2016 - Lisboa - Portugal

PAGE 2016: Clinical Applications
David McDougall

The Impact of Model Selection for Personalised Dosing

David AJ McDougall (1,2), Jenny Martin (3), E Geoffrey Playford (4,5), Bruce Green (2)

(1) School of Pharmacy, University of Queensland, Brisbane, Australia; (2) Model Answers Pty Ltd, Brisbane, Australia; (3) School of Medicine and Population Health, University of Newcastle, New South Wales, Australia; (4) School of Medicine, University of Queensland, Brisbane, Australia; (5) Infection Management Services, Princess Alexandra Hospital, Brisbane, Australia.

Objectives: Model based personalised dosing (MBPD) is a form of individualised therapy where patient specific pharmacokinetic (PK) or pharmacodynamic parameters are estimated in real-time. Little research has been conducted to evaluate how the choice of model impacts dose recommendations. Voriconazole, a triazole antifungal with nonlinear kinetics is the motivating example used in this research. Model selection for voriconazole MPBD is potentially critical, as many different structural models are present in the literature. The aim of this work was to assess the impact of model miss-specification on dose recommendations and clinical outcomes.

Methods: Five reduced miss-specified population models were developed from a published model by removing key structural components. [1] Parameters for the reduced models were developed using the stochastic simulation and estimation functionality in PsN.[2] The dose adjustments required to reach a target concentration in 100 simulated subjects were determined using the empirical Bayes estimates.  The expected plasma concentration that would have resulted from the dose recommendations was derived from the simulated PK parameters. Logistic regression models linking exposure to clinical response[3] and neurotoxicity[4] were applied to the predicted plasma concentrations to assess the probability of clinical outcomes.

Results: Removing CYP2C19 genotype as a covariate on clearance resulted in similar dose recommendations to the required doses with minimal impact on the plasma concentrations achieved and percentage of subjects within the therapeutic range. The models with only linear clearance performed poorly, recommending large doses that would have resulted in toxic exposure.  The probability of clinical success was similar for all the models. The probability of neurotoxicity was lowest when the model contained non-linear clearance.  The median probability of neurotoxicity increased 4-8 fold when a model with linear clearance was used.    

Conclusion: Structurally miss-specified clearance had a large impact on plasma concentrations and the likelihood of toxicity.  This is relevant for voriconazole as several published models have only linear clearance.[5-9] Removing genotype was of little importance given the probability of clinical response and neurotoxicity was comparable.  Once plasma concentrations become available, the benefit of genotype in MBPD is of little benefit. 



References
[1] McDougall DA, Martin J, Playford EG, et al. Determination of a suitable voriconazole pharmacokinetic model for personalised dosing. J Pharmacokinet Pharmacodyn 2015
[2] Lindbom L, Pihlgren P, Jonsson EN. PsN-Toolkit--a collection of computer intensive statistical methods for non-linear mixed effect modeling using NONMEM. Computer methods and programs in biomedicine 2005;79(3):241-57.
[3] Troke PF, Hockey HP, Hope WW. Observational study of the clinical efficacy of voriconazole and its relationship to plasma concentrations in patients. Antimicrob Agents Chemother 2011;55(10):4782-8.
[4] Pascual A, Csajka C, Buclin T, et al. Challenging recommended oral and intravenous voriconazole doses for improved efficacy and safety: population pharmacokinetics-based analysis of adult patients with invasive fungal infections. Clin Infect Dis 2012;55(3):381-90.
[5] Wang T, Chen S, Sun J, et al. Identification of factors influencing the pharmacokinetics of voriconazole and the optimization of dosage regimens based on Monte Carlo simulation in patients with invasive fungal infections. J Antimicrob Chemother 2013.
[6] Pascual A, Csajka C, Buclin T, et al. Challenging recommended oral and intravenous voriconazole doses for improved efficacy and safety: population pharmacokinetics-based analysis of adult patients with invasive fungal infections. Clin Infect Dis 2012;55(3):381-90.    
[7] Han K, Capitano B, Bies R, et al. Bioavailability and population pharmacokinetics of voriconazole in lung transplant recipients. Antimicrob Agents Chemother 2010;54(10):4424-31
[8] Nomura K, Fujimoto Y, Kanbayashi Y, et al. Pharmacokinetic-pharmacodynamic analysis of voriconazole in Japanese patients with hematological malignancies. Eur J Clin Microbiol Infect Dis 2008;27(11):1141-3.
[9] Han K, Bies R, Johnson H, et al. Population pharmacokinetic evaluation with external validation and Bayesian estimator of voriconazole in liver transplant recipients. Clin Pharmacokinet 2011;50(3):201-14.


Reference: PAGE 25 (2016) Abstr 5738 [www.page-meeting.org/?abstract=5738]
Oral: Clinical Applications
Click to open PDF poster/presentation (click to open)
Top