An epidemiological study of the clinical impact of pharmacokinetic anticonvulsant drug interactions based on serum drug level analysis.

The impact of pharmacokinetic anticonvulsant drug interactions on prescribing patterns and serum drug level distribution in a routine clinical setting was evaluated in a population of 848 patients chronically treated with phenytoin, phenobarbital, carbamazepine and valproic acid (either alone or as two-drug combinations) and referred for therapeutic drug monitoring for the first time. While dosages of each drug did not differ significantly between monotherapy and polytherapy patients, significant differences in serum level distribution were found. The proportion of patients with suboptimal serum carbamazepine and valproic acid levels (less than 4 and less than 50 micrograms/ml, respectively) was much greater in the polytherapy than in the monotherapy groups, probably as a consequence of induction of carbamazepine and valproic acid metabolism by combined anticonvulsants. Conversely, the proportion of phenobarbital levels above the upper limit of the optimal range (40 micrograms/ml) was greater among patients receiving phenytoin in combination than among patients taking phenobarbital alone, presumably as a result of phenytoin-induced inhibition of barbiturate metabolism. The therapeutic implications of these findings are discussed.