Content area
Abstract
Various known physiological and biological changes can affect the pharmacokinetics of drugs in pregnant patients, and these changes can compromise efficacy, safety and toxicity of these drugs. Due to ethical and practical considerations, it is not possible to directly evaluate these changes in the pharmacokinetic profile of drugs during pregnancy. Alternative approaches that predict drug disposition during pregnancy that are practical and easy to apply are highly desirable. To address this need, this dissertation research applied physiologically based pharmacokinetic (PBPK) modeling and simulation to predict and describe systemic drug exposure during pregnancy. The models incorporated known time-dependent changes in the physiology, ontogeny, and biology during pregnancy to describe alterations in drug disposition during pregnancy. Systemic exposure predictions of several probe compounds as pregnancy matures were conducted.
All modeling and simulation exercises were performed using GastroPlus TM Version 8.0.002 (Simulations Plus, Inc., Lancaster, CA, USA). PBPK models were used to predict or describe systemic exposure in pregnant women as pregnancy progresses (i.e., across trimesters), based on physiologic changes during pregnancy that can impact ADME pathways. A simple and readily implementable multi-tissue\organ whole-body PBPK model was utilized. The structural model incorporated baseline physiologic data from healthy non-pregnant women. Subsequently, changes in key pregnancy-related physiological parameters (e.g., total body weight, total body water, cardiac output, plasma volume, red blood cell volume, glomerular filtration rate, creatinine clearance, and uterine blood flow) as well as change in cytochrome P450 activity were introduced for each trimester of pregnancy, based on literature data. The models were then applied to a series of reference drugs that differed with respect to the primary clearance pathways: metformin (renal excretion), oseltamivir carboxylate (renal excretion), caffeine (CYPIA2 metabolism), nevirapine (CYP3A4 + CYP2B6 metabolism), lopinavir (CYP3A4 metabolism) and tacrolimus (CYP3A4 metabolism).
For each probe compound, plasma concentration-time profiles were simulated for the medications across each population (non-pregnant women, lst , 2nd, and 3rd trimester pregnant women). The predicted systemic exposure metrics (Cmax, AUC) were compared to published clinical data. Parameter sensitivity analyses (PSA) were performed to identify the critical parameters that most influenced model parameters (Cmax, and AUC -0t). Population estimates (90% confidence intervals) were estimated for the pharmacokinetic parameters were generated through model simulations involving 2500 subjects using a virtual population approach. The PBPK model-simulated pharmacokinetic profiles for test medications were in agreement with observed clinical data for the changes in exposure (AUC and Cmax) during pregnancy. The fold error (EE) was calculated based on the ratio of clinical observed and model predicted PK parameter values. The fold error for the base PBPK model was less than 0.15 (15%) for all the reference drugs. PSA validate the key parameters, which were responsible for the observed changes in systemic exposure during pregnancy. For population estimates, the pharmacokinetic parameters were comparable to clinical observed values for all the probe drugs. The ratio of observed and predicted values ranged from 0.85 to 1.15, indicating, that the PBPK modeling approach was useful in predicting drug pharmacokinetics for drugs that are extensively metabolized as well as really excreted medications in pregnant women during each trimester.
In summary, this dissertation research demonstrated that PBPK modeling represents a potentially useful tool to help establish dosing guidelines that ensure safe and effective drug therapy for pregnant patients. During drug development, PBPK modeling can inform the design of clinical studies and aid in anticipation of potential exposure changes in pregnant women for compounds a priori.
