Subtle Cognitive Impairments in Patients with Long-Term Cure of Cushing’s Disease

Page 1

J. Clin. Endocrinol. Metab. published online Apr 6, 2010; , doi: 10.1210/jc.2009-2032  

A. M. Middelkoop, Alberto M. Pereira and Johannes A. Romijn Jitske Tiemensma, Nieke E. Kokshoorn, Nienke R. Biermasz, Bart-Jan S. A. Keijser, Moniek J. E. Wassenaar, Huub

  Disease

Subtle Cognitive Impairments in Patients with Long-Term Cure of Cushing's

Society please go to: http://jcem.endojournals.org//subscriptions/ or any of the other journals published by The EndocrineJournal of Clinical Endocrinology & MetabolismTo subscribe to

Copyright © The Endocrine Society. All rights reserved. Print ISSN: 0021-972X. Online

Page 2

Subtle Cognitive Impairments in Patients withLong-Term Cure of Cushing’s Disease

Jitske Tiemensma, Nieke E. Kokshoorn, Nienke R. Biermasz, Bart-Jan S. A. Keijser,Moniek J. E. Wassenaar, Huub A. M. Middelkoop, Alberto M. Pereira,and Johannes A. Romijn

Departments of Endocrinology and Metabolism (J.T., N.E.K., N.R.B., B.-J.S.A.K., M.J.E.W., A.M.P., J.A.R.)and Neurology (H.A.M.M.), Leiden University Medical Center, 2300 RC Leiden, The Netherlands

Context and Objective: Active Cushing’s disease is associated with cognitive impairments. Wehypothesized that previous hypercortisolism in patients with Cushing’s disease results in irrevers-ible impairments in cognitive functioning. Therefore, our aim was to assess cognitive functioningafter long-term cure of Cushing’s disease.

Design: Cognitive assessment consisted of 11 tests, which evaluated global cognitive functioning,memory, and executive functioning.

Patients and Control Subjects: We included 74 patients cured of Cushing’s disease and 74 controlsmatched for age, gender, and education. Furthermore, we included 54 patients previously treatedfor nonfunctioning pituitary macroadenomas (NFMA) and 54 controls matched for age, gender,and education.

Results: Compared with NFMA patients, patients cured from Cushing’s disease had lower scores onthe Mini Mental State Examination (P � 0.001), and on the memory quotient of the WechslerMemory Scale (P � 0.050). Furthermore, patients cured from Cushing’s disease tended to recallfewer words on the imprinting (P � 0.013), immediate recall (P � 0.012), and delayed recall (P �

0.003) trials of the Verbal Learning Test of Rey. On the Rey Complex Figure Test, patients cured fromCushing’s disease had lower scores on both trials (P � 0.002 and P � 0.007) compared with NFMApatients. Patients cured from Cushing’s disease also made fewer correct substitutions on the Letter-Digit Substitution Test (P � 0.039) and came up with fewer correct patterns on the Figure FluencyTest (P � 0.003) compared with treated NFMA patients.

Conclusions: Cognitive function, reflecting memory and executive functions, is impaired inpatients despite long-term cure of Cushing’s disease. These observations indicate irreversibleeffects of previous hypercortisolism on cognitive function and, thus, on the central nervoussystem. These observations may also be of relevance for patients treated with high-dose ex-ogenous glucocorticoids. (J Clin Endocrinol Metab 95: 0000 – 0000, 2010)

Cushing’s disease is characterized by excessive exposureto cortisol. Despite curative treatment, cardiovascu-

lar morbidity and mortality remain increased in these pa-tients (1, 2). In addition, despite long-term cure of Cush-ing’s disease, these patients have persistent physical andpsychological complaints, associated with decreased qual-ity of life parameters (3).

Patients with active Cushing’s disease and Cushing’ssyndrome have cognitive impairments, especially in thememory domain. Previous studies reported impairmentsin memory, visual and spatial information, reasoning, ver-bal learning, and language performance (4–10). Struc-tures important in cognitive functioning, like the hip-pocampus and cerebral cortex, are rich in glucocorticoid

ISSN Print 0021-972X ISSN Online 1945-7197Printed in U.S.A.Copyright © 2010 by The Endocrine Societydoi: 10.1210/jc.2009-2032 Received September 22, 2009. Accepted March 8, 2010.

Abbreviations: HADS, Hospital Anxiety and Depression Scale; MMSE, Mini Mental StateExamination; MQ, memory quotient; NFMA, nonfunctioning pituitary macroadenoma.

O R I G I N A L A R T I C L E

E n d o c r i n e C a r e

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 1

J Clin Endocrin Metab. First published ahead of print April 6, 2010 as doi:10.1210/jc.2009-2032

Copyright (C) 2010 by The Endocrine Society

Page 3

receptors and are therefore particularly vulnerable tothe glucocorticoid excess present in Cushing’s disease(7). Starkman et al. (11) reported that 27% of the pa-tients with active Cushing’s syndrome fell outside the95% confidence intervals for normal subject hippocam-pal formation volume and that hippocampal formationvolume and performance on cognitive tests were positivelyrelated. In accordance, many other studies in humans andanimal models have documented that prolonged, in-creased endogenous or exogenous exposure to glu-cocorticoids may have long-lasting adverse effects onbehavioral, psychiatric, and cognitive functions, due tofunctional and, over time, structural alterations in specificbrain target areas including the hippocampus (11–16). Af-ter treatment, all patients in the study by Starkman et al.(17) showed an increase in hippocampal formation vol-ume, and half of the patients also showed an increase incognitive function test scores. In contrast, other studiesfound no improvements in cognitive functioning within 1yr after treatment (18, 19). Some studies reported im-paired cognitive functioning in patients with treated Cush-ing’s disease (6, 18–20). However, these studies includedonly small numbers of subjects (n � 35), and patients weretested relatively shortly (i.e. within the first 12–18 months)after cure of Cushing’s disease. Therefore, it is presentlyunclear to which extent impairments in cognitive func-tioning remain present in patients with much longer du-ration of cure of Cushing’s disease.

We hypothesized that previous hypercortisolism in pa-tients with Cushing’s disease results in irreversible impair-ments in cognitive functioning. Therefore, we evaluatedcognitive functioning in patients after long-term cure forCushing’s disease and compared these data with those ofage- and sex-matched controls as well as with those ofpatients treated for nonfunctioning pituitary macroad-enomas (NFMA) and matched controls.

Subjects and Methods

SubjectsWe included four groups of subjects: 1) patients cured from

Cushing’s disease and 2) gender-, age-, and education-matchedcontrol subjects and 3) patients previously treated for NFMAand 4) gender-, age-, and education-matched control subjects.The inclusion of these additional control groups was necessarybecause patients with Cushing’s disease and NFMA differ withregard to age and gender.

We invited all patients in remission after treatment for Cush-ing’s disease in our institution to participate (n � 153). Eachpatient was asked to provide a control person of comparable age,gender, and education. Patients and their controls were evalu-ated at the same time. Patients who did not respond were en-couraged by phone to participate. The response rate was 93%.Eighty-five patients were willing to participate, of whom 74 pa-

tients actually participated in all cognitive tests. Fifty-seven pa-tients preferred not to participate, whereas 11 patients did notrespond. The characteristics of patients who participated in thetests and those who did not participate were carefully compared.There were no differences in clinical characteristics between bothgroups. Reasons for not participating were distance to our in-stitution, participation in other studies, old age, and debilitatingdisease.

The diagnosis of Cushing’s disease had been established byclinical signs and symptoms and by biochemical tests includingincreased urinary excretion rates of free cortisol, decreased over-night suppression by dexamethasone (1 mg) and, since 2004,elevated midnight salivary cortisol values in addition to sup-pressed ACTH levels. All patients were treated by transsphenoi-dal surgery, if necessary followed by repeat surgery and/or ra-diotherapy. Cure of Cushing’s disease was defined by normalovernight suppression of plasma cortisol levels (�100 nmol/liter) after administration of dexamethasone (1 mg) and normal24-h urinary excretion rates of cortisol (�220 nmol/24 h). Hy-drocortisone independency was defined as a normal cortisol re-sponse to CRH or insulin tolerance test. Patients were followedat our department with yearly intervals, and pituitary hormonesubstitution was prescribed in accordance with the results ofyearly evaluation. Persistent cure of Cushing’s disease was doc-umented by normal values of a dexamethasone (1 mg) suppres-sion test, urinary cortisol excretion rates, and midnight salivarycortisol levels before participation in the current study.

In addition, we invited 132 patients with NFMA to partici-pate in the study and to provide a control person (see above). Theresponse rate was 94%. Fifty-four had undergone transsphenoi-dal surgery and participated in all cognitive tests. There were nodifferences in clinical characteristics between participants andnonparticipants.

