QJM Advance Access originally published online on July 30, 2007
QJM 2007 100(9):567-573; doi:10.1093/qjmed/hcm065
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The effect of hormone replacement therapy on cognitive function in post-menopausal women
From the 1Gaucher Clinic, Shaare Zedek Medical Center, Jerusalem, 2NeuroTrax Corporation, Teaneck, USA, and 3Department of Obstetrics and Gynecology, Hebrew University–Hadassah Medical Center, Mt Scopus, Jerusalem, Israel
Address correspondence to Dr D. Elstein, Gaucher Clinic, Shaare Zedek Medical Center, One Bazak Road, Jerusalem 91031, Israel. email: elstein{at}szmc.org.il
Received 6 March 2007 and in revised form 24 May 2007
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Background: Despite interest in causes of dementia in older persons, particularly in post-menopausal women, it is unclear whether hormone replacement therapy (HRT) is a risk factor.
Aim: To assess cognitive function in post-menopausal women with high educational status receiving HRT, compared to non-users.
Design: Cognitive functioning was assessed with in women aged 55–60 years with at least university-level education, using the Mindstreams system, a computerized cognitive battery with multiple domains.
Results: Of 165 women meeting the inclusion/exclusion criteria, 82 women (49.7%) declined participation. Of the remaining 83, 40 (48.2%) had never received HRT; the remainder was divided into women receiving 5–9 years HRT (n = 29)versus those with 
10 years HRT (n = 11). There were no statistically significant differences between HRT users and non-users in global scores or sub-domains of cognitive functioning, and no difference between those women receiving HRT for 5–9 years vs. 10 years.
Discussions: Long-term HRT does not appear to impair cognitive functioning in highly-educated women. Recommendations regarding post-menopausal HRT should be made on an individual basis.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
There is increasing concern about dementia among the elderly, particularly since impairment of cognitive functioning is variable, and is not necessarily an inevitable result of aging. It may therefore be argued that certain lifestyle decisions and/or genetic predispositions may affect cognition as individuals age, or conversely, that some interventions might prevent or halt cognitive decline. Among the populations investigated are post-menopausal women receiving hormone replacement, either as oestrogen alone or in combination with progesterone, for whom there might be some neuroprotection because of long-term therapy.1
At the turn of the century, meta-analyses of research and epidemiological studies suggested that hormone replacement therapy (HRT) might have an ameliorative effect on cognition in post-menopausal women.2,3 However, subsequent reports of the Women's Health Initiative Memory Study (WHIMS; WHISCA) in very large cohorts4–6 stated that HRT increased the risk of cognitive impairment and dementia in elderly women. Combination therapy (oestrogen plus progesterone) had a deleterious effect on verbal memory, but a trend to improvement in figural memory, with prolonged use only.6 But oestrogen, with or without progesterone, appeared to increase the risk of cognitive dysfunction in older women.7 In a study contemporary with WHIMS of paired-matched post-menopausal older women with long-term use of oestradiol patches, episodic memory and mental ‘flexibility’ was worse in women receiving HRT than in controls.8 However, a more recent study found exactly the opposite: HRT users had improved episodic memory and verbal fluency.9 Further, in recent follow-up studies of 5 and 6 years of HRT, in hormone users versus non-users10 and even when including a group of ‘irregular’ users of HRT,11 the results did not show any significant difference among the groups on all standard measures of cognitive functioning. It has also been recently reported that while there was no difference in cognitive scores between those receiving HRT and non-users, the risk of impairment was less in short-term users.12
Finally, the WHIMS studies reported that the inimical effect of HRT in women older than 65 years was greater among women with lower cognitive function at initiation of therapy.5 An earlier study2 had concluded exactly the opposite: HRT induces the greatest protective effect in older women with low education, while a recent study9 reported improved cognitive function independent of age and education.
