prime cohort until they died or were classified as having dementia, or until follow-up ended. Proxy interviews were attempted when women could not be interviewed by telephone. Proxies were also interviewed as a component of our multistage dementia classification procedure.

For proxy interviews, women were asked to identify people they saw at least once a month who knew them well. With more than one proxy option, the person living in closest physical proximity was the proxy interview priority. The proxy interview criteria resulted in usual priority for the spouse, followed by a child, friend, or sibling. Contact frequency and study participant knowledge were confirmed during proxy interviews.

Death ascertainment used linkage with the California death index and hospital discharge data (28), plus proxy reports. Copies of death certificates were obtained; mentions of dementia were identified as dementia cases.

Woman-years of follow-up were calculated from the date of the baseline interview to the date of the interview that resulted in dementia classification, the date of loss to follow-up, or December 31, 2003, if the woman was alive, nondemented, and still participating.

Classification of cognitive status
At each annual follow-up, dementia outcome was classified as 1) no cognitive impairment or minimal cognitive impairment, 2) cognitive impairment without definite evidence of dementia (uncertain), or 3) dementia. Women classified with dementia were not followed actively and were censored in the analysis.

The methods used for multistage classification of cognitive status and the results of comparisons with "gold standard" dementia assessment (which included in-person examination by a neurologist and neuropsychological testing), conducted at the University of Southern California Alzheimer's Disease Research Center, have been published elsewhere (29). Sensitivity in comparing dementia with no dementia using the gold standard was 0.83, and specificity was 1.0.

In the first stage, the TICSm (27) was administered. Women with a TICSm score greater than 27 were classified as having "no or minimal cognitive impairment" for that year. Women with a TICSm score of 27 or less were selected for the second stage of assessment, wherein the Telephone Dementia Questionnaire (TDQ) (30) was administered to a proxy.

The TDQ collects information about the subject's cognitive function in four domains: memory, fluency, comprehension, and orientation. The TDQ results were reviewed independently by three investigators, who were blinded to hormone use. A predefined protocol was used to classify the second-stage TDQ result: no or minimal impairment, dementia, or uncertain. Classification of a woman as having dementia required memory deficits, impairment in at least two other cognitive domains, and functional impairment. The criteria for dementia in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (31), require only one other cognitive deficit in addition to memory. More stringent classification criteria (requiring two other cognitive domains) were applied in this study because some of the cognitive domains assessed in person cannot be assessed telephonically. The independent classifications were reviewed in a consensus conference. Women with consensus classification of "dementia" or "no or minimal impairment" were given that classification for that interview year. Women with consensus classification of "uncertain" at the second stage, based on the TDQ, and women with a low TICSm score for whom a proxy TDQ could not be completed proceeded to the third stage of assessment.

In the third stage, medical records were reviewed by reviewers who were blinded to hormone therapy use. Women were given a final classification of "dementia" at this stage if there was a diagnosis of dementia recorded in the medical record and a classification of "cognitive impairment but dementia uncertain" if the medical record mentioned cognitive impairment but not dementia. If there was no mention of these conditions in the medical record, the women with TDQ data were given a final classification of "cognitive impairment but dementia uncertain" for that year. Women without a completed TDQ were given a classification for that year of "no or minimal impairment."

Medical records were used to subclassify dementia cases according to whether the dementia had a known cause (vascular, Lewy body, Parkinson's disease, etc.). If none of these causes were mentioned in medical records, the woman was classified as having dementia without a known cause, an approximation of Alzheimer's disease.

Analyses
Descriptive statistics were generated for demographic and self-reported medical condition variables (stroke, myocardial infarction, diabetes mellitus, hypertension, and Parkinson's disease). Chi-squared tests were done to test for statistical significance in the analysis of nonresponse. The Kaplan-Meier procedure was used to estimate the probability of dementia-free survival by hormone therapy use. The log-rank test was used to assess the statistical significance of differences in dementia-free survival. The Cox proportional hazards model was used to estimate crude and age-adjusted hazard ratios, and hazard ratios were adjusted for other potential confounders. The regression models included self-reported variables found to be strongly related to dementia in the literature (age and education) and other available variables that were associated in our data set. The variables in the final, fully adjusted model were forced. Exact 95 percent confidence intervals were calculated for all hazard ratio estimates. A p value less than 0.05 was considered statistically significant.

