Practical guidelines for plasma concentration monitoring of anticonvulsants are necessary to ensure the appropriate and cost-effective use of these drugs, The guidelines should be based on clinically relevant considerations and recommendations for the drug treatment of epilepsy, including the choice of drugs, Immunoassays are the methods of choice in many laboratories engaged in therapeutic drug monitoring, This assay type has several advantages over other currently used methods, being a precise, reproducible and rapid method for determination of anticonvulsants in micro samples. However, it does not have the same screening potential as chromatographic methods, which can also be used for determination of less common anticonvulsants and new drugs. For several anticonvulsants. there is a more or less well defined therapeutic plasma concentration range. This range, however, must not be strictly interpreted, because many of the studies on which these ranges are based involved patients with severe epilepsy who were treated with several anticonvulsants. Nevertheless, most patients are optimally treated with a drug when steady-state plasma concentrations are maintained within that range. Further studies are necessary to determine the place of plasma concentration monitoring of the newer anticonvulsants. There are several pitfalls in plasma concentration monitoring of anticonvulsants. In practice, the use of plasma drug concentrations requires considerable interpretative skills, a common problem being the interpretation of the therapeutic range. In recent years, plasma concentration monitoring in general has attracted criticism, with suggestions that drug concentrations are being measured unnecessarily or interpreted incorrectly. It is, therefore, of utmost importance that these measurements be requested only on the basis of sound clinical judgement. This will enable therapeutic drug monitoring to be kept as a valuable tool to be used when attempting to treat individual patients most effectively and with the fewest adverse effects. The development and tc sting of new anticonvulsants should include an evaluation of plasma concentration monitoring as early as possible in the course of development. Specially designed clinical studies should be used, preferably in patients receiving monotherapy.