Pituitary function was assessed at yearly intervals. In patientswho were glucocorticoid dependent after treatment, recovery ofthe pituitary-adrenal axis was tested twice a year. The dose ofhydrocortisone was on average 20 mg/d divided into two to threedosages. After withdrawal of hydrocortisone replacement for24 h, a fasting morning blood sample was taken for the mea-surement of serum cortisol concentrations. Patients with serumcortisol concentration less than 120 nmol/liter were consideredglucocorticoid dependent, and hydrocortisone treatment was re-started. Patients with serum cortisol levels of 120–500 nmol/literwere tested by ACTH stimulation tests (250 �g). A normal re-sponse to ACTH stimulation was defined as a stimulated cortisolhigher than550nmol/liter. In case the cortisol response toACTHwas normal the patients were tested by insulin tolerance test orCRH stimulation test. In case the cortisol responses to these testswere less than 550 nmol/liter, hydrocortisone treatment was re-started. Evaluation of GH deficiency was done by insulin toler-ance test or arginine-GHRH test only in patients under the ageof 70 yr and only after at least 2 yr of remission. Patients with aninadequate stimulation of GH by one of these tests were treatedwith recombinant human GH, aiming at IGF-I levels between 0and �2 SD values. In addition, the twice-yearly evaluation con-sisted of measurement of free T4 and testosterone (in male pa-tients). If results were below the lower limit of the respectivereference ranges, substitution with L-T4 and/or testosterone wasstarted. In the case of amenorrhea and low estradiol levels inpremenopausal women, estrogen replacement was provided. Pa-tient and treatment characteristics were collected from the pa-tient records.

2 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 4

Twelve percent of the controls were treated for hypertensionwith appropriate blood pressure control (i.e. �140/90 mmHg)without evidence of hypertensive organ damage. Four percent ofthe controls were treated for type 2 diabetes mellitus with gly-cosylated hemoglobin levels less than 7% and without evidenceof organ damage.

Inclusion criteria for the current study were age older than 18yr and remission defined by strict biochemical criteria at the timeof study. Patients with present or previous drug or alcohol abuseor with neurological problems, not related to Cushing’s diseaseor NFMA, were excluded. The protocol was approved by theMedical Ethics Committee, and written informed consent wasobtained from all subjects.

Study designA single study visit was planned, during which each subject of

the two patient groups and the two control groups underwentanamnesis and performed the cognitive tests.

Cognitive evaluationEleven cognitive tests were to be completed to assess the full

spectrum of cognitive functioning. A functional classificationwas used to subdivide the tests into the cognitive domains globalcognitive functioning, memory, and executive functioning (21).

To measure global cognitive functioning, the Mini MentalState Examination (MMSE) was used. This is a 30-point ques-tionnaire to assess cognition, with a higher score reflecting betterperformance (22).

Memory was measured with the Wechsler Memory Scale,resulting in a memory quotient (MQ) based on scores in var-ious subscales (23). The Verbal Learning Test of Rey, to mea-sure verbal memory and learning, consists of three trials.Number of correctly recalled words was counted for each trial(24). The Rey Complex Figure, which measures drawing andvisual memory, consists of two trials. A higher score indicatesbetter visual memory (25).

Executive functioning was measured with the Trail MakingTest (26), which measures psychomotor functioning andvisuoconceptual tracking. Time used for both tests and numberof mistakes were counted. The Stroop Color-Word Test (27)measures interference. Number of correct and wrong responseswere counted. The Letter-Digit Substitution Test (28) measuresmental flexibility and speed of information processing. Numberof correctly substituted letters and errors within 60 sec werecounted. The Digit-Deletion Test measures selective attentionand concentration. Number of correctly deleted digits and thenumber of missed digits in 3 min were counted. The Figure Flu-ency Test measures the ability to produce new figures and as-sesses nonverbal mental flexibility and fluency (29). The numberof correct figures, percentage of repeats, and percentage ofwrong figures were counted. The FAS Test employs the letters F,A, and S and measures verbal mental flexibility and fluency (30).The number of correctly produced words and percentage of re-peats and errors were counted. Furthermore, the SynonymsSubtest of the Groninger Intelligence Test-2 was used, with ahigher score indicating better performance (31).

The Hospital Anxiety and Depression Scale (HADS) was usedto measure anxiety and depression. The HADS consists of 14items on a 4-point scale. Both anxiety and depression subscalescores range from 0–21 points. Higher scores indicate more se-vere anxiety and/or depression. A total score higher than 13

points on both subscales together is used to characterize subjectsas anxious or depressed (32, 33).

Statistical analysisData were analyzed using SPSS for Windows version 16.0.2

(SPSS Inc., Chicago, IL). All data are reported in tabular form,expressed as mean � SD. The primary analysis comprised thecomparison of the results between patients cured from Cushing’sdisease and their matched controls and between the patients withNFMA and their matched controls. Groups were compared us-ing a general linear mixed model, with the matched patient-control couples as random factor. Secondary analysis comprisedthe comparison of results in relation to patient and treatmentcharacteristics. To compare patients treated for Cushing’s dis-ease and for NFMA, mean and SD scores for each cognitive testwere calculated for each control group, and subsequently, Z-scores were calculated for each patient group in relation to theirappropriate control group. A general linear model was used tocompare the Z-scores, with postoperative additional radiother-apy, hydrocortisone usage, and hypopituitarism as fixed factors.Independent variables affecting cognitive functioning in patientscured from Cushing’s disease were explored by stepwise linearregression analysis. The standardized �-coefficients of this anal-ysis were reported. To check the appropriateness of assumptionsfor each statistical analysis, we used Levene’s test, Durbin-Watson test, histograms, and scatter plots. All assumptions weremet, except for the independence assumption for parametricdata. We therefore used nonparametric tests to analyze the clin-ical characteristics of patients vs. controls (McNemar test, Fried-man ANOVA, and Wilcoxon signed-ranks test). The level ofsignificance was set at P � 0.05.

Results

Patient characteristics

Patients treated for Cushing’s disease (Table 1)All 74 patients were treated by transsphenoidal sur-

gery, and 20 patients (27%) received additional radio-therapy because of persistent disease after surgery. Themean duration of remission was 13 � 13 yr (range 1–51yr). The number of years in remission was calculated fromthe date of curative transsphenoidal surgery or, in case ofpersistent postoperative disease, from the date of the nor-malization of the biochemical tests after postoperative ra-diotherapy. Any degree of hypopituitarism was present in43 patients (58%), and hydrocortisone replacement wasgiven to 38 patients (51%). There were no differencesbetween patients and controls with respect to age, gender,and education. We asked all patients whether they expe-rienced limitations with respect to memory and/or exec-utive functioning. Sixty-two percent reported memoryproblems, and 47% reported problems in executivefunctioning.

All patients with Cushing’s disease also completed theHADS questionnaire. The mean scores for the depression

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 3

Page 5

subscale were 5.6 � 4.7 and for the anxiety subscale 5.0 �4.7, resulting in total HADS scores of 10.5 � 8.8. This iswell below the cutoff score of 13 (32, 33), which indicatesthat there is, on average, no clinical depression or anxietyin this cohort of Cushing’s disease patients.

Patients treated for nonfunctioning pituitarymacroadenomas (Table 2)

All patients (n � 54) were treated by transsphenoidalsurgery, and 24 patients (44%) received postoperative ra-diotherapy. Fifty patients (93%) required treatment forpituitary insufficiency, and hydrocortisone replacementtherapy was given to 31 patients (57%). There were nodifferences between patients and controls with respect toage, gender, and education. Thirty-nine percent reported

memory problems, and 24% reported problems in exec-utive functioning.

Cognitive function

Patients with Cushing’s disease vs. matched controls(Table 3)

Patients with long-term cure of Cushing’s disease didnot perform worse on measures of global cognitive func-tioning. However, these patients showed a lower MQ onthe Wechsler Memory Scale compared with controls (P �0.015), especially in the subtests concentration (P �0.023), visual memory (P � 0.013), and associative learn-ing (P � 0.023). Furthermore, patients recalled fewerwords than controls in the immediate and delayed recall

TABLE 1. Clinical characteristics of patients cured from Cushing’s disease and matched controls

Cushing’s disease (n � 74) Matched controls (n � 74) P valueGender (male/female) 13/61 13/61 1.00Age (yr), mean � SD 52 � 13 52 � 13 0.26Education (n)

Low 29 28 0.28Average 19 22High 26 24

Transsphenoidal surgery (n) 74 (100%) NA NAPostoperative radiotherapy (n) 20 (27%) NA NADuration of remission (yr), mean � SD 13 � 13 NA NADuration of follow-up (yr), mean � SD 16 � 12 NA NAHypopituitarism (n)

Any axis 43 (58%) NA NAGH 26 (35% NA NALH/FSH 19 (26%) NA NATSH 24 (32%) NA NAADH 11 (15%) NA NA

Hydrocortisone substitution (n) 38 (51%) NA NA

NA, Not applicable.

