| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLE |
From the Department of Psychiatry (M.S.); Neuropsychology Division, Department of Psychiatry (B.G., S.B.); Department of Biostatistics, School of Public Health (M.A. Schork); and the Department of Internal Medicine (D.S.), University of Michigan, Ann Arbor, MI.
Address reprint requests to: Monica N. Starkman, MD, Department of Psychiatry, University of Michigan, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0840. Email: starkman{at}umich.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Forty-eight patients with a first episode of acute, untreated Cushing’s disease and 38 healthy control subjects were studied.
RESULTS: Scores for four of five verbal IQ subtests were significantly lower in patients with Cushing’s disease; their scores were significantly lower for only one nonverbal performance IQ subtest (block design). Verbal, but not visual, learning and delayed recall at 30 minutes were significantly decreased among patients with Cushing’s disease. Although verbal delayed recall was significantly lower in these patients, the retention index (percentage), which compares the amount of initially learned material to that recalled after the delay, was not significantly decreased. There was no significant association between depression scores and cognitive performance. A higher degree of cortisol elevation was associated with poorer performance on several subtests of learning, delayed recall, and visual-spatial ability.
CONCLUSIONS: Chronically elevated levels of glucocorticoids have deleterious effects on particular domains of cognition. Verbal learning and other verbal functions seem more vulnerable than nonverbal functions. The results suggest that both the neocortex and hippocampus are affected.
Key Words: Cushing’s disease, • cortisol, • cognition, • learning, • neuropsychological tests.
Abbreviations: ACTH = adrenocorticotropic hormone;; CD = Cushing’s disease;; CS = Cushing’s syndrome;; SCL-90-R = 90-item Symptom Checklist–Revised;; WAIS-R = Wechsler Adult Intelligence Scale–Revised;; WMS = Wechsler Memory Scale.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Recent advances in neurobiology have heightened interest in the effects of glucocorticoid steroids on the brain and behavior. Animal studies indicate that the hippocampus, the brain structure critical for learning and memory, is especially vulnerable to both a deficiency of glucocorticoids (1, 2) and elevation of glucocorticoids (3, 4). A variety of experimental models show that exposure of animals to chronic social stress, chronic experimental stress, or administration of exogenous glucocorticoid induces changes in hippocampal pyramidal cell morphology and pyramidal cell loss (3–5) and impairs memory performance (6).
Spontaneous CS includes the major subtypes of CD resulting from pituitary hypersecretion of ACTH and pituitary ACTH–independent spontaneous CS such as that resulting from an adrenal adenoma. The disorder represents a unique human model to characterize the relationship between cognitive performance and chronic exposure to elevated levels of glucocorticoids. Patients with spontaneous CS experience chronic elevations of endogenous cortisol for periods of months to years before diagnosis and treatment. Their symptom profile characteristically includes complaints of difficulties with concentration and memory that occur in the absence of overt clouding of consciousness (7, 8). We previously found that patients with spontaneous CS exhibited variability in hippocampal formation volume as measured by magnetic resonance imaging of the brain. In 27% of patients with CS, hippocampal formation volume fell outside the 95% confidence interval for normal hippocampal volume given in the literature. There was an association between reduced hippocampal formation volume and lower scores for verbal learning and memory. Hippocampal formation volume was also negatively correlated with plasma cortisol levels, suggesting an association among elevated cortisol levels, reduced hippocampal formation volume, and memory dysfunction (9).
A limited number of studies have begun to define the specific pattern of cognitive deficits that develops in spontaneous CS. An initial study documented that patients with spontaneous CS exhibit cognitive abnormalities as measured by standard bedside tests of cognition (10). In another study, patients with active spontaneous CS were found to have variable degrees of neuropsychological impairment. Moderate to severe deficits were found in 34% of patients with spontaneous CS when language and nonlanguage test results were compared with standardized norms. Mild deficits (based on normative comparisons) were reported in 29% of these patients, and 37% had equivocal or no neuropsychological deficits (11). Although intriguing, this study was limited by the lack of a control group and the inclusion of both major subgroups of spontaneous CS, CD and pituitary ACTH–independent spontaneous CS. Although both of these clinical conditions involve elevated cortisol levels, they differ in etiology and peripheral ACTH levels. Another study of patients with spontaneous CS subsequently reported deficits in both language and nonlanguage tasks as well as in nonverbal memory functions (12). One study examined only patients with CD, the most common form of spontaneous CS, and found selective impairment of learning and recall functions (13). This study also was limited because the comparison group consisted of neurology inpatients believed to have no direct central nervous system involvement in their disorders but who were diagnosed with conditions (eg, pain syndrome) whose symptoms or treatments could have adversely affected cognitive function.
