Abstracts

Rationale: To develop a combined dose-response model for levetiracetam immediate-release (LEV IR) and LEV extended-release (LEV XR) in adults suffering from partial onset seizures. To undertake simulations for comparing the reduction in seizure frequency between 1000 mg once daily (o.d.) LEV XR and 500 mg twice daily (b.i.d.) LEV IR. Methods: Dose-response analysis was carried out using data from four (LEV IR) and one (LEV XR) randomized, double-blind, placebo-controlled, efficacy trials of adjunctive levetiracetam in patients with partial epilepsy. The trials consisted of an 8- or 12-week prospective baseline followed by an up-titration period (IR only), and a stable-dose evaluation period of 12 to 24 weeks at 1000, 2000, 3000 or 4000 mg/day b.i.d. for LEV IR and 1000 mg/day o.d. for LEV XR. Modeling of seizure counts was performed by nonlinear mixed effects modeling in NONMEM VI with the Laplace estimation method. Individual seizure frequency was modeled as a Poisson process, expressed as a function of baseline seizures, drug treatment, placebo effect and subject-specific random effects. The Mixture function was used for partitioning the population in two sub-groups of patients experiencing decreased or increased seizure frequency compared with baseline. In the first group, the drug effect was modeled using an Emax dose-response function on top of the placebo effect, whereas the change in seizure frequency in non-improving patients was dose-independent. Distributions of reduction in seizure frequency between LEV IR, LEV XR and placebo were compared by simulating 40,000 patients per arm. Results: The combined IR (950 patients) and XR (153 patients) data were well described by a mixture model of improving and deteriorating patients. The proportion of patients improving on LEV appeared to be higher for XR than IR (88% vs 73%). The dose-response relationship in patients improving on LEV was adequately fitted by an Emax model, with an ED50 of 1617 mg/day (XR) vs 2139 mg/day (IR). The placebo effect was also higher for XR than IR, with 68% vs 58% of improving patients. A daily dose of 1000 mg levetiracetam XR is predicted to result in a reduction in seizure frequency of ≥49% in half of all patients. At the same dose, a reduction in seizure frequency of ≥27% in half of all the patients is predicted for IR. However, a parallel shift is apparent in the placebo group for the two formulations. Treatment effect, defined as the difference between median reduction from baseline in seizure frequency under placebo and under LEV 1000 mg/day, is predicted to be 16% for both IR and XR formulations. Conclusions: A combined dose-response model on levetiracetam IR and XR formulations adequately describes seizure counts data. The treatment effect over placebo is predicted to be similar for levetiracetam extended-release 1000 mg/day once daily and levetiracetam immediate-release formulation 1000 mg/day b.i.d intake. Study funded by UCB.