Cruz JP(1,2), Aniceto N(1,2), Ribeiro AC(3), Paixão P(1), Morais JG(1).
(1) iMed.UL, Faculty of Pharmacy, University of Lisbon. (2) Laboratory for Molecular Diagnosis of Infeccious diseases, Santa Maria Hospital, Lisbon, Portugal. (3) CiiEM, Egas Moniz Institute for Health Sciences.
Objectives: Efavirenz, used in the treatment of HIV, belongs to the Non-Nucleoside Reverse Transcriptase Inhibitors class.1 Currently, great variability is observed in the population pharmacokinetics (popPK) which hinders therapeutic management, and liver metabolism has been considered one of the main sources for this variability.2 In order to establish the determining factors affecting popPK variability, we aim to construct a population pharmacokinetic-pharmacogenetic (PKPG) model by nonlinear mixed effects modeling (NLMEM) from sparse data, and accounting for CYP2B6 516G>T mutation (CYP2B6_SNP). Bayesian estimation of the popPK parameters was also performed.
Methods: For model construction a dataset of 15 patients with single-point high drug concentrations, demographic and biochemical data, as well as SYP2B6_SNP characterization was used. This was done using MONOLIX 4.2.2. A one compartment open model with first order absorption was used as the structural model, parameterized in terms of Clearance (CL/F) and Volume of distribution (V/F). The covariate model was built by fitting the data to each available variable and selecting the ones that improved model fitting. The selected variables were added in a stepwise manner. In parallel, Bayesian estimations of CL/F and V/F were achieved using PKS systems3.
Results: The covariate model included CYP2B6 mutation and total bilirubin (Bil) as covariates of CL/F, both significant to the model. However no significant covariate was found for V/F. The final popPK parameters were 68.3 L ± 20, and 4.32 L.h-1 ± 0.81 for V/F and CL/F respectively, and the covariate model for CL was: CL= -0.291*(CYP2B6_SNP)+0.971*Bil. Interindividual and residual variability in CL/F was improved in the covariate model as compared to the basic model (13% to 8%, and 26.6 to 12.6%, respectively). Good data fitting was demonstrated with the goodness of fit diagnostic plots.4 Bayesian estimates of CL/F and V/F were 5.9 L.h-1 ± 2.4 and 282 L ± 0.5, respectively.
Conclusions: The model was improved with Bil and CYP2B6_SNP as covariates, demonstrating the important role of this enzyme in the pharmacokinetics of Efavirenz. The covariate model partially addressed the data variability. Bayesian estimation yielded similar CL/F values to those from the NLME basic model (5.45 L.h-1). However, V/F values from the former and the latter (96.1 L) were very different, which could stem from the high impact of initial Bayesian estimates.
References:
[1] Béthune M-P. Non-nucleoside reverse transcriptase inhibitors (NNRTIs), their discovery, development, and use in the treatment of HIV-1 infection: A review of the last 20 years (1989–2009). Antiviral Research. 2013. 85(1):75-90.
[2] Fabbianni M, Giambenedetto SD, et al. Pharmacokinetic variability of antiretroviral drugs and correlation with virological outcome: 2 years of experience in routine clinical practice J. Antimicrob. Chemother. 2009. 64 (1): 109-117.
[3] Lacarelle B, Pisano P, et al. Abbott PKS system: a new version for applied pharmacokinetics including Bayesian estimation. Int J Biomed Comput. 1994. 36(1-2):127-30.
[4] Mould DR and Upton RN. Basic Concepts in Population Modeling, Simulation and Model-Based Drug Development – Part 2: Introduction to Pharmacokinetic Modeling Methods. CPT Pharmacometrics Syst. Pharmacol. 2013. 2, e38.
Reference: PAGE 23 (2014) Abstr 3251 [www.page-meeting.org/?abstract=3251]
Poster: Drug/Disease modeling - Infection