Program

PO2-13-5

Predictive performance of physiologically-based pharmacokinetic model for amiodarone and its active metabolite: evaluation using therapeutic drug monitoring data from Japanese patients

[Speaker] Naoaki Hashimoto:1,2
[Co-author] Kosuke Doki:1,2, Kazutaka Aonuma:3, Masato Homma:1,2
1:Pharmacy, University of Tsukuba hospital, Japan, 2:Department of Pharmaceutical Sciences, Graduate School of Comprehensive Human Sciences, University of Tsukuba, Japan, 3:Department of Cardiology, Graduate School of Comprehensive Human Sciences, University of Tsukuba, Japan

Backgrounds: Therapeutic drug monitoring (TDM) of amiodarone (AMD) and its active metabolite, mono-desethyl-amiodarone (DEA), is practically conducted to avoid adverse effects because of large inter-individual variabilities in their pharmacokinetics, especially in a long-term administration. Physiologically-based pharmacokinetic (PBPK) model for AMD and DEA was currently developed to predict the pharmacokinetics [1]. Although the PBPK model was well-validated using observed plasma concentration data during a 14-week treatment period, it was not evaluated in patients administered for a long-term.
Aim: We evaluated predictive performance of the PBPK model for AMD and DEA in long-term administration of AMD using practical TDM data from Japanese patients.
Methods: PBPK modelling and simulation was employed based on the pre-validated compound files (AMD and DEA) [1] and population file (Japanese) in the Simcyp Simulator (v16.1; Certara). The simulation was performed in virtual individuals receiving multiple-dose of oral AMD with or without loading doses during a year. The simulated concentrations were compared with TDM data from 45 Japanese patients (male/female: 33/12, 64±11 years) receiving multiple-dose of AMD (200 mg/day with and without loading doses of 400 mg/day). Predictive performance was evaluated as frequency of observed concentration data between 5th/95th percentiles of predicted concentrations.
Results: Observed plasma AMD and DEA concentrations from TDM data were 618±344 and 443±188 in short-term (0-91 days, n=38), and 1180±523 and 892±285 in long-term (>91 days, n=68), respectively. The predicted plasma AMD concentrations were successfully recovered using the PBPK model, and consistent with practical TDM data regardless administration period: predictive performance was 95% in short-term and 97% in long-term. The predicted plasma DEA concentrations were successfully recovered using the PBPK model in short-term, whereas were underestimated in long-term compared with practical TDM data: predictive performance was 97% in short-term and 63% in long-term.
Conclusion: Although the PBPK model successfully predicted plasma AMD concentrations, it underestimated plasm DEA concentration in long-term. Further investigation is required to examine other factors influencing pharmacokinetic profile of DEA for long-term administration of AMD.
References:
[1] Chen Y, et al. Drug Metab Dispos (2015), 43: 182-189

Advanced Search