The present study was undertaken to confirm and extend these earlier investigations. We limited our sample to patients with CD and healthy control subjects with a normal level of cortisol. We hypothesized that cognition, particularly in the functions of learning and memory, is adversely affected in patients with untreated CD. We also hypothesized that severity of cognitive dysfunction is associated with the degree of elevation of peripheral cortisol levels. Finally, because depression is often present in patients with CD (7, 14), the possible association between cognitive and affective domains of behavior was investigated by examining the relationship between severity of depression and degree of cognitive impairment.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Control subjects.
Control subjects were recruited from concurrent studies that included neuropsychological testing. All available control subjects with no history of head injury, no current or past medical or psychiatric illness, no medication use, and within the age and education range of the CD patients were included, resulting in a sample of 30 women and 8 men. Mean age for the control subjects (36.7 ± 13.9 years) was not significantly different from that of the CD patients (t = -0.03, NS). Mean education in the control subjects (13.6 ± 2.0 years) also was not significantly different from that of patients with CD (t = -0.92, NS). Gender representation was not significantly different (
2 = 0.003, NS).
The study was approved by the University of Michigan’s Institutional Review Board for Medical Experimentation, and all patients and control subjects gave written informed consent before inclusion in the study.
Neuropsychological Tests
The WAIS-R (15) and WMS (16, 17) were used as the primary cognitive measures of attention, verbal and visual memory functions, verbal ability, and visual-motor functioning. These tests were used to generate three summary scores: verbal IQ, performance IQ (both from the WAIS-R), and memory quotient (from the WAIS-R). Listed below are descriptions of subtests used in this study, including references associating particular tasks with brain structure and functions. It should be recognized, however, that such neuropsychological measures often do tap functions subserved by several brain regions.
The WAIS-R is divided into two sections, verbal and performance. The following verbal subscales were used in this study:
Information.
The information subtest measures general knowledge from long-term experience by requiring a person to answer questions associated with common information. Positron emission tomography studies have demonstrated increases in cerebral glucose metabolism during performance of this test; these increases occur predominantly in the left hemisphere, but mild increases have been observed in the right hemisphere (18).
Comprehension.
The comprehension subtest measures verbal reasoning by having a person answer questions associated with practical social knowledge and social judgment, including proverb definitions. This test has been associated with generalized left hemisphere functioning (18).
Vocabulary.
This subtest measures general mental ability by having the patient give definitions for up to 40 words of increasing difficulty presented both orally and visually. Increased glucose metabolism has been found to occur predominantly in and around the left temporal lobe during performance of this test (18).
Similarities.
This subtest measures verbal concept formation by requiring a person to explain the similarity between pairs of common words. Similarities scores have been associated with predominantly the left temporal and frontal regions (18, 19).
Arithmetic.
The arithmetic subtest measures the ability of subjects to solve mental arithmetic problems. Strong, generalized left hemisphere increases in cerebral metabolism have been associated with this task (18).
The following nonverbal performance subtests measuring visual-spatial ability were used in the study:
Digit symbol.
This subscale is a test of psychomotor speed and problem-solving requiring code substitution based on pairs of letters and numbers. Glucose metabolism studies indicate bilateral increases in posterior cerebral regions during test performance with somewhat greater involvement of the right side (18).
Picture arrangement.
This visuospatial reasoning and planning subtest requires subjects to arrange cartoon pictures to tell a logical, coherent story. This subtest has been associated with parietal functioning bilaterally (18), although right hemisphere lesions seem to be more detrimental to performance (20).
Block design.
This test of visuospatial organization requires subjects to use blocks to reproduce two-dimensional, abstract, colored patterns. It has been associated with integrity of posteroparietal regions, particularly those involving the right cerebral hemisphere (18).
