Abstracts

Rationale: Data gathered during routine clinical use of a drug are an important companion to data from randomised clinical trials, and analysing retention over time in subgroups can help to inform the optimal use of an antiepileptic drug (AED). We collected observational data of perampanel (PER) use in routine clinical practice from investigators across Europe, to explore safety, retention, and seizure outcomes in a large dataset. Methods: Patient-level data from 44 sites across Europe were merged into a single dataset. One-year retention rate, 1-year seizure-freedom rate, and incidence of adverse events (AEs) have been reported elsewhere. Survival analysis was conducted to explore retention over time on PER and Kaplan-Meier survival curves were generated to visually examine the impact of baseline characteristics on survival probability (retention). Results: The full analysis set comprised 2396 people prescribed PER for focal and generalized seizures. Of N=2332 with retention data, 49.5% continued to take PER at 1 year; and 9.2% of N=803 with seizure freedom data were seizure free at 1 year (Rohracher et al. IEC 2017). Survival curves were generated for: individuals with fast vs slow PER titration (fast=2mg every 2 weeks or quicker); early vs late PER add-on (early=2 or fewer previous AEDs); number of concomitant AEDs (≤3 vs >3); and mode of action of concomitant AEDs. Separation between the curves was seen only for titration rate (N=959, Fig 1) and timing of PER add-on (N=1871, Fig 2). There was a tendency for higher survival probability with slow vs fast titration, but fast titration (N=830) was 6-times more common than slow (N=129). Survival probability was also higher with early vs late add-on, but late add-on (N=1720) was over 10-times more common than early (N=151). Multivariate analysis will be conducted to further explore predictors of outcomes with PER. Conclusions: These data are consistent with clinical experience emerging over the past 5 years with PER. To maximise the likelihood of people persisting with PER, titration should be slow (2mg increments at intervals of >2 weeks) rather than fast. Survival probability may also be higher when PER is added earlier in the treatment pathway (after failure of only 1 or 2 AEDs, vs 3 or more prior AEDs), although this is confounded by a less refractory epilepsy population at this treatment stage. Significance: These results warrant a change in the way PER is initiated (slow titration, earlier add-on), to increase the likelihood that individuals will remain on PER treatment over the long term (≥12 months). Funding: none