| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Epidemiology and Biostatistics, Erasmus University Medical School (S.K., L.J.L., H.A.P.P., A.H., M.M.B.B.), and the Department of Internal Medicine, University Hospital Rotterdam (F.H.d.J., H.A.P.P., S.W.J.L.), Rotterdam; the Department of Chronic Diseases and Environmental Epidemiology, National Institute of Public Health and the Environment (L.J.L.), Bilthoven; and the Julius Center for Patient Oriented Research, Utrecht University (R.P.S.), Utrecht, The Netherlands
Address all correspondence and requests for reprints to: Dr. S. Kalmijn, Department of Epidemiology and Biostatistics, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
Hypothalamo-pituitary-adrenal (HPA) axis overactivity, which is related to stress and possibly also to aging, leads to increased cortisol levels (3). Sustained cortisol exposure in rodents and primates results in damage of the hippocampus (4). In human studies, increased cortisol levels and HPA axis overactivity have been associated with cognitive impairment and dementia (3, 5, 6, 7, 8, 9, 10, 11). Only a few studies have investigated the longitudinal association between cortisol and cognitive function. Lupien et al. (5) found a relationship between increases in cortisol levels over time and cognitive impairment in a small sample of healthy elderly. Among 194 healthy participants of the MacArthur Field Study, an association was observed between increasing cortisol levels and cognitive decline in women, but not in men (6).
DHEA and its sulfate, DHEAS, are regarded as markers of aging (12). In animal experiments DHEAS has been shown to enhance neuronal and glial survival and differentiation in culture (13, 14), and injection of DHEAS into the brains of mice improved long term memory (13). DHEAS may act directly on the hippocampus, as it has been shown to enhance the magnitude of hippocampal primed burst (2) and long term potentiation (15). There are a number of case-control studies on the association between DHEAS and Alzheimer’s disease; however, these yielded conflicting results (16, 17, 18). In one prospective population-based study, there was no consistent association between DHEAS and subsequent cognitive function (19).
We examined the cross-sectional and longitudinal effects of peripheral levels of the steroid hormones cortisol and DHEAS on cognitive impairment and decline in a sample of participants from the prospective population-based Rotterdam Study.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
The Rotterdam Study is a single center prospective population-based study (20) designed to investigate determinants of chronic disabling diseases in the elderly. The conduct of the study was approved by the medical ethics committee of Erasmus University, and written consent was obtained from all participants. All residents of Ommoord, a suburb in Rotterdam, aged 55 yr or over, including those living in homes for the elderly, were invited to participate. The baseline examinations started in May 1990 and continued until June 1993. Of the 10,275 eligible subjects, 7,983 (78%) agreed to participate. During a home visit, trained interviewers administered a questionnaire, covering, among other areas, socio-demographic background, medical history, and medication use. This was followed by 2 clinical examinations at the research center, including neuropsychological testing. The follow-up examination started in September 1993 and lasted until December 1994. Of the 7,215 subjects who were still alive, 6,315 (88%) agreed to participate.
For the present additional examination, a random sample (n = 219) was taken of participants from the Rotterdam Study, who had completed the baseline examination in the preceding 6 months, were between 55–80 yr old, and had no history of psychiatric or endocrine diseases, including diabetes mellitus treated with medication. There were no differences in age, sex, or education between our sample and the other participants of the Rotterdam Study in the same age range and without dementia or known diabetes mellitus. The mean baseline Mini-Mental State Examination (MMSE) score was higher in our sample [28.1 (SD = 1.6) vs. 27.6 (SD = 1.9); P = 0.002].
