Novel Mechanistic Physiologically Based Pharmacokinetic Model to Predict Renal Clearance and CYP2D6 Activity Across Pregnancy

Abstract ID 95096 Poster Board 231 Pregnancy induces physiological changes that impact the pharmacokinetics (PK) of drugs. Interpretation and prediction of observed PK changes during pregnancy is challenging via static methods due to the interconnected dynamic alterations. For example, pregnancy results in an approximately 40% increase in cardiac output and kidney blood flow, leading to changes in renal clearance (CLr) of drugs but also in distribution rates and hepatic clearance. Additionally, the increased activity of cytochrome P450 2D6 (CYP2D6) and 3A4 (CYP3A4) during pregnancy increases drug clearance resulting in altered pharmacological effects. Despite these known changes, there is a gap in systematic physiologically based pharmacokinetic (PBPK) modeling of pregnancy in simulating changes in CLr mechanisms during pregnancy. This is important as CLr is a major clearance pathway for may drugs used during pregnancy and urinary ratios are often used as markers of pregnancy mediated metabolic clearance changes. This study aims to comprehensively understand pregnancy-induced alterations in CLr via PBPK modeling and to address the translation of urinary ratios of CYP2D6 substrates to quantitation of the induction of CYP2D6 activity during pregnancy. Pregnancy-related physiological changes for the third trimester were collected from the literature. Existing literature suggests an elevation in urine pH during pregnancy and altered urine flow and water reabsorption. These changes were incorporated into a previously published mechanistic kidney model and a novel pregnancy-mechanistic kidney model was developed using Matlab Simulink. The pregnancy-mechanistic kidney model was incorporated into a full PBPK model, and a parent-metabolite PBPK model to simulate plasma concentration profiles for the selected drugs. A 100% induction in CYP2D6 activity was incorporated in the liver. The full model was validated fusing dextromethorphan (DEX) and its metabolite dextrorphan (DOR), metoprolol, clonidine, and atenolol as model drugs. These drugs were selected due to known changes in CLr during pregnancy and CYP2D6-mediated metabolism. Drug-specific parameters were verified against observed CLr in nonpregnant populations. Considering different physicochemical properties of the drugs such as pKa, permeability, and fraction unbound in plasma together with the model incorporated physiological changes during pregnancy, changes in CLr and plasma concentrations-versus-time profiles were simulated. The mechanistic kidney model was successfully developed with fold-error (predicted CLr /observed CLr) and absolute fold error consistently below 2 for the selected drugs for nonpregnant and pregnant populations. The observed alterations in maternal plasma concentration profiles of the model drugs and their metabolites were predicted within predefined acceptance criterion. In addition, the plasma and urinary DOR/DEX metabolic ratios were successfully simulated across pregnancy. Significant impact of altered urine pH was predicted for this metabolic ratio suggesting that changes in urine pH during pregnancy may confound use of urinary metabolic markers for assessment of liver CYP2D6 induction. The developed model provides a comprehensive approach to predicting drug disposition during pregnancy and can be applied to predict changes in CLr for drugs that are renally cleared but have not been previously studied during pregnancy.