Trial duration and follow-up.

The chronic nature of most seizure disorders requires that antiepileptic drugs (AEDs) be administered for many years. Therefore, new drugs should be evaluated over a period sufficient to assess potential development of tolerance as well as long-term safety. While patients and prescribers desire extensive long-term data prior to registration, too stringent requirements would discourage industry's investment into new treatments, and a compromise between these stakes should be sought. In add-on studies, it is recommended that patients completing double-blind assessment be offered indefinite open-label treatment, and that safety data be collected prospectively. Monotherapy trials in refractory patients include presurgical and conversion to monotherapy studies. Duration of assessment in these studies rarely exceeds few weeks, and endpoints relate to seizure deterioration rather than improvement; this surrogate measure of efficacy is not necessarily predictive of clinical usefulness and, in view of ethical concerns, the rationale for these studies should be questioned. There may be scope, however, for alternative study designs whereby refractory patients are converted to monotherapy and followed up for several months. Active-control long-term monotherapy trials in newly diagnosed patients represent the mainstay for evaluating efficacy and safety. In patients with partial or generalized tonic-clonic seizures, most trials conducted to date had a duration of 6-12 months, but this period may be insufficient to assess seizure freedom rates at optimized dosages, and at least one 2-year trial would be desirable. With respect to open-label follow-up, many protocols require that the randomization code not be broken until the database has been sealed. Patients completing blind assessment are often forced through unblinding procedures which involve tapering of trial medication and institution of an open-label therapy which may differ, in dosage or type of medication, from that taken previously. In view of ethical concerns, the rationale for this practice should also be questioned.