TABLE 2. Clinical characteristics of the patients treated for nonfunctioning pituitary macroadenomas and matchedcontrols

NFMA (n � 54) Matched controls (n � 54) P valueGender (male/female) 30/24 30/24 1.00Age (yr), mean � SD 61 � 11 59 � 11 0.06Education (n)

Low 18 21 0.90Average 21 15High 15 18

Transsphenoidal surgery (n) 54 (100%) NA NAPostoperative radiotherapy (%) 24 (44%) NA NADuration of follow up (yr), mean � SD 15 � 12 NA NAHypopituitarism (n)

Any axis 50 (93%) NA NAGH 40 (74%) NA NALH/FSH 32 (59%) NA NATSH 33 (61%) NA NAADH 6 (11%) NA NA

Hydrocortisone substitution (%) 31 (57%) NA NA

NA, Not applicable.

4 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 6

trials of the Verbal Learning Test of Rey (P � 0.001 onboth trials). In accordance, patients scored lower thancontrols in the delayed trial of the Rey Complex Figure(P � 0.040).

In tests assessing the executive functioning domain, theLetter-Digit Substitution Test showed that patients sub-stituted fewer letters than controls (P � 0.026). Further-more, patients deleted fewer digits (P � 0.035) on theDigit-Deletion Test and produced more repeated patternson the Figure Fluency Test (P � 0.045) when comparedwith controls.

When patients with short-term (�10 yr, mean 4 � 2 yr,range 1–8 yr) and long-term (�10 yr, mean 24 � 11 yr,

range 11–51 yr) remission were compared, only a singletest result was significantly different between these twogroups. Patients with short-term remission had a higherpercentage of errors on the FAS than those in the long-termremission group (3.1 vs. 0.9%, P � 0.012).

Patients treated for NFMA vs. matched controls(Table 4)

Patients treated for NFMA did not perform worse onmeasures of global cognitive functioning. In tests assessingthe memory domain, there were some differences betweenpatients and controls. Patients scored lower on the subtest

TABLE 3. Cognitive outcomes: patients cured from Cushing’s disease vs. matched controls

Cushing’s disease (n � 74) Matched controls (n � 74) P valueGlobal cognitive function

MMSE score 27.9 (1.9) 28.3 (2.0) 0.173Memory

Wechsler Memory ScaleMQ 109.0 (16.8) 115.6 (15.6) 0.015Information 5.8 (0.4) 5.9 (0.4) 0.677Orientation 4.9 (0.2) 5.0 (0.2) 0.701Concentration 7.0 (2.0) 7.7 (1.4) 0.023Logical memory 6.3 (3.2) 7.1 (3.2) 0.118Digit span 9.9 (1.9) 10.3 (1.8) 0.231Visual memory 8.1 (3.0) 9.2 (3.4) 0.013Associative learning 16.0 (3.4) 17.2 (2.8) 0.023

Verbal Learning Test of ReyImprinting, total 5.8 (2.1) 6.3 (2.2) 0.154Immediate, total 9.4 (2.7) 11.0 (2.3) 0.000Delayed, total 7.5 (3.0) 9.4 (3.2) 0.000

Rey Complex Figure TestImmediate 17.2 (6.0) 18.9 (6.7) 0.063Delayed 16.7 (6.3) 18.6 (6.8) 0.040

Executive functioningTrail Making Test

Trail A, time 0.4 (0.3) 0.4 (0.4) 0.889Trail A, errors 0.1 (0.3) 0.2 (0.4) 0.111Trail B, time 1.3 (1.3) 1.2 (0.9) 0.415Trail B, errors 0.7 (1.7) 0.7 (2.2) 0.911

Stroop Color-Word TestInterference, total 39.8 (11.0) 42.0 (10.6) 0.220Interference, mistakes 0.3 (0.8) 0.2 (0.5) 0.297

Letter-Digit Substitution TestNo. correct 31.6 (8.0) 34.2 (7.9) 0.026No. errors 0.1 (0.2) 0.1 (0.3) 0.555

Digit-Deletion TestNo. correct 376.5 (102.2) 409.7 (91.3) 0.035No. missed 5.0 (5.2) 4.3 (4.8) 0.385

Figure Fluency TestNo. patterns 62.0 (23.4) 66.9 (22.7) 0.164% Repeats 9.0 (11.3) 6.2 (5.7) 0.045% Errors 17.2 (12.3) 16.9 (13.1) 0.866

FASNo. correct 33.1 (14.8) 36.2 (13.3) 0.168% Repeats 1.8 (2.8) 1.2 (2.5) 0.131% Errors 2.1 (3.8) 1.5 (4.8) 0.391

Synonyms subtest of the GroningerIntelligence test

4.5 (1.9) 4.5 (1.8) 0.134

Data are mean (SD).

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 5

Page 7

associative learning of the Wechsler Memory Scale (P �0.032) when compared with controls.

In tests assessing executive functioning, there was a dif-ference between patients and controls on the Trail MakingTest. Patients needed more time on Trail A and B and mademore errors on Trail A when compared with controls (P �0.001, P � 0.035, and P � 0.019, respectively). Further-more, patients had a lower total score on the Stroop Color-Word Test (P � 0.045).

When patients with short-term (�10 yr) and long-term(�10 yr) duration of follow-up were compared, patientswith long-term follow-up scored worse on the GroningerIntelligence Test (P � 0.003) and made more errors on thefirst trail of the Trail Making Test (P � 0.001).

Comparison of Z-scores between patients curedfrom Cushing’s disease and patients treated fornonfunctioning pituitary macroadenomas (Table 5)

Patients cured from Cushing’s disease performed worseon the MMSE, which measures global cognitive function-ing, compared with patients treated for NFMA (P �0.001). The observed difference between the two patientgroups in the MMSE is most likely clinically not very rel-evant. Apparently, patients with long-term cure of Cush-ing’s disease do not suffer from impaired global cognitivefunctioning, because there were no differences comparedwith their matched controls.

In the memory domain, patients cured from Cushing’sdisease had a lower MQ measured with the Wechsler

TABLE 4. Cognitive outcomes: patients cured from NFMA vs. matched controls

NFMA (n � 54) Matched controls (n � 54) P valueGlobal cognitive function

MMSE score 28.9 (1.1) 28.4 (1.4) 0.053Memory

Wechsler Memory ScaleMQ 118.2 (16.9) 118.1 (13.9) 0.965Information 5.9 (0.3) 5.9 (0.3) 0.693Orientation 4.9 (0.2) 5.0 (0.2) 0.651Concentration 7.6 (1.8) 7.3 (1.8) 0.526Logical memory 7.4 (3.3) 7.5 (2.6) 0.884Digit span 10.0 (1.6) 10.0 (1.9) 0.869Visual memory 8.9 (3.5) 8.6 (3.1) 0.618Associative learning 15.7 (3.0) 16.9 (2.6) 0.032

Verbal Learning Test of ReyImprinting, total 5.2 (1.9) 5.0 (1.9) 0.550Immediate, total 9.2 (2.9) 9.8 (2.2) 0.163Delayed, total 6.9 (3.4) 7.5 (2.7) 0.278

Rey Complex Figure TestImmediate 19.6 (6.6) 19.4 (5.8) 0.850Delayed 19.2 (6.4) 19.1 (5.8) 0.911

Executive functioningTrail Making Test

Trail A, time 0.6 (0.39) 0.4 (0.18) 0.001Trail A, errors 0.3 (0.5) 0.1 (0.3) 0.019Trail B, time 1.4 (0.69) 1.2 (0.6) 0.035Trail B, errors 0.5 (0.9) 0.4 (0.9) 0.724

Stroop Color-Word TestInterference, total 36.2 (9.1) 38.9 (8.8) 0.045Interference, mistakes 0.1 (0.4) 0.1 (0.4) 1.00

Letter-Digit Substitution TestNo. correct 31.2 (8.5) 32.7 (6.7) 0.230No. errors 0.0 (0.2) 0.1 (0.4) 0.309

Digit-Deletion TestNo. correct 366.3 (88.7) 389.2 (76.9) 0.148No. missed 3.9 (4.3) 3.7 (4.1) 0.771

Figure Fluency TestNo. patterns 47.9 (22.9) 53.0 (22.1) 0.597% Repeats 9.2 (9.8) 8.5 (9.3) 0.840% Errors 19.8 (11.4) 20.5 (16.3) 0.780

FASNo. correct 33.6 (13.4) 34.6 (11.7) 0.680% Repeats 1.1 (2.0) 1.9 (2.6) 0.082% Errors 2.3 (6.0) 2.2 (3.3) 0.919

Synonyms subtest of the GroningerIntelligence Test

5.0 (1.9) 5.0 (1.8) 0.958

Data are mean (SD).