Blood measurements
Blood was obtained after an overnight fast at the research center between 0800–0900 h and was allowed to coagulate for 30 min. Serum was separated by centrifugation and quickly frozen in liquid nitrogen. In addition, an overnight dexamethasone suppression test was performed (21). Participants were given a tablet of 1 mg dexamethasone and were instructed to take this at 2300 h. The next morning, at the same time as on the previous day, a fasting blood sample was obtained in the same manner. Measurements included cortisol and corticosteroid-binding globulin (CBG) before and after dexamethasone, and DHEAS before dexamethasone. CBG levels were estimated using a double antibody RIA with murine monoclonal antibodies against human CBG and radioiodinated CBG as the tracer (Medgenix Diagnostics, Brussels, Belgium). Variation coefficients for this assay were below 12%. Cortisol and DHEAS were estimated using Coat-A-Count kits provided by Diagnostic Products (Los Angeles, CA). These kits use immobilized antibodies against the respective steroids and radioiodinated tracers. The specificity of the antibodies was high. Of the endogenous steroids, 11-deoxycortisol showed the highest cross-reaction in the cortisol assay (11.4%); for the DHEAS assay, estrone sulfate had the highest cross-reactivity (0.56%). Variation coefficients for these assays were less than 12.5% for cortisol and less than 6.0% for DHEAS. Two measures of cortisol were examined, i.e. total cortisol and an estimation of free circulating cortisol, using a formula that included cortisol, CBG, and albumin concentrations, according to the method of Södergard et al. (22) The correlation coefficient of free cortisol with total cortisol was 0.61. Suppression of total cortisol after dexamethasone treatment can be regarded as a measure of negative feedback sensitivity of the HPA axis to glucocorticoids. Because the cortisol level after dexamethasone administration was not normally distributed, we used the natural logarithm for further calculations. We also investigated the ratio of free cortisol over DHEAS, because DHEAS may antagonize the actions of cortisol (23, 24, 25).
MMSE
Global cognitive function was tested with the Dutch version of the 30-point MMSE during the (first) visit to the research center (26). It was administered by specially trained research assistants. The MMSE includes questions on orientation to time and place, registration, attention and calculation, recall, language, and visual construction. This screening test was originally created for a clinical setting (26) and is extensively used in epidemiological studies (27). Although it tests mainly cortical functions, these are important to daily functioning and are severely affected in dementia. If less than 4 individual items (of 20) were not answered by the subject, these were rated as errors (28). If a subject did not answer 4 or more individual items, the total MMSE score was considered missing. Cognitive impairment was defined as a score below 26 (29), and cognitive decline as a drop in the MMSE score of more than 1 point/yr (approximately >1 SD). The mean follow-up time between the first and second MMSEs was 1.9 yr (SD = 0.23).
Other measurements
A number of factors were considered as possible confounding variables: age, sex, and education, classified into four levels: completed primary education; lower vocational or general education; intermediate vocational or general education; and higher vocational training, college, or university (Unesco, Paris, France). For all analyses on cortisol we adjusted for symptoms of depression. Symptoms of depression were assessed at the follow-up examination with translated versions of the Center for Epidemiologic Studies Depression Scale (CES-D) (30) in the first 99 participants and the Hospital Anxiety and Depression Scale (HADS) (31) in the last 46 participants. In 44 participants, there was no (complete) depression scale available; these subjects were significantly older and had a lower baseline MMSE score, but the measures of cortisol and DHEAS were not different from those in subjects with data on the depression scale. We used the standard cut-off of 16 or above for the CES-D (30) and a cut-off of 8 or above for the HADS (after exclusion of the anxiety items) (31) to identify subjects with depressive symptoms. Body mass index (weight/height2) was considered a possible confounding factor, because low food intake and low body weight might induce relative resistance to the dexamethasone suppression test, and a low body mass index has been related to dementia (32). In the analyses of cognitive decline, we also adjusted for the baseline MMSE score.
Statistical analysis
One subject who was using hydrocortisone and one subject who had a greatly elevated DHEAS level (15.1 µmol/L) were excluded from the analyses. Complete information on cognitive function at baseline was available for 189 subjects. Follow-up data on cognitive decline were available for 169 subjects. For the analyses of cortisol, estrogen users (n = 2) and users of antiepileptics (n = 4) were excluded, as these medications can influence cortisol and/or CBG levels and sensitivity to dexamethasone.
