| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
University Departments of Growth and Reproduction (T.H.J., K.M.M.) and Clinical Genetics (M.S.), Rigshospitalet, DK-2100 Copenhagen, Denmark; Department of Health Psychology (C.P.L.R., E.L.M.), Institute of Public Health, University of Copenhagen, DK-2100 Copenhagen, Denmark; and The Kinsey Institute for Research in Sex, Gender, and Reproduction (J.M.R.), Indiana University, Bloomington, Indiana 47405
Address all correspondence and requests for reprints to: Katharina M. Main, University Department of Growth and Reproduction, GR-5064, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. E-mail: katharina.main{at}rh.hosp.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to investigate cognitive function in CAH women.
Design: This was a case-control study.
Setting: This study was conducted at a tertiary center for pediatric endocrinology at the University Hospital of Copenhagen.
Participants: Thirty-five Danish CAH women (age 17–51 yr) were included, and participation rate was 84%. Control women were recruited through the Danish Civil Registration System and matched on age and education.
Main Outcome Measures: An abbreviated form of the Wechsler Adult Intelligence Scale was used, i.e. full-scale intelligence quotient (IQ; five of 11 subtests), which included three of six verbal IQ subtests and two of five performance IQ subtests.
Results: A significantly lower IQ was found in CAH patients compared with controls with respect to mean full-scale IQ (84.5 vs. 99.1; P < 0.001), mean verbal IQ (86.6 vs. 97.3; P < 0.001), and mean performance IQ (85.7 vs. 101.3; P < 0.001). The salt-wasting CAH group had lower IQ scores than the simple-virilizing CAH group, which reached significance for mean total IQ (81.2 vs. 92.8, P = 0.04) and mean verbal IQ (84.7 vs. 95.5, P = 0.05), and additionally, lower scores than the late-onset CAH group, which reached significance for performance IQ (mean 81.5 vs. 96.2, P = 0.02).
Conclusions: Impaired cognitive function was observed in patients with CAH, especially in salt-wasting CAH. These intriguing findings may reflect adverse effects of hyponatremic episodes, suboptimal postnatal hormone replacement therapy or prenatal adrenal androgen excess, and the potential psychosocial consequences of the disorder.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CAH subjects. This was an adult follow-up study of females diagnosed with CAH in childhood or adolescence at the University Hospital of Copenhagen (Rigshospitalet) in the period 1953–2003 [n = 37, 35 completed intelligence quotient (IQ) testing]. Patients with a previous contact to the departments of pediatrics, plastic surgery, pediatric surgery, or gynecology were recruited by letter. The participation rate was 84%.
The patients with CYP21 mutations were grouped into salt wasters (SWs; n = 19, mean age, 31.2 yr; range, 19–46 yr), simple virilizers (SVs; n = 6, 34.6 yr, 23–51), and late-onset (LO) CAH (n = 5, 25.5 yr, 19–36). Patients with steroidogenic acute regulatory protein (StAR) deficiency (n = 3), CYP21 deficiency diagnosed in adolescence (n = 1), and 17-hydroxylase deficiency (n = 1) were combined in a single group (28.8 yr, 17–49 yr). Patients with CYP21 deficiency were separated by mutation severity, i.e. null, A, B, and C mutations (3), salt-wasting status, and clinical picture (Tables 1
and 2). In each of the SW, LO, and StAR groups, a pair of siblings was included.
|
|
Diagnosis and sex assignment.
Birth data were available in 24 patients. Except for the SV group, one patient from each subgroup had a birth weight less than 2500 g. The SW patients were diagnosed at a median age of 0 yr (range, 0–0.3 yr), the SV patients at 0 yr (0–1.2 yr), the LO patients at 8.5 yr (5.3–10.0 yr), and the patients in the mixed group at 0.17 yr (0–16.1 yr). As shown in Tables 1
and 2, seven SW females were misassigned as boys at birth; six were reassigned as females during the 1st month of age and one patient at 3 months. Nine SW females and two SV females were born with ambiguous genitalia. Thus, in the salt-wasting group, 84% had severely virilized genitalia at birth.
Hyponatremia and hypoglycemia.