Object assembly.
In this visual construction subtest, subjects are asked to assemble four common objects from presented jigsaw pieces. This test involves primarily posterior parietal areas, particularly of the right cerebral hemisphere (18, 21).
Picture completion.
This subtest reflects the ability to discriminate important from unimportant fine visual detail. Subjects are asked to identify an important missing detail in presented line drawings. This task is associated with bilateral increases in cerebral metabolism (right greater than left) (18).
The following subtests from the WMS (Russell modification) were used. Although newer versions of the WMS now exist and are being used in our current studies, only this version was available at the initiation of this project. We continued to use this scale to provide consistency of methodology in all available subjects.
Mental control.
This subtest measures concentration and mental tracking. It consists of asking the patient to count backward, recite the alphabet, and count by three’s under time pressure.
Memory span.
This subtest is an attention measure consisting of immediate recall of a series of digits, both forwards and backwards. Cerebral metabolic activity during this test is highest in bilateral, anterior dorsal regions (18), and test performance is more vulnerable to left hemisphere lesions than to either right hemisphere or diffuse damage (20).
Logical memory.
This subtest measures the ability to freely learn and recall "ideas" presented in two literary passages read aloud to the subject.
Associate learning.
This task requires the patient to listen to paired associations of words read aloud and then to recall the correct response to stimulus words over three trials.
Visual reproduction.
In this nonverbal learning and memory task, subjects are asked to first study a geometric pattern for 10 seconds, after which time the item is removed and the subject must reproduce the design.
In the last three subtests, patients are instructed that they will be asked to recall the material after a 30-minute delay from initial presentation. For these tests, the following were examined: immediate recall or initial learning, delayed recall at one-half hour, and percentage of delayed recall, representing retention of newly learned material. The latter score is calculated as delayed recall divided by immediate recall, multiplied by 100 (17). Initial learning and delayed recall scores measure time-limited memory storage (22, 23). The percentage of delayed recall, as a measure of forgetting, characterizes the decay of successfully encoded material. It is a measure of a component of memory different from the processes represented by initial learning and delayed recall scores (24). Although likely to be subserved by a variety of brain regions, learning and memory are functions considered to be predominantly dependent on the hippocampus. Functional magnetic resonance imaging studies indicate that the hippocampus is sensitive to encoding and is activated during associate learning paradigms (25), and functional neuroimaging as well as animal lesion and observations in humans with impairments indicate that the hippocampus is involved in memory retrieval (26, 27).
The orientation and personal information subtests of the WMS were not used for formal statistical analysis to reduce the number of comparisons, because almost all CD patients and control subjects had perfect scores on these two subtests.
Two motor tasks also were examined in this study: tapping speed and strength of grip, both averaged over the dominant and nondominant hands. These measures were used to explore the possibility that group differences in cognitive test performance might relate to either decreased effort due to the depression seen in many patients with CD or weakness related to health status.
Neuropsychological tests were administered to patients with CD and to control subjects by trained examiners from the University of Michigan Neuropsychology Division in the same suite of rooms. CD patients were tested immediately before their diagnostic admission. Because all patients with the potential diagnosis of CS received neuropsychologic testing, at the time of testing the examiners had no knowledge of whether the patients would ultimately be confirmed to have the disease. The same examiners evaluated both patients and control subjects. Although all CD patients completed the full neuropsychological protocol, some of the control subjects did not undergo some WAIS-R subtests because of protocol constraints in the ongoing studies for which they had been initially recruited. The total number of individuals available for each subtest analysis is included in Table 1.
|
Measures of Cortisol
Urinary free cortisol and total plasma cortisol (protein-bound plus free) were determined by radioimmunoassay with Coat-a-Count kits (Diagnostics Products Corporation, Los Angeles, CA). The interassay coefficient of variation ranged from 4.0% to 6.4%, the sensitivity was 0.2 µg/dl, and recovery ranged from 91% to 100%. The specificity was 11.4% cross-reactivity with the natural steroid 11-deoxycortisol; 46% and 3.1% with the synthetic steroids prednisolone and prednisone, respectively; and less than 1.0% with all other steroids tested.
