Population Pharmacokinetics of Rimonabant in Obesity
Jean-Marie MARTINEZ, David FABRE & Vanaja KANAMALURU
sanofi-aventis
Objectives: Rimonabant is an orally active selective antagonist for the cannabinoid CB1 receptor which has been developed in the treatment of multiple cardiometabolic risk factors in overweight/obese patients. The objective of this analysis was to develop and qualify a PopPK model for rimonabant and to investigate the influence of several covariates on the PK of rimonabant in the population enrolled in phase III.
Methods: The PopPK analysis was performed using NONMEM on plasma samples from more than 3000 obese patients enrolled in the two long term Phase III studies (RIO Europe and RIO North America) and receiving rimonabant for two years as once daily 5 or 20 mg doses. Different structural PK models (one- and two- compartments models, with and without lag time) and different residual error models (additive, proportional, power, and combined) were evaluated. The interindividual error model was exponential for all parameters. The relationship between individual estimates and covariates was then investigated. Selected covariates were added individually (forward selection method) to the model and tested for statistical significance. Before validation, a model verification was performed by an examination of the goodness of fit plots, mean weighted residuals and by estimation of the several quality criteria such as bias, precision or Average Fold Error. Once the PopPK model predictive ability validated (using different approaches such as Prediction errors, Bayesian estimation, validation from parameters, the individual PK parameter estimates and some derived parameters were calculated for all patients using the final model.
Results: A two-compartment model best fitted the data (~14000 concentrations obtained from more than 3000 patients). Of the covariates evaluated, African ethnicity showed an influence on rimonabant clearance, with 1.75‑fold higher values in black than in non-black obese patients. Peripheral and central volumes of distribution of rimonabant were dependent on body weight, with higher values in more obese patients. Central volume of distribution of rimonabant was influenced by age, with higher values in older patients. The PK of rimonabant in obese patients did not depend on sex, renal function or dose over the dose range 5-20 mg. Inter-patient variabilities in rimonabant clearance, central and peripheral volumes were about 46, 62 and 18%, respectively. The residual variability (intraindividual variability) ranged from 16% (at the LOQ value, 5 ng/mL) to 9% for the maximal observed concentration.
Conclusions: A PopPK model was developed and validated for rimonabant data obtained from obese patients. It showed good agreement with observed plasma concentrations and was used to assess exposure parameters (AUC0-24, Cmax and Cmin at steady state) for each patient.