In the SW group, 11 patients (58%) had verified hyponatremic episodes with serum sodium levels less than 130 mmol/liter (normal range, 135–145 mmol/liter) during infancy and early childhood (Tables 1 and 2). If all degrees of sodium depletion were included (i.e. all 
131 nmol/liter or described as hyponatremia in medical records), a total of 16 patients (84%) have had hyponatremic episodes. One SW patient had hypoglycemia with convulsions. In the mixed group, all StAR patients had suffered from hyponatremia during childhood, one also with a hypoglycemic episode without convulsions. In the SV and LO groups, neither hyponatremia nor hypoglycemia was noted in early childhood.
Corticosteroid treatment. The mean doses of hydrocortisone (HC) per square meter in the first 2 yr of life were grouped as following: 1) less than 10 mg HC/m2, 2) 10–15 mg HC/m2, 3) 15–30 mg HC/m2, and 4) more than 30 mg HC/m2. Data on body surface area and glucocorticoid treatment were available from 13 of 25 records in SW CAH, grouping one patient as B, three patients as C, and nine patients as D. In the SV group, data on five patients were available, grouping one patient as C and four patients as D. In the mixed group, data on three patients were present, grouping two patients as B and one as C. None of the LO patients had substitution therapy before 2 yr of age.
Comorbidity. Significant current somatic comorbidity existed in one SW female diagnosed with cystic fibrosis. In infancy, another SW female had unilateral ureteronephrectomia performed due to a malformed kidney and hydronephrosis. Five SW patients were medicated with selective serotonin reuptake inhibitors. Of these, four were treated for depression and one for both depression and obsessive-compulsive disorder.
Hospitalizations. The median number of hospitalizations in the SW group was seven (range, one to 25), for the SV group six (range, three to 11), for the LO group one (range, zero to 6), and for the mixed group seven (range, three to 13).
Controls. A control group was recruited through the Danish Civil Registration System. For each patient, 20 women living in the Copenhagen area and born in the same month and year as the patient were contacted by a letter containing a description of the study and questions on educational level (i.e. qualifications achieved, total number of completed school years, and vocational training). The response rate for the potential controls was 38%, corresponding to a mean response from 7.5 controls per patient. The woman with the closest match on education was selected for pair-wise matching with the index patient. None of the controls had any adrenal diseases.
Assessment program
Patients and controls were administered a medical interview, anthropometrical measurements, a gynecological examination, a psychological interview, questionnaires on personality, sexual and social functioning, cognitive tests, and hormone analyses.
Intelligence testing
The participants completed five subtests from the Wechsler Adult Intelligence Scale (WAIS) (4), which provides three IQs: verbal, performance, and full-scale IQs. Wechsler IQs are defined to have a population mean of 100 (SD = 15) and classified in the categories: IQ less than or equal to 69, mentally defective; 70–79, borderline; 80–89, dull normal; 90–109, average; 110–119, bright normal; 120–129, superior; and more than or equal to 130, very superior. Three of six verbal subtests and two of five performance subtests were administered. The three verbal subtests were information (29 questions on general knowledge), similarities (13 questions requiring the subject to describe how two seemingly dissimilar items might in fact be similar, i.e. examples of the same concept), and vocabulary (40 words to be defined by the subject). The two performance subtests were block design (combining 10 sets of four or nine blocks to match patterns presented on cards) and picture arrangement (eight sets of small pictures to be arranged in a logical sequence to tell a story). Two trained examiners administered the WAIS test in approximately 45 min. No systematic difference in test results of control persons was found between the two observers (P = 0.31). The examiners were not blinded with respect to the diagnoses of the included patients.
Standard procedures were used to derive IQs, i.e., raw scores were converted to scaled scores and sums of scaled scores converted to IQs, using norms based on the WAIS results of 200 Danish 20- to 33-yr-old women representative for their generation of Danes with respect to educational level and included in the Prenatal Development Project (5). Thus, the IQ norms used in the present study differed from standard IQ norms, both because they were exclusively based upon women and because they were derived from the same five subtests administered to the CAH group, instead of all the 11 subtests, comprising the full WAIS. However, correlations between the above-mentioned five subtests and all 11 subtests in the sample from the Prenatal Development Project were 0.96, 0.96, and 0.90 for full-scale, verbal, and performance IQs, respectively.