6 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 8

Memory Scale compared with patients with NFMA (P �0.050) in the subscales concentration (P � 0.017) andvisual memory (P � 0.006). On the Verbal Learning Testof Rey, patients cured from Cushing’s disease recalledfewer words in the imprinting (P � 0.013), the immediaterecall (P � 0.012), and the delayed recall trials (P � 0.003)compared with NFMA patients. Furthermore, on the ReyComplex Figure, patients with cured Cushing’s diseasescored worse on both trials (P � 0.002 and P � 0.007,respectively) when compared with NFMA patients.

In tests measuring executive function, patients curedfrom Cushing’s disease made fewer correct substitutions

on the Letter-Digit Substitution Test (P � 0.039) and cameup with fewer correct patterns on the Figure Fluency Test(P � 0.003) compared with treated NFMA patients.

Factors associated with cognitive function inpatients with Cushing’s disease

As expected, age and educational level were associatedwith the outcomes of almost all cognitive tests, whereasgender was not. Potential factors of influence, includinghypopituitarism, hydrocortisone dependency, durationof remission, and additional radiotherapy, were addedin the stepwise linear regression model with adjust-

TABLE 5. Cognitive function: comparison between patients with Cushing’s disease and patients with NFMA by Z-scores, calculated for each patient group by comparison with their own matched control groups

Z-scores Cushing’s disease (n � 74) Z-scores NFMA (n � 54) P valueGlobal

MMSE score �0.21 (�0.4 to 0.0) 0.34 (0.1–0.5) 0.001Memory

Wechsler Memory ScaleMQ �0.42 (�0.7 to �0.2) 0.01 (�0.3 to 0.3) 0.050Information �0.07 (�0.3 to 0.2) 0.07 (�0.2 to 0.4) 0.109Orientation �0.07 (�0.3 to 0.2) �0.10 (�0.4 to 0.2) 0.960Concentration �0.47 (�0.8 to �0.1) 0.12 (�0.2 to 0.4) 0.017Logical memory �0.26 (�0.5 to 0.0) �0.03 (�0.4 to 0.3) 0.073Digit span �0.21 (�0.5 to 0.1) 0.03 (�0.2 to 0.3) 0.230Visual memory �0.31 (�0.5 to �0.1) 0.09 (�0.2 to 0.4) 0.006Associative learning �0.41 (�0.7 to �0.1) �0.43 (�0.7 to �0.1) 0.293

Verbal Learning Test of ReyImprinting, total �0.23 (�0.5 to 0.0) 0.11 (�0.2 to 0.4) 0.013Immediate, total �0.70 (�1.0 to �0.4) �0.27 (�0.6 to 0.1) 0.012Delayed, total �0.60 (�0.8 to �0.4) �0.21 (�0.6 to 0.1) 0.003

Rey Complex Figure TestImmediate �0.25 (�0.5 to 0.0) 0.03 (�0.3 to 0.3) 0.002Delayed �0.27 (�0.5 to �0.1) 0.02 (�0.3 to 0.3) 0.007

Executive functionTrail Making Test

Trail A, time 0.02 (�0.2 to 0.2) 1.16 (0.6–1.8) 0.081Trail A, errors �0.21 (�0.4 to �0.1) 0.49 (0.1 to 0.9) 0.167Trail B, time 0.14 (�0.2 to 0.5) 0.36 (0.0–0.7) 0.939Trail B, errors 0.02 (�0.2 to 0.2) 0.06 (�0.2 to 0.3) 0.480

Stroop Color-Word TestInterference, total �0.21 (�0.4 to 0.0) �0.31 (�0.6 to 0.0) 0.823Interference, mistakes 0.21 (�0.1 to 0.6) 0.00 �0.3 to 0.2) 0.270

Letter-Digit Substitution TestNo. correct �0.33 (�0.6 to �0.1) �0.22 (�0.6 to 0.1) 0.039No. errors �0.08 (�0.3 to 0.1) �0.15 (�0.3 to 0.0) 0.722

Digit-Deletion TestNo. correct �0.37 (�0.6 to �0.1) �0.30 (�0.6 to 0.0) 0.359No. missed 0.15 (�0.1 to 0.4) 0.05 (�0.2 to 0.3) 0.053

Figure Fluency TestNo. patterns �0.22 (�0.5 to 0.0) �0.10 (�0.4 to 0.2) 0.003% Repeats 0.49 (0.0–0.1) 0.04 (�0.2 to 0.3) 0.215% Errors 0.03 (�0.2 to 0.2) �0.04 (�0.2 to 0.2) 0.757

FASNo. correct �0.24 (�0.5 to 0.0) �0.09 (�0.4 to 0.2) 0.423% Repeats 0.26 (�0.1 to 0.3) 0.03 (�0.5 to 0.5) 0.067% Errors 0.13 (0.0 � 0.3) 0.03 (�0.5 to 0.5) 0.735

Synonyms subtest of the GroningerIntelligence Test

�0.24 (�0.5 to 0.0) 0.01 (�0.3 to 0.3) 0.215

Data are Z-scores mean (95% confidence interval).

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 7

Page 9

ments for age and education. We calculated regressioncoefficients for test outcomes that were associated withduration of remission, which might indicate the poten-tial for improvement.

Global cognitive functioning was not associated withany of the variables. In the memory domain, the WechslerMemory Scale MQ was positively associated with dura-tion of remission (� � 0.276; P � 0.017). In the executivefunction domain, the number of missed digits on the Digit-Deletion Test was positively associated with duration ofremission (� � 0.245; P � 0.041) and additional radio-therapy (� � 0.361; P � 0.002). Furthermore, the numberof correct patterns in the Figure Fluency Test was nega-tively associated with hypopituitarism (� � �0.278; P �0.012) and hydrocortisone dependency (� � �0.230; P �0.040). The percentage of mistakes in the Figure FluencyTest was positively associated with hydrocortisone depen-dency (� � 0.224; P � 0.048). The percentage of mistakeson the FAS Test was inversely associated with duration ofremission (� � �0.254; P � 0.034).

There was a significant correlation between the out-come on the Wechsler Memory Scale (MQ) and dura-tion of remission (r � 0.236; P � 0.049). There was alsoa significant correlation between the number of misseddigits on the Digit-Deletion Test and duration of remis-sion (r � 0.245; P � 0.041), and the percentage ofmistakes on the FAS and duration of remission (r ��0.254; P � 0.034).

Discussion

This study demonstrates that cognitive function is im-paired in patients despite long-term cure of Cushing’s dis-ease. These patients reported impairments in memory indaily life, which was confirmed by cognitive functioningtests. The performance was decreased in certain aspects ofexecutive functioning and several memory tasks com-pared with matched controls. These impairments were notmerely related to pituitary disease in general, because thesepatients with long-term cure of Cushing’s disease also re-vealed impaired cognitive function compared with pa-tients previously treated for NFMA. These observationsindicate irreversible effects of previous hypercortisolismon cognitive function and, thus, on the central nervoussystem.

The outcomes of the cognitive tests are in general af-fected by many factors, including age, gender, and edu-cational level. Because the controls and patients were per-fectly matched, these potentially confounding factors didnot influence our results or conclusions. We do not thinkthat our results can be explained to a large extend by thedifference in gender distribution between both patient

groups. First, patients with long-term cure of Cushing’sdisease had impaired cognition compared with gender-matched controls. Second, we used Z-scores derived fromthe comparisons between patients and appropriatelymatched controls to compare the patients with curedCushing’s disease with patients treated for NFMA, be-cause the gender differences were too large between thesetwo patients groups to justify a direct comparison.

Several clinical characteristics influenced outcomeparameters. Hypopituitarism was associated with mildlyimpaired executive functioning. Hydrocortisone depen-dency and additional radiotherapy were negativelyassociated with memory and executive functioning,whereas the duration of remission positively influencedmemory and executive functioning. These findings do notinvalidate our conclusions, because these factors were alsopresent in patients treated for NFMA, who in general hadbetter performances compared with the patients curedfrom Cushing’s disease.

