Experimental studies and controlled clinical testing of valproate and vigabatrin

‐ Gamma‐aminobutyric acid (GABA) is the most important inhibitory transmitter, quantitatively, in the CNS. Evidence exists that decreased GABAergic neurotransmission may play a role in some forms of epilepsy. Consequently, manipulating the GABA system may be a therapeutic possibility in the treatment of this disease. Inhibition of the major GABA degrading enzyme, GABA‐transaminase (GABA‐T), seems to be the most promising approach. Currently, 2 antiepileptic drugs, valproate (VPA) and vigabatrin, gamma‐vinyl GABA (GVG), are available, which are supposed to inhibit the degradation of GABA. Both drugs cause an increase in the total concentration of GABA in the brain, but to a different extent. VPA produces a moderate elevation, which seems to be the result of a marked increase in the transmitter‐related GABA pool, while the pronounced elevation in GABA concentration observed during treatment with GVG seems to be caused mainly by an increase in the non‐transmitter‐related (glial) GABA pool. In order to investigate this apparently differential influence of VPA and GVG on the GABA system, a number of studies were undertaken in selectively cultured astrocytes and neurons from mice. For both drugs neuronal GABA‐T proved far more sensitive with regard to inhibition than glial GABA‐T. In order to obtain a more direct measure of a potential GABAergic mechanism of action of VPA and GVG, synaptic release of endogenous GABA was determined after culturing neurons in the presence of clinically relevant concentrations of the drugs. GVG caused a significant increase in GABA release, even at concentrations as low as 25 μM. For VPA only the highest of the investigated concentrations (300 μM) augmented GABA release. It is concluded that the antiepileptic effect of GVG seems to be caused by a direct GABAergic mechanism of action. For VPA an influence on the GABA system may play a role in the antiepileptic effect of the drug. However, the lack of definite data on human brain levels of VPA after chronic treatment, combined with evidence that VPA exhibits a number of other effects that may be relevant for its antiepileptic properties, makes the interpretation of a GABAergic mechanism of action difficult. Controlled clinical trials have been increasingly applied within all areas of medicine. In 1982 a survey of the literature identified 29 studies of antiepileptic drugs, where the design involved randomization, the double‐blind principle and a statistical analysis of the results. A cross‐sectional analysis of these trials demonstrated a number of methodological problems, such as insufficient use of washout periods in crossover designs, heterogeneous classification of seizures and infrequent use of statistics in the evaluation of side effects. On the basis of controlled trials demonstrating the antiepileptic effect of VPA and GVG, methodological considerations in defining so‐called therapeutic serum levels are discussed. The majority of studies that have formed the basis for establishing such levels seem methodologically insufficient, since the individual patient has only infrequently been evaluated at different drug concentrations. A study of VPA, comprising a double‐blind design with multiple crossover, investigated patients at pre‐established serum levels in a randomized sequence. This approach favours the view that comparison of the clinical effect of different serum levels of one drug corresponds methodologically to a comparison of the clinical effect of different drugs. Perspectives for the future development of new antiepileptic drugs and for the clinical testing of these compounds are presented.