Summary: Despite the current trend toward monotherapy, polytherapy in children with epilepsy is still common. A drug combination is advantageous only if it achieves a higher efficacy: toxicity ratio (therapeutic index) or if its antiepileptic spectrum is wider. Studies of brain concentrations of antiepileptic drugs have so far shown that a higher efficacy toxicity ratio is not achieved by most combinations. Problems are associated with drug combinations. First, numerous pharmacokinetic interactions are documented. These interactions, which can be associated with significant changes in blood levels at a given dose, make frequent measurements and dosage readjustments necessary. They can also alter the concentration of active metabolites or the free fraction of a drug. Second, toxicity can be assumed to be at least partially cumulative, since reduction in polytherapy has been shown to be associated with a reduction in side effects. Third, the therapeutic range appears to depend on whether a drug is taken alone or in combination, so that polytherapy confuses the interpretation of serum drug measurements. Fourth, the presence of more than one drug will add to the difficulty in evaluating the efficacy or side effects of any single drug. Finally, a pharmacodynamic interaction between valproate and several other antiepileptics, particularly the barbiturates, can lead to a stuporous state. Transition from polytherapy to monotherapy is much more difficult to achieve than the opposite.