| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
New York State Psychiatric Institute and Department of Psychiatry (H.F.L.M.-B., C.D.), Columbia University, New York, New York 10032; and New York Presbyterian Hospital/Weill Medical College of Cornell University (S.W.B., A.D.C., J.S.O., M.I.N.), New York, New York 10021
Address all correspondence and requests for reprints to: H. F. L. Meyer-Bahlburg, Dr. rer. nat., Department of Psychiatry, Columbia University, 1051 Riverside Drive, NYSPI Unit 15, New York, New York 10032. E-mail: meyerb{at}childpsych.columbia.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The first report of successful suppression of adrenal androgen production in an affected female fetus by the administration of dexamethasone (DEX) to the mother appeared in 1984 (3), and numerous reports on the medical outcome have been published since (1, 2, 4, 5, 6, 7, 8). Prenatal treatment of a CAH-risk pregnancy must begin soon after diagnosis of pregnancy because sexual differentiation of the genitalia occurs in human fetuses between 9 and 13 wk gestation (9, 10) and must continue until chorionic villus sampling or amniocentesis can be performed. In CAH-affected females, treatment is continued until birth. Appropriately treated females are born with normal or only slightly virilized external genitalia. Adverse treatment effects in the mother include hyperglycemia, excessive weight gain, chronic epigastric pain, irritability, nervousness, fatigue, and increased facial hair growth (1, 2, 4, 5, 6, 8, 11).
Relatively little is known about the effects of early-prenatal and chronic DEX exposure on the later development of the child. In preterm infants treated with corticosteroids for bronchopulmonary dysplasia antenatally, follow-up data from three randomized trials (up to 3, 6, and 12 yr of age, respectively) showed increased survival after such treatment but no differences between corticosteroid-treated and untreated children in growth and development and on a variety of psychometric tests (12). However, these studies often used corticosteroids other than DEX in a very short course of treatment (for a few hours before delivery and/or shortly after birth), which takes place considerably later in development than the DEX treatment for CAH-risk pregnancies. In the context of the latter, DEX itself in the doses recommended does not appear to be teratogenic (1, 4, 6, 8). Treatment with less than 25 µg/kg·d does not appear to cause growth retardation, and early follow-up studies reported normal growth and development (1). More recently, Lajic et al. (5) evaluated the physical development (by chart review) of 44 prenatally DEX-treated children from at-risk pregnancies including five girls with severe CAH. In comparison with matched controls, the majority of the DEX-treated children showed normal prenatal and postnatal growth, but several adverse events such as failure to thrive and delayed psychomotor development were seen among the DEX-treated infants.
Animal studies have shown lasting effects of perinatal corticosteroid treatment on the nervous system and behavior in rodents and primates, including various aspects of both social and cognitive functioning (13, 14, 15, 16, 17). The majority of studies use doses of one or two orders of magnitude above the standard dose recommended for the treatment of pregnancies at risk for CAH [0.02 mg/kg·d in three divided doses (8)], which raises questions about their applicability to the standard DEX treatment of CAH risk pregnancies, but a recent report indicated that even a single, low dose (0.05 mg/kg) of DEX used in late rat gestation can disrupt the transcription factors that regulate brain cell differentiation (18). The hypothalamic-pituitary-adrenal axis has a major role in the regulation of memory consolidation (19) involving the hippocampus and amygdala. The hippocampus in particular is a structure vital to learning and memory and possesses high concentrations of glucocorticoid receptors (20). There are recent reports that high levels of corticosteroids associated with stress in rodents and primates and posttraumatic stress disorder, Cushing’s syndrome, and depression in humans cause loss of hippocampal volume (13, 15, 21), although CRH appears to be involved as well (22). Because of such findings and because of other concerns about potential long-term physiological effects, especially the potential for a permanent modification of hypothalamic-pituitary-adrenal axis activity later in life, a number of recent reports (5, 23, 24, 25, 26) advocate the careful evaluation of the use of DEX in human pregnancies for possible cognitive and behavioral side effects, among others.
The current study examined the effects of DEX on motor and cognitive development in a large sample. [A preliminary report on nearly the first third of the sample of this study was presented at a 1999 conference (28).]
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The main source of participating children was the prenatal-diagnosis database of one of the authors (M.I.N.) (8), to which new children, including both DEX exposed and DEX unexposed, were added throughout the data collection period. In addition, 80 children with CAH (age, 0–18 yr), part of a CAH sample for medical studies that is not included in the prenatal-diagnosis database, were available as a source for recruiting additional DEX-untreated girls with CAH.
