(T-134) Integrating Physiologically Based Pharmacokinetic (PBPK) Modeling and Strain-Specific Exposure-Efficacy Requirements to Refine the Dosage Regimens for Voriconazole in a Pediatric Cancer Population

Sarsvat Patel, PhD – Associate Director, Quantitative Pharmacology, Repare Therapeutics; Jonathan Constance, PhD – Assistant Professor, Department of Pediatrics, Spencer Fox Eccles School of Medicine, University of Utah

Objectives: Voriconazole is a triazole antifungal used to treat fungal infections in children with cancer. However, previous research reports by our group1 and others2 have shown that voriconazole pharmacokinetics (PK) is highly variable in this population, and many patients fail to achieve therapeutic target levels. Subtherapeutic levels of voriconazole in this immunocompromised population can be fatal. This study aims to develop a pediatric PBPK model that incorporates observed interpatient variability as well as cytochrome P450 (CYP) enzyme ontogeny. The simulated data was compared to the antifungal pharmacodynamics (PD) of voriconazole against multiple organisms to optimize the dose regimen in this population.

Methods: Initially, an adult PBPK model of voriconazole that incorporated hepatic clearance by CYP2C19, CYP2C9, and CYP3A4 enzymes was developed. The adult model was verified with oral (10 studies) and intravenous (4 studies) PK datasets of voriconazole in adults. The PBPK modeling was performed using PK Sim® software, and the PD analysis was performed using R. The adult model was then extrapolated to children with and without cancer using ontogeny equations of CYP maturation. The pediatric PBPK model was then verified using the observed PK data of voriconazole in children with (N=14) and without cancer (N=11). For both models, bias and precision were evaluated by calculating the average fold error (AFE) and absolute average fold error (AAFE), respectively. The final pediatric PBPK model of voriconazole was used to simulate the levels of voriconazole in virtual pediatric patients (N=1000) across the age continuum using dosing regimens used in children with cancer. The probability of target attainment (PTA) for different pre-clinical and clinical PK/PD indices of efficacy for voriconazole was considered (fAUC24/MIC ≥25, Ctrough > 2 µg/mL, and Ctrough, SS/MIC >2) to understand the overall benefit in each pediatric sub-population group.

Results: The adult PBPK model successfully predicted the PK of voriconazole after IV and oral administration. The adult model was successfully extrapolated to the pediatric model by incorporating ontogeny equations of CYP maturation. For both adult and pediatric models, more than 90% of adult observed PK data were captured within the 90% prediction interval of the model-simulated data. The AFE and AAFE values for observed versus predicted voriconazole concentrations in adult and pediatric models were within 0.5 and 2. The outcomes of the PBPK model-based simulations indicated that the pediatric dosing regimens need to be tailored based on the child’s age and the specific infectious fungal species involved.

Conclusions: A PBPK model to predict the PK of voriconazole in children with cancer was developed and verified. In clinical settings, the infectious fungal species must be considered. Additionally, further research with large sample sizes is necessary to validate the current findings.

Citations: 1. Biltaji E et al. American College of Clinical Pharmacology (ACCP) Annual Meeting, Sep 17-19, 2017, San Diego, CA. (2017)
2. Tucker L, et al. J Pediatr Pharmacol Ther. 20 17-23. (2015)