Application of a Physiologically Based Pharmacokinetic Approach to Predict Theophylline Pharmacokinetics Using Virtual Non-Pregnant, Pregnant, Fetal, Breast-Feeding, and Neonatal Populations

Abstract

Perinatal pharmacology is influenced by a myriad of physiological variables that are changing dynamically. The influence of these covariates has not been assessed systemically. The objective of this work was to use theophylline as a model drug and to predict its pharmacokinetics before, during (including prediction of the umbilical cord level), and after pregnancy as well as in milk (after single and multiple doses) and in neonates using a physiological-based pharmacokinetic (PBPK) model. Neonatal theophylline exposure from milk consumption was projected in both normal term and preterm subjects. Predicted infant daily doses were calculated using theophylline average and maximum concentration in the milk as well as an estimate of milk consumption. Predicted concentrations and parameters from the PBPK model were compared to the observed data. PBPK predicted theophylline concentrations in non-pregnant and pregnant populations at different gestational weeks were within 2-fold of the observations and the observed concentrations fell within the 5th−95th prediction interval from the PBPK simulations. The PBPK model predicted an average cord-to-maternal plasma ratio of 1.0, which also agrees well with experimental observations. Predicted postpartum theophylline concentration profiles in milk were also in good agreement with observations with a predicted milk-to-plasma ratio of 0.68. For an infant of 2 kg consuming 150 ml of milk per day, the lactation model predicted a relative infant dose (RID) of 12 and 17% using predicted average (Cavg,ss) and maximum (Cmax,ss) concentration in milk at steady state. The maximum RID of 17% corresponds to an absolute infant daily dose of 1.4 ± 0.5 mg/kg/day. This dose, when administered as 0.233 mg/kg every 4 h, to resemble breastfeeding frequency, resulted in plasma concentrations as high as 3.9 (1.9–6.8) mg/L and 2.8 (1.3–5.3) (5th−95th percentiles) on day 7 in preterm (32 GW) and full-term neonatal populations.

Use of In Vitro Dynamic Colon Model (DCM) to Inform a Physiologically Based Biopharmaceutic Model (PBBM) to Predict the In Vivo Performance of a Modified-Release Formulation of Theophylline

“…Initially, Abduljalil et al’s [ 37 ] theophylline model (which is also available in Simcyp’s compounds library) was used, and its predictivity was assessed against Staib et al’s [ 38 ] in vivo data of the theophylline solution administered directly to the colon. The only modification was to block the absorption of theophylline in the stomach and small intestine (SI) (by setting the absorption rate scalar (ARS) to 0.001, as well as setting the transit times in these two regions to 0.001 h), as explained in Section 2.2.1 .…”

“…The data of Staib et al [ 38 ] were not included in the clinical studies of Abduljalil et al [ 37 ] to develop and verify their PBPK model. However, their model showed good predictivity across an array of clinical studies involving oral administration of theophylline solution.…”

“…The first one was to develop a bespoke fit-for-purpose model using the Staib data; however, it would have been challenging to extrapolate this model to other clinical studies and predict the exposure of theophylline following administration of MR formulations. The second option was to use the PK parameters of theophylline and the ARS and transit times for the stomach and SI based on the model developed by Abduljalil et al [ 37 ], and optimize the ARS in the colon so that the predicted AUC of theophylline exposure after direct administration in the ascending colon (AUC AC ) fell within the in vivo-observed range of 79–87% of the predicted AUC in the control study (AUC C ) [ 38 ]. Clear et al [ 40 ] also showed that the AUC of an IR tablet of theophylline delivered directly to the colon using an Intelisite ® capsule was ~85% of the AUC obtained after oral administration of the same IR tablet.…”

“…All simulations were carried out using the healthy volunteer population. The PK parameters of theophylline ( Table S1 ) were derived from the non-pregnant populations in Abduljalil et al [ 37 ].…”

A perspective on the current use of the phase distribution model for predicting milk‐to‐plasma drug concentration ratio

“… 4 This mistake was corrected by the authors in a later publication and a later version of the PBPK software. 6 …”