The place of saliva in antiepileptic drug monitoring.

It has previously been shown that saliva phenytoin concentration bears a constant relationship to plasma free concentration whether protein binding of phenytoin is normal or disturbed by other drugs, pregnancy, renal failure, or hypoalbuminaemia. The present work examines the relationship between saliva (S), plasma free (F), and plasma total (P) concentrations of other anticonvulsants in 100 epileptic patients. Mean S/P ratios were for phenobarbitone 0.37 (r = 0.95), primidone 0.95 (r = 0.87), and carbamazepine 0.27 (r = 0.94). A highly significant correlation of S with F was found for these drugs, more significant than the correlation of S with P for carbamazepine in patients receiving multiple anticonvulsant drugs. Saliva valproate, however, had no predictive value for P or F. No binding to saliva proteins was demonstrated for any drug. Data for in vitro binding to plasma proteins was in good agreement with ex vivo data. Saliva is therefore a valid medium for monitoring treatment with phenobarbitone, primidone, and carbamazepine, as well as phenytoin.