An interesting study in younger women with menopausal symptoms undergoing blood-oxygenation-level-dependent functional magnetic resonance imaging (BOLD fMRI) documented increased frontal system activity with oestrogen therapy and improved executive functioning.13 Thus, among these and other studies, many memory systems14 have been variously tested, but the results are still equivocal and open to interpretation. Indeed, one of us15 has recommended individual risk-benefit evaluations to determine the needs and expectations of older post-menopausal women vis a vis HRT. In coronary artery disease, it appears from animal and human studies that HRT is capable of preventing atherosclerosis but cannot stop the process once it has started.15 Similarly, the WHIMS study cannot address the role of HRT in preventing, delaying, or reducing the incidence of dementia, as women involved in the WHIMS were 65 years and older, and it is assumed that in this age group even women considered healthy might already have clinically undetected decline in cognitive function resulting from atherosclerosis. It is questionable whether any preventive effects of HRT could be achieved in elderly patients exposed to long-term oestrogen deficiency. Consequently, the WHIMS study should be considered as a trial to study the efficacy of hormones in secondary prevention rather than primary prevention of dementia.
Because of these equivocal results in the literature, this study was designed to assess cognitive function using an innovative computerized system (Mindstreams, NeuroTrax Corp.)16 among post-menopausal women with high educational status receiving oestrogen plus progesterone therapy from the peri-menopausal period for varying periods, and to determine whether combined HRT affects sub-scores of cognitive function relative to non-users and if length of therapy is an independent variable. The novelty of the system can be framed by comparison to traditional neuropsychological testing, which is expensive, time-consuming (usually several hours), and requires a trained professional to administer. In contrast, Mindstreams is inexpensive, brief (under one hour), and requires only a technician to supervise administration, as the tests are essentially self-administered. Because traditional neuropsychological tests are administered by a person, they are more likely to be influenced by administrator bias. Traditional neuropsychological testing also generally measures only accuracy. Mindstreams measures both accuracy and response time on a millisecond timescale and is thus more sensitive for detecting subtle changes in cognition. Other innovations of the system are automatic and immediate scoring, including calculation of normalized (standardized) scores, and immediate report generation. Unlike other computerized systems, Mindstreams is suitable not only for research but also for widespread clinical use, requiring no special equipment, only a standard personal computer and an internet connection. Finally, the system benefits from minimal ceiling and floor effects due to adaptive testing designs. and automatically incorporates alternate forms to minimize learning on follow-up.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Participants
All medical records in a large Medical-Centre-based clinic for post-menopausal women were surveyed for those who matched the inclusion and exclusion criteria. We excluded women who: (i) were known to suffer from dementia or were treated for cognitive decline; (ii) had clinical depression; (iii) had had a stroke in the past 6 months; (iv) had undergone hysterectomy and/or oophorectomy; or (v) had had HRT for <5 years. Inclusion criteria were chosen that strictly defined a specific population: women aged 55–60 years, of Ashkenazi Jewish ethnicity, with at least a university/college education level.
Of 3500 medical records, 3280 women (93.7%) did not meet all inclusion and exclusion criteria; of the remaining 165 women, 82 (49.7%) declined to participate in the study. Of the 83 who agreed to participate, 40 (48.2%) had never received any HRT. The remaining group of 43 included women who had received oral combined HRT for at least 5 years. The treatment was started within a year of menopause; most women (87.5%) reported using 17-ß-oestradiol + NETA progesterone (oral or transdermal); of these 29 women had 5–9 years; 11 women had 10 or more years of usage.
All women who agreed to participate signed an informed consent form and were asked to complete a demographic questionnaire and take the Mindstreams cognitive evaluation. All participants were told that the results would be made available to them immediately should they desire to have them.
Cognitive testing
Cognitive functioning used the Mindstreams system, a computerized cognitive battery for detection of mild impairment.16 Mindstreams generates single-trial data consisting of information related to each individual trial including: stimulus type, onset and offset of the stimulus (in ms), expected response type, actual response type, and time of response. Processing of Mindstreams data incorporates automatic quality control mechanisms designed to maximize the validity of the cognitive data. For example, there are continuous checks of the local testing computer for adequate performance throughout the testing session, as required for accurate timing measurements.