Educational level and current employment of participants and parents
Information on the participants’ education and current employment was collected by a questionnaire on socioeconomic status (SES). From items describing type and total number of completed school years, vocational training, current employment, and number of employees, classes between 1 (highest) and 7 (lowest) were allocated (6). Parental vocational training and preponderant type of employment were scored in three categories: no education/employment, academic training/employment, and other.
Statistical analysis
Statistical analyses were conducted using SPSS for Windows 13.0 (SPSS Inc., Chicago, IL) by nonparametric tests. The Wilcoxon matched pairs signed rank sum test was used in comparisons of patients and matched controls and in comparisons of verbal and performance IQs in the same patient. Mann-Whitney and Kruskal-Wallis tests were used in analyses of differences between patient subgroups.
Approvals
Approvals were granted from the local ethical committee (01-051/01, 11-075/03), and the Danish Data Registry (2001-41-0689). The study was performed according to the Helsinki Declaration II after obtaining a written consent.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). The SES score in the LO group was marginally higher than in the controls (P = 0.08), whereas no differences were observed for the other subgroups vs. their matched controls (P = 0.18–1.0). There were no differences between patients and matched controls with respect to maternal education (P = 0.16–1.0), maternal employment (P = 0.32–1.0), paternal education (P = 0.10–0.76), or paternal employment (P = 0.10–1.0).
|
). For the verbal IQ, the mean in the patients was 86.6 (64–107), whereas it was 97.3 (70–132) for the controls (P < 0.001). For the performance IQ, the CAH patients obtained a mean of 85.7 (62–127) and the controls a mean of 101.3 (range 73–122), P < 0.001.
|
Statistical analyses were repeated with exclusion of the six siblings one at a time to evaluate the potential effects of correlated observations. These analyses showed essentially the same results (data not shown).
The SW group had significantly lower total and verbal IQs than the SV group (P = 0.04 and 0.05), and a nonsignificant trend was observed for performance IQ (P = 0.08). The SWs had a significantly lower performance IQ than the LO group (P = 0.02) but did not differ significantly in full-scale or verbal IQs (P = 0.10 and 0.45). No differences were found between the SW group and the mixed group (P = 0.30–1.0). The SV group and the LO group did not differ in any of the IQs (P = 0.33–0.93), but the SV group had significantly higher total IQ (P = 0.03) and verbal IQ (P = 0.05) than the mixed group. The LO group had marginally significant higher full-scale IQ and performance IQ (P = 0.06, resp.) than the mixed group, whereas no differences were observed for the verbal IQ (P = 0.15).
IQ differences between patients and matched controls were of similar magnitude (P = 0.07).
The SW and SV patients had higher verbal IQs than performance IQs, although not significantly so (P = 0.23 and 0.46, respectively), whereas the opposite finding was observed in the LO and the mixed group (P = 0.10 and 0.23, respectively). No significant difference in performance IQ vs. verbal IQ was shown in the controls (P = 0.09). No IQ differences were demonstrated between the five selective serotonin reuptake inhibitor-treated SW patients vs. the remaining 14 SW patients (P = 0.75–0.89 for the three IQs).
Patients with verified hyponatremic crises (n = 14) vs. all other CAH patients (n = 21) revealed significantly lower IQs for total IQ (mean 78.6 vs. 88.4, P = 0.02) and for verbal IQ (mean 79.9 vs. 91.0, P = 0.006) but not for performance IQ (mean 82.9 vs. 87.6, P = 0.21).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Cognitive function in CAH patients has been evaluated in several studies with discrepant findings. Money and Lewis (7) initially observed higher than average IQ in a sample of 70 patients, but in a subsample of seven families, no IQ differences were found between patients and 10 nonaffected siblings. In subsequent studies, the increased IQs were confirmed, but no differences between the IQs of patients and their sibling were demonstrated (8, 9). Explanations for the observed IQ advantage in CAH patients included excess androgen levels or some aspects of the genetics related to the syndrome (7), but the findings may be related to a number of methodological problems in addition to outdated test norms: selective sampling bias including an underrepresentation of SW patients, an overrepresentation of women (presumable due to the more pronounced clinical picture), or a favoring of higher IQ parents bringing their affected children to specialized medical attention, in addition to outdated norms and changed referral practices (1, 8, 9, 10).