Table 6 summarizes all studies on the effect of Cush-ing’s disease and syndrome on cognitive functions, includ-ing the effects of treatment. Our observations extend thoseof previous studies. Four previous studies studied cogni-tive functioning in treated Cushing’s disease patients, witha total of 98 patients and 77 controls. In the first study,patients with treated Cushing’s disease (n � 27) showedimprovement of verbal fluency and recall within 18months of follow-up, whereas brief attention did notchange. This indicates that some but not all of the effectsof previous glucocorticoid excess are reversible (20). Thesecond study showed that there were no differences be-tween patients (n � 33) and matched controls in IQ duringactive disease and 12 months after treatment. There was,however, a positive relation for some subscales of the IQtest and recovery of the hypothalamic-pituitary-adrenalaxis. There was also a negative association between someIQ subscales and duration of disease (18). The third studyshowed that 1 yr after surgical treatment, high levels ofcortisol caused long-lasting impairments in attention,visuospatial processing, memory, reasoning, and verbalfluency in patients with Cushing’s syndrome (n � 13) (19).Furthermore, the last study observed that patients withCushing’s disease (n � 25) showed selective impairmentsin memory functions. After treatment, the eight patientswho were retested showed amelioration of these memoryimpairments (6). Our study indicates that patients withlong-term cure of Cushing’s disease have impaired scoresof memory and to a lesser extent in executive functionscompared with both matched controls and treated NFMApatients. Our study differs in several respects from theprevious studies. First, the number of patients included inour study was relatively large compared with the previous

8 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 10

studies. Second, the duration of cure was very long in ourstudy compared with previous studies. Third, we com-pared the patients with long-term cure of Cushing’s dis-ease both with matched controls and with patients previ-ously operated for NFMA. From the studies summarizedin Table 6, including our present study, the notion emergesthat active Cushing’s disease is associated with cognitiveimpairment and that treatment of Cushing’s disease re-sults in some but not complete recovery of cognitiveimpairment.

Several other studies evaluated the effects of pituitaryadenomas, including ACTH-producing adenomas, onexecutive functioning and memory but did not specifythe differences between different pituitary adenomas(34–36). Therefore, these studies do not permit any con-clusion with respect to the specific effects of Cushing’sdisease compared with the effects of other pituitary ade-nomas on cognitive function.

Prolonged glucocorticoid excess modifies neurotrans-mitter function and neuronal structure of the central ner-vous system (7, 11). In rodents, chronic exposure to highlevels of glucocorticoids impairs hippocampal long-termpotentiation (12) and decreases hippocampal synapticplasticity (13). In humans, endogenous active Cushing’sdisease is associated with cognitive impairment (6, 7, 20).The hippocampus is one of the most sensitive structures inthe brain for glucocorticoids and is crucial in cognitivefunction (37). The persistent impairments in cognitivefunction in patients with previous Cushing’s disease mightbe explained by irreversible effects of previous glucocor-ticoid excess on the central nervous system, especially thehippocampus. Additional studies, including functionalmagnetic resonance imaging and postmortem analyses ofthe central nervous system, are required to evaluate theeffects of previous glucocorticoid excess on brain areas ofinterest.

Patients with long-term cure of Cushing’s disease are aunique, monofactorial model to study the long-term ef-fects of glucocorticoid exposure. The results of the currentstudy may also apply to patients previously treated withhigh-dose glucocorticoids for nonendocrine diseases. Inaddition, the results might also be of relevance for patientswith chronically increased glucocorticoid levels in condi-tions like depression (38, 39).

In the review process of the manuscript, there was con-cern with respect to the presentation of the data withoutadjustments for multiple comparisons. Simply defined,these adjustments test for no effects in all the primaryendpoints undertaken vs. an effect in one or more of thoseendpoints. This is a difficult methodological issue becausethere are divergent views on the need for statistical ad-justment for multiplicity. This is also reflected in the

Lancet papers by Schulz and Grimes (40, 41), who advo-cate a restrictive approach toward adjustments for mul-tiple comparisons. If we consider our own data and if weassume that the differences would mostly reflect false-pos-itive results, it is to be expected that the positive significantresults would have been randomly distributed among thedifferent variables. However, this is not the case, as shownin Tables 3 and 4. Moreover, there are several argumentsthat cortisol excess can indeed cause irreversible effects onthe central nervous system (see above). We designed thisstudy inourpatients cured fromCushing’sdiseasewith theprimary aim to evaluate cognitive function in detail, inview of the documented abnormalities in previous studiesand those observed in experimental animal studies. In-deed, the main results of our study point toward similaradverse effects of previous Cushing’s disease documentedin previous studies, although these had a different studydesign. According to Schulz and Grimes (40, 41), statis-tical adjustments somewhat rescue the positive results ofscattershot analyses. However, we performed a targetedevaluation and analysis focused on cognitive function re-lated to previous Cushing’s disease rather than a scatter-shot analysis of cognitive functions in general. Therefore,in our opinion, our data should not be neglected merelybecause of the absence of adjustments for multiple com-parisons. Moreover, this would carry the serious risk ofmissing an important association between previous Cush-ing’s disease and cognitive impairments.

A limitation of the present study was the cross-sectionalstudy design. Consequently, we do not have any informa-tion on premorbid functions, the effects of active Cush-ing’s disease, and the extent of reversibility of the dis-turbed parameters. Nonetheless, these limitations do notinvalidate our observations that patients with long-termcure have subtle impairments in cognitive function com-pared with matched controls and with patients treatedsimilarly for NFMA. It might be argued that potential biasmay have been introduced by the selection of the controlsby the patients. In previous studies, we used similarly se-lected controls and compared the responses of thesematched controls with those obtained from publishedDutch control populations for several questionnaires (in-cluding HADS, Nottingham Health Profile, Multidimen-sional Fatigue Index, and Short Form) (3, 42, 43). In gen-eral, the conclusions obtained in the matched controlsubjects were in agreement with the literature-based ref-erence data. In the present study, the self-selection ofcontrols enabled a perfect match for an additional pa-rameter, i.e. socioeconomic status, an important deter-minant of the outcomesof thequestionnaires, inaddition toage, gender, and education. Moreover, we used the samemethod of selection of controls for both groups of patients.

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 9

Page 11

TAB

LE6.

Ove

rvie

wof

stud

ies

onco

gniti

vefu

nctio

nin

patie

nts

with

Cus

hing

’sdi

seas

e

Au

tho

r,ye

arN

um

ber

of

sub

ject

sG

end

er(m

/f)

Ag

e(y

r),

mea

n�

SDA

ctiv

e/tr

eate

dM

eth

od

sO

utc

om

esW

hela

net

al.,

1980

(4)

35C

ushi

ng’s

synd

rom

e7/

2835

�N

AA

ctiv

eM

ichi

gan

Neu

rops

ycho

logi

cal

Test

batt

ery

13pa

tient

sha

dno

sign

sof

neur

opsy

chol

ogic

alde

ficits

,10

patie

nts

had

few

and

mild

defic

its,8

had

mod

erat

ean

dm

ore

freq

uent

sign

sof

impa

irmen

t,an

d4

had

freq

uent

and

mar

ked

defic

its.T

hese

impa

irmen

tsw

ere

mor

efr

eque

ntan

dse

vere

inno

nver

balv

isua

l-ide

atio

nal

and

visu

alm

emor

yfu

nctio

ns.

Gra

ttan

-Sm

ithet

al.,

1992

(34)

10C

ushi

ng’s

synd

rom

e,27

chro

mop

hobe

A,1

5pr

olac

tinom

a,13

acro

meg

aly,

21di

sfig

ured

ordi

sabl

edin

patie

nts

NA

NA

Trea

ted,

dura

tion

ofre

mis

sion

not

give

n

Rey-

Ost

errie

thC

ompl

exFi

gure

,Con

trol

led

Ora

lW

ord

Ass

ocia

tion

Test

,Tr

ailM

akin

gTe

st,W

echs

ler

Mem

ory

Scal

e,W

arrin

gton

Reco

gniti

onM

emor

yTe

stFa

ces,

New

Adu

ltRe

adin

gTe

st

Not

spec

ified

for

Cus

hing

’sdi

seas

e/sy

ndro

me.

Ove

rall,

patie

nts

with

pitu

itary

aden

omas

expe

rienc

edim

pairm

ent

ofm

emor

yan

dex

ecut

ive

func

tion,

whi

chw

asno

tre

late

dto

size

orty

peof

tum

oror

the

effe

cts

oftr

eatm

ent.

Star

kman

etal

.,19

92(1

1)

12C

ushi

ng’s

synd

rom

e,1

norm

alsu

bjec

t

2/10

37�

14Tr

eate

d,du

ratio

nof

rem

issi

onno

tgi

ven

Mag

netic

reso

nanc

eim

agin

g,W

echs

ler

Mem

ory

Scal

e-Ru

ssel

lMod

ifica

tion,

Trai

lM

akin

gTe

st

Hip

poca

mpa

lfor

mat

ion

volu

me

fell

outs

ide

the

95%

conf

iden

cein

terv

alfo

r27

%of

the

patie

nts.