Outcomes were censored at December 31, 2003, the date of the last contact if lost to follow-up, or the date of classification with dementia. The outcome, nondementia, included both women with a final classification of "no or minimal cognitive impairment" and women with a final classification of "cognitive impairment but dementia uncertain." That is, the contrast in the main analysis was between dementia and the combined group of women with no or minimal cognitive impairment and women with cognitive impairment.

The main analyses employed information on hormone therapy use as determined by prescription, with nonusers as the referent group. Analyses were also carried out taking into account both prescription information and self-reported information on hormone therapy use at baseline. Information about when hormone therapy started and age at menopause was available only by self-report. Age at menopause was defined as the self-reported age at which menstrual periods stopped. This information was used to assess the specific hypothesis that women who initiated hormone therapy near the menopause were uniquely protected against the development of dementia.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
The mean age at baseline of the 2,906 women included in this analysis was 78.7 years (standard deviation (SD), 3.2). A total of 1,519 women were hormone therapy users by prescription at baseline; 1,387 were nonusers by prescription. The mean self-reported duration of hormone therapy at baseline was 30.5 years for users of estrogen (SD, 10.8) and 23.2 years for users of estrogen/progestin (SD, 10.9). The mean self-reported age at initiation of hormone therapy was 48.3 years for estrogen users (SD, 10.7) and 54.9 years for estrogen/progestin users (SD, 10.7). More than 90 percent of the estrogen used was conjugated equine estrogen; the preponderance of progestin was medroxyprogesterone acetate.

Incident dementia was identified in 283 women. Of these dementia cases, 185 were based on the TDQ, 88 on medical records, and 10 on death certificates.

Both crude and adjusted hazard ratios for dementia were elevated in relation to higher age, lower educational attainment, and history of stroke, diabetes, or Parkinson's disease (table 1). The crude and age-adjusted hazard ratios for dementia in prescription estrogen users were 1.15 (95 percent confidence interval (CI): 0.89, 1.49) and 1.13 (95 percent CI: 0.87, 1.46), respectively. The crude and age-adjusted hazard ratios for dementia in prescription estrogen/progestin users were 1.25 (95 percent CI: 0.90, 1.73) and 1.31 (95 percent CI: 0.94, 1.82), respectively.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Numbers of dementia cases, dementia incidence, and hazard ratios for dementia among female members of a Southern California health plan, according to sociodemographic characteristics and prescription-determined hormone use at baseline, 1999–2003

 
Dementia-free survival showed no significant association with hormone therapy use (figure 2; log-rank test: p = 0.34). The curves for dementia-free survival did not converge or diverge over time.


Figure 2
View larger version (10K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 2. Kaplan-Meier plot of dementia-free survival according to prescription-determined hormone use at baseline, Southern California, 1999–2003.

 
After adjustment for age, education, and self-reported medical history variables, the hazard ratios for dementia were 1.23 (95 percent CI: 0.94, 1.59) for prescription estrogen users and 1.34 (95 percent CI: 0.95, 1.89) for prescription estrogen/progestin users (table 2).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Incidence of dementia and adjusted hazard ratios for dementia among female members of a Southern California health plan, by prescription-determined hormone use, 1999–2003

 
Of the 283 dementia cases, 191 were classified as dementia without known cause, corresponding to probable Alzheimer's disease based on standard criteria (30). The adjusted hazard ratios for dementia without known cause were 1.22 (95 percent CI: 0.89, 1.68) for prescription estrogen users and 1.41 (95 percent CI: 0.94, 2.12) for prescription estrogen/progestin users. The adjusted hazard ratios for dementia with known cause (mostly vascular (n = 92)) were 1.18 (95 percent CI: 0.75, 1.86) for prescription estrogen users and 1.31 (95 percent CI: 0.72, 2.39) for prescription estrogen/progestin users.

