Letizia Carrara (1), Marylore Chenel (1, 2*), Ludwig Vincent (3**), Matthieu Jacobs (3), Marie-Cécile Secouard-Faure(4***), Amelie Moreau (5), Adrien Tessier (1)
(1) Clinical Pharmacometrics - Quantitative Pharmacology, Servier, Suresnes, France (2) Translational Pharmacometrics - Quantitative Pharmacology, Servier, Suresnes, France (3) Translational Pharmacometrics - Quantitative Pharmacology, Servier, Orléans, France (4) Clinical Pharmacokinetics - Quantitative Pharmacology, Servier, Suresnes, France (5) Drug Interaction & In Vitro Technology - DMPK, Servier, Orléans, France (*) currently employed by Pharmetheus AB, Sweden (**) currently employed by Certara UK Limited, Simcyp Division, Sheffield, UK (***) currently employed by Certara, Paris, France
Introduction:
Physiologically Based PharmacoKinetics (PBPK) models have become an important tool in drug development. They are increasingly used for various applications, including prospective predictions of pharmacokinetics (PK) differences in populations in which clinical data are limited, or to mechanistically describe absorption processes and the effect of food on drug absorption.
However, challenges remain in predicting complex situations in which the impact of several factors is investigated simultaneously.
Drug S PK is fully characterized in adults, but limited clinical data are available in pediatrics. Using drug S as a paradigm drug, here we present an approach to predict drug PK in pediatrics in fasted and fed states, to ultimately inform recommendation on drug administration in children by applying a PBPK model built on adult data only.
Objectives: a) to develop a PBPK model to mechanistically describe the PK of drug S in adults in fasted and in fed states, b) to apply the model to prospectively predict PK in fasted and in fed states in pediatrics (2-18 years old), to assess magnitude of the food effect and c) to support recommendation on drug administration in children.
Methods:
Analyses were performed using Simcyp® simulator V18R2.
1) PBPK model building and qualification using data of adult healthy volunteers (HV)
Plasma concentrations and urinary excretion data collected following intravenous (IV) administration of different dosages of drug S in adult HV were used to identify clearance and distribution processes. Absorption was modelled using plasma concentration data following oral administration in fasted state. The Advanced Dissolution Absorption and Metabolism (ADAM) model was applied to describe the absorption in the Gastro-Intestinal (GI) tract. The human jejunal effective permeability Peff,man was calculated with a first order approach using the formula
Peff,man=(R/2)*Ka
Where Ka is the absorption constant [s-1] and R [cm] is the mean radius of the intestine of the population under study.
2) Simulation of food effect in adults
PK data collected following a food effect study in adult HV were used to assess the ability of the model to predict drug absorption in fed state. Absorption-related parameters were refined using a middle-out approach based on both literature findings1 and observed data.
3) Prospective prediction of PK and food effect in pediatrics
The model developed in 1) and 2) was applied to prospectively predict drug PK in pediatric population in fasted and in fed states. Peff,man was corrected for the mean intestinal radius of the pediatric (sub)population under study. Ka was not rescaled.
4) Simulation of food staggering
An intermittent fasted/fed state physiology in the GI tract was mimicked to simulate different time intervals between drug administration and food intake.
Simcyp® HV population was used in 1), 2) and 4) (food staggering could not be simulated in pediatrics). 2 customized pediatric subpopulations with different body weight distributions were simulated via the ‘fixed individual trial design’ option in 3).
Results:
A minimal PBPK model with enzymatic and renal clearances and ADAM model for absorption was developed. The model well captured the plasma concentration-time profiles and the amount of drug excreted unchanged in the urine following IV and oral administrations of different doses of drug S in adult HV in both fasted and fed states, and was therefore deemed appropriate to prospectively predict drug PK in fasted and in fed states in pediatrics.
Simulations showed a similar trend in adults and in pediatrics: food did not affect AUC, but Cmax with a maximum predicted fold change (fed/fasted state) of 0.54 and of 0.73 for adults and pediatrics, respectively.
Food staggering simulations in adult were deemed predictive for pediatrics, and showed a significant decrease in Cmax when food was taken within 15 minutes after drug administration. A prolonged food effect was predicted when the drug is taken after the meal (up to 3 hours).
These results support the recommendation to i) assume the drug in fasted state (at least 3h after the meal) and ii) avoid food ingestion for 15 minutes following drug intake.
Conclusions:
PBPK modelling was used to investigate food effect in children. Simulation results could be used to support the registration without undertaking pediatric food effect study.
References:
[1] Bonner JJ, Burt H, Johnson TN, Whitaker MJ, Porter J, Ross RJ. Development and verification of an endogenous PBPK model to inform hydrocortisone replacement dosing in children and adults with cortisol deficiency. European Journal of Pharmaceutical Sciences. 2021 Oct 1;165:105913.
Reference: PAGE 30 (2022) Abstr 10002 [www.page-meeting.org/?abstract=10002]
Poster: Drug/Disease Modelling - Absorption & PBPK