Furt

herm

ore,

ther

ew

asan

asso

ciat

ion

betw

een

redu

ced

hipp

ocam

palf

orm

atio

nvo

lum

ean

dlo

wer

scor

eson

verb

alle

arni

ngan

dm

emor

yte

sts.

Mar

tigno

niet

al.,

1992

(5)

24C

ushi

ng’s

dise

ase,

24m

atch

edco

ntro

ls16

/8N

AA

ctiv

edi

seas

e,7

patie

nts

wer

ere

test

ed6

mon

ths

afte

rtr

eatm

ent

Att

entio

n,m

emor

y,la

ngua

ge,v

isuo

-spa

tial,

and

logi

cala

bilit

ies

Patie

nts

with

activ

edi

seas

esh

owed

impa

irmen

tin

verb

alan

dno

nver

bal

epis

odic

mem

ory.

The

7re

test

edpa

tient

ssh

owed

asi

gnifi

cant

reco

very

ofve

rbal

mem

ory.

(Con

tinue

d)

10 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 12

TAB

LE6.

Con

tinue

d

Au

tho

r,ye

arN

um

ber

of

sub

ject

sG

end

er(m

/f)

Ag

e(y

r),

mea

n�

SDA

ctiv

e/tr

eate

dM

eth

od

sO

utc

om

esM

auri

etal

.,19

93(6

)25

Cus

hing

’sdi

seas

e,60

norm

alsu

bjec

ts(in

patie

nts)

8/17

36�

14A

ctiv

e,8

patie

nts

wer

ere

test

ed6

mon

ths

afte

rtr

eatm

ent

Logi

cM

emor

y,Se

rial

Lear

ning

Test

,Dig

itSp

an,

Vis

ualR

epro

duct

ion,

Rave

n’s

Col

ored

Prog

ress

ive

Mat

rices

,Dig

itSy

mbo

lSub

stitu

tion

Test

,Si

mila

ritie

s,C

ance

llatio

nTa

sk,T

rail

Mak

ing

Test

,W

ord

Flue

ncy,

Stre

et’s

Com

plet

ion

Test

Patie

nts

with

Cus

hing

’sdi

seas

esh

owed

sele

ctiv

eim

pairm

ent

ofm

emor

yfu

nctio

ns.T

here

wer

eno

defic

itsin

othe

rco

gniti

vefu

nctio

ns.T

he8

rete

sted

patie

nts

show

edsi

gnifi

cant

amel

iora

tion

ofm

emor

yde

ficits

.

Peac

eet

al.,

1997

(35)

36pi

tuita

rytu

mor

patie

nts,

36co

ntro

ls13

/23

NA

27tr

eate

d(m

ean

dura

tion

sinc

esu

rger

y9

�6

yr),

9un

trea

ted

Nat

iona

lAdu

ltRe

adin

gTe

st,

Dig

itSp

an,A

udito

ryV

erba

lLe

arni

ngTe

st,S

tory

Reca

ll,Re

cogn

ition

Mem

ory

Test

for

Face

s,St

roop

Neu

rops

ycho

logi

cal

Scre

enin

gTe

st,C

ontr

olle

dO

ralW

ord

Ass

ocia

tion

Test

,Blo

ckD

esig

n,Tr

ail

Mak

ing

Test

Not

spec

ified

for

Cus

hing

’sdi

seas

e/sy

ndro

me.

Patie

nts

show

edsi

gnifi

cant

impa

irmen

tsin

exec

utiv

efu

nctio

ning

and

som

ewha

tim

pairm

ent

inm

emor

y.Ra

diot

hera

pydi

dno

tap

pear

tobe

asso

ciat

edw

ithco

gniti

vefu

nctio

ning

.

Peac

eet

al.,

1998

(36)

10C

ushi

ng’s

dise

ase,

22N

FMA

,21

prol

actin

oma,

9ac

rom

egal

y,7

cran

ioph

aryn

giom

a,23

heal

thy

cont

rols

NA

NA

Trea

ted,

dura

tion

ofre

mis

sion

not

give

n

Nat

iona

lAdu

ltRe

adin

gTe

st,

Dig

itSp

an,A

udito

ryV

erba

lLe

arni

ngTe

stof

Rey,

Stor

yRe

call,

Reco

gniti

onM

emor

yTe

stFa

ces,

Con

trol

led

Ora

lWor

dA

ssoc

iatio

n,Bl

ock

Des

ign,

Trai

lMak

ing

Test

Not

spec

ified

for

Cus

hing

’sdi

seas

e/sy

ndro

me.

Surg

ical

lytr

eate

dpa

tient

ssu

ffer

edfr

omim

pairm

ents

inm

emor

yan

dex

ecut

ive

func

tions

.Th

ere

wer

eno

sign

ifica

ntne

gativ

eef

fect

sas

soci

ated

with

radi

othe

rapy

.

(Con

tinue

d)

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 11

Page 13

TAB

LE6.

Con

tinue

d

Au

tho

r,ye

arN

um

ber

of

sub

ject

sG

end

er(m

/f)

Ag

e(y

r),

mea

n�

SDA

ctiv

e/tr

eate

dM

eth

od

sO

utc

om

esFo

rget

etal

.,20

00(7

)19

Cus

hing

’sdi

seas

e,19

mat

ched

cont

rols

18/1

47�

11A

ctiv

eV

isua

lTar

get

Det

ectio

n,St

roop

Test

,Dig

itSy

mbo

lSu

bstit

utio

nTe

st,T

rail

Mak

ing

Test

,Jud

gem

ent

ofLi

neO

rient

atio

n,Be

llsTe

st,

Hoo

per

Vis

ual

Org

aniz

atio

n,G

ollin

Figu

res,

Bloc

kD

esig

n,O

bjec

tA

ssem

bly,

Vis

ual

Repr

oduc

tion-

Cop

y,C

alifo

rnia

Ver

balL

earn

ing

Test

,Log

ical

Mem

ory,

Dig

itSp

an,V

isua

lMem

ory,

Bent

onV

isua

lRet

entio

nTe

st,C

ompr

ehen

sion

,Pi

ctur

eC

ompl

etio

n,Pi

ctur

eA

rran

gem

ent,

Arit

hmet

ic,

Sim

ilarit

ies,

Rave

n’s

Stan

dard

Mat

rices

,V

ocab

ular

yLe

tter

Cat

egor

yFl

uenc

y

Patie

nts

with

Cus

hing

’sdi

seas

esc

ored

wor

seon

alls

tand

ardi

zed

neur

opsy

chol

ogic

alte

sts,

com

pare

dw

ithth

eirm

atch

edco

ntro

ls,pa

rtic

ular

lyin

proc

esse

sin

volv

ing

sele

ctiv

eat

tent

ion

and

visu

alco

mpo

nent

s.IQ

scor

esof

patie

nts

wer

ew

ithin

the

norm

alra

nge.

Dor

nan

dC

erro

ne,

2000

(18)

33C

ushi

ng’s

synd

rom

e,17

mat

ched

cont

rols

5/28

36�

9A

ctiv

ean

d12

mon

ths

afte

rtr

eatm

ent

Wec

hsle

rA

dult

Inte

llige

nce

Scal

e-Re

vise

d,Pr

ofile

ofM

ood

Stat

es,S

ympt

omC

heck

list

90-R

evis

ed

Ther

ew

ere

nogr

oup

diff

eren

ces

inco

gniti

vefu

nctio

nac

ross

time.

Ther

ew

as,h

owev

er,a

tren

dfo

rpa

tient

sw

ithC

ushi

ng’s

synd

rom

eto

have

low

erIQ

scor

esat

base

line.

For

som

eIQ

subs

cale

s,th

ere

was

apo

sitiv

ere

latio

nw

ithre

cove

ryof

the

HPA

axis

and

ane

gativ

eas

soci

atio

nw

ithdu

ratio

nof

dise

ase.

(Con

tinue

d)

12 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 14

TAB

LE6.

