Summary: Purpose: When monotherapy with antiepileptic drugs (AEDs) fails, combination therapy is tried in an attempt to improve effectiveness by improving efficacy, tolerability, or both. We reviewed the available studies (both animal and human) on AED polytherapy to determine whether AEDs can be selected for combination therapy based on their mechanisms of action, and if so, which combinations are associated with increased effectiveness. Because various designs and methods of analysis were used in these studies, it was also necessary to evaluate the appropriateness of these approaches.
Methods: Published papers reporting on AED polytherapy in animals or humans were identified by Medline search and by checking references cited in these papers.
Results: Thirty-nine papers were identified reporting on two-drug AED combinations. Several combinations were reported to offer improved effectiveness, but no uniform approach was used in either animal or human studies for the evaluation of pharmacodynamic drug interactions; efficacy was often the only end point.
Conclusions: There is evidence that AED polytherapy based on mechanisms of action may enhance effectiveness. In particular, combining a sodium channel blocker with a drug enhancing GABAergic inhibition appears to be advantageous. Combining two GABA mimetic drugs or combining an AMPA antagonist with an NMDA antagonist may enhance efficacy, but tolerability is sometimes reduced. Combining two sodium channel blockers seems less promising. However, given the incomplete knowledge of the pathophysiology of seizures and indeed of the exact mechanisms of action of AEDs, an empirical but rational approach for evaluating AED combinations is of fundamental importance. This would involve appropriate testing of all possible combinations in animal models and subsequent evaluation of advantageous combinations in clinical trials. Testing procedures in animals should include the isobologram method, and the concept of drug load should be the basis of studies in patients with epilepsy.