• antituberculosis drug;
  • non-linear mixed-effect modelling;
  • population pharmacokinetics;
  • rifampicin


What is known and Objective:  Rifampicin (RIF) shows wide variability in its pharmacokinetics. The purpose of this study was to develop and validate a population pharmacokinetic model to characterize the inter- and intra-individual variability in pharmacokinetic parameters of RIF in Mexican patients.

Methods:  Ninety-four patients receiving antituberculosis therapy participated in this prospective study. Plasma concentration-time data were described using a one-compartment model with lag time, absorption and first-order elimination. The potential influence of demographic and clinical characteristics of the patients, and the pharmaceutical formulation (A, B, C and D) on the pharmacokinetics parameters, was evaluated by non-linear mixed-effect modelling (nonmem). Seventy-seven additional patients participated in the validation of the model.

Results and Discussion:  The final population pharmacokinetic model obtained was as follows: apparent clearance CL/F = 8·17 L/h (1·40 as high for males), apparent distribution volume Vd/F = 50·1 L (1·29 as high for males), absorption rate constant KaA = 0·391/h, KaB,C,D = 2·70/h, relative bioavailability F= 0·468, FB,C,D = 1, lag time in the absorption phase Tlag = 0·264 h. The final model improved the precision on the parameter estimates (CL/F, Vd/F and Ka by 31·9%, 16·7% and 92·9%, respectively). The residual variability was 27·3%.

What is new and Conclusion:  Gender was associated with changes in CL/F and Vd/F whereas the pharmaceutical formulation was associated with changes in F and altered the Ka. The validation data set showed that the model could be used in clinical practice for Bayesian dose adjustment of RIF in TB patients.