In contrast, a study conducted in 1994 of 22 prenatally virilized female CAH patients (16 SW and six SV patients; mean age, 22.7 yr) reported significantly lower intelligence in the patients than in controls (11). Our findings, based on individually matched controls and IQ norms derived from the results of 200 normal females, are in agreement with this most recent study.
Various factors may explain our observations. Psychosocial consequences related to CAH may affect intellectual development during childhood and adolescence. Furthermore, the current psychological states of the patients may affect test performance; e.g. cognitive impairment has been reported in depressed patients (12), and attention deficits associated with depression may affect test performance directly or via test motivation. However, in this study, the patients treated for depression did not show additional IQ impairment.
Lower IQ scores in SW patients compared with non-SW patients have previously been reported (7, 8, 10, 11, 13). It has been hypothesized that this phenomenon may reflect brain injury from episodes of hypotension and hyponatremia (13). In support of this hypothesis, lower IQ scores were primarily observed in patients with previous hyponatremic episodes, and not to the same degree in patients without these episodes. Neonatal screening is an important factor in reducing salt-wasting crises, but because this examination is not yet established in Denmark, none of the patients in the present sample were diagnosed via this screening procedure.
Exposure to androgens in utero has been proposed to delay left cerebral hemisphere maturation with a subsequent risk of developing learning disabilities (14). Both increased rate of learning disabilities in CAH patients (13, 15, 16) as well as unaffected learning ability (17) have been reported. Nass and Baker (13) reported SVs more learning disabled than SWs, despite the fact that the SW females had the lowest IQ. Thus, cerebral damage due to hypotension and hyponatremia was suggested to override the androgen effect, resulting in general muting of IQ, rather than causing a specific learning disability. Lower IQs in the SWs, compared with SVs, were also found in the present study. Female CAH fetuses have been reported to be exposed to much higher levels of adrenal androgens in amniotic fluid than normal male fetuses (18). Thus, increased intrauterine exposure to androgens may also contribute to our findings because we found the lowest test scores in the SWs. However, this factor cannot explain the cognitive findings seen in the LO and the mixed group of patients.
Prenatal androgen exposure on the developing brain has also been proposed to result in increased spatial performance because there are well-known gender differences in spatial tests (19, 20). Previous studies have shown higher spatial ability in girls exposed to high prenatal testosterone levels in utero (assessed in amniotic fluid between 14 and 20 wk gestation) (21) and lower spatial ability in men with idiopathic hypogonadotropic hypogonadism (22). Furthermore, an enhanced spatial ability has been reported in CAH females compared with unaffected female relatives (23). Higher performance IQ than verbal IQ in females with CAH has been reported previously (24), although not retrieved in other studies (17, 25, 26). In the present study, enhanced spatial ability in the prenatally most virilized females was not documented. It is possible that deficits in general intelligence prevented detection of any enhancement in spatial abilities in the CAH patients or that our sample size was too small to detect the effects.
Decreased glucocorticoid production (in the SW and SV CAH), disruption of the hypothalamic-pituitary-adrenal axis, and difficulties with adequate substitution therapy regime may also contribute to impaired cognitive function. Glucocorticoids are important for normal maturation of the developing central nervous system. In addition to exerting a tonic inhibitory control on the hypothalamic-pituitary-adrenocortical-axis (27) and being important to learning and memory consolidation, the hippocampus possesses high concentrations of corticosteroid receptors (28). Additionally, exposure to excess cortisol can induce adverse effects, including hippocampal atrophy with memory dysfunction in adult patients with Cushing’s disease (29). In a magnetic resonance imaging-based study (30), classic CAH children were found to have hippocampal volumes similar to their healthy counterparts. However, in the CAH children, decreased size of the amygdala, a structure important in mediation of emotions (31), was observed. In another study, white matter abnormalities or temporal lobe atrophy were found by magnetic resonance imaging in one third of SW and SV CAH patients despite normal neurological examination (32), the most likely explanation for this finding being excess glucocorticoid treatment. During the first 2 yr of life, the patients in our study were treated with different types of glucocorticoids, including dexamethasone, cortisone, cortisone acetate, prednisone, and prednisolone. The mean glucocorticoid doses were calculated as minimum doses and reflected only the treatment doses given in steady phases of the disease. Thus, a considerable number of patients had been treated with excessive doses of glucocorticoids in infancy, compared with the current recommendations of 10–15 mg HC/m2/d (33). Cerebral injury as a result of steroid treatment may be an important factor especially in SW patients. It is yet unknown whether the current treatment recommendations will prevent adverse effects on brain development.
