In the context of a prospective study on the adverse effects of anti-epileptic drugs on fetal outcome, we evaluated our experience with prenatal diagnosis by ultrasonography and alpha-fetoprotein (AFP) determination in amniotic fluid. We compared these results with AFP values in maternal serum obtained prior to amniocentesis. From November 1985 to July 1990, amniocentesis at 16-18 weeks of gestation was performed in 267 pregnancies of 237 different women using anti-epileptic drugs. Among 92 pregnancies with maternal valproic acid use, five (including one concordantly affected monozygotic twin-pair) were terminated because of a spina bifida aperta, all prenatally diagnosed by AFP determination and acetylcholinesterase electrophoresis in amniotic fluid. The maternal serum AFP level was raised (> or = 2.5 multiples of the median (MOM) for singleton pregnancies and > or = 4.5 MOM for twin pregnancies) in only two of these five affected pregnancies. We emphasize that maternal serum AFP levels may be unreliable for prenatal screening for fetal neural tube defects in women taking valproate and recommend that amniocentesis and fetal ultrasound examination should be offered directly.

Exposure to antiepileptic drug (AED) treatment in utero occurs in 1 of every 250 newborns. The absolute risk of major malformations in these infants is about 7-10%, approximately 3-5% higher than in the general population. Specific risk factors include high maternal daily dosage or serum concentrations of AED, low folate levels, polytherapy, and generalized seizures during pregnancy. Adverse pregnancy outcomes, including congenital heart malformations, facial clefts, spina bifida aperta, hypospadias, growth retardation, and psychomotor and mental retardation, are associated with, although not necessarily caused by, AED exposure. Specific cognitive defects, hypertelorism, and nail hypoplasia can be causally related to specific AED exposures. To prevent teratogenic side effects, the prospective mother should be treated with AEDs only when absolutely necessary. Monotherapy with the AED that is most effective in the lowest possible daily dose (divided into at least two or three administrations) should be prescribed. High-dose folate supplementation (4-5 mg/day) reduces the risk of a neural tube defect in a child whose sibling had such a defect, but its impact on the specific teratogenic risks of AEDs is unknown. A substantial proportion of fetal malformations may be secondarily prevented by prenatal diagnosis, consisting of a fetal structural ultrasound examination at weeks 18 and 20 of gestation and, with VPA or CBZ administration, an alpha 1-fetoprotein analysis of amniotic fluid at week 16. Determination of a specific defect prevention strategy depends largely on parental attitudes toward prenatal diagnosis and termination of pregnancy, which should be discussed before conception.(ABSTRACT TRUNCATED AT 250 WORDS)

Summary: One in every 250 newborns is exposed to antiepileptic drugs (AEDs) in utero. Various studies have attributed a teratogenic effect to these AEDs, mainly consisting of major malformations, minor anomalies, intrauterine or postnatal growth failure, and psychomotor retardation. Prospective studies confirm the increased risk of major malformations. The absolute risk of 7–10% is about 3–5% higher than that in the general population. Barbiturates and phe‐nytoin (PHT) are particularly associated with congenital heart malformations, facial clefts, and some other malformations. Valproate (VPA) and carbamazepine (CBZ) are associated predominantly with spina bifida aperta (1–2 and 0.5–1.0% risk, respectively) and hypospadias. Bilateral radial aplasia is a rare but specific effect of VPA. Several studies identified additional risk factors, i.e., high daily AED dosage, high maternal serum AED concentrations, low folate levels, or po‐lytherapy [phenobarbital (PB) plus primidone (PRM) plus PHT or CBZ plus VPA plus PB with or without PHT]. The few prospective studies controlled for socioeconomic factors or that considered parental findings indicate that risk of specific cognitive defects rather than risk of overall mental retardation may be increased, that early growth retardation is followed by a catch‐up growth to normal, and that ocular hypertelorism and nail hypoplasia are the only minor anomalies causally related to PHT exposure. However, no final conclusions can be made. Genetic predisposition to the teratogenic side effects of AEDs plays a role, codetermining the recurrence risk if the woman has previously given birth to a child with a major malformation. The molecular genetic basis of this predisposition is unclear. No tests are yet available for identifying parents or fetuses at particularly high risk. Prevention of teratogenic AED side effects is possible by critical evaluation, before pregnancy, of the woman's need for AED therapy. If AED therapy cannot be avoided, it should be monotherapy with the lowest possible dosage. High peak levels should be avoided by dividing the daily dosage into at least two or three doses. The efficacy of folate supplementation alone in reducing teratogenic risks is unclear. In cases of obvious folate deficiency, treatment is required, but concomitant vitamin B12 deficiency should first be excluded or treated. Prenatal diagnosis should be offered. It may consist of fetal ultrasound examination during Week 18–20, and of afetoprotein analysis of amniotic fluid obtained during Week 16 if the mother is receiving VPA or CBZ. This should be discussed with the parents before the pregnancy. The choice of medication is made not only by balancing the therapeutic advantages and teratogenic risks of each AED or AED combination but also by considering the parental attitudes toward prenatal diagnosis and termination of pregnancy.

We have studied 52 pregnancies in epileptic women taking long-term valproate and have measured the concentrations of the parent compound and 13 of its metabolites by gas chromatography-mass spectrometry in amniotic fluid, maternal serum, and 24 h maternal urine samples. All metabolites of valproate present in the serum could also be detected in the amniotic fluid, although at much lower concentrations. Amniotic fluid concentrations of valproate and several of its metabolites ((E) delta 2-valproate, (2E,3'E) delta 2,3'-valproate, and 3-keto-valproate) correlated with total valproate concentrations as well as with unbound valproate concentrations in maternal serum. We suggest that the amniotic fluid acts as a deep compartment, with slow appearance and disappearance of valproate and its main metabolites. The data further suggest that during the first and early second trimesters of pregnancy the beta-oxidation of valproate decreases. In pregnancies associated with fetal neural tube defects (n = 5) significantly higher daily doses of valproate were used compared with normal pregnancies (n = 47). This resulted in higher concentrations of valproate in maternal serum. However, the metabolite patterns in maternal serum, 24 h urine samples, and amniotic fluid did not show any significant differences in pregnancies with neural tube defects.