Compared with women classified as nonusers by both prescription and self-report, the adjusted hazard ratios in women classified as estrogen or estrogen/progestin users by both methods were 1.07 (95 percent CI: 0.79, 1.44) for estrogen and 1.32 (95 percent CI: 0.91, 1.91) for estrogen/progestin (table 3). The adjusted hazard ratio for dementia in women classified as users of either prescription estrogen or prescription estrogen/progestin who stated that they had never used hormones when interviewed at baseline was 1.64 (95 percent CI: 0.94, 2.88). These women had filled at least one prescription for hormone therapy in each of the 7 years prior to interview.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Incidence of dementia and adjusted hazard ratios for dementia in users of estrogen or estrogen plus progestin, according to source of information on hormone use, Southern California, 1999–2003*

 
Table 4 shows hazard ratios for dementia according to time of hormone therapy initiation in relation to menopause based on self-reported information on both hormone therapy and menopause. Women whose age of initiation of hormone use could not be determined (n = 265) were analyzed as a separate group. Compared with women who reported never using hormone therapy at baseline, adjusted hazard ratios were less than 1 (a protective association) for the hormone therapy groups with known ages of starting hormone therapy and menopause; the 95 percent confidence intervals were wide and included 1. In women whose age at initiation of hormone therapy could not be determined, the adjusted hazard ratio for dementia was 1.54 (95 percent CI: 1.05, 2.27).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Incidence of dementia and adjusted hazard ratios for dementia according to self-reported hormone use, by timing of the start of hormone use in relation to menopause, Southern California, 1999–2003

 
Medical records for women with dementia who self-reported never using hormone therapy but were classified as hormone therapy users by prescription (n = 17) revealed evidence of "dementia bias" (selective underreporting of hormone therapy use in women with dementia) (19). Fifteen of the 17 women were verified as hormone therapy users at baseline interview.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Our findings do not provide support for the contention that there is an overall protective effect of estrogen or estrogen/progestin use against the development of dementia recognized after age 75 years. The number of dementia cases was large (n = 283), and the lower bounds of the 95 percent confidence intervals were consistently near 1 for both estrogen and estrogen/progestin. The study was designed on the basis of a formal sample size calculation to have 80 percent power to detect a reduction in dementia risk of 0.8 or more, based on a two-sided test with alpha equal to 0.05. Recruitment goals were met, and dementia rates were higher than expected. However, the width of the 95 percent confidence intervals for the hazard ratio estimates for estrogen and estrogen/progestin encompassed the possibility of an overall protective effect of long-term estrogen and estrogen/progestin use of 10 percent, in spite of a large number of dementia cases. An adverse effect of combined estrogen/progestin cannot be ruled out.

Prior epidemiologic studies of Alzheimer's disease and hormone therapy yielded relative risk estimates ranging from 0.18 (14) to 2.38 (1). Many studies lacked rigorous methods for ascertaining hormone therapy use and/or incident dementia. One study used proxy respondents to obtain information on hormone therapy use for cases and controls (16). In another study, classification of dementia outcome was based only on death certificates (10).

The Upper Manhattan (11) and Cache County (20) studies were carefully conducted prospective cohort studies that examined dementia incidence in relation to hormone therapy use. In the Upper Manhattan Study, the relative risk of incident dementia was 0.40 in hormone users. Only 156 women in the study were exposed to estrogen, and the finding was based on only nine cases of dementia in estrogen users (11). In the Cache County Study, Zandi et al. (20) reported an adjusted hazard ratio of 0.59 (95 percent CI: 0.36, 0.96) for incident Alzheimer's disease in women who had ever used hormone therapy (not separated by use of estrogen and estrogen/progestin). This estimate was also based on a very small number of Alzheimer's disease cases (n = 93). The lower risk of incident Alzheimer's disease among hormone therapy users was confined to former users, where the overall hazard ratio for incident dementia (n = 9) was 0.33 (95 percent CI: 0.15, 0.65); hazard ratios were 0.58, 0.32, and 0.17 for durations of use of <3 years, 3–10 years, and >10 years, respectively. For current hormone therapy users, the hazard ratios for incident Alzheimer's disease (n = 26) were 1.08 overall (95 percent CI: 0.59, 1.91) and 2.41, 2.12, and 0.55 for durations of use of <3 years, 3–10 years, and >10 years, respectively. Zandi et al. suggested that their findings might indicate a limited window of time following menopause in which hormone exposure reduces Alzheimer's disease risk, as Marder and Sano (21) have also hypothesized.

