Abstracts

Rationale: Seizure misclassification in randomized clinical studies of antiepileptic drugs (AEDs) can negatively impact trial outcomes. To ensure homogeneity and accuracy of seizure classification in clinical trials, The Epilepsy Study Consortium has created a seizure training program for study investigators and site coordinators to enhance the understanding of seizure types and provide independent verification of seizure classification. This prospective training has previously been used in AED clinical trials, including PREVAIL, a recently completed global, phase 3 study evaluating the efficacy and safety of USL255, once-daily extended-release topiramate, as adjunctive therapy in subjects with partial-onset seizures (NCT01142193).Methods: The PREVAIL training program included a seizure identification video providing in-depth descriptions of seizure types and advice for patient interviews. The video was moderated by a Consortium member at the investigator meetings and posted online for those unable to attend. After the video, a quiz was administered and the answers discussed in real time. PREVAIL was the first trial to implement this quiz with the requirement for retraining if a passing score ( 70%) was not achieved. Additionally, each site was required to submit a Seizure Identification Form (SIF) for all subjects, which included descriptions of seizures by the subject and/or caregiver, and the investigator s classification(s) of seizure(s). The SIF was reviewed independently by the Consortium immediately after the Screening Visit and was to be approved prior to randomization. To ensure seizures were captured correctly in the case report forms (CRFs) on subsequent study visits, misclassifications were communicated to the study team to allow for subject retraining (as necessary). CRFs were checked by the sponsor/CRO to confirm that classifications matched those approved by the Consortium. Additionally, subject diaries were assessed by centralized CRO reviewers to ensure qualifying seizure type and counts met protocol-specified inclusion criteria.Results: The Epilepsy Consortium reviewed and reconciled 308 SIFs; only 1 subject was deemed unsuitable for PREVAIL. Eleven subjects (3.57%) had 12 misclassified seizures; a rate lower than previously reported by the Consortium (14% [2010], 10% [2013]) for other trials. During PREVAIL, complex partial seizures (CPS) were the most commonly misclassified, representing 8 (66.7%) of the 12. Five CPS were incorrectly identified as secondarily generalized tonic-clonic seizures (SGTC), 2 as simple partial seizures (SPS) w/motor signs, and 1 as SPS w/out motor signs. The 4 remaining misclassifications included 2 cases each of SGTC and SPS w/motor signs.Conclusions: Improved investigator training, in conjunction with independent expert review by The Epilepsy Study Consortium, resulted in lower rates of seizure misclassifications in the USL255 PREVAIL phase 3 clinical trial. Supported by Upsher-Smith Laboratories, Inc.