Con

tinue

d

Au

tho

r,ye

arN

um

ber

of

sub

ject

sG

end

er(m

/f)

Ag

e(y

r),

mea

n�

SDA

ctiv

e/tr

eate

dM

eth

od

sO

utc

om

esSt

arkm

anet

al.,

2001

(8)

48C

ushi

ng’s

dise

ase,

38he

alth

yco

ntro

ls11

/37

37�

14A

ctiv

eW

echs

ler

Adu

ltIn

telli

genc

eSc

ale-

Revi

sed,

Wec

hsle

rM

emor

ySc

ale,

Sem

istr

uctu

red

Clin

ical

Inte

rvie

w,H

amilt

onD

epre

ssio

nRa

ting

Scal

e

Patie

nts

with

Cus

hing

’sdi

seas

esc

ored

wor

seon

4of

the

5ve

rbal

IQsu

btes

ts,

and

one

nonv

erba

lper

form

ance

IQsu

btes

tcom

pare

dw

ithco

ntro

ls.V

erba

lle

arni

ngan

dde

laye

dre

call

wer

esig

nific

antly

decr

ease

din

patie

nts.

Forg

etet

al.,

2002

(19)

13C

ushi

ng’s

synd

rom

e2/

1147

�13

1yr

afte

rsu

rgic

altr

eatm

ent

See

Forg

etet

al.,

2000

(7)

Com

pare

dw

ithob

serv

atio

nsbe

fore

trea

tmen

t(7

),th

ere

wer

eno

impr

ovem

ents

inco

gniti

vefu

nctio

ning

inpa

tient

sw

ithC

ushi

ng’s

synd

rom

e.St

arkm

anet

al.,

2003

(17)

24C

ushi

ng’s

dise

ase

4/20

34�

13A

ctiv

ean

d16

�9

mon

ths

afte

rsu

rger

y

Mag

netic

reso

nanc

eim

agin

g,W

echs

ler

Mem

ory

Scal

e-Ru

ssel

lMod

ifica

tion,

Sele

ctiv

eRe

min

ding

Test

,D

elay

edM

emor

y(p

arag

raph

san

dpa

ired

wor

ds),

Voc

abul

ary,

Arit

hmet

ic,S

ympt

oms

Che

cklis

t-90

All

patie

nts

show

edan

incr

ease

inhi

ppoc

ampa

lfor

mat

ion

volu

me

afte

rtr

eatm

ent,

and

18pa

tient

ssh

owed

anin

crea

sein

caud

ate

head

volu

me,

and

52–

61%

ofth

epa

tient

ssh

owed

anin

crea

sein

lear

ning

scor

e,w

here

as61

%of

the

patie

nts

show

edan

incr

ease

inm

emor

ysc

ore.

For

voca

bula

ry,2

5%ha

da

bett

ersc

ore,

and

50%

had

anin

crea

sed

scor

efo

rar

ithm

etic

.H

ook

etal

.,20

07(2

0)27

Cus

hing

’sdi

seas

e4/

2339

�13

Act

ive,

and

3–5,

6–1

2,13

–18

mon

ths

afte

rtr

eatm

ent

Busc

hke

Sele

ctiv

eRe

min

ding

Test

,Dig

itSp

an,V

erba

lFl

uenc

y,Sy

mpt

omC

heck

list

90-R

evis

ed,

mag

netic

reso

nanc

eim

agin

g

Old

eran

dyo

unge

rpa

tient

ssh

owed

com

para

ble

leve

lsof

cogn

itive

dysf

unct

ion

befo

retr

eatm

ent,

with

youn

ger

patie

nts

show

ing

am

ore

rapi

dim

prov

emen

t.V

erba

lflu

ency

and

reca

llsh

owed

reco

very

,but

brie

fat

tent

ion

did

not.

The

impr

ovem

ent

inve

rbal

reca

llw

asas

soci

ated

with

ade

crea

sein

cort

isol

leve

lsan

dan

incr

ease

inhi

ppoc

ampa

lfor

mat

ion

volu

me

1yr

afte

rtr

eatm

ent. (C

ontin

ued)

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 13

Page 15

TAB

LE6.

Con

tinue

d

Au

tho

r,ye

arN

um

ber

of

sub

ject

sG

end

er(m

/f)

Ag

e(y

r),

mea

n�

SDA

ctiv

e/tr

eate

dM

eth

od

sO

utc

om

esM

icha

udet

al.,

2009

(9)

10C

ushi

ng’s

synd

rom

e,10

age-

mat

ched

heal

thy

cont

rols

,10

olde

rco

ntro

ls

2/8

44�

7A

ctiv

eSe

eFo

rget

,200

0(7

)Th

eag

e-m

atch

edco

ntro

lgro

uppe

rfor

med

bett

erth

anth

epa

tient

sw

ithC

ushi

ng’s

synd

rom

ean

dol

der

cont

rols

on9

of23

test

s.Th

eol

der

cont

rols

and

patie

nts

perf

orm

edsi

mila

rlyon

thes

ete

sts.

This

sugg

ests

that

hype

rsec

retio

nof

gluc

ocor

ticoi

dsha

san

agin

g-lik

eef

fect

onco

gniti

vefu

nctio

ning

inpa

tient

sw

ithC

ushi

ng’s

synd

rom

e.Le

ón-C

arrió

net

al.,

2009

(10)

15C

ushi

ng’s

synd

rom

e,15

heal

thy

cont

rols

0/15

39�

14A

ctiv

eSi

mpl

eLe

tter

Can

cella

tion

Test

,Con

ditio

nalL

ette

rC

ance

llatio

nTe

st,T

ower

ofH

anoi

,Str

oop

Test

,Lur

ia’s

Mem

ory

Wor

ds-R

evis

ed

Patie

nts

show

edsi

gnifi

cant

atte

ntio

nal-

depe

nden

tw

orki

ngm

emor

yde

ficits

and

impa

irmen

tin

dela

yed

reca

llta

sk.

Ther

ew

ere

nodi

ffer

ence

sbe

twee

npa

tient

san

dco

ntro

lson

basi

cat

tent

iona

land

exec

utiv

efu

nctio

ning

.Pr

esen

tst

udy

74C

ushi

ng’s

dise

ase,

74m

atch

edco

ntro

ls,

54N

FMA

,54

mat

ched

cont

rols

13/6

152

�13

13�

13yr

inre

mis

sion

MM

SE,W

echs

ler

Mem

ory

Scal

e,V

erba

lLea

rnin

gTe

stof

Rey,

Rey

Com

plex

Figu

reTe

st,T

rail

Mak

ing

Test

,St

roop

Col

or-W

ord

test

,Le

tter

-Dig

itSu

bstit

utio

nTe

st,D

igit-

Del

etio

nTe

st,

Figu

reFl

uenc

y,FA

S,G

roni

nger

Inte

llige

nce

Test

Syno

nym

sSu

btes

t

Patie

nts

with

long

-ter

mcu

reof

Cus

hing

’sdi

seas

esc

ored

sign

ifica

ntly

wor

seon

test

sm

easu

ring

mem

ory

and

toa

less

erex

tent

onte

sts

mea

surin

gex

ecut

ive

func

tioni

ngco

mpa

red

with

both

mat

ched

cont

rols

and

patie

nts

trea

ted

for

NFM

A.T

hese

obse

rvat

ions

indi

cate

irrev

ersi

ble

effe

cts

ofpr

evio

ushy

perc

ortis

olis

mon

cogn

itive

func

tion.

HPA

,Hyp

otha

lam

ic-p

ituita

ry-a

dren

al;N

A,n

otav

aila

ble;

m,m

ale;

f,fe

mal

e.

14 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000

Page 16

Even though the selection procedure may have inducedsome, but unknown, bias, the data indicate that there weredifferences in outcome parameters between both groups ofpatientswith the similar selectionmethodof controls.There-fore, the outcomes are not a consequence of the study designor the selection procedure of the control subjects but, rather,of the long-term consequences of Cushing’s disease.

In summary, there are subtle impairments in cognitivefunction in patients during long-term follow-up after cureof Cushing’s disease compared with NFMA patients andmatched controls. The greatest impairment was present inmemory, although executive functioning was also af-fected. This impairment in cognitive function after treat-ment of Cushing’s disease is not merely the result of pitu-itary disease in general and/or its treatment but includesspecific elements most likely caused by the irreversible ef-fects of previous glucocorticoid excess on the central ner-vous system.

Acknowledgments

Address all correspondence and requests for reprints to:J. Tiemensma, MSc., Department of Endocrinology and Met-abolic Diseases C4-R, Leiden University Medical Center, P.O. Box9600, 2300 RC Leiden, The Netherlands. E-mail: [email protected].

Disclosure Summary: The authors have nothing to disclose.