The mean age at initiation of hormone therapy for women in our study (48.3 years for estrogen and 54.9 years for estrogen/progestin) was somewhat higher than would be expected if women initiated use right at the time of menopause. The self-reported data on age at hormone therapy initiation may have been approximate, given the long time between the event and the interview. Because the precise date of initiation of hormone therapy relative to age at menopause cannot be known, an effect of exposure beginning at menopause on dementia prevention cannot be precluded. However, 70 percent of women in the current study who reported both age at menopause and age at initiation of hormone use started hormone therapy within a 10-year protective "window of opportunity" near the menopause. Thus, a substantial majority of women in our study who took hormones were first exposed to them at or near the time of menopause. The overall null results of our analysis of self-reported information on hormone therapy and menopause timing do not support the hypothesis that using hormone therapy during a critical period at or around menopause prevents dementia.

This study overcame a number of limitations of prior epidemiologic studies of dementia and hormone therapy. It included a large number of elderly women with a very long average duration of hormone use. The number of incident cases in both hormone users and nonusers was large (threefold larger than in the Cache County Study). Although the validation study sample was small, the method of classification was formally evaluated and the results were published (29).

Other problems with design, outcome assessment, and exposure assessment are potential limitations to consider in this and other studies of dementia and hormone therapy. From a design perspective, there is a challenge in the use of self-reported data in epidemiologic studies of dementia: Missing self-reported data were associated with an increased hazard ratio for dementia for a number of variables, including hormone use and menopause. The increased risk of dementia associated with missing data may be a consequence of lack of exposure recall in people whose memory impairment cannot be detected using the instruments available for screening subjects in large epidemiologic studies. Excluding people with missing data from the analysis in epidemiologic studies of dementia in order to remedy the problem of missing data would decrease the numbers of dementia cases available for analysis and therefore reduce power. Such an exclusion has the potential to cause bias in the incidence and hazard ratio estimates in the remaining population. For example, if smoking were truly a cause of dementia and dementia caused people to fail to recall that they smoked, investigators estimating the hazard ratio for dementia in smokers based on subjects with a known smoking status might find a protective effect of smoking that was spurious.

Additional potential limitations arise from the assessment of outcome. First, attrition due to dropout and death may have introduced bias. Dropout was greatest between baseline and the first annual follow-up interview. Hormone therapy users at baseline were less likely than nonusers to drop out. The possibility that differential dropout of nonusers biased the estimate of dementia risk cannot be excluded. Deaths posed an additional problem in that women could have died with dementia before the study was able to detect it. Second, Alzheimer's disease has been the focus of most prior epidemiologic studies of hormones and dementia. The outcome category in this study, dementia without known cause, approximated possible Alzheimer's disease, but the inability to definitively identify women with Alzheimer's disease was a limitation. Finally, bias in the estimates of risk may have resulted from the selection of the time scale, and misclassification of cognitive impairment was evaluated (32). Additional analyses were carried out using age as the time variable (as compared with time in the study) (33) and combining dementia and cognitive impairment as outcomes. It is encouraging that hazard ratio estimates were similar and did not change the conclusions regarding dementia risk (data not shown).

The assessment and classification of hormone therapy use presented additional potential for bias. For the primary analyses, we classified hormone therapy on the basis of prescriptions filled at baseline and in the 7 years prior to baseline. Self-reported hormone therapy data were analyzed and compared with prescription data at baseline but not in subsequent years. Changes in self-reported hormone therapy use showed that at the last follow-up in 2003, only 572 (29 percent) of the women classified as hormone therapy users by prescription at baseline reported current hormone therapy use. Nonetheless, we did not perform analyses of change in use to avoid two potential biases. First, a downward bias (toward benefit and away from 1) in the estimated hazard ratio for dementia among continued hormone therapy users would arise if women who developed dementia stopped using hormone therapy selectively on the advice of their physician or because of an inability to comply with directions for medication use. Second, a downward bias in the estimated hazard ratio for dementia among continued hormone therapy users would arise if women who developed dementia were less likely to report continued hormone therapy use during follow-up because of "dementia bias" (a selective lack of recall of hormone therapy among women with dementia). Use of computer-stored information about prescriptions filled for hormone therapy after baseline enabled us to avoid the latter bias (due to self-reporting problems) but did not remedy any bias due to selective cessation of hormone therapy with dementia development. It is reassuring, however, that the analysis of the self-reported data yielded the same conclusion as the analysis based on prescription information.