Accuracy outcome parameters are calculated as the average accuracy (i.e. percent correct) across all trials included in a particular level of a test. Similarly, response time outcome parameters are calculated as the average response time across all correct responses in a level. Composite score outcome parameters are computed as average accuracy divided by average response time for a level, in an attempt to account for the speed-accuracy trade-off. Standard deviations of response time outcome parameters are computed as the standard deviation of the response times across all correct responses in a level. Other test-specific outcome parameters (e.g. inter-tap interval for the Finger Tapping test) are generated in a similar manner, as appropriate for the test.
Performance in study individuals was compared with that of women of similar age and years of education, drawn from a database of Mindstreams data collected in controlled research studies.16,17
Mindstreams data were uploaded to the NeuroTrax central server, where data processing occurred, during which aggregate outcome parameters were computed from raw single-trial data. Outcome parameters were calculated using automatic algorithms blind to diagnosis/group. To permit averaging performance across different types of outcome parameters (e.g. accuracy, reaction time), each parameter was normalized (standardized) according to age and years of education and fitted to an IQ-style scale (mean 100, SD 15).
The following are brief descriptions of the tests administered. Go-NoGo Response Inhibition test: timed continuous performance test during which responses are made to large coloured stimuli that are any colour but red. Verbal Memory: ten pairs of words (the study set) are presented, followed by a recognition test in which one member (the target) of a previously presented pair appears, together with a list of four candidates for the other member of the pair. There are four immediate repetitions and one delayed repetition after 10 minutes. Problem Solving test: pictorial puzzles of gradually increasing difficulty are presented. Each puzzle consists of a 2 x 2 array containing three black-and-white geometric forms with a certain spatial relationship among them and a missing form. Participants must choose the best fit for the fourth (missing) form from among six possible alternatives. Stroop Interference test: Timed test of response inhibition and set shifting adapted from the well established paper-based test. For example, in the ‘No Interference [Meaning]’ phase, the task is to choose the colour named by a word presented in white letter-colour. In the final [‘Interference’] phase, participants choose the letter-colour of a word that names a different colour. Non-Verbal Memory: this is similar to the test of verbal memory, except that geometric figures are used instead of words. Finger Tapping: participants must tap on the mouse button with their dominant hand. Catch Game: a test of motor planning requiring hand-eye coordination and rapid responses. Subjects ‘catch’ a ‘falling object’ by moving a ‘paddle’ horizontally on the computer screen so that it can be positioned directly in the path of the falling object. Accuracy is weighted by level of difficulty so that the participant gets more points for catching the object at a faster versus slower rate. Staged Information Processing Speed test: this test comprises three levels of information processing load: single digits, two-digit arithmetic problems (e.g. 5–1), and three-digit arithmetic problems (e.g. 3 + 2–1). For each of the three levels, stimuli are presented at three different fixed rates, incrementally increasing as testing continues. Participants are instructed to respond as quickly as possible by pressing the left mouse button if the digit or result is less than or equal to 4 and the right mouse button if it is greater than 4. Verbal Function test: pictures of common objects are presented; in the first phase, the word that best rhymes with the name of the object must be selected from among four choices; in the second phase, the name of the picture must be selected. Visual Spatial Processing test: Computer-generated scenes containing a red pillar are presented. Participants must select the view of the scene from the vantage point of the red pillar.