The study was limited to children younger than 12 yr of age from families who resided within the continental United States and had at least one English-speaking parent. During the data collection period, a total of 827 potential child participants younger than 12 yr of age were identified, and initial contact letters mailed to their mothers. Fifty-eight children were found to be ineligible [27 because of language, seven due to death of the child (age at death ranging from 2 wk to 5.5 yr; cause of death: in five CAH-salt wasting patients, probable adrenal crises; in one of two non-CAH children, questionable sudden infant death syndrome; in the other, unknown), four because of residence outside the United States, three because of age over 12 yr, and 17 because of other reasons]. For 124 children, the address proved to be invalid, and address tracing was unsuccessful during the data collection period. Of the 645 eligible and traceable children, mothers refused participation for 49, 98 agreed but did not return questionnaires, three returned inappropriately completed questionnaires, and eight returned their questionnaires too late for inclusion. Thus, the data for 487, or 75.5% of the 645 eligible and traceable children, were included in the data analysis.
Assessments
Mothers were mailed a set of questionnaires tailored to the age group of their child, which included four widely used, normed standard questionnaires as measures of cognitive and motor development. The manuals or initial journal articles describing the four questionnaires contain detailed information on scale reliability, validity, and norms.
The Kent Infant Development Scale (KIDS) (29) is a 252-item questionnaire designed for the age group 0 to 15 months and yields age-based normalized standard scores for five developmental subscales and a composite (see Table 2
). The Revised Prescreening Developmental Questionnaire (RPDQ or Revised Denver) (30) includes four age-specific forms with a total of 105 items that in combination are designed to cover the age group 0 months to 6 yr. For the current study, we used the age-based delay score (no delay, one delay, two or more delays). The Child Development Inventory (CDI) (31) has 270 yes/no items for the age group 15 months to 6 yr, which are summed up to yield the eight domain scales and an overall summary scale, general development (see Table 3). For our analyses here, we used the five-point categorization of delays for each developmental domain ranging from 0 (above the mean for age) to 4 (>2 SD below the mean for age).
|
|
Information on demographic variables such as ethnicity, age, and parental education [coded according to Hollingshead (33) and averaged over both parents] was obtained from the parents. Because the majority of patients received their treatment by local physicians under long-distance consultation by one of the authors (M.I.N.) without being on an experimental treatment protocol, the recording practices of physicians were very variable, and so was, given the mobility of the population, the availability of detailed records and the cooperation of various health professionals in providing basic data such as onset and end of DEX treatment in terms of gestational week (which would permit the calculation of the number of weeks of prenatal exposure to DEX) and gestational age at birth.
Procedures
Through their treating obstetrician/pediatrician or the consulting physician, mothers were initially mailed a survey-information sheet with the request to complete and return an answer sheet. If a mother agreed to participate, she was mailed the set of questionnaires appropriate for the age of the child in question that she was asked to return in stamped return envelopes after completion. Repeated follow-up calls were conducted by study staff to increase the rate of return. Mothers received an honorarium of $5 for returning the answer sheet and $20 for completing the questionnaires. Written informed consent was obtained from all participating mothers after the appropriate institutional review boards had approved the study.
Analysis
Because the effects of CAH and the protocol for prenatal DEX treatment differ between boys and girls, all statistical analyses for DEX effects were first performed separately among boys without CAH, boys with CAH, girls without CAH, and girls with CAH. Because the analyses had to be done for age-limited subsamples as defined by each questionnaire, the resulting sample sizes were often small. To increase statistical power, the analyses were repeated after combining boys and girls into all CAH and all non-CAH groups, and finally also after combining in addition children with and without CAH into a total sample. Although this strategy resulted in a very large number of comparisons, we decided not to use Bonferroni corrections for controlling the number of comparisons, to avoid overlooking potentially important side effects.
The basic comparisons were between DEX-exposed and unexposed children, initially by t test for continuous scales and by 
2 for binary scales. To control for potentially confounding demographic variables, the initial comparisons were followed by stepwise hierarchical regressions for continuous scales and by stepwise hierarchical logistic regressions for binary scales. For all comparisons, a block of demographic variables was entered first, with a selection criterion of P 
0.10 (so that only demographic variables meeting this criterion entered the regression model). The DEX exposed/nonexposed main effect variable was forced into the equation second. The block of demographic variables included age and mean parental education for comparisons involving sex- and CAH-specific groups (e.g. non-CAH boys); age, mean parental education, sex, and ethnicity for comparisons involving the CAH-specific subgroups (e.g. non-CAH boys and girls combined); and age, mean parental education, sex, ethnicity (white vs. other), and CAH status for comparisons involving the total sample.