The present study broadens the available information about the relation between hormone therapy use and dementia beyond prior epidemiologic studies. It included a large number of incident cases of dementia in a population of women with histories of very long-term use, beginning in many cases near the menopause, thereby improving the chance of detecting an association should one exist. Thus, these results add to the growing evidence that long-term use of hormone therapy does not help prevent the development of dementia.


   ACKNOWLEDGMENTS
 
This study was funded by National Institute on Aging grant AG14745.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Goodman Y, Bruce AJ, Cheng B, et al. Estrogens attenuate and corticosterone exacerbates excitotoxicity, oxidative injury, and amyloid β-peptide toxicity in hippocampal neurons. J Neurochem (1996) 66:1836–44.[Web of Science][Medline]
  2. Singer CA, Rogers KL, Dorsa DM. Modulation of Bcl-2 expression: a potential component of estrogen protection in NT2 neurons. Neuroreport (1998) 9:2565–8.[Web of Science][Medline]
  3. Heyman A, Wilkinson WE, Stafford JA, et al. Alzheimer's disease: a study of epidemiological aspects. Ann Neurol (1984) 15:335–41.[CrossRef][Web of Science][Medline]
  4. Amaducci LA, Fratiglioni L, Rocca WA, et al. Risk factors for clinically diagnosed Alzheimer's disease: a case-control study of an Italian population. Neurology (1986) 36:922–31.[Abstract/Free Full Text]
  5. Broe GA, Henderson AS, Creasey H, et al. A case-control study of Alzheimer's disease in Australia. Neurology (1990) 40:1698–707.[Abstract/Free Full Text]
  6. Graves AB, White E, Koepsell TD, et al. A case-control study of Alzheimer's disease. Ann Neurol (1990) 28:766–74.[CrossRef][Web of Science][Medline]
  7. Brenner DE, Kukull WA, Stergachis A, et al. Postmenopausal estrogen replacement therapy and the risk of Alzheimer's disease: a population-based case-control study. Am J Epidemiol (1994) 140:262–7.[Abstract/Free Full Text]
  8. Henderson VW, Paganini-Hill A, Emanuel CK, et al. Estrogen replacement therapy in older women. Comparisons between Alzheimer's disease cases and nondemented control subjects. Arch Neurol (1994) 51:896–900.[Abstract/Free Full Text]
  9. Mortel KF, Meyer JS. Lack of postmenopausal estrogen replacement therapy and the risk of dementia. J Neuropsychiatry Clin Neurosci (1995) 7:334–7.[Abstract/Free Full Text]
  10. Paganini-Hill A, Henderson VW. Estrogen replacement therapy and risk of Alzheimer disease. Arch Intern Med (1996) 156:2213–17.[Abstract/Free Full Text]
  11. Tang MX, Jacobs D, Stern Y, et al. Effect of oestrogen during menopause on risk and age at onset of Alzheimer's disease. Lancet (1996) 348:429–32.[CrossRef][Web of Science][Medline]
  12. Kawas C, Resnick S, Morrison A, et al. A prospective study of estrogen replacement therapy and the risk of developing Alzheimer's disease: The Baltimore Longitudinal Study of Aging. Neurology (1997) 48:1517–21.[Abstract/Free Full Text]
  13. Lerner A, Koss E, Debanne S, et al. Smoking and oestrogen-replacement therapy as protective factors for Alzheimer's disease. Lancet (1997) 349:403–4.[Web of Science][Medline]
  14. Baldereschi M, Di Carlo A, Lepore V, et al. Estrogen-replacement therapy and Alzheimer's disease in the Italian Longitudinal Study on Aging. Neurology (1998) 50:996–1002.[Abstract/Free Full Text]
  15. Marder K, Tang MX, Alfaro B, et al. Postmenopausal estrogen use and Parkinson's disease with and without dementia. Neurology (1998) 50:1141–3.[Abstract/Free Full Text]
  16. Slooter AJ, Bronzova J, Witteman JC, et al. Estrogen use and early onset Alzheimer's disease: a population-based study. J Neurol Neurosurg Psychiatry (1999) 67:779–81.[Abstract/Free Full Text]
  17. Waring SC, Rocca WA, Petersen RC, et al. Postmenopausal estrogen replacement therapy and risk of AD: a population-based study. Neurology (1999) 52:965–70.[Abstract/Free Full Text]
