Development of Physiologically Based Pharmacokinetic/Pharmacodynamic Model for Indomethacin Disposition in Pregnancy

Abstract

Findings of a recent clinical study showed indomethacin has lower plasma levels and higher steady-state apparent clearance in pregnant subjects when compared to those in non-pregnant subjects reported in separate studies. Thus, in the current work we developed a pregnancy physiological based pharmacokinetic/pharmacodynamic (PBPK/PD) model for indomethacin to explain the differences in indomethacin pharmacokinetics between pregnancy and non-pregnancy. A whole-body PBPK model with key pregnancy-related physiological changes was developed to characterize indomethacin PK in pregnant women and compare these parameters to those in non-pregnant subjects. Data related to maternal physiological and biological changes were obtained from literature and incorporated into the structural PBPK model that describes non-pregnant PK data. Changes in indomethacin area under the curve (AUC), maximum concentration (Cmax) and average steady-state concentration (Cave) in pregnant women were predicted. Model-simulated PK profiles were in agreement with observed data. The predicted mean ratio (non-pregnant:second trimester (T2)) of indomethacin Cave was 1.6 compared to the observed value of 1.59. In addition, the predicted steady-state apparent clearance (CL/Fss) ratio was almost similar to the observed value (0.46 vs. 0.42). Sensitivity analysis suggested changes in CYP2C9 activity, and to a lesser extent UGT2B7, as the primary factor contributing to differences in indomethacin disposition between pregnancy and non-pregnancy. The developed PBPK model which integrates prior physiological knowledge, in vitro and in vivo data, allowed the successful prediction of indomethacin disposition during T2. Our PBPK/PD model suggested a higher indomethacin dosing requirement during pregnancy.

Fig 1. Structure of pregnancy physiologically based pharmacokinetic (p-PBPK) model.

Fig 2. Predicted and observed plasma concentration-time profiles of indomethacin following the administration of a single oral dose of 25 and 50mg.

The solid (50 mg) and dashed (25 mg) lines represent predicted mean indomethacin profile (A), and multiple oral dose of 25mg three times daily, the solid line represent predicted mean indomethacin profile (B) to non-pregnant subjects. Mean observed data are overlaid (○ 25mg dose profile; ● 50mg dose profile) [22]. The green and blue shaded areas represent the 90% confidence interval for the simulated data, and error bars represent ±SD.

Fig 3. Predicted and observed plasma concentration–time profiles of indomethacin following chronic oral dosing of 25 mg four times daily during the second trimester (T₂).

The solid line represents predicted mean indomethacin profile during T₂. Mean observed data [21] are represented by the symbol (○). The green shaded area represents the 90% confidence interval for the simulated data, and error bars represent ± SD.

Fig 4. Sensitivity analysis to evaluate mechanism(s) primarily contributes to differences in indomethacin levels in pregnancy.

Contribution of changes in metabolism (CYP2C9 and UGT2B7 activities), plasma protein binding (PB), glomular filtration rate (GFR), and volume of distribution (V_d) to variation in C_ave (Black columns) and CL/F_ss (Grey columns) during pregnancy.

Fig 5. Simulated and observed PD effect-time profiles for indomethacin presented as % decrease in PGEM (13, 14-dihydro-15-ketoprostaglandin E2) plasma concentration vs. time after single oral administration of 25mg.

The solid line represents predicted mean indomethacin profile in non-pregnant subjects. The dashed line represents predicted mean indomethacin profile in pregnant subjects. Mean observed data are overlaid for 25 mg dose in non-pregnant subjects [39]. The green and blue shaded areas represent the 90% confidence interval for the simulated data, and error bars represent ± SD.