Research Article: First-Principles and Empirical Approaches to Predicting In Vitro Dissolution for Pharmaceutical Formulation and Process Development and for Product Release Testing

Date Published: February 21, 2019

Publisher: Springer International Publishing

Author(s): Nikolay Zaborenko, Zhenqi Shi, Claudia C. Corredor, Brandye M. Smith-Goettler, Limin Zhang, Andre Hermans, Colleen M. Neu, Md Anik Alam, Michael J. Cohen, Xujin Lu, Leah Xiong, Brian M. Zacour.


This manuscript represents the perspective of the Dissolution Working Group of the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ) and of two focus groups of the American Association of Pharmaceutical Scientists (AAPS): Process Analytical Technology (PAT) and In Vitro Release and Dissolution Testing (IVRDT). The intent of this manuscript is to show recent progress in the field of in vitro predictive dissolution modeling and to provide recommended general approaches to developing in vitro predictive dissolution models for both early- and late-stage formulation/process development and batch release. Different modeling approaches should be used at different stages of drug development based on product and process understanding available at those stages. Two industry case studies of current approaches used for modeling tablet dissolution are presented. These include examples of predictive model use for product development within the space explored during formulation and process optimization, as well as of dissolution models as surrogate tests in a regulatory filing. A review of an industry example of developing a dissolution model for real-time release testing (RTRt) and of academic case studies of enabling dissolution RTRt by near-infrared spectroscopy (NIRS) is also provided. These demonstrate multiple approaches for developing data-rich empirical models in the context of science- and risk-based process development to predict in vitro dissolution. Recommendations of modeling best practices are made, focused primarily on immediate-release (IR) oral delivery products for new drug applications. A general roadmap is presented for implementation of dissolution modeling for enhanced product understanding, robust control strategy, batch release testing, and flexibility toward post-approval changes.

Partial Text

Orally administered solid dosage forms (tablets and capsules) constitute a large fraction of pharmaceutical products. These formulations are designed to release the active pharmaceutical ingredient (API) through the patient’s gastrointestinal (GI) tract in a prescribed manner. Understanding the in vivo mechanism of API release and absorption is a key objective to streamline and optimize the development of orally administered drug products. Dissolution testing is an in vitro laboratory performance test that assesses how efficiently a drug is released from its dosage form. During drug development, dissolution profiles have been used to understand the impact of formulation composition and process parameters on the in vitro release of API. Dissolution testing also plays an important role in the context of science- and risk-based process development, validation, evaluation of post-approval formulation changes to drug product quality, assessment of bioequivalence, and as a surrogate for in vivo drug release. In manufacturing, in vitro dissolution has been used routinely as a quality control (QC) release test to ensure batch-to-batch manufacturing consistency or quality. It has become an integral part of regulatory filings worldwide, with the expectation to serve as a QC tool to detect critical quality attribute (CQA) changes that affect in vivo release leading to exposure (i.e., bioperformance). It is also part of regulatory requirements to apply for a waiver of in vivo bioequivalence studies based on predictive in vitro/in silico methods (post-approval changes, new strengths, formulation modifications) (1–5).

The development of predictive in vitro dissolution modeling applies the same principles as that of traditional experimental dissolution methods, with the same criteria for rejecting non-bioequivalent batches (i.e., dissolution behavior shown to correspond to an unacceptable deviation in expected bioperformance). In development of predictive dissolution models, empirical and first-principles-based approaches have been documented for a range of intended purposes. Combinations of the two, such as the use of first-principles-derived parameters as inputs for an empirical model and vice versa, are also common. Here, we categorize modeling approaches based on the type of model used for decision-making, regardless of the input source. For example, a hybrid model using first-principles-determined parameters as inputs for an empirical model for decision-making is considered as an empirical modeling approach, and vice versa. Figure 1 provides a cartoon description of the progression and use of dissolution modeling across a drug product development timeline. With project progression, as the amount of data and maturity of knowledge increase, so do the model capabilities and predictive power. Early in development, with a dearth of data, first-principles-based models are created to aid in formulation development and process screening. With increased data and knowledge, these models inform or mature into data-driven predictive models to enable RTRt and QC testing.Fig. 1Description of the progression and use of dissolution modeling across a drug product development timeline

The paper illustrates a general strategy for predictive in vitro dissolution model development and a general roadmap (Fig. 10) for its implementation for enhanced product understanding, robust control strategy, batch release testing, and flexibility for post-approval changes. The selection among various modeling approaches based on product and process understanding is demonstrated at different phases of drug development. Early in the research and development process, when little data and process understanding is available, first-principles models can be used to provide guidance for formulation and process development. As greater amounts of data and understanding are generated, a first-principles-based and data-driven empirical approach becomes affordable for linking material attributes and process conditions to the drug product dissolution profile in order to enable a predictive dissolution model for batch or real-time release. Post-approval changes can utilize the same framework, relying on first principles to understand the effect of the change and on existing or extended empirical dissolution models for product release with minimal additional experimental burdens.Fig. 10General roadmap for predictive in vitro dissolution model development