Population plasma pharmacokinetics of ¹¹C-flumazenil at tracer concentrations

Objective: The objectives of the study were to develop a population pharmacokinetic model for ¹¹C-flumazenil at tracer concentrations, to assess the effects of patient-related covariates and to derive an optimal sampling protocol for clinical use. Methods: A population pharmacokinetic model was developed using nonlinear mixed effects modelling (NONMEM) with data obtained from 51 patients with either depression or epilepsy. Each patient received ∼370 MBq (1-4 μg) of ¹¹C-flumazenil. The effects of selected covariates (gender, weight, type of disease and age) were investigated. The model was validated using a bootstrap method. Finally, an optimal sampling design was established. Results: The population pharmacokinetics of tracer quantities of ¹¹C-flumazenil were best described by a two compartment model. Type of disease and weight were identified as significant covariates (P < 0.002). Mean population pharmacokinetic parameters (percent coefficient of variation) were: CL 1530 ml_ min^-1 (6.6%), V₁ 24.8 × 10³ mL (3.8%), V₂ 27.3 × 10³ mL (5.4%), and Q 2510 mL min^-1 (6.5%). CL was 20% lower in patients with epilepsy, and the influence of weight on V₁ was 0.55% kg ^-1. For the prediction of the AUC, a combination of two time points at t = 30 and 60 min post injection was considered optimal (bias -0.7% (95% CI -2.2 to 0.8%), precision 5.7% (95% CI 4.5-6.9%)). The optimal sampling strategy was cross-validated (observed AUC = 296 MBql^-1 min^-1 (95% CI 102-490), predicted AUC = 288 MBql^-1 min^-1 (95% CI 70-506)). Conclusions: The population pharmacokinetics of tracer quantities of ¹¹C-flumazenil are well described by a two-compartment model. Inclusion of weight and type of disease as covariates significantly improved the model. Furthermore, an optimal sampling procedure may increase the feasibility and applicability of ¹¹C-flumazenil PET.