Pharmacokinetic data from 48 children who were taking valproic acid were analyzed by multiple step‐wise linear regression. Children who were receiving enzyme‐inducing antiepileptic drugs(n= 27) had greater(p<0.01) clearances, elimination rates, and dosage requirements and greater(p<0.05) variability in pharmacokinetic values than patients receiving monotherapy. Age and polytherapy explained most of the interpatient variability in total (r2= 0.80;p<0.001) and intrinsic (r2= 0.77;p<0.001) clearances and the elimination rate (r2= 0.61;p<0.002). Free fraction variability was related to valproate concentration and phenobarbital (r2= 0.47;p<0.001). Distribution volume variance was associated with free fraction (r2= 0.48;p<0.001). The effect of age and polytherapy on valproate clearance is primarily attributable to changes in metabolism rather than in protein binding. Valproic acid dosage requirements are greater and more variable for children who are receiving other enzyme‐inducing antiepileptic drugs.Clinical Pharmacology and Therapeutics(1993)53,22–29; doi:10.1038/clp