The diagnosis, as well as the prediction, of idiosyncratic adverse drug reactions (ADR) is challenging. Idiosyncratic ADRs are manifold, possibly exhibit long latencies, and the correlation between administered dose and ADR incidence is not apparent at first sight. They are often related to a patient-specific metabolic phenotype that alters the individual drug exposure. Therefore, patients could directly benefit from an accurate diagnosis and prediction of an idiosyncratic ADR by taking into account personalised treatment regimens. Currently, this is hard to achieve both during drug development and daily health care without complex, invasive, or time-consuming testing [1]. An easy-to-perform, minimally-invasive but still reliable test for individual ADR predisposition is requisite to increase patient safety, for example, by assessing the metabolic capacity through pharmacokinetics (PK) analysis of a patient after taking a probe drug.
Methods:
The presented approach is based on physiologically-based pharmacokinetic (PBPK) modelling to simulate drug concentration-time profiles and assess PK variability in virtual populations. The drug-specific PBPK models and the virtual populations were built with PK-Sim from the Open Systems Pharmacology Suite (OSPS) [2]. Pharmacokinetic data and information on enzyme variability were gathered from the literature. Analyses on variability were performed using R and the OSPS toolbox for R.
Results:
We built four drug-specific whole-body PBPK models for the substances contained in the flu medication Frenadol® (FRE) to evaluate the use of FRE as a probe drug for a metabolic phenotyping test [3]. After qualifying the PBPK models for the four active ingredients caffeine, dextromethorphan, acetaminophen (APAP), and chlorpheniramine, as well as their main metabolites with literature PK data, we simulated virtual populations with different phenotypes caused by a) varied biometrics and b) additionally varied enzyme expression levels. After analyzing the PK effect induced by these sources of variability, suitable sampling time points and molecules for assessing the metabolic phenotype of each enzyme were proposed. For evaluating the activity of CYP1A2, CYP2D6, CYP3A4, UGT1A6, and possibly CYP2E1, the blood levels of the molecules paraxanthine, dextromethorphan, and its metabolites dextrorphan and 3-hydroxymorphinan, as well as the APAP metabolites APAP-glucuronide, and APAP-cysteine should be measured. According to the simulations, the most appropriate time point for a concurrent analysis is between three and five hours after FRE administration [4].
Conclusions:
The presented approach based on PBPK modelling can be used to guide a metabolic phenotype test with FRE and design a clinical study with diligent sampling time points and molecules. Performing such a phenotyping test could increase patient safety in a fast, reliable, and minimally-invasive manner. Thereby, our analyses aid personalised treatment decisions and timely predictions and preventions of idiosyncratic ADRs.
Vanessa Baier, Annika R.P. Schneider, Hanna Kreuzer, José V. Castell, Lars M. Blank, Lars Kuepfer
https://www.page-meeting.org/default.asp?abstract=10125
Objectives:
The diagnosis, as well as the prediction, of idiosyncratic adverse drug reactions (ADR) is challenging. Idiosyncratic ADRs are manifold, possibly exhibit long latencies, and the correlation between administered dose and ADR incidence is not apparent at first sight. They are often related to a patient-specific metabolic phenotype that alters the individual drug exposure. Therefore, patients could directly benefit from an accurate diagnosis and prediction of an idiosyncratic ADR by taking into account personalised treatment regimens. Currently, this is hard to achieve both during drug development and daily health care without complex, invasive, or time-consuming testing [1]. An easy-to-perform, minimally-invasive but still reliable test for individual ADR predisposition is requisite to increase patient safety, for example, by assessing the metabolic capacity through pharmacokinetics (PK) analysis of a patient after taking a probe drug.
Methods:
The presented approach is based on physiologically-based pharmacokinetic (PBPK) modelling to simulate drug concentration-time profiles and assess PK variability in virtual populations. The drug-specific PBPK models and the virtual populations were built with PK-Sim from the Open Systems Pharmacology Suite (OSPS) [2]. Pharmacokinetic data and information on enzyme variability were gathered from the literature. Analyses on variability were performed using R and the OSPS toolbox for R.
Results:
We built four drug-specific whole-body PBPK models for the substances contained in the flu medication Frenadol® (FRE) to evaluate the use of FRE as a probe drug for a metabolic phenotyping test [3]. After qualifying the PBPK models for the four active ingredients caffeine, dextromethorphan, acetaminophen (APAP), and chlorpheniramine, as well as their main metabolites with literature PK data, we simulated virtual populations with different phenotypes caused by a) varied biometrics and b) additionally varied enzyme expression levels. After analyzing the PK effect induced by these sources of variability, suitable sampling time points and molecules for assessing the metabolic phenotype of each enzyme were proposed. For evaluating the activity of CYP1A2, CYP2D6, CYP3A4, UGT1A6, and possibly CYP2E1, the blood levels of the molecules paraxanthine, dextromethorphan, and its metabolites dextrorphan and 3-hydroxymorphinan, as well as the APAP metabolites APAP-glucuronide, and APAP-cysteine should be measured. According to the simulations, the most appropriate time point for a concurrent analysis is between three and five hours after FRE administration [4].
Conclusions:
The presented approach based on PBPK modelling can be used to guide a metabolic phenotype test with FRE and design a clinical study with diligent sampling time points and molecules. Performing such a phenotyping test could increase patient safety in a fast, reliable, and minimally-invasive manner. Thereby, our analyses aid personalised treatment decisions and timely predictions and preventions of idiosyncratic ADRs.