Regulatory Experience with In Vivo In Vitro Correlations (IVIVC) in New Drug Applications

Abstract

In the past two decades, in vitro in vivo correlation (IVIVC) has been considered an important tool for supporting biowaivers, setting dissolution acceptance criteria, and more recently in the Quality by Design (QbD) framework promoting the establishment of clinically meaningful drug product specifications using dissolution as the endpoint. Based on our review experience at the FDA, for the purposes of this article, we analyzed the current state of regulatory submissions containing IVIVC approaches and discussed the successes and failures from the perspectives of study design to methodology. In the past decade, the overall acceptance rate of the IVIVC submissions is about 40%. Moreover, the number of IVIVC studies seen in the submissions per year is not increasing. Establishing clinically meaningful drug product specifications through the linkages between the identified critical quality attributes and in vivo performance is key for developing a quality drug product. To achieve this goal, there is an imminent need for addressing the issues behind a low success rate in IVIVC development. The results from the current analysis revealed that special considerations should be taken in areas such as (1) selection of appropriate number/kind of formulations for IVIVC development/validation, (2) construction of exploratory plots to guide model building and selection, (3) investigation of the reasons of inconclusive predictability, (4) improvement on the quality and richness of the data, and (5) avoidance of over parameterization. The development and incorporation of biopredictive dissolution methods and the use of non-conventional approaches, including mechanistic/physiologically based approaches, should be explored to increase the likelihood of IVIVC success.

Developing Quantitative In Vitro–In Vivo Correlation for Fenofibrate Immediate-Release Formulations With the Biphasic Dissolution-Partition Test Method

“…As either individual or mean pharmacokinetic profiles can be deconvoluted by the W-N method, we chose to use mean plasma drug concentrationetime profiles based on scientific literature on this subject and recommendations in the studies. [31][32][33][34] Calculation of fraction of absorptionetime profiles was performed with the use of the W-N equation through Excel (Microsoft). The elimination rate constant (k e ) was estimated from the least square analysis of the terminal portion of the FEN pharmacokinetic profile by using the equation:…”

An updated