Standardized outcome parameters that measure similar cognitive functions were averaged to produce seven ‘index scores’, each summarizing performance in a single cognitive domain.18 The following index scores were computed: memory (mean accuracies for learning and delayed recognition phases of Verbal and Non-Verbal Memory tests); executive function (composite scores (accuracy divided by reaction time) for the interference phase of the Stroop test and Go-NoGo test, mean weighted accuracy for ‘Catch’ Game); visual-spatial (mean accuracy for Visual Spatial Processing test); verbal (weighted accuracy for verbal rhyming test (part of Verbal Function test)); attention (mean reaction times for the Go-NoGo test, a non-interference phase of the Stroop test, and a low-load stage of Staged Information Processing Speed test; mean standard deviation of reaction time for the Go-NoGo test; mean accuracy for a medium-load stage of Information Processing test); information processing (composite scores (accuracy divided by reaction time) for various low- and medium-load stages of the Staged Information Processing Speed test); motor skills (mean time until first move for ‘Catch Game,’ mean inter-tap interval and standard deviation of inter-tap interval for Finger Tapping test).
A Global Cognitive Score (GCS) reflecting general cognitive status was computed as the average of the index scores. Index scores and the GCS served as dependent measures for the present study.
Ethics Committee approval was obtained for the study, and informed consent was obtained from all participants.
Statistical analyses
The 
2 test (with Fisher's exact test) was used to compare groups with regard to demographic (independent) variables such as marital status, smoking, alcohol use, etc. The t-test was used to test the effect of these variables on the cognitive scores and the effect of duration of therapy on scores. The t-test for equality of means (Mann-Whitney U test for non-parametric data) was used to compare age, years of education, BMI, time between menopause and advent of HRT, duration of HRT, and comparing cognitive scores between groups. Unless specified, all tests were two-tailed and p < 0.05 was considered significant. For more than two groups, the ANOVA with multiple pair-wise procedure (post-hoc) was used.
Power calculation
The minimum mean differences (M2-M1) in Mindstreams summary measures associated with small, medium, and large effects19 required for power of 0.80 (p < 0.05, two-tailed) were calculated. According to the method described by Howell,20 196 subjects would be needed to detect a small effect, 32 subjects for a medium effect, and 13 subjects for a large effect. Thus, only 13 subjects would be necessary for power of 0.80 (p < 0.05, two-tailed) in a study of an experimental agent anticipated to result in a minimum change of eight normalized units in overall Mindstreams performance (i.e. Global Cognitive Score); 32 subjects would be sufficient to detect a change of five units, and 196 subjects would be required to detect a change of two units.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Table 1 presents the results of the demographic data of the participants. There were no statistically significant differences between groups in any of the demographic variables, including between HRT users and non-users who complained of menopausal symptoms.
|
Table 2 presents between-groups (HRT vs. non-HRT) results for Mindstreams index scores and the GCS. There were no statistically significant differences between the groups for any of the scores.
|
Table 3 presents Mindstreams GCS score as a function of demographic data. Again, there were no statistically significant between group differences in GCS for demographic variables. There was also no statistically significant difference in GCS according to duration of HRT (5–9 years vs.
10 years).
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
The literature regarding the effect of any hormonal treatment on cognitive functioning in post-menopausal women is inconclusive, in the sense that one cannot conclude whether, when taking into account only the variables of menopausal symptoms and cognition, the risk-benefit ratio is towards an overall benefit. There is the basic question of whether HRT affects cognition in any of its forms, since there are serious and large cohort studies variously proving that there are beneficial effects, selective effects, or no effects, or that HRT induces dysfunction and cognitive decline. The effect of the duration of therapy and the importance of educational status of the woman prior to menopause are also unresolved.