Statistical Analysis
Analysis of covariance was used as a conservative approach to test for group differences because the sample sizes varied for the different individual subtest comparisons. In every comparison between patients and control subjects, age and education were used as covariates. For each major set of comparisons, the summary scores for each of the three major domains (verbal IQ, performance IQ, and WMS memory quotient) were first examined with a target family error rate set at 
= 0.05. For domains that were statistically significant, post hoc examinations of the individual subtests comprising the domain were performed. Bonferroni adjustments were made to the p values for each individual test within the domain. Although the Bonferroni adjustment technique was chosen a priori as the primary method of analysis, non–Bonferroni-adjusted significance levels also are presented for review. Because the sample was predominantly female, analyses were repeated using women only. The results were essentially similar and therefore are not reported. Two-group comparisons were also recalculated using nonparametric statistics to ensure that significance was not affected by outlier or distribution effects. Because these results were essentially identical to the parametric analyses, they are not reported here. To evaluate the relationship between cortisol levels and performance on the neuropsychological tests, partial correlations were computed, adjusting for possible effects of education and age. Glucocorticoid values were log-transformed. To examine the role of depression in the CD group, similar partial correlations were computed for neuropsychological test scores and measures of depression from the SCL-90-R and Hamilton scales. The CD group also was divided into two categories on the basis of the depressed mood score (depressed mood present or depressed mood absent). These two groups were then compared with one another for performance on the neuropsychological tests. To examine any medication effects, t tests and a one-factor analysis of variance were used. Significance levels were set at p < .05 for comparisons related to mood and medications.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
From the WMS subtests, the two tests that reflect attention and working memory, mental control and digit recall, did not meet Bonferroni criteria for significant differences between CD and control subjects. Values for the WMS tasks reflecting verbal learning (logical memory and associate learning) were significantly lower in the CD patients as compared with control subjects. In contrast, the visual learning score (visual reproduction) for CD patients was not significantly different from that of control subjects. Delayed recall of both verbal subtests was significantly lower in CD patients. In contrast, delayed recall of visual material was not significantly different between the two groups. No group differences were found for retention of newly learned material as measured by percentage of delayed recall for either the verbal or visual subtests. Just as with the intellectual subtests, although not all comparisons reached significance based on Bonferroni correction, individual mean scores for the WMS measures were generally lower for the CD group as compared with the control group.
The next set of analyses was conducted to examine the potential role of depression in predicting performance among the CD group. As anticipated, patients in the CD group were significantly more depressed than control subjects (Table 1). To examine whether the differences observed in cognitive function could be explained, in part, by a relative lack of effort or weakness in CD patients with depression, tapping speed and strength of grip were examined. No significant differences between CD patients and control subjects in performance of these tests were observed, suggesting no difference in effort applied or other interference from depression with manual performance (Table 1). To examine the association between cognitive functioning and depression in the CD group, correlations were performed between neuropsychological test scores and levels of depressed mood based on the SCL-90-R Depression Scale score and the modified Hamilton Depression Scale score. There were no significant correlations for the three major composites (verbal IQ, performance IQ, or WMS memory quotient) with either of these two depression scales (all p > .5 except for verbal IQ and modified Hamilton score, p > .14). We then examined within the CD group whether patients with and without depression, based on the observer-rated, semistructured interview, differed from each other in cognitive function. Three patients with depressed mood rated subthreshold were excluded from the analysis. Subjects with depressed mood absent (N = 26) or present (N = 19) were not significantly different from each other in age or education. There were no significant differences between the two mood groups for verbal IQ, performance IQ, or the WMS memory quotient (p > .30). A similar result was obtained when the analyses were rerun after adding the patients with subthreshold depressed mood to the patients with no depressed mood.
The potential confounding effect of medication was examined in the CD patients. There were no significant differences in any of the three major domains examined for patients taking "no medication" vs. those taking "any medications" (p > .48). Similarly, analyses of variance examining each of the three domains by four subgroups (ie, no drugs, antihypertensives only, antihypertensives plus other medications, and other medications only) were not significant (p > .83). We also explored duration of illness as a contributing variable. None of the three cognitive composites demonstrated significant correlations with duration of illness in months (p > .39).
