This study aimed to evaluate the pharmacokinetics and pharmacodynamics of oral levetiracetam therapy in drug refractory adult epileptic outpatients, as well as factors affecting them. Concentration-time data were collected at steady state, while seizure recurrence was monitored for 13 months. Non-linear mixed effects modeling was applied, and covariates assessed included weight, height, age, daily dose and creatinine clearance.Plasma concentrations of levetiracetam were best described by a one-compartment pharmacokinetic model (V/F = 34.7 L) with first-order absorption (ka = 0.616 h-1) and clearance (CL/F = 3.26 L/h). Patient's CrCL was found to significantly affect levetiracetam clearance (beta = 0.795). Time to seizure occurrence followed an exponential distribution and the mean time to seizure occurrence was estimated Te = 22.08 days. Seizure rate per month followed a Poisson distribution, while mean seizure rate per month was estimated λ = 1.33. Daily dose significantly affected the mean estimated time to seizure (beta = -2.2) and the mean monthly seizure rate (beta = 2.27) in a reverse way. Using discrete time Markov chains, it was shown that the transition probability from focal seizures to focal to bilateral tonic-clonic is significantly altered in relation to patient's CrCL.Simulations showed that dose should be adjusted in relation to CrCL, while low doses of levetiracetam are more effective for seizure control. Modeling and simulation in every-day clinical practice may provide significant information for the optimization of seizure control using well-known agents.
Keywords: Population modeling; epilepsy; levetiracetam; personalized therapy; pharmacodynamics; pharmacokinetics.