The clinical development of a new AED follows a stepwise approach. The preclinical data, including data on pharmacologic actions, toxicology, pharmacokinetics, and interaction profile, should be known before starting studies in humans. Initial human safety and pharmacokinetic data are usually obtained in healthy volunteers. For ethical reasons, the patient population selected for the initial clinical trials is usually represented by subjects with epilepsy who have continuing seizures in spite of several attempts to control the disorder with appropriate AEDs, used in combination and at optimized dosages. Therefore, the first application usually concerns the add-on indication in refractory partial seizures in adults, based on results of randomized, parallel-group, double-blind, adjunctive-therapy, placebo-controlled trials. A sufficient number of adolescents (over 12 years of age) should be recruited in the confirmatory studies. It is useful at this stage to submit a global development plan, showing which extensions are envisioned. In particular, the preclinical profile or some clinical exploratory studies (e.g., a double-blind cross-over study after an open enrichment phase) may suggest that the test drug has a potential for usefulness in other seizure types (e.g., seizure types associated with generalized epilepsies) or other syndromes. Once evidence for a beneficial effect has emerged from studies in adults, it is recommended to implement pediatric studies as planned in the clinical development plan and before registration of the product in adults (see guidelines on investigation of medicinal products in the pediatric population⁶). It may also be useful for the applicant to define at an early stage plans regarding the monotherapy indication, and which patient populations will be investigated for this purpose. A policy regarding elderly patients is also of importance. The development of parenteral formulations is encouraged. Besides bioequivalence and safety data, the demonstration of efficacy in status epilepticus may be an important objective with such formulations, although very little regulatory experience has been acquired in this particular domain.

Patients included in the clinical trials should be classified according to the International Classification of Seizures and, when possible, the International Classification of Epilepsies and Epileptic Syndromes. It is recommended to evaluate the whole spectrum of effectiveness of the test drug, not only according to the type of seizures (e.g., partial vs. generalized seizures), but also to the type of seizures associated with a specific epileptic syndrome. It is acknowledged that patients with partial epilepsies usually represent the initial target, since they are the most prevalent and a substantial percentage of them are not well controlled. Efficacy, however, should be evaluated for all the seizure types present in this condition, including simple, complex partial, and secondarily generalized tonic–clonic (GTC) seizures. A pooled analysis across several studies, optimally prespecified, may permit assessment of efficacy in the less frequent seizure types.

The other syndromes should be explored separately, which is the case for idiopathic generalized epilepsies and symptomatic/cryptogenic generalized epilepsies, including some syndromes specific for infancy and childhood (e.g., West syndrome or infantile spasms, Lennox-Gastaut syndrome, Doose syndrome, Dravet syndrome). Addressing these syndromes will require evaluating the efficacy of an agent separately on the different seizure types present in the given condition (e.g., spasms, GTC seizures, absences, and myoclonic, tonic, or atonic seizures). The impact upon the other clinical features of the disease, such as cognitive outcome, should also be addressed. The global antiepileptic efficacy of an agent in a specific epileptic syndrome can only be claimed when efficacy has been shown for all seizure types associated with that syndrome.

In newly diagnosed patients, who often have a low number of seizures, it may be difficult to identify the seizure type or the syndromic context at the time of inclusion. Some apparently primarily GTC seizures may in fact be secondary to unrecognized partial seizures, particularly in patients with nocturnal seizures. Other GTC seizures may occur in an unrecognized context of idiopathic generalized epilepsy, in the absence of typical electroencephalographic (EEG) traits or other seizure types. The choice of including or excluding undetermined GTC seizures should be made in relation to the expected efficacy spectrum of the test drug, and of the comparator in case of active control trials. Some of these undetermined seizures may be classified later on in the trial, and it might be important to see whether response for these reclassified seizures or syndromes are compatible with the efficacy spectrum of the test product and of the comparator.

Pediatric studies should be conducted as early as the development of the test product allows, in order to avoid an excessive delay in obtaining a marketing authorization for children compared to the marketing authorization for adults. The pediatric needs in epilepsy were assessed by the CHMP Pediatric Expert Group,² which identified the general need for studies of AEDs in refractory epilepsies. Existing data are generally insufficient in two particular domains: (a) pharmacokinetic, efficacy, and safety data of the newer AEDs in refractory epilepsies in young children, particularly under the age of 3 to 4 years, and (b) long-term safety including effects on cognitive function and brain maturation.