The current study compared a group of relatively young women with (at least) university-level education who had been exposed to combination HRT for at least 5 years, with good compliance and good control of menopausal symptoms (as assessed by one of us, DH-C, their primary physician for the entire duration) relative to women who had never been exposed to HRT (and who were seen for equal periods of time). Only women with more than 5 years of therapy were chosen, to remove the possible confounder of short-term therapy and/or poor compliance. Recent studies have suggested that appropriately high hormone levels, as achieved by good compliance to a regimen of therapy, may have a small but definite impact on cognitive performance.21 Thus, this study specifically addressed only highly-educated, ethnically comparable, relatively young women who began HRT shortly after menopause and were highly compliant with their regimen. These limiting criteria may have decreased overall enrolment into the study, but on the other hand, should also have made the results less amorphous. The choice of women with only high educational status was to eliminate the possible confounder of years of education. The decision to only include relatively young women was to preclude the confounder of an overly long interval between menopause and institution of therapy. Years elapsed from actual menopause may affect some functions such as executive function, more than other functions.22
The format of the study was a computerized evaluation of various cognitive domains as well as an overall score of function that is compared to a normative sample of similar age and education. This computerized system reduces investigator bias. The overall conclusion of this study is that there were no statistically significant differences between the two groups with regard to cognitive function as assessed by this format. Similarly, there were no statistically significant differences in global scores or in any of the sub-domains of cognitive functioning between groups. There was also no difference between those women who received HRT for more than 5 but less than 10 years, versus more than 10 years. Although the latter groups were smaller than the overall cohort (because they were limited to those receiving NETA), these latter findings corroborate similar studies in older women.23
Based on this study, one cannot extrapolate to women with low educational status, or to women with very short-term or poor compliance to HRT. However, the use of a younger cohort of women may be an important feature of the current design. In rat studies of the effect of hormone therapy on cues requisite for the Morris water maze, treatment reversed hippocampal markers of age-related memory impairment but importantly, a higher dosage of hormone was required in the older rats.24 Therefore, age at initiation of HRT may be important in terms of retaining cognitive status. Our results are in agreement with recent studies of a similar format but in women whose educational level was not controlled.25
We evaluated several domains of cognitive function, and compared the responses among those exposed to HRT immediately at the start of menopause for at least five years with those untreated, and our study groups were well-defined, relatively young, high-functioning, and age-matched. Recent studies with equivocal results have recommended more sensitive tools for testing cognition,26 and our use of the Mindstreams tests may represent the advent of such sensitive, multi-domain assessment to studies of this type. Our results do not contradict the recent studies suggesting that compliance, age at onset of HRT, educational level and the choice of assessment tool are all important. They suggest that HRT may not be inimical to cognitive function in (at least some) highly-educated post-menopausal women, and that individualized HRT remains worth consideration.
|
Acknowledgements |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
This study was performed in partial fulfillment of requirements for an MD degree by Revital Lavi at the Hebrew University-Hadassah Medical School, Jerusalem Israel. Glen M. Doniger and Ely Simon are paid employees of the Neurotrax Corporation; none of the other authors have disclosures relevant to this study. No funding was received for this study.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
1. Deroo BJ, Korach KS. Estrogen receptors and human disease. J Clin Invest (2006) 116:561–70.[CrossRef][Web of Science][Medline]
2. Hogervorst E, Williams J, Budge M, Riedel W, Jolles J. The nature of the effect of female gonadal hormone replacement therapy on cognitive function in post-menopausal women: a meta-analysis. Neuroscience (2000) 101:485–512.[CrossRef][Web of Science][Medline]
3. Sano M. Understanding the role of estrogen on cognition and dementia. J Neural Transm Suppl (2000) 59:223–9.[Medline]
4. Shumaker SA, Legault C, Kuller L, 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.
5. Espeland MA, Rapp SR, Shumaker SA, et al. Conjugated equine estrogens and global cognitive function in postmenopausal women: Women's Health Initiative Memory Study. JAMA (2004) 291:2959–68.
6. Resnick SM, Maki PM, Rapp SR, et al. Effects of combination estrogen plus progestin hormone treatment on cognition and affect. J Clin Endocrinol Metab (2006) 91:1802–10.
7. Craig MC, Maki PM, Murphy DG. The Women's Health Initiative Memory Study: findings and implications for treatment. Lancet Neurol (2005) 4:190–4.[Web of Science][Medline]
8. File SE, Heard JE, Rymer J. Trough oestradiol levels associated with cognitive impairment in post-menopausal women after 10 years of oestradiol implants. Psychopharmacology (Berl) (2002) 161:107–12.[CrossRef][Medline]
9. Yonker JE, Adolfsson R, Eriksson E, Hellstrand M, Nilsson LG, Herlitz A. Estrogen therapy and cognition: a 6-year single-blind follow-up study in postmenopausal women. Neurology (2006) 67:706–9.
