Strategies for Determining Correct Cytochrome P450 Contributions in Hepatic Clearance Predictions: In Vitro-In Vivo Extrapolation as Modelling Approach and Tramadol as Proof-of Concept Compound

Physiologically-based Pharmacokinetic Modeling to Assess the Impact of CYP2D6-Mediated Drug-Drug Interactions on Tramadol and O-Desmethyltramadol Exposures via Allosteric and Competitive Inhibition

Tramadol is an opioid medication used to treat moderately severe pain. Cytochrome P450 (CYP) 2D6 inhibition could be important for tramadol, as it decreases the formation of its pharmacologically active metabolite, O‐desmethyltramadol, potentially resulting in increased opioid use and misuse. The objective of this study was to evaluate the impact of allosteric and competitive CYP2D6 inhibition on tramadol and O‐desmethyltramadol pharmacokinetics using quinidine and metoprolol as prototypical perpetrator drugs. A physiologically based pharmacokinetic model for tramadol and O‐desmethyltramadol was developed and verified in PK‐Sim version 8 and linked to respective models of quinidine and metoprolol to evaluate the impact of allosteric and competitive CYP2D6 inhibition on tramadol and O‐desmethyltramadol exposure. Our results show that there is a differentiated impact of CYP2D6 inhibitors on tramadol and O‐desmethyltramadol based on their mechanisms of inhibition. Following allosteric inhibition by a single dose of quinidine, the exposure of both tramadol (51% increase) and O‐desmethyltramadol (52% decrease) was predicted to be significantly altered after concomitant administration of a single dose of tramadol. Following multiple‐dose administration of tramadol and a single‐dose or multiple‐dose administration of quinidine, the inhibitory effect of quinidine was predicted to be long (≈42 hours) and to alter exposure of tramadol and O‐desmethyltramadol by up to 60%, suggesting that coadministration of quinidine and tramadol should be avoided clinically. In comparison, there is no predicted significant impact of metoprolol on tramadol and O‐desmethyltramadol exposure. In fact, tramadol is predicted to act as a CYP2D6 perpetrator and increase metoprolol exposure, which may necessitate the need for dose separation.

The use of PBPK modeling across the pediatric age range using propofol as a case

The project SAFEPEDRUG aims to provide guidelines for drug research in children, based on bottom-up and top-down approaches. Propofol, one of the studied model compounds, was selected because it is extensively metabolized in liver and kidney, with an important role for the glucuronidation pathway. Besides, being a lipophilic molecule, it is distributed into fat tissues, from where it redistributes into the systemic circulation. In the past, both bottom-up (Physiologically based pharmacokinetic, PBPK) and top-down approaches (population pharmacokinetic, popPK) were applied to describe its pharmacokinetics (PK). In this work, a combination of the two was used to check their performance to describe PK in children and neonates (both term and preterm) using propofol as a case compound. First, in vitro data was generated in human liver microsomes and recombinant enzymes and used to develop an adult PBPK model in Simcyp^®. Activity adjustment factors (AAFs) were calculated to account for differences between in vitro and in vivo enzyme activity. Clinical data were analyzed using a 3-compartment model in NONMEM. These data were used to construct a retrograde PBPK model and for qualification of the PBPK models. Once an accurate in vivo clearance was obtained accounting for the contribution of the different metabolic pathways, the resulting PBPK models were challenged with new data for qualification. After that, the constructed adult PPBK model for propofol was extrapolated to the pediatric population. Both the default built-in and in vivo derived ontogeny functions were used to do so. The models were qualified by comparing their predicted PK parameters to published values, and by comparison of predicted concentration-time profiles to available clinical data. Clearance values were predicted well, especially when compared with values obtained from trials where long-term sampling was applied, whereas volume of distribution was lower compared to the most common popPK model predictions. Concentration-time profiles were predicted well up until and including the preterm neonatal population. In this work, it was thus shown that PBPK can be used to predict the PK up to and including the preterm neonatal population without the use of pediatric in vivo data. This work adds weight to the need for further development of PBPK models, especially regarding distribution modeling and the use of in vivo derived ontogeny functions.