  18. Seshadri S, Zornberg GL, Derby LE, et al. Postmenopausal estrogen replacement therapy and the risk of Alzheimer disease. Arch Neurol (2001) 58:435–40.[Abstract/Free Full Text]
  19. Petitti DB, Buckwalter JG, Crooks VC, et al. Prevalence of dementia in users of hormone replacement therapy as defined by prescription data. J Gerontol A Biol Sci Med Sci (2002) 57:M532–8.[Abstract/Free Full Text]
  20. Zandi PP, Carlson MC, Plassman BL, et al. Hormone replacement therapy and incidence of Alzheimer disease in older women: The Cache County Study. JAMA (2002) 288:2123–19.[Abstract/Free Full Text]
  21. Marder K, Sano M. Estrogen to treat Alzheimer's disease: too little, too late? So what's a woman to do? (Editorial). Neurology (2000) 54:2035–7.[Free Full Text]
  22. Henderson VW, Paganini-Hill A, Miller BL, et al. Estrogen for Alzheimer's disease in women: randomized, double-blind, placebo-controlled trial. Neurology (2000) 54:295–301.[Abstract/Free Full Text]
  23. Wang PN, Liao SQ, Liu RS, et al. Effects of estrogen on cognition, mood, and cerebral blood flow in AD: a controlled study. Neurology (2000) 54:2061–6.[Abstract/Free Full Text]
  24. Mulnard RA, Cotman CW, Kawas C, et al. Estrogen replacement therapy for treatment of mild to moderate Alzheimer disease: a randomized controlled trial. Alzheimer's Disease Cooperative Study. JAMA (2000) 283:1007–15.[Abstract/Free Full Text]
  25. Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women. The Women's Health Initiative Memory Study: a randomized controlled trial. JAMA (2003) 289:2654–62.
  26. Shumaker SA, Legault C, Kuller C, et al. Conjugated equine estrogens and incidence of probable dementia and mild cognitive impairment in postmenopausal women. Women's Health Initiative Memory Study. JAMA (2004) 291:2947–58.[Abstract/Free Full Text]
  27. Welsh KA, Breitner JCS, Magruder-Habib KM. Detection of dementia in the elderly using telephone screening of cognitive status. Neuropsychiatry Neuropsychol Behav Neurol (1993) 6:103–10.[Web of Science]
  28. Chen W, Petitti DB, Geiger AM. Mortality following development of breast cancer while using oestrogen or oestrogen plus progestin: a computer record-linkage study. Br J Cancer (2005) 93:392–8.[CrossRef][Web of Science][Medline]
  29. Crooks VC, Clark L, Petitti DB, et al. Validation of multi-stage telephone-based identification of cognitive impairment and dementia. BMC Neurol (2005) 5:8. (Electronic article).[CrossRef][Medline]
  30. Gallo JJ, Breitner JC. Alzheimer's disease in the NAS-NRC Registry of aging twin veterans, IV. Performance characteristics of a two-stage telephone screening procedure for Alzheimer's dementia. Psychol Med (1995) 25:1211–19.[Web of Science][Medline]
  31. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, fourth edition, text revision. (2000) Washington, DC: American Psychiatric Association.
  32. Glymour MM. Invited commentary: when bad genes look good—APOE*E4, cognitive decline, and diagnostic thresholds. Am J Epidemiol (2007) 165:1239–46.[Abstract/Free Full Text]
  33. Thiebaut ACM, Benichou J. Choice of time-scale in Cox's model analysis of epidemiologic cohort data: a simulation study. Stat Med (2004) 23:3803–20.[CrossRef][Web of Science][Medline]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 JWatch Women's HealthHome page
Does Hormone Therapy Prevent Postmenopausal Dementia?
Journal Watch Women's Health, May 22, 2008; 2008(522): 5 - 5.
[Full Text]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
167/6/692    most recent
kwm362v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Petitti, D. B.
Right arrow Articles by Chui, H. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Petitti, D. B.
Right arrow Articles by Chui, H. C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?