10. O'Hara R, Schroder CM, Bloss C, et al. Hormone replacement therapy and longitudinal cognitive performance in postmenopausal women. Am J Geriatr Psychiatry (2005) 13:1107–10.[CrossRef][Web of Science][Medline]
11. Alhola P, Polo-Kantola P, Erkkola R, Portin R. Estrogen therapy and cognition: a 6-year single-blind follow-up study in postmenopausal women. Neurology (2006) 67:706–9.
12. Bagger YZ, Tanko LB, Alexandersen P, Qin G, Christiansen C. PERF Study Group. Early postmenopausal hormone therapy may prevent cognitive impairment later in life. Menopause (2005) 12:12–17.[CrossRef][Web of Science][Medline]
13. Joffe H, Hall JE, Gruber S, et al. Estrogen therapy selectively enhances prefrontal cognitive processes: a randomized, double-blind, placebo-controlled study with functional magnetic resonance imaging in perimenopausal and recently postmenopausal women. Menopause (2006) 13:411–22.[CrossRef][Web of Science][Medline]
14. Budson AE, Price BH. Memory dysfunction. N Engl J Med (2005) 352:692–9.
15. Haimov-Kochman R, Hochner-Celnikier D. Are there second thoughts about the results of the WHI study? Acta Obstet Gynecol Scand (2006) 85:387–93.[Web of Science][Medline]
16. Dwolatzky T, Whitehead V, Doniger GM, et al. Validity of the Mindstreams computerized cognitive battery for mild cognitive impairment. J Mol Neurosci (2004) 24:33–44.[CrossRef][Web of Science][Medline]
17. Schweiger A, Doniger GM, Dwolatzky T, Jaffe D, Simon ES. Reliability of a novel computerized neuropsychological battery for mild cognitive impairment. Acta. Neuropsychologica (2003) 1:407–13.
18. Elstein D, Guedalia J, Doniger GM, Antebi V, Arnon Y, Simon E, Zimran A. Computerized cognitive testing in patients with type I Gaucher disease: effects of enzyme replacement and substrate reduction. Genet Med (2005) 7:124–30.[Web of Science][Medline]
19. Cohen J. Statistical power analysis for the behavioral sciences (1988) 2nd. Academic Press: NewYork.
20. Howell DC. Statistical methods for psychology (1997) 4th. Belmont CA: Duxbury.
21. Yonker JE, Adolfsson R, Eriksson E, Hellstrand M, Nilsson LG, Herlitz A. Verified hormone therapy improves episodic memory performance in healthy postmenopausal women. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn (2006) 13:291–307.[Medline]
22. Elsabagh S, Hartley DE, File SE. Cognitive function in late versus early postmenopausal stage. Maturitas (2006) Epub ahead of print.
23. MacLennan AH, Henderson VW, Paine BJ, et al. Hormone therapy, timing of initiation, and cognition in women aged older than 60 years: the REMEMBER pilot study. Menopause (2006) 13:28–36.[CrossRef][Web of Science][Medline]
24. Foster TC, Sharrow KM, Kumar A, Masse J. Interaction of age and chronic estradiol replacement on memory and markers of brain aging. Neurobiol Aging (2003) 24:839–52.[CrossRef][Web of Science][Medline]
25. Grigorova M, Sherwin BB. No differences in performance on test of working memory and executive functioning between healthy elderly postmenopausal women using or not using hormone therapy. Climacteric (2006) 9:181–94.[CrossRef][Web of Science][Medline]
26. Shah S, Bell RJ, Savage G, et al. Testosterone aromatization and cognition in women: a randomized, placebo-controlled trial. Menopause (2006) 13:600–8.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||