Issue link: https://www.e-digitaleditions.com/i/1509086
20 Pharmaceutical Technology ® Trends in Formulation 2023 eBook PharmTech.com Drug De velopment Case study #1: Impact of move to new manufacturing site on BE of drug product The sponsor's decision to move production of one of its products to another site resulted in FDA requir- ing that it demonstrate BE bet ween the t wo prod- ucts. Rather than requiring them to run a costly and time-consuming clinical BE study, the agency was open to leveraging PBPK, if t he model was suf f i- ciently verified and scientifically justifiable (2). This complex solid oral product had several unique characteristics that needed to be accounted for in the PBPK model. The particle size distribution was bi- modal and changed over time compared to bio-batch. To simulate the different particle size distributions, the Particle Population Balance (PPB) model was used, which recognizes that each particle in a formulation can be different and that this type of heterogeneity can have an impact on drug dissolution. By account- ing for populations of differently sized particles in a formulation, PPB simulated a realistic dissolution rate, which was validated with prior clinical BE data. The model also helped assess sensitivity to critical product quality attributes and establish safe space. In addition to the particle size complexity, the drug itself had a complex PK profile, including multiple peaks, enterohepatic recycling (EHR) considerations, and food effect. To address these challenges, a PBPK food staggering model was used to ensure that the relevant physiological parameters were adjusted as patients transitioned bet ween fasted-fed and fed- fasted states throughout the day after ingesting the formulation. Historically, PBPK models assumed ei- ther fasted or fed states throughout the simulation, but a subject may not be in a fasted or fed state con- tinuously during the day. The food staggering model allows simulation of feeding and f luid intake events and provides a more realistic simulation of drug ab- sorption from a given formulation. The ADAM model includes dynamic changes in bile salt concentration in response to a feeding event, f luid volume dynam- ics in the GI tract with intake of f luid, and time-de- pendent intestinal and stomach pH models. These are incorporated mechanistically and quantitatively into the final extrapolation between the in-vitro data from the new formulation compared to the historical in-vivo clinical data. The PBPK model helped t he sponsor to rapid ly move production to the new site without performing a clinical study. Case study #2: Defining clinically relevant dissolution specifications using PBPK Per FDA, "PBPK absorption modeling and simulation [M&S] has served as an impactful tool in establishing the relationship of critical quality attributes (CQAs) including formulation variables, specifically in-vitro dissolution, to the in-vivo performance. One common application of PBPK absorption M&S from the perspec- tive of drug product quality is to support the establish- ment of clinically relevant dissolution specifications (CRDS), resulting in many cases in wider acceptance criteria and therefore regulatory flexibility (3). The sponsor had developed a PBPK model for de- termining DDIs and other clinical predictions for elagolix, a non‐peptide, oral gonadotropin‐releas- ing hormone receptor antagonist for endometriosis that was approved in 2018. That base model was re- fined to include the mechanistic absorption model (ADAM) with in-vitro dissolution profiles as input to simulate the dissolution from the formulation. The model was validated by assessing its predictive per- formance by comparing the simulated PK profile and PK parameters with clinical PK of the Phase III formu- lation and the formulation to be marketed, as shown in Figure 2 (4). The simulated area under the concentration-time c u r ve f r om t i me z er o e x t r apol ate d to i n f i n it y (AUC inf ), maximum observed plasma concentration (C ma x ), and time of the maximum observed plasma concentration (T max ) were within 80–125% of observed data for both formulations showing acceptable pre- dictive performance of the IVIVE approach used with the PBPK model. Original dissolution specifications for the elagolix tablets were set and justified based on physicochemical properties of the drug substance as well as release and stability data. However, some of the batches showed slower dissolution. The PBPK model was used to perform a sensitivity analysis of up to 50% slower dissolution profiles on the PK of elagolix. The model showed that up to a 30% decrease in dissolution rate would yield PK profiles that are bioequivalent to the reference commercial formulation. FDA accepted the PBPK model to demonstrate that lots with up to 30% slower release rates compared Biosimulation, which accelerates drug development, is actively used across the industry, and encouraged by FDA and other global regulators.