Ari Brekkan Viggosson, Andrew C. Hooker, Mats O. Karlsson, Siv Jönsson
Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden
Objectives: Monoclonal antibodies (mAbs) may display target mediated drug disposition (TMDD) [1] when target binding notably alters the disposition of the drug. In this work TMDD models, combined with optimal design methods, were employed to evaluate the consequences of reduced sampling designs for mAbs. The aim was to i) optimize reduced sampling schedules for TMDD models and ii) determine the consequences of reduced designs on parameter precision, precision of free target level predictions at certain time-points and on dose choice for future studies.
Methods: The quasi steady-state (QSS) and quasi equilibrium (QE) approximations [2, 3, 4] of the full TMDD model were used to describe the disposition and interaction of a drug with a soluble target. Sampling schedules for pharmacokinetic trials featuring mAbs were evaluated, reduced with respect to the number of samples, subjects in the trials and/or trial length and optimized using PopED [5]. Expected parameter imprecision was evaluated and used to obtain population predictions of free target levels given each design. The reduced designs were compared to the original designs with respect to efficiency, parameter uncertainty and imprecision of free target levels at certain time-points. The Ds-criterion was used in the optimization and calculation of efficiency to focus on fixed-effect parameters. Based on population predictions of free target levels the likelihood of making an erroneous conclusion regarding dose selection was calculated.
Results: Reduction of the total number of samples from 1872 to 1440 in the QE model and from 624 to 288 in the QSS model did not decrease the Ds-efficiency of the designs below 90%. Substantial reduction in information content (Ds-efficiency ≤ 60%) resulted in precision of free target predictions at 14 days of 4.49-22.21% (root-mean-squared error) over a dose range, compared with 3.31-17.77% for the original design. Designs with 69% fewer samples than the original were 33% more likely to result in an erroneous dose choice to reach target suppression. Reducing the amount of samples by 23% did not affect the dose choice at an 80% power level.
Conclusions: Rich sampling designs for mAbs may be superfluous depending on the purpose of the study. Parameter uncertainty and imprecision in prediction of target levels did not always increase for substantially reduced designs. The risk of making an erroneous dose choice for future studies was marginally increased for reduced designs.
References:
[1] Keizer RJ, Huitema AD, Schellens JH, Beijnen JH. Clinical pharmacokinetics of therapeutic monoclonal antibodies. Clinical pharmacokinetics 2010; 49: 493-507.
[2] Gibiansky L, Gibiansky E. Target-mediated drug disposition model for drugs that bind to more than one target. Journal of pharmacokinetics and pharmacodynamics 2010; 37: 323-46.
[3] Hayashi N, Tsukamoto Y, Sallas WM, Lowe PJ. A mechanism-based binding model for the population pharmacokinetics and pharmacodynamics of omalizumab. British journal of clinical pharmacology 2007; 63: 548-61.
[4] Mager DE, Krzyzanski W. Quasi-equilibrium pharmacokinetic model for drugs exhibiting target-mediated drug disposition. Pharmaceutical research 2005; 22: 1589-96.
[5] Nyberg J, Ueckert S, Stromberg EA, Hennig S, Karlsson MO, Hooker AC. PopED: an extended, parallelized, nonlinear mixed effects models optimal design tool. Computer methods and programs in biomedicine 2012; 108: 789-805.
Reference: PAGE 23 (2014) Abstr 3179 [www.page-meeting.org/?abstract=3179]
Poster: Methodology - Study Design