The next set of analyses examined the association between peripheral cortisol levels and neuropsychological dysfunction in the CD patients using partial correlations with age and education as covariates. Of the intellectual summary scores, performance IQ was significantly negatively correlated with urinary free cortisol concentration (r = -0.38, p < .008) and verbal IQ was not. For the summary WMS memory quotient, there was a strong trend for a negative correlation with urinary free cortisol (r = -0.27, p < .06). Of the individual performance IQ subtests, urinary free cortisol was significantly negatively correlated with block design (r = -0.54, p< .0001), object assembly (r = -0.50, p < .001), and digit symbol (r = -0.31, p < .04). Of the WMS subtests, urinary free cortisol was significantly negatively correlated with associate learning (r = -0.47, p < .001) and associate learning delayed recall (r = -0.35, p < .01). These analyses were repeated with depression scores added as an additional covariate, and the results were virtually identical. For all the above analyses, using mean plasma cortisol concentrations yielded substantially similar results.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The most robust and consistent differences between CD patients and control subjects were found across measures of verbal intellectual skills and in verbal learning and recall. This suggests a particular sensitivity of verbal as compared with visuospatial functions in CD. Verbal deficits are consistent with the clinical cognitive complaints reported by patients with CD and CS (30). In contrast to the findings with verbal intellectual measures, relatively less robust and inconsistent differences were found for intellectual measures of visuospatial functioning. This pattern of more marked involvement of verbal skills stands in contrast to that seen in organic brain diseases such as delirium and dementia, which are predominantly characterized by early and increased vulnerability in nonverbal, visual-motor functions or generalized impairment of both verbal and visual functions (31, 32). This also differs from the pattern often cited in aging, suggesting enhanced vulnerability of visual-spatial functions with increasing age (33–35).
One visuospatial, nonverbal task, block design, was notable in that it did show a significant, robust difference between CD and control subjects. This timed subtest requires the subject to analyze a pictured, two-dimensional, colored pattern into its component parts and then use colored blocks to resynthesize the pattern and reproduce the abstract design. This task also requires coordinated movement for successful completion, but findings from other measures of motor function rule out praxic disturbances as explanatory of the results. Further highlighting the particular vulnerability of this task in CD, of all the tasks in the study, block design had the highest partial correlation with urinary free cortisol, which explained more than 25% of the variance.
New learning of verbal information, whether from verbally presented paragraphs or word pairs, was also significantly decreased in CD. In contrast, initial learning of visual information (visual reproduction) was not significantly reduced after Bonferroni correction. In addition, although verbal delayed recall scores were significantly lower in the CD group as compared with the control group, visual delayed recall scores were not lower after Bonferroni correction, further highlighting the relatively greater decrement of verbal, as compared with visual, functions. Such a dichotomy has also been observed in posttraumatic stress disorder. In these patients, verbal learning and recall were significantly impaired, but visual learning and recall were not (36). One possible explanation for the presence of such a dichotomy in our findings, and those of others, is that it may reflect differences in the difficulty of tests used to examine verbal and visual functions. However, it may reflect a biologic difference in verbal and visual learning and thus a specific pathologic consequence in CD.
Our findings with initial learning and delayed recall measures are, on the whole, consistent with the one published study examining cognition of patients with CD (13). In that study, too, significantly lowered scores were found in measures of verbal learning and recall (WMS immediate and delayed recall). Those investigators reported visual learning as also being impaired, although a Bonferroni correction for multiple comparisons was not applied and this difference from control subjects was among the least robust of their findings.
An interesting finding in our study was that the percentage of delayed recall for both verbal and visual material was not decreased in CD patients. This measure takes initial learning into account and examines an adjusted rate of "forgetting" in subjects. We found that once material was learned, the percentage of retention one-half hour after learning was not different from that of control subjects. This finding seems consistent with results from the other study of CD patients mentioned above. Although percentage of delayed recall was not specifically examined by the investigators in that study, the published tables reveal similar modest declines from initial learning scores in both patients and control subjects. This finding in CD is different from neuropsychological observations in patients with generalized dementias, such as Alzheimer’s disease, in which rapid forgetting is a hallmark (37, 38). Our finding is similar to that for subjects with alcohol dependence, in whom deficits are apparent in both initial learning and delayed recall measures, but whose rates of forgetting do not differ significantly from those of control subjects (39). In patients with psychotic major depression, too, inefficiency in learning new material yet normal retention of this learned material for one-half hour have been demonstrated (40). Our results suggest that chronic elevation of cortisol has less effect on retrieval at one-half hour than on initial learning. The mechanisms and regions involved in retrieval after this short time period are unclear. They are likely different from mechanisms involved in longer-term consolidation and retention (41).