References

1. Colao A, Pivonello R, Spiezia S, Faggiano A, Ferone D, Filippella M,Marzullo P, Cerbone G, Siciliani M, Lombardi G 1999 Persistenceof increased cardiovascular risk in patients with Cushing’s diseaseafter five years of successful cure. J Clin Endocrinol Metab 84:2664–2672

2. Dekkers OM, Biermasz NR, Pereira AM, Roelfsema F, van AkenMO, Voormolen JH, Romijn JA 2007 Mortality in patients treatedfor Cushing’s disease is increased, compared with patients treatedfor nonfunctioning pituitary macroadenoma. J Clin EndocrinolMetab 92:976–981

3. van Aken MO, Pereira AM, Biermasz NR, van Thiel SW, HoftijzerHC, Smit JW, Roelfsema F, Lamberts SW, Romijn JA 2005 Qualityof life in patients after long-term biochemical cure of Cushing’s dis-ease. J Clin Endocrinol Metab 90:3279–3286

4. Whelan TB, Schteingart DE, Starkman MN, Smith A 1980 Neuro-psychological deficits in Cushing’s syndrome. J Nerv Ment Dis 168:753–757

5. Martignoni E, Costa A, Sinforiani E, Liuzzi A, Chiodini P, MauriM, Bono G, Nappi G 1992 The brain as a target for adrenocorticalsteroids: cognitive implications. Psychoneuroendocrinology 17:343–354

6. Mauri M, Sinforiani E, Bono G, Vignati F, Berselli ME, Attanasio R,Nappi G 1993 Memory impairment in Cushing’s disease. Acta Neu-rol Scand 87:52–55

7. Forget H, Lacroix A, Somma M, Cohen H 2000 Cognitive declinein patients with Cushing’s syndrome. J Int Neuropsychol Soc6:20–29

8. Starkman MN, Giordani B, Berent S, Schork MA, Schteingart DE

2001 Elevated cortisol levels in Cushing’s disease are associated withcognitive decrements. Psychosom Med 63:985–993

9. Michaud K, Forget H, Cohen H 2009 Chronic glucocorticoid hy-persecretion in Cushing’s syndrome exacerbates cognitive aging.Brain Cogn 71:1–8

10. Leon-Carrion J, Atutxa AM, Mangas MA, Soto-Moreno A, PumarA, Leon-Justel A, Martín-Rodriguez JF, Venegas E, Domínguez-Morales MR, Leal-Cerro A 2009 A clinical profile of memory im-pairment in humans due to endogenous glucocorticoid excess. ClinEndocrinol (Oxf) 70:192–200

11. Starkman MN, Gebarski SS, Berent S, Schteingart DE 1992 Hip-pocampal formation volume, memory dysfunction, and cortisollevels in patients with Cushing’s syndrome. Biol Psychiatry 32:756 –765

12. Foy MR, Stanton ME, Levine S, Thompson RF 1987 Behavioralstress impairs long-term potentiation in rodent hippocampus. BehavNeural Biol 48:138–149

13. Bodnoff SR, Humphreys AG, Lehman JC, Diamond DM, Rose GM,Meaney MJ 1995 Enduring effects of chronic corticosterone treat-ment on spatial learning, synaptic plasticity, and hippocampal neu-ropathology in young and mid-aged rats. J Neurosci 15(1 Pt 1):61–69

14. Bourdeau I, Bard C, Forget H, Boulanger Y, Cohen H, Lacroix A2005 Cognitive function and cerebral assessment in patients whohave Cushing’s syndrome. Endocrinol Metab Clin North Am 34:357–369, ix

15. Fietta P, Fietta P, Delsante G 2009 Central nervous system effects ofnatural and synthetic glucocorticoids. Psychiatry Clin Neurosci 63:613–622

16. Brown ES 2009 Effects of glucocorticoids on mood, memory, andthe hippocampus. Treatment and preventive therapy. Ann NY AcadSci 1179:41–55

17. Starkman MN, Giordani B, Gebarski SS, Schteingart DE 2003 Im-provement in learning associated with increase in hippocampal for-mation volume. Biol Psychiatry 53:233–238

18. Dorn LD, Cerrone P 2000 Cognitive function in patients with Cush-ing syndrome: a longitudinal perspective. Clin Nurs Res 9:420–440

19. Forget H, Lacroix A, Cohen H 2002 Persistent cognitive impairmentfollowing surgical treatment of Cushing’s syndrome. Psychoneu-roendocrinology 27:367–383

20. Hook JN, Giordani B, Schteingart DE, Guire K, Giles J, Ryan K,Gebarski SS, Langenecker SA, Starkman MN 2007 Patterns of cog-nitive change over time and relationship to age following successfultreatment of Cushing’s disease. J Int Neuropsychol Soc 13:21–29

21. Lezak MD 1995 Neuropsychological assessment. 3rd ed. NewYork: Oxford University Press

22. Folstein MF, Folstein SE, McHugh PR 1975 “Mini-mental state.” Apractical method for grading the cognitive state of patients for theclinician. J Psychiatr Res 12:189–198

23. Wechsler D, Stone CP 1945 Wechsler Memory Scale. New York:Psychological Corporation

24. Rey A 1958 L’examin Clinique en Psyhcologie. Paris: Presses Uni-versitaires de France

25. Rey A 1941 L’examen psychologique dans les cas d’encephalopathietraumatique. Arch Psychol 28:286–340

26. Reitan R 1956 Trail making test: manual for administration, scor-ing, and interpretation. Bloomington, IN: Indiana University

27. Stroop J 1935 Studies of interference in serial verbal reactions. J ExpPsychol 18:643–662

28. van der Elst W, van Boxtel MP, van Breukelen GJ, Jolles J 2006 TheLetter Digit Substitution Test: normative data for 1,858 healthyparticipants aged 24–81 from the Maastricht Aging Study (MAAS):influence of age, education, and sex. J Clin Exp Neuropsychol 28:998–1009

29. Regard M, Strauss E, Knapp P 1982 Children’s production on verbaland non-verbal fluency tasks. Percept Mot Skills 55:839–844

30. Benton AL, Hamsher Kd 1976 Multilingual Aphasia Examination.Iowa City, IA: University of Iowa

J Clin Endocrinol Metab, June 2010, 95(6):0000–0000 jcem.endojournals.org 15

Page 17

31. Luteijn F, Ploeg FAEvd 1983 Manual Groninger Intelligence Test.Lisse, The Netherlands: Swets, Zeitlinger

32. Spinhoven P, Ormel J, Sloekers PP, Kempen GI, Speckens AE, VanHemert AM 1997 A validation study of the Hospital Anxiety andDepression Scale (HADS) in different groups of Dutch subjects. Psy-chol Med 27:363–370

33. Zigmond AS, Snaith RP 1983 The hospital anxiety and depressionscale. Acta Psychiatr Scand 67:361–370

34. Grattan-Smith PJ, Morris JG, Shores EA, Batchelor J, Sparks RS1992 Neuropsychological abnormalities in patients with pituitarytumours. Acta Neurol Scand 86:626–631

35. Peace KA, Orme SM, Thompson AR, Padayatty S, Ellis AW,Belchetz PE 1997 Cognitive dysfunction in patients treated for pi-tuitary tumours. J Clin Exp Neuropsychol 19:1–6

36. Peace KA, Orme SM, Padayatty SJ, Godfrey HP, Belchetz PE 1998Cognitive dysfunction in patients with pituitary tumour who havebeen treated with transfrontal or transsphenoidal surgery or medi-cation. Clin Endocrinol (Oxf) 49:391–396

37. McEwen BS 2008 Central effects of stress hormones in health and

disease: understanding the protective and damaging effects of stressand stress mediators. Eur J Pharmacol 583:174–185

38. Yehuda R 2002 Post-traumatic stress disorder. N Engl J Med 346:108–114

39. Wolkowitz OM, Burke H, Epel ES, Reus VI 2009 Glucocorti-coids: mood, memory, and mechanisms. Ann NY Acad Sci 1179:19 – 40

40. Schulz KF, Grimes DA 2005 Multiplicity in randomised trials I:endpoints and treatments. Lancet 365:1591–1595

41. Schulz KF, Grimes DA 2005 Multiplicity in randomised trials II:subgroup and interim analyses. Lancet 365:1657–1661

42. Biermasz NR, van Thiel SW, Pereira AM, Hoftijzer HC, van HemertAM, Smit JW, Romijn JA, Roelfsema F 2004 Decreased quality oflife in patients with acromegaly despite long-term cure of growthhormone excess. J Clin Endocrinol Metab 89:5369–5376

43. Dekkers OM, van der Klaauw AA, Pereira AM, Biermasz NR,Honkoop PJ, Roelfsema F, Smit JW, Romijn JA 2006 Quality of lifeis decreased after treatment for nonfunctioning pituitary macroad-enoma. J Clin Endocrinol Metab 91:3364–3369

16 Tiemensma et al. Cognitive Impairments in Cushing’s Disease J Clin Endocrinol Metab, June 2010, 95(6):0000–0000