What clinical trial designs have been used to test antiepileptic drugs and do we need to change them?

Designs used to evaluate the efficacy and safety of antiepileptic drugs (AEDs) have evolved considerably over the years. A major impulse to develop methodologically sound randomised controlled trials dates back to the Kefauver-Harris Drug Amendment of 1962, through which the US congress introduced the requirement of substantial evidence for proof of efficacy in a new drug application. The mainstay for the initial approval of most new AEDs has been, and still is, the placebo-controlled adjunctive therapy trial, which evolved over the years from the cross-over to the parallel-group design. In the early days, when few AEDs were available, enrolment of patients into these trials was relatively easy and prolonged placebo exposure could be justified by lack of alternative treatment options. With more than 20 drugs now available to treat epilepsy, however, exposing patients to placebo or to a potentially ineffective investigational agent faces practical and ethical concerns. Recruitment difficulties have led sponsors to markedly increase the number of trial sites, but there is evidence that this may adversely affect the ability to differentiate between effective and ineffective treatments. Methodological and practical difficulties are also encountered with monotherapy trials. Because regulatory guidelines for monotherapy approval differ between Europe and the US, sponsors need to pursue separate and costly development programs on the two sides of the Atlantic. Moreover, the scientific validity of the monotherapy trial paradigms currently used in Europe (the non-inferiority design) and in the US (the conversion to monotherapy design with historical controls) has been questioned. This article will review these issues in some detail and discuss how trial designs and regulatory approval processes may evolve in the future to address these concerns.