Abstracts

Rationale: To determine the population pharmacokinetics (PK) of brivaracetam (BRV), and the exposure-response relationship between BRV concentration and daily seizure counts in the adjunctive treatment of partial-onset seizures in adult subjects with epilepsy.Methods: Data from three Phase-3 adequate, well-controlled efficacy trials of BRV were used (N01252, N01253 and N01358). A population PK model with first order absorption and elimination components was fit to the data using NONMEM. The effect of body weight on BRV clearance and distribution volume was implemented using allometric scaling with estimated exponents. A population exposure-response model was developed to describe the relationship between BRV plasma concentration and daily seizure counts. The model described daily seizure counts using a negative binomial distribution taking previous day seizures into account, and using a mixture model to separate a 'placebo like' and a 'responder' subpopulation. A covariate analysis was performed to investigate factors influencing the effectiveness of BRV.Results: The population PK model provided an excellent description of the data, as documented using goodness of fit plots and a visual predictive check (VPC). BRV population clearance and distribution volume were 3.58 L/h and 48.1 L, respectively. Co-administration of enzyme inducers carbamazepine, phenytoin, and phenobarbital decreased BRV exposure (with 95% confidence interval) by -25.8% (-28.2, -23.4), -21.1% (-25.3, -16.7), and -19.3% (-25.0, -13.0), respectively. The population exposure-response model (1549 patients on BRV or placebo) was capable of simulating the observed trial outcomes as shown by VPCs (Figure 1). Covariate analysis indicated that levetiracetam reduced the additional effect of BRV to close to zero, and identified baseline seizure frequency as a strong predictor for being assigned to the mixture-model responder population. The probabilities for ending up in the mixture-model responder population were estimated to be 54.8% at 4 seizures/28 days, 29.3% at 9 seizures/28 days (the median baseline seizure frequency), and 0.8% at 168 seizures/28 days. No other significant covariates influencing BRV effectiveness were detected. Simulation allowed characterization of the dose-response curve suggesting that the maximum response is obtained by BRV 150–200 mg/day (Figure 2).Conclusions: The population exposure-response model allowed the pharmacostatistical description of the relationship between BRV exposure and daily seizure counts reduction from baseline in adjunctive treatment of partial-onset seizures. It supports the effective therapeutic dose range of 50–200 mg/day. Baseline antiepileptic drugs other than levetiracetam did not influence the response. UCB supported