Treatment duration was correlated with each outcome scale across all subgroups of DEX-exposed children, partialing out for each outcome scale the same demographic variables (age, mean parental education, sex, CAH status, and ethnicity) that were selected for inclusion in the respective stepwise hierarchical regression.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. In terms of ethnicity, 82% of the children were white, 9% Hispanic, and the rest included a variety of ethnic groups. The mean age was 5.55 yr (SD =; 3.46) and the mean midparental education index (Hollingshead) (33) was 5.38 (SD = 1.11) corresponding to partial college education. Few children came from low socioeconomic background. Among the children younger than 6 yr of age, the DEX-exposed children were on average half a year younger than the unexposed (P 0.002, t test) and included significantly more white children (86 vs. 76%, P 0.046, Fisher’s Exact Test). Among the older children, the DEX-exposed were approximately 11 months younger, on average (P 0.001). Duration of prenatal DEX exposure, where available, averaged 29 wk for girls with CAH, 14 wk for control girls, 8 wk for control boys, and 7 wk for boys with CAH. Gestational age at birth, where available, did not differ between the DEX-exposed and -unexposed children.
|
) or the various subsamples. On the RPDQ, which covers the age range 0–5.99 yr, 120 DEX-exposed and 134 unexposed children were compared. None of the comparisons on the delay scale integrated over the four age-based versions of the questionnaire showed statistically significant differences, neither by initial t test nor by stepwise hierarchical regression in the total sample or the subgroups (data not shown). The CDI covers the age range from 15 months to 5.99 yr. Eighty-nine DEX-exposed and 126 unexposed children were compared on each of the nine scales. None of the 63 initial t tests was significant (data not shown). Comparisons by stepwise hierarchical regression also did not show any significant differences in either the total sample (Table 3) or the various subgroups (data not shown). Table 4 shows the partial correlations of the duration of DEX exposure to the questionnaire scales. This analysis was limited to the sample of DEX-exposed children for whom we had data on the duration of their exposure to DEX. For the 20 children with such data for the KIDS, none of the partial correlations reached 0.05 significance (although all five subscales and the full scale uniformly showed an association of longer DEX duration with lower developmental scores). For the 89 children with DEX duration data for the RPDQ, the correlation of DEX duration with the RPDQ delay scale was practically zero. For the 75 children with data on both DEX duration and the CDI, longer DEX duration was significantly associated with fewer developmental delays on three subscales (whereas four of the remaining five subscales and the composite scale showed the same direction of association but without reaching significance). Applying Bonferroni corrections for multiple correlations rendered the three coefficients not significant (data not shown).
|
2 and the stepwise hierarchical (logistic) regressions (data not shown). |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The lack of adverse findings is probably not due to technical aspects of the analysis, namely the use of several demographic control variables in samples of moderate size. Control variables entered a regression model only if they were significantly associated with the dependent scale. For example, 13 of the 18 main regressions for the KIDS included only one or no covariates, which should not render the statistical tests invalid. Furthermore, an examination of the means and SD showed that the groups were very similar, as reflected in the nonsignificant regression P values. Group comparisons by t test with no covariates also were all nonsignificant.
This study has four significant limitations: 1) It was not a randomized clinical trial; 2) we did not test the children with psychological or neuropsychological tests but used mother-administered screening instruments, albeit well standardized; 3) the study response rate of 75.5% compares favorably with the response rates of typical postal questionnaire surveys, largely due to our repeated follow-up contacts with the families, but approximately a quarter of the eligible children with a known address did not participate; and 4) subsample sizes were often small and the corresponding analyses thereby of low statistical power. Despite these limitations, we conclude that the data obtained here make it unlikely that there are marked adverse effects of early-prenatal DEX treatment on motor and cognitive development, if they exist at all. However, replication of the study on other samples appears desirable, and a study of ours involving direct cognitive testing of DEX-exposed children and controls is in progress.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: CAH, Congenital adrenal hyperplasia; CDI, Child Development Inventory; DEX, dexamethasone; KIDS, Kent Infant Development Scale; RPDQ, Revised Prescreening Developmental Questionnaire.
Received July 18, 2002.
Accepted October 29, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Hirvikoski, A. Nordenstrom, T. Lindholm, F. Lindblad, E. M. Ritzen, A. Wedell, and S. Lajic Cognitive Functions in Children at Risk for Congenital Adrenal Hyperplasia Treated Prenatally with Dexamethasone J. Clin. Endocrinol. Metab., February 1, 2007; 92(2): 542 - 548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Forest Recent advances in the diagnosis and management of congenital adrenal hyperplasia due to 21-hydroxylase deficiency Hum. Reprod. Update, November 1, 2004; 10(6): 469 - 485. [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 |