Differences in baseline characteristics according to cognitive
impairment and according to steroid hormone levels above and below the
median were tested with the Mann-Whitney test for continuous variables
and the 
2 test for categorical variables. Logistic
regression was used to estimate the odds ratios (OR) and 95%
confidence intervals (CI) for the risk of cognitive impairment and
decline. The independent variables of interest were total cortisol,
free cortisol, the natural logarithm of cortisol after dexamethasone
administration, DHEAS, and the free cortisol/DHEAS ratio. We included
confounding variables in the model. A missing indicator variable was
created for the missing values for the depressive symptoms. We also
investigated whether there was effect modification by gender (33) by
including the product term as a covariate in the model. All tests were
two sided, and P < 0.05 was considered statistically
significant. Data analyses were performed using BMDP statistical
software (BMDP Statistical Software, Inc., Los Angeles, CA).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
Subjects with cognitive impairment were older, more often had symptoms
of depression at follow-up, and had lower DHEAS concentrations and
higher ratios of cortisol to DHEAS (Table 1
). Subjects with a total cortisol level
above the median (506.5 nmol/L) had a lower body mass index than those
with a total cortisol level below the median [25.8 (SD =
3.8) vs. 27.1 (SD = 3.4); P =
0.01]. Subjects with a free cortisol concentration above the median
were older and less often female (Table 2). Participants with DHEAS levels below
the median were more often female (66% vs. 34%;
P < 0.001).
|
|
). The OR
for the association with cognitive impairment according to a 1
SD increase in the free cortisol level (SD =
30.3) was 1.5 (95% CI, 0.9–2.4). CBG levels were not associated with
cognitive function (data not shown). The natural logarithm of cortisol
after dexamethasone treatment was associated with a significantly
increased risk of cognitive decline (adjusted OR per SD
increase = 1.5; 95% CI, 1.0–2.3). The DHEAS level was inversely,
but not significantly, related to cognitive impairment (OR = 0.5;
95% CI, 0.2–1.1) and decline (OR = 0.6; 95% CI, 0.4–1.1). As
expected, the ratio of cortisol over DHEAS was significantly associated
with cognitive impairment (OR = 1.8; 95% CI, 1.0–3.2).
Adjustment for body mass index did not change these estimates.
Exclusion of subjects with depressive symptoms (n = 14)
essentially yielded the same estimates, but CIs were wider because of
the smaller sample size (results not shown). There were no significant
interactions between the variables of interest and sex.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
Methodological issues
It could be argued that selection bias may have affected the validity of our results. Participants were healthy and relatively young, and follow-up duration was short, leading, on the average, to only a small drop in the MMSE score. Combined with the small sample size, this would only impede the detection of a significant modest association. Furthermore, the 20 subjects in our sample without a MMSE score at follow-up (due to death or nonresponse) had a significantly lower baseline MMSE score and were older than subjects who were not lost to follow-up, but there were no differences in cortisol or DHEAS levels, making selective loss to follow-up less likely.
The MMSE was used to assess cognitive function. Although the MMSE is developed as a global clinical screening test, it is a valid and reliable indicator of cognitive impairment, with a test-retest reliability generally between 0.80–0.95 (35). The MMSE was not originally created to measure cognitive decline and may be less sensitive to small changes in cognitive function. However, the reliability of the change in the MMSE after 1 yr or more in patients with dementia was 0.74, which is reasonable (36).
Another methodological issue is that we may not have adequately adjusted for depression. Depression is strongly associated with cognitive impairment (37) and also with increased cortisol levels and diminished suppression of cortisol after dexamethasone treatment (38). Thus, depression is potentially an important confounder or intermediate. The use of two different questionnaires might not have been optimal. The CES-D was designed for epidemiologic research, whereas the HADS was designed for a clinical setting. Both have been validated, but the CES-D is probably more reliable, especially in this setting (30). Furthermore, we adjusted for symptoms of depression at follow-up and not at baseline. Thus, there may be residual confounding in our results for cognitive impairment, but less in our results for cognitive decline.
Cortisol and HPA axis overactivity
Our results confirm the findings of the few prospective studies that have investigated the association between cortisol and cognition (5, 6). A prospective study among 19 healthy elderly subjects showed a relationship between increases in basal cortisol levels over time and 2 out of 17 cognitive measures, i.e. explicit memory and selective attention (5). Although the researchers used a conservative test for significance, they did not adjust for age or other possible confounders. Among 194 healthy elderly participants from the MacArthur Field Study, an association was observed between a rise in cortisol levels and a decline in memory performance, but only in women (6). No associations were found with abstraction and spatial ability.
It has been shown that HPA axis overactivity is more frequent among patients with Alzheimer’s disease (10). Also, in early dementia, a decreased sensitivity to glucocorticoid feedback, using low doses of dexamethasone, has been shown. (11) In addition, cortisol has been associated with severity of dementia and cognitive decline in demented patients (8, 9). Furthermore, patients with Cushing’s syndrome more often present with cognitive impairment (39). Finally, one of the side-effects of treatment with synthetic corticosteroids is cognitive deterioration (40).