The typical correlation between parent and offspring IQ has been reported to be 0.42 (34); ideally, CAH patients and controls should be matched on parental IQ. This was not possible in the present study, but we observed no significant differences between patients and controls on measures of parental education and employment status. High correlations are usually observed between intelligence and education (35); consequently, intelligence differences should not be expected between patients and controls well-matched on educational level. If the primary objective of the study had been to investigate intelligence in CAH patients, it would be a methodological error to match on educational level because this may remove most of the variance of interest. Because patients and controls were matched on education and no SES differences were observed, the lower IQ scores found in the patients did not reflect lower SES levels.
In an effort to understand the paradox of substantial intelligence differences despite matching on education, we analyzed an alternative one- to eight-point measure of education, combining information on achieved school qualifications and vocational training. The rescored measure of educational level only showed significant differences between SW patients and controls. In a previous Danish study (35), a correlation between educational level and full-scale IQ of 0.64 was observed, whereas we observed comparatively low correlations of 0.37 and 0.36 in the patients and controls, respectively. This may possibly reflect the small sample size but also problems with our measure of education. However, the rescored measure of educational level is very similar to measures used successfully in other Danish studies (36), and the explanation of the apparent discrepancy between intelligence and educational achievement in our CAH patients remains an open question.
Observers were not blinded during the IQ testing. However, intelligence testing was primarily included as confounder for primary outcome variables such as quality of life. The null hypothesis was no expected differences between patients and controls. Thus, we believe that our design did not introduce a systematic bias. Our study population may not be representative for all CAH patients because they were recruited from a university hospital and may be more severely affected. Our participation rate was high, but we cannot exclude selection bias. However, there is typically a positive association between IQ and participation rate. Thus, it is unlikely that selection bias explains the low intelligence observed in CAH patients.
Conclusion
We observed intelligence deficits in CAH women, especially in salt-wasting CAH. The findings may partly reflect psychosocial consequences of CAH. However, depending on the type of CAH, various factors including prenatal exposures to low levels of glucocorticoids and high levels of androgens and suboptimal hormone replacement including hyponatremia may contribute to the impaired cognitive functions observed. Similar studies on male CAH patients may help in further evaluation of androgens’ role on cognitive function.
Furthermore, because low intelligence may have a considerable negative effect on quality of life, our findings strengthen the necessity of including assessment of cognitive function in follow-up examinations of patients with CAH and in their health care programs.
|
Footnotes |
|---|
None of the authors have potential financial conflicts of interest to declare.
First Published Online January 31, 2006
Abbreviations: CAH, Congenital adrenal hyperplasia; HC, hydrocortisone; IQ, intelligence quotient; LO, late onset; SES, socioeconomic status; StAR, steroidogenic acute regulatory protein; SV, simple virilizer; SW, salt waster; WAIS, Wechsler Adult Intelligence Scale.
Received August 31, 2005.
Accepted January 20, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  O. Inozemtseva, E. Matute, and J. Juarez Learning Disabilities Spectrum and Sexual Dimorphic Abilities in Girls with Congenital Adrenal Hyperplasia J Child Neurol, August 1, 2008; 23(8): 862 - 869. [Abstract] [PDF]
|
|
|
|
 |
|
 D. P. Merke Approach to the Adult with Congenital Adrenal Hyperplasia due to 21-Hydroxylase Deficiency J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 653 - 660. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. H Johannsen, C. P L Ripa, E. L Mortensen, and K. M Main Quality of life in 70 women with disorders of sex development Eur. J. Endocrinol., December 1, 2006; 155(6): 877 - 885. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
