Prenatal diagnosis and treatment of 21-hydroxylase deficiency has been utilized for over 15 years and is appropriate in families where a previous family member has been affected (Figure 5).(73) Dexamethasone (20 μg/kg/day in 3 divided doses) is administered to the pregnant mother before the 9th week of gestation, or ideally before the 7th week, blind to the sex or affected status of the fetus. This suppresses excess adrenal androgen secretion and prevents virilization should the fetus be an affected female. As the urogenital sinus formation begins at approximately the ninth week of gestation, (74) prenatal treatment must begin as soon as the pregnancy is confirmed in order to prevent virilization of external genitalia in the affected female. Dexamethasone is used because it crosses the placenta, crossing from the maternal to the fetal circulation.
Figure 5. Algorithm depicting prenatal management of pregnancy in families at risk for a fetus with 21-OHD. (6)
Prenatal diagnosis by DNA analysis (75, 76) requires chorionic villus sampling in approximately the 9th week of gestation. Amniotic fluid cells obtained by amniocentesis in the 16th - 20th week of gestation may also be used for DNA analysis. If the fetus is determined to be a male upon karyotype or an unaffected female upon DNA analysis, treatment is discontinued. Otherwise, treatment is continued to term. This is the first instance of an inborn metabolic disorder to be successfully treated prenatally. The largest human studies have found that dexamethasone administered in proper doses at or before the 9th week of gestation is effective in reducing virilization in the genetic female so that genitoplasty was not needed in the majority of cases (Figure 6). (6, 77, 78)
Figure 6. Genitalia of prenatally treated (bottom) and untreated (top) siblings with Classical CAH owing to 21-hydroxylase deficiency. The prenatally untreated girl (top) has virilized genitalia with an enlarged clitoris and scrotalization of the labia majora (Prader IV). The prenatally treated sib was born with normal female genitalia.
Since prenatal treatment became available, the majority of studies have proven it to be effective and safe for both mother and child, provided patients and physicians adhere to the recommended therapeutic protocol. (6) Some results from animal studies, in which excess glucocorticoid dosages were used, suggest there are maternal complications of glucocorticoid treatment.(79) However, results from one of the largest studies in humans, indicate there are no enduring maternal complications (i.e. edema, striae, excessive weight gain, spontaneous abortion).(6) In fact, all maternal complications disappeared after delivery and all mothers contacted (save three, who did not desire another child) stated that they would repeat their care with dexamethasone if they were pregnant again.
To date, the majority of studies show no major effect of glucocorticoid therapy on fetal birth weight, growth or development; treated newborns have been shown to reach developmental milestones at the appropriate ages. (6, 77, 78) In a study by Lajic et al., (78) in which only 44 pregnancies were studied as compared to 595 reported by New et al.,(80) although weight gain by the mother increased in the first trimester, overall weight gain was not different from untreated. There was no increase in spontaneous abortions. In monkeys, degeneration of the hippocampus was observed with administration of dexamethasone to the mother. However, the dose given was 200 times that used in human prenatal treatment. Cognition and postnatal growth reported by Lajic et al. were normal in humans. A long term follow-up study of 44 children treated prenatally in Scandinavia demonstrated normal pre- and post natal growth compared to matched controls.(78) In our study, prenatally treated newborns also did not differ in weight, length, or head circumference from untreated, unaffected newborns.(80, 81)
In one study by Forest et al., only 2 of 37 dexamethasone-treated newborns were abnormally small, and in both cases, it was undetermined if it was treatment-related or a result of underlying factors (i.e. the parents being undersize).(77) Studies undertaken with monkeys only demonstrated fetal growth retardation in the range of 40-160 g/kg/d, a dose 2-8 times that of the highest human dose. (77) However, the largest human studies have shown that birth weight, birth length and head circumference were not significantly different in the offspring of dexamethasone treated pregnancies and those not treated; the average difference in birthweight was only 300 grams, a clinically insignificant value. (6, 77, 82, 83)
Miller and Seckl raise the specter of prenatal dexamethasone therapy leading to hypertension in adulthood. (79) It is not clear, however, that hypertension is the result of low birth weight or large placental size, as previous studies arrive at contradictory conclusions. One study with 449 subjects concluded that there was no correlation with adult blood pressure and birth weight alone, but that low birth weight and large placental weight together were predictors of adult high blood pressure.(84) The highest adult blood pressures were found in those who were small babies born with large placentas. (84) In another study of 1511 subjects, in which they adjusted for height and body mass index, Whincup et al. showed that birth weight had a smooth graded inverse relation with blood pressure in 9-11 year olds. (85) In Whincup’s report, subjects with low birth weight and high placental weight did not have particularly high mean blood pressures, contradicting the former Barker report.(84) Thus, it is difficult to correlate low birth weight with hypertension, as these studies examine low birth weight due to malnutrition and placental data is often not included. (84, 86-89)
Miller and Seckl also state the possibility that dexamethasone prenatal treatment contributes to fetal death, but this claim is unsupported by other authors.(90) In the 1993 report by Forest et al., there is strong evidence to the contrary.(83) In this study, there was 9% incidence of spontaneous abortion of 64 treated pregnancies, while 14% incidence in untreated pregnancies, suggesting that fetal death and spontaneous abortion cannot be attributed to dexamethasone treatment in every case.
Despite the claim that surgical reparation is necessary in CAH patients, (90) in the large experience of our CAH patients who were prenatally treated with dexamethasone, none to date have required reparative surgery of the genitalia as long as compliance has been assured. Cases of unsuccessful treatment are almost always correlated with a late treatment start date, too small a dosage of dexamethasone, or interrupted treatment. Therefore, the solution for effective outcome is to monitor it meticulously. When treatment is properly administered, genitoplasty has not been necessary. It should be clear that there are no long-term studies to date of surgical correction of ambiguous genitalia.
It has also been reported previously that children treated prenatally with dexamethasone are adversely affected cognitively and developmentally. (91) However, the discussion in the paper in which Trautman et al. state that the differences they found in treated versus untreated were not necessarily negative changes, but differences. Trautman et al. also state that these results should be considered preliminary until they can be replicated by a larger sample. A recent survey of 174 prenatally dexamethasone-exposed children, ages 1 month to 12 years, (including 48 with CAH) compared to 313 unexposed children (including 195 with CAH) found no differences in cognitive or motor development between the two groups.(92) Therefore, we believe that proper prenatal treatment of fetuses at risk for CAH can be considered effective and safe. Thus, long-term follow-up studies are necessary to evaluate cognition, gender, temperament, and handedness (an indicator of prenatal androgen effect) in children and adults prenatally treated in order to evaluate the long-term consequences of prenatal dexamethasone treatment. These studies are being prepared for publication. The oldest known patient is now 20 years of age, and she suffers no cognitive or developmental defects (New, pers. com.).