The relation between cortisol and cognitive function may be due to different effects of cortisol on the hippocampus. In experimental animal studies, administration of corticosterone led to a decrease in the number of pyramidal cells in the hippocampus (4), soma shrinkage (41), and a reduction in the number and length of dendritic branches (42). In patients with Alzheimer’s disease and Cushing’s syndrome, higher cortisol levels were associated with lower hippocampal volumes (43, 44). Therefore, stress-induced HPA axis overactivity and increased cortisol levels may cause hippocampal damage and, subsequently, cognitive decline. On the other hand, hippocampal damage, i.e. as a result of Alzheimer pathology, may lead to HPA axis overactivity and increased cortisol levels, because hippocampal neurons are involved in the negative feedback of glucocorticoid secretion. This is described as the glucocorticoid cascade of aging (45). Of course, both sequences of events may also proceed at the same time.
DHEAS
There has been one previous prospective population-based study that investigated the relationship between DHEAS and subsequent cognitive impairment, in which no consistent association was found (19). The follow-up duration was quite long (16 yr), and subjects with lower DHEAS levels at baseline had a higher mortality, suggesting that there may have been selective survival bias. There have been a number of case-control studies on DHEAS levels and Alzheimer’s disease. Some found that Alzheimer’s disease patients had lower DHEAS levels than controls (16, 46), but this could not be confirmed by others (17, 18). These studies were mostly based on small numbers of patients and controls, thereby increasing the possibility of false negative results. A small pilot study among patients with major depression showed improvement of selective memory functions after DHEA administration (47).
Serum levels of DHEAS are the highest of all steroids in humans.
Cerebrospinal fluid levels of DHEA and DHEAS correlated
significantly with peripheral concentrations (48). There are a number
of plausible mechanisms for an association between DHEAS and cognitive
function. Animal experiments showed that DHEAS can enhance neuronal and
glial survival and differentiation in culture (13, 14), perhaps because
DHEAS acts as a 
-aminobutyric acid antagonist (49). Furthermore,
DHEAS can improve memory in mice (13), which may be due to a direct
effect of DHEAS on hippocampal neurons, through enhancement of the
magnitude of the hippocampal primed burst (2) and long term
potentiation (15) in rats. Another explanation for the association with
cognition may be that DHEA can be transformed into
estrogens in peripheral tissues. (50) The risk of Alzheimer’s disease
has been shown to be reduced in women using estrogen compared with
those who did not (51). Furthermore, DHEA may act as a
glucocorticoid antagonist, as described for a number of systems
(23, 24, 25). Therefore, a high ratio of cortisol over DHEAS may especially
lead to hippocampal damage (15) and thus to cognitive impairment.
Indeed, the present study showed a significant association between the
cortisol/DHEAS ratio and cognitive impairment. Finally, DHEAS may be
regarded as a marker of general health status and could therefore be
related to cognitive function.
|
Conclusion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
Received September 30, 1997.
Revised June 12, 1998.
Accepted June 19, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. K. Lee, T. A. Glass, M. J. McAtee, G. S. Wand, K. Bandeen-Roche, K. I. Bolla, and B. S. Schwartz Associations of Salivary Cortisol With Cognitive Function in the Baltimore Memory Study Arch Gen Psychiatry, July 1, 2007; 64(7): 810 - 818. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. van Gelder, M. Tijhuis, S. Kalmijn, S. Giampaoli, A. Nissinen, and D. Kromhout Marital Status and Living Situation During a 5-Year Period Are Associated With a Subsequent 10-Year Cognitive Decline in Older Men: The FINE Study. J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2006; 61(4): P213 - P219. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Kaufman and A. Vermeulen The Decline of Androgen Levels in Elderly Men and Its Clinical and Therapeutic Implications Endocr. Rev., October 1, 2005; 26(6): 833 - 876. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Fonda, R. Bertrand, A. O'Donnell, C. Longcope, and J. B. McKinlay Age, Hormones, and Cognitive Functioning Among Middle-Aged and Elderly Men: Cross-Sectional Evidence From the Massachusetts Male Aging Study J. Gerontol. A Biol. Sci. Med. Sci., March 1, 2005; 60(3): 385 - 390. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Wood, J. M. Cline, M. S. Anthony, T. C. Register, and J. R. Kaplan Adrenocortical Effects of Oral Estrogens and Soy Isoflavones in Female Monkeys J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2319 - 2325. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Lovas, G. Gebre-Medhin, T. S. Trovik, K. J. Fougner, S. Uhlving, B. G. Nedrebo, O. L. Myking, O. Kampe, and E. S. Husebye Replacement of Dehydroepiandrosterone in Adrenal Failure: No Benefit for Subjective Health Status and Sexuality in a 9-Month, Randomized, Parallel Group Clinical Trial J. Clin. Endocrinol. Metab., March 1, 2003; 88(3): 1112 - 1118. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. W. J. Lamberts The Endocrinology of Aging and the Brain Arch Neurol, November 1, 2002; 59(11): 1709 - 1711. [Full Text] [PDF]
|
|
|
|
 |
|
 B. Breuer, C. Martucci, S. Wallenstein, A. Likourezos, L. S. Libow, A. Peterson, and B. Zumoff Relationship of Endogenous Levels of Sex Hormones to Cognition and Depression in Frail, Elderly Women Am J Geriatr Psychiatry, June 1, 2002; 10(3): 311 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. Riggs, S. Khosla, and L. J. Melton III Sex Steroids and the Construction and Conservation of the Adult Skeleton Endocr. Rev., June 1, 2002; 23(3): 279 - 302. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Richards, R. Hardy, D. Kuh, and M. E. Wadsworth Birthweight, postnatal growth and cognitive function in a national UK birth cohort Int. J. Epidemiol., April 1, 2002; 31(2): 342 - 348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kalmijn, J. A. M. J. L. Janssen, H. A. P. Pols, S. W. J. Lamberts, and M. M. B. Breteler A Prospective Study on Circulating Insulin-Like Growth Factor I (IGF-I), IGF-Binding Proteins, and Cognitive Function in the Elderly J. Clin. Endocrinol. Metab., December 1, 2000; 85(12): 4551 - 4555. [Abstract] [Full Text]
|
|
|
|
|
|
P. J. Hunt, E. M. Gurnell, F. A. Huppert, C. Richards, A. T. Prevost, J. A. H. Wass, J. Herbert, and V. K. K. Chatterjee Improvement in Mood and Fatigue after Dehydroepiandrosterone Replacement in Addison's Disease in a Randomized, Double Blind Trial J. Clin. Endocrinol. Metab., December 1, 2000; 85(12): 4650 - 4656. [Abstract] [Full Text]
|
|
|
|
|
|
G. A. Laughlin and E. Barrett-Connor Sexual Dimorphism in the Influence of Advanced Aging on Adrenal Hormone Levels: The Rancho Bernardo Study J. Clin. Endocrinol. Metab., October 1, 2000; 85(10): 3561 - 3568. [Abstract] [Full Text]
|
|
|
|
 |
|
 S. D. Moffat, A. B. Zonderman, S. M. Harman, M. R. Blackman, C. Kawas, and S. M. Resnick The Relationship Between Longitudinal Declines in Dehydroepiandrosterone Sulfate Concentrations and Cognitive Performance in Older Men Arch Intern Med, July 24, 2000; 160(14): 2193 - 2198. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Bergendahl, A. Iranmanesh, T. Mulligan, and J. D. Veldhuis Impact of Age on Cortisol Secretory Dynamics Basally and as Driven by Nutrient-Withdrawal Stress J. Clin. Endocrinol. Metab., June 1, 2000; 85(6): 2203 - 2214. [Abstract] [Full Text]
|
|
|
|
 |
|
 S. Kiechl, J. Willeit, E. Bonora, S. Schwarz, and Q. Xu No Association Between Dehydroepiandrosterone Sulfate and Development of Atherosclerosis in a Prospective Population Study (Bruneck Study) Arterioscler. Thromb. Vasc. Biol., April 1, 2000; 20(4): 1094 - 1100. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Parker Jr., S. M. Slayden, R. Azziz, S. L. Crabbe, G. A. Hines, L. R. Boots, and S. Bae Effects of Aging on Adrenal Function in the Human: Responsiveness and Sensitivity of Adrenal Androgens and Cortisol to Adrenocorticotropin in Premenopausal and Postmenopausal Women J. Clin. Endocrinol. Metab., January 1, 2000; 85(1): 48 - 54. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