Several factors may have contributed to lower cognitive performance among CD patients in this study. Decrements in more complex aspects of attention may underlie, at least in part, some of the decrements in more general cognitive ability. CD patients did perform at a lower level than normal subjects on a task requiring complex functions of concentration and mental manipulation (ie, arithmetic), although performance on other attention, concentration, and working memory tasks, such as counting backwards or reciting the alphabet (ie, mental control or digit span), did not meet Bonferroni criteria to demonstrate significant differences between the CD and control groups. Acute anxiety due to a diagnostic admission at the time of testing could have negatively impacted attention, but difficulty with attention and concentration is a chronic symptom frequently reported by patients with CD (30). Altered concentration of steroids other than cortisol, such as androgens and estrogens, may also be affecting cognitive performance in CD patients.
Two factors are unlikely to have contributed to the decrements in cognitive function. The higher prevalence of depression among CD patients in the present study did not explain their decreased cognitive function. As expected, the patients with CD were more depressed. However, CD patients showed no decrement in effort as measured by the motor tests (grip strength and tapping). There was also no significant association between severity of depressed mood or their depressive syndrome scores and cognitive performance. This indicates that the behavioral domains of mood and cognition can be affected independently of each other in response to hypercortisolemia. A second possibility is that the CD group, by chance, may have been simply less intelligent than the control subjects. However, there was no statistically significant difference in education, and all analyses used education as one of the covariates. In addition, the information subtest, which relies heavily on long-term retention and general fund of knowledge, was not significantly different, arguing against the possibility of marked premorbid differences.
Our results are consonant with those in other human populations exposed chronically to elevated levels of cortisol. In patients with Alzheimer’s disease, elevated plasma cortisol concentrations were associated with declines in cognitive performance 1 year later (42). Normal elderly subjects studied over a 4-year period, whose basal cortisol levels had risen over time, showed a marked decline on tasks measuring selective attention and explicit declarative memory, compared with those without such progressive elevation (43). Normal nonelderly subjects who received several days of stress-level exposure to cortisol had decreases in immediate and delayed verbal declarative memory but not in nonverbal memory or attention (44). With respect to percentage of delayed recall, here, too, inspection of the published means revealed no relative difference in the small, similar declines from initial to delayed recall between subjects exposed to excessive cortisol and control subjects.
In our study cortisol concentration showed a significant correlation or strong trend with two of the summary domains examined (WMS memory quotient and performance IQ) and was significantly correlated with subtests reflecting verbal learning and visuospatial functioning. Although patients with CD were significantly different from control subjects on verbal intellectual measures, the verbal IQ and its subtest scores did not show a significant correlation with cortisol concentration. One possibility for the lack of correlation may be a "ceiling effect" such that even mildly elevated cortisol levels result in deleterious consequences on verbal aspects of cognition. Another possibility stems from the fact that the cortisol level at the time of testing most probably does not accurately reflect the degree of chronic exposure of the brain to elevated glucocorticoids, because it is unlikely that a uniform degree of hypercortisolemia was present in each patient over time. Nonverbal tasks may be more sensitive to the actual level of cortisol at the time of testing. Lastly, it is possible that the subtests that were correlated with cortisol concentration reflect biological specificity. Block design, the nonverbal task that was significantly lower in CD patients, was also the nonverbal task most robustly correlated with cortisol concentration. Of the subtests of the WMS, both learning and delayed recall of paired associates were significantly lower in CD patients and also significantly correlated with cortisol concentrations.
