Developmental neurotoxicity of anticonvulsants: human and animal evidence on phenytoin.

Most epileptic women delivering children each year take anticonvulsants throughout pregnancy. The teratogenic potential of anticonvulsants is most notable for phenytoin, trimethadione, valproic acid, and carbamazepine. This review focuses on the human and animal evidence for the teratogenicity of phenytoin, with emphasis on neurobehavioral end points. The Fetal Hydantoin Syndrome (FHS) consists of craniofacial defects and any two of the following: pre/postnatal growth deficiency, limb defects, major malformations, and mental deficiency. Available data suggest a prevalence of FHS of 10-30% in infants of women ingesting 100-800 mg/kg of phenytoin during the first trimester or beyond. Unfortunately, data on neurobehavioral development in FHS children is limited. Animal models of FHS have been developed and those focusing on neurobehavioral effects are reviewed. Phenytoin produces multiple behavioral dysfunctions in rat offspring at subteratogenic and nongrowth retarding doses. These behaviorally teratogenic doses produce maternal serum phenytoin concentrations in rats comparable to those found in humans. The dysfunctions in rats are dose-dependent and exposure-period-dependent, but independent of nutritional, maternal rearing, or seizure disorder confounds. Effects include vestibular dysfunction, hyperactivity and deficits in learning and memory. General comparability between the human and animal findings for phenytoin are apparent, however, difficulties with existing studies prevent precise comparisons. Animal studies have not dealt satisfactorily with the potential contribution of epileptic disease state to the FHS, with fetal brain drug concentration determinations, a complete dose-effect range, effects in multiple species (although limited nonhuman primate data exist), site of CNS injury, and the comparability of end points assessed. Human studies have not dealt satisfactorily with issues of the need for prospective study designs, separation of the effects of different anticonvulsants, or adequate long-term follow-up of cases, especially with attention to neuropsychological assessment.