Suein Choi (1,2), Seunghoon Han (1,2), Dong-Seok Yim (1,2)
(1) PIPET (Pharmacometrics Institute for Practical Education and Training), College of Medicine, The Catholic University of Korea; (2) Department of Pharmacology, College of Medicine, the Catholic University of Korea
Objectives: In many clinical trial situations involving a multiple dose administration, the steady-state full pharmacokinetic (PK) study is emphasized to predict PK parameter while the trough samplings are only used for compliance check and often omitted for various reasons such as the inconvenience of the subjects and increase of total research budget. However, particularly for drug with long half-life, multiple dose full PK data obtained during a single inter-dose interval (II) after 3 half-lives may be insufficient to properly estimate the PK. In this study, using stochastic simulation and estimation method, we were to evaluate the contribution of trough concentration data for PK evaluation in such situations, and to obtain the most informative sampling timepoint with athe appropriate number of sampling
Methods: PK parameters and inter- and intra individual variation were obtained from the bioequivalence study data of amlodipine using nonlinear mixed effect modeling software NONMEM. Amlodipine PK model was assessed by goodness-of-fit plots and precision of parameters were evaluated by bootstrap analysis. Once amlodipine PK model was developed, stochastic simulation and estimation were performed using the predicted PK profile to evaluate the efficiencies of different trough sampling scenarios, which was performed just before the next daily dosing. Amlodipine PK profile was simulated with 64 different sampling scenarios and simulated data were generated 1000 times for each scenario. RSE (Relative standard error (%)) and the average relative bias (%) estimated for each scenario were evaluated to select the most effective scenario to predict the PK parameters with minimum effort.
Results: Amlodipine PK was well described by the 2-compartment PK model with first-order absorption followed by zero-order absorption with lag time and linear elimination. (Oral clearance(CL/F) = 36.4 L/hr, Central volume(Vc/F) = 1150 L, Peripheral volume(Vp/F) = 910 L, Intercomparmental clearance (Q/F) = 118 L/hr, Absorption rate constant of first-order absorption(Ka) = 0.563/hr, Lag time of first-order absorption (Alag1) = 0.51 hr, Lag time of zero-order absorption(Alag2) = 3.85 hr, Fraction amount for first-order absorption (F1) = 0.762, Absorption duration of zero-order absorption(D2) = 2.36 hr) Visual inspection including goodness-of-fit plots and bootstrap result showed that PK model and parameters were predicted adequately. Based on the estimated RSE and average relative bias value for each scenario, the result showed that, for one trough sampling, sampling at 48hr or 72hr are recommended (RSE(%) = 23.7%, average relative bias(%) = 3.8%). For two trough samplings, sampling at 48hr and 96hr is recommended (RSE(%) = 17.3%, average relative bias(%) = 2.1%). Adding trough sampling after 2 half-lives did not improve the prediction, and three times of trough sampling did not show significant improvement from two times of trough sampling when samplings were performed at adequate timepoint such 48hr and 96hr.
Conclusions: The results implied that there were more informative timepoints to estimate accurate PK parameters in the multiple dose studies. When this timepoint can be predicted at the stage of protocol development through similar approach, an efficient clinical trial can be conducted with minimized sample numbers for the drug.
References:
[1] Huang X H et al. Saudi Pharmaceutical Journal (2014) 22, 63–69
[2] Svensson E M et al. The AAPS Journal, Vol. 18, No. 1, January 2016 [3] Karlsson K E et al. The AAPS Journal, Vol. 13, No. 1, March 2011
[4] FAULKNER J K et al. Br. J. clin. Pharmac. (1986), 22, 21-25
Reference: PAGE 28 (2019) Abstr 9195 [www.page-meeting.org/?abstract=9195]
Poster: Study Design