While using physiologically-based pharmacokinetic modeling (PBPK) in determining virtual bioequivalence (VBE) has significantly increased over the past few years, this type of mechanistic biosimulation still offers enormous potential to supplant expensive and time-consuming clinical studies. The paper referenced below addresses our increased knowledge of oral drug absorption and the applicability of the Simcyp Simulator to assess virtual BE between two formulations in the same subject cohort, under different conditions, and different populations. The paper highlights a practical way of estimating within-subject variability in GI physiology that are poorly understood in the absence of direct measurements of parameters in human subjects.
Achieving the 90% Confidence Interval Requirement
Clinical BE studies are typically designed to demonstrate similarity in the systemic exposure of two products containing the same active drug substance. Commonly, these studies use a cross-over design where each subject receives both the test and reference drug. The cross-over approach can reduce the impact of variability on BE outcome, and that can cause both false positives and negatives in achieving the required 90% confidence interval (CI). While the impact of physiological variability between subjects on detecting formulation differences can be mitigated by cross-over studies, the possibility of falling outside the 80-125% confidence interval exists as shown in Figure 1. For example, the amplitude of CIs in BE studies depends not only on the number of enrolled subjects (between-subject variability), but also on the within-subject variability (WSV) in the rate and extent of bioavailability that is determined by the drug as well as its formulation attributes and their interplay with the WSV in physiology, particularly those of GI tract.
Simcyp Simulator Includes Within Subject Variability in VBE Studies
The Simcyp Simulator is frequently used to predict clinical outcomes in untested populations. For VBE, the Simulator has a unique capability of simulating both between-subject (inter) and within-subject (intra) variability to assess those outcomes. Bego et al. 2022 describe two paths toward demonstrating that capability (figure 2).
within-subject variability (WSV)
- Path B1: Use of available between-subject variability (BSV) in lieu of WSV in physiological parameters as a conservative measure
To understand the risk of BE failure between two formulations, a sufficiently verified PBPK model could be used to simulate cross-over VBE trials using hybrid WSV coefficients. Using the ‘fixed trial design’ option in Simcyp, the individual subject demographics are used to create virtual subjects for a virtual crossover BE assessment. Two trials of subjects can be simulated for each formulation, leveraging BSV data when WSV is not available assuming WSV is as high as BSV. Such approach has merit if the interest is to understand biopharmaceutics risk, however, this approach can over-estimate the crossover BE variability if the objective is to simulate realistic BE.
- Path B2: Propagating an array of WSV in physiology and eliminating the sets that are incompatible with observed WSV in pharmacokinetics
Occasionally, drugs are classified as “highly variable drugs” due to the sensitivity of a particular drug or formulation towards physiological variability. Thus, calling drugs “highly variable” is misleading. The drug and formulations are generally tightly controlled during manufacturing, and the formulation is consumed by the human volunteers. Still some drugs and formulations lead to more variable PK than others because of sensitivity and interaction of specific drugs with human physiological variability. Absorption is largely driven by drug and formulation properties, and variability in PK of different drugs/formulations reflects the sensitivity towards different physiological parameters. However, the underlying GI physiology and its variability should be independent of the drug/formulation administered (unless the drug itself is acting on a certain GI parameter, for example by increasing GI motility or by increasing stomach pH). Recognizing that partial or fully replicated clinical trial designs have more ability to assess WSV, a local and global sensitivity analyses was conducted in Simcyp to understand the most sensitive GI physiology parameters and co-variation’s impact on the PK of Posaconazole in the same set of subjects. Specifically, GI parameters that are influential for the systematic exposure of the drug and sensitivities around the underlying physiological variabilities was determined. The results enabled the PBPK modeling of different WSV combinations in the most influential GI parameters, which were then compared to the observed WSV in PK to eliminate incompatible combinations of WSV in underlying GI physiological variations. The authors were able to arrive at the most plausible combination of WSV parameters for GI physiology that explained the observed WSV in PK parameters of Posaconazole. Following these encouraging results from a proof-of-concept study, if we analyze multiple drugs/formulations with varying sensitivities to GI parameters, we would be able to exclude more and more incompatible physiological WSV combinations leading to narrowing of the true WSV space of physiological parameters. In this case study, we were able to look only at GI parameters’ variability pertaining to this specific drug and formulation were sensitive to. In the future, another drug and formulation could show sensitivity to physiological WSV which it is interacting more. For example, acids and bases would interact differently to gastric pH, but that would not change the variability of gastric pH itself. Thus, analyzing a wider set of compounds and formulations would allow us to derive realistic sets of physiological WSV that are consistent in explaining the observed PK variability via PBPK modeling.
To directly measure the WSV in GI parameters, large sets of repeat measurement studies need to be conducted in human volunteers. Such studies can be expensive, time -consuming and would require a collaborative effort between various labs with expertise in measuring specific GI physiological parameters. Such measurements should be done ideally on the same occasion in the same set of subjects to understand the co-variations of the measured parameters. Given the challenges involved, it is highly unlikely that such clinical measurements would become available in near future. This analysis demonstrates that the Simcyp Simulator can be used to estimate realistic WSV of physiological parameters within the GI tract, enabling crossover VBE studies, and expanding confidence in the use of PBPK for replacing or reducing clinical BE studies. To learn more about this topic, please read this publication, Bego et al. 2022.
Reference:
Bego, M, Patel, N, Christofoletti, R, Rostami, A; “Proof of Concept in Assignment of Within-Subject Variability During Virtual Bioequivalence Studies: Propagation of Intra-Subject Variationin Gastrointestinal Physiology Using Physiologically Based Pharmacokinetic Modeling,” The AAPS Journal (2022) 24:21, DOI: 10.1208/s12248-021-00672-z