With regard to neural mechanisms and brain region involvement underlying our findings in CD, glucocorticoid receptors are widely distributed in the brain, making involvement of a wide variety of anatomic regions possible. The decrements in verbal intellectual skills, for example, suggests particular involvement of the neocortex. Because learning and memory are dependent on the hippocampus, our results indicating decrements in verbal learning and delayed recall are consistent with the emerging view of the hippocampus as a brain structure especially vulnerable to the effects of stress and elevated levels of glucocorticoids (Refs. 3–6 and 45–49).
In conclusion, the findings of the present study extend the evidence in humans indicating that chronically elevated levels of glucocorticoids have particular deleterious effects on cognition. We have recently demonstrated increases in hippocampal formation volume after treatment of CD (50). Longitudinal studies examining the reversibility of the cognitive deficits reported here are in progress to confirm and extend these results.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication April 11, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Z LeWinn, L. R Stroud, B. E Molnar, J. H Ware, K. C Koenen, and S. L Buka Elevated maternal cortisol levels during pregnancy are associated with reduced childhood IQ Int. J. Epidemiol., December 1, 2009; 38(6): 1700 - 1710. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Tsolaki, F. Kounti, and S. Karamavrou Severe Psychological Stress in Elderly Individuals: A Proposed Model of Neurodegeneration and Its Implications American Journal of Alzheimer's Disease and Other Dementias, April 1, 2009; 24(2): 85 - 94. [Abstract] [PDF]
|
|
|
|
 |
|
 S. L. Wright and C. Persad Distinguishing Between Depression and Dementia in Older Persons: Neuropsychological and Neuropathological Correlates J Geriatr Psychiatry Neurol, December 1, 2007; 20(4): 189 - 198. [Abstract] [PDF]
|
|
|
|
 |
|
 S Appenzeller, A D Carnevalle, L M Li, L T L Costallat, and F Cendes Hippocampal atrophy in systemic lupus erythematosus Ann Rheum Dis, December 1, 2006; 65(12): 1585 - 1589. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Lekwauwa, D. McQuoid, and D. C. Steffens Hippocampal Volume Is Associated With Physician-Reported Acute Cognitive Deficits After Electroconvulsive Therapy J Geriatr Psychiatry Neurol, March 1, 2006; 19(1): 21 - 25. [Abstract] [PDF]
|
|
|
|
 |
|
 J. R. Lindsay, T. Nansel, S. Baid, J. Gumowski, and L. K. Nieman Long-Term Impaired Quality of Life in Cushing's Syndrome despite Initial Improvement after Surgical Remission J. Clin. Endocrinol. Metab., February 1, 2006; 91(2): 447 - 453. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Gold, I. Dziobek, K. Rogers, A. Bayoumy, P. F. McHugh, and A. Convit Hypertension and Hypothalamo-Pituitary-Adrenal Axis Hyperactivity Affect Frontal Lobe Integrity J. Clin. Endocrinol. Metab., June 1, 2005; 90(6): 3262 - 3267. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. P. Merke, J. N. Giedd, M. F. Keil, S. L. Mehlinger, E. A. Wiggs, S. Holzer, E. Rawson, A. C. Vaituzis, C. A. Stratakis, and G. P. Chrousos Children Experience Cognitive Decline Despite Reversal of Brain Atrophy One Year After Resolution of Cushing Syndrome J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2531 - 2536. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. O'Brien, A. Lloyd, I. McKeith, A. Gholkar, and N. Ferrier A Longitudinal Study of Hippocampal Volume, Cortisol Levels, and Cognition in Older Depressed Subjects Am J Psychiatry, November 1, 2004; 161(11): 2081 - 2090. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. C. Sandeep, J. L. W. Yau, A. M. J. MacLullich, J. Noble, I. J. Deary, B. R. Walker, and J. R. Seckl From The Cover: 11{beta}-Hydroxysteroid dehydrogenase inhibition improves cognitive function in healthy elderly men and type 2 diabetics PNAS, April 27, 2004; 101(17): 6734 - 6739. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. PORTER, P. GALLAGHER, J. M. THOMPSON, and A. H. YOUNG Neurocognitive impairment in drug-free patients with major depressive disorder The British Journal of Psychiatry, March 1, 2003; 182(3): 214 - 220. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
