Therapeutic drug monitoring: a comprehensive and critical review of analytical methods for anticonvulsive drugs.

In the last 20 years, there has been considerable improvement in the determination of anticonvulsive drugs in body fluids. Gas-liquid chromatography, high-performance liquid chromatography and immunological methods (radio-, enzyme-, fluoro-, and nephelo-immunoassays) have progressively supplanted spectrophotometry and thin-layer chromatography. As the number of publications shows, these methods have included a great variety of procedures. At present, gas-liquid chromatography and enzyme-immunoassays are routinely performed, awaiting a wider spread of liquid chromatography and other immunological techniques. This paper is a comprehensive review of the analytical literature on the determination of phenobarbitone, primidone, phenytoin, carbamazepine, valproic acid, ethosuximide, clonazepam, and some of their metabolites in physiological fluids. Gas-liquid chromatographic methods are more particularly reviewed. In order to facilitate the choice between these procedures, a critical selection of the techniques is given and recommendations are made. Emphasis is laid on technical problems encountered with the assays, as well as the need for a rigorous analytical assessment and internal and external quality controls. This review is completed by considerations on the determination of the free drug fraction and on sample collection and storage.

Dose-dependency of the ratio between carbamazepine serum level and dosage in patients with epilepsy.

Carbamazepine serum levels have been determined by gas-liquid chromatography in 24 children and 26 adults with epilepsy on chronic carbamazepine treatment. A significant correlation has been found between carbamazepine steady-state levels and doses per kilogram body weight in both children (p less than 0.01) and adults (p less than 0.05). This relationship is characterized by a significant decline in the level/dose ratio with the doses for adults (p less than 0.001) and, to a lesser extent, for children (p less than 0.05). These results are consistent with a dose-dependent bioavailability.