Clinical applications of physiologically-based pharmacokinetic modeling: Perspectives on the advantages and challenges.

To the Editor: Pharmacotherapeutic strategies used by clinicians and clinical pharmacists are often based on available guidelines, their observations including druglevel monitoring results, and trained experience. In clinical settings, Bayesian estimation, based on a population pharmacokinetic (PK) model, has been the cornerstone to support dose adjustment using routine drug monitoring results. However, the physiologically based PK (PBPK) model can improve our mechanistic understanding of the important factors underlying the large PK variability, which is observed by descriptive and population PK analyses, in a systematic manner. The complementary utility of these pharmacometric approaches will be able to enhance the empirical decision-making process (ie, educated guesswork) with knowledge-driven theoretical simulations. In this letter, we propose the next-generation paradigm that will use PBPK modeling to better understand clinical PK variability in a comprehensive and systematic manner and also discuss the potential applications of PBPK modeling in clinical settings. PBPK modeling is a quantitative technology that integrates multiple drug physicochemical and human physiological parameters to predict the disposition of drugs. The PK simulations are prospective activities that can help us identify factors that contribute to the PK variability observed in patients. PBPK modeling is predominantly used in the pharmaceutical industry during drug development to assess the risk of drug–drug interaction (DDI) and predicts the PKs of drugs in special populations, such as pediatric and organ impairment subjects.1 The utility of PBPK modeling has expanded to an area close to a clinical setting beyond the drug development stage. PBPK modeling can also be used to simulate PK behavior in a “virtual” patient with characteristics identical to those of a “realworld” patient. The pharmaceutical industry provides health care professionals with general guidelines on drug treatment in the package insert. These guidelines are useful resources that contain representative PK profiles and DDI information that are obtained from clinical trials in standard healthy adults and limited patient populations. However, clinicians encounter a wide variety of patients with various disease states, age ranges, and drug therapies, which alter drug disposition. The quantitative information on the PKs for the intended patient population may not be available and may not be easily extrapolated. PBPK models can be used to provide quantitative information by simulating the PK profiles of drugs in such special populations. This letter highlights 2 potential uses of PBPK modeling: (1) the support of dosing regimens for DDI and their potential risk and (2) exploration of the mechanistic insights into the variability of drug PKs among patients (Fig. 1). In a clinical setting, patients with various disease states often require multiple medications. This polypharmacy is problematic because it increases the risk of DDI incidence. PBPK models can simulate specific dosing regimens that may include the dosing amount, number of doses, and route of administration. In addition, the PBPK platform can simultaneously model multiple drugs for DDI simulations.2 These characteristics allow us to generate potential dosing regimens that minimize DDI risk by changing dose timing and adjusting the dosage. This prospective approach will be beneficial in supporting clinical decisions before starting a new therapy. PBPK modeling can also be used to explore factors that contribute to the PK variability among patients. A PBPK model is designed with multiple drugand population-specific parameters, which enable us to mechanistically account for patient-specific pathophysiological changes due to age, genetic factors, and disease states at varying levels of severity.3 Therefore, the PBPK model can provide PK simulations altered by physiological changes observed in patients (eg, diseased organ function or reduced cardiac output) and identify potential covariates that may impact the PK. The potential clinical application of PBPK modeling is promising. However, there are challenges that need to be overcome for practical usage. The predictability of PBPK simulations, which relies on the quality of drug-specific and populationspecific information available, is the most integral part. Recent scientific progress regarding in vitro experimental systems has improved our knowledge on drugspecific information. However, anatomical and physiological information still requires continual investigation and evaluation, especially for ageand disease-dependent parameters in special populations, which include pediatric, elderly, and disease patients (eg, developmental changes in brain transporter expression and P450 enzyme abundance in cancer patients). To obtain such information, interdisciplinary collaborations among clinical, analytical, and basic scientists are essential. Recently, the FDA allowed the pharmaceutical industry to include PBPK simulation results as reference information in the package insert.4 The simulation results are mainly related to DDI risk assessments and genotyperelated PK parameter estimates in a rare disease population. One future challenge is to include a posterior assessment of the PBPK-simulated results using actual concentration measurements [eg, therapeutic drug monitoring (TDM)] as a part of the postmarketing assessment of drugs. Furthermore, to fill in the gap between virtual subjects and actual patients, we need to scrutinize patient The salary of C. Emoto and that of T. Fukuda was partly supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under Award Number R21HD095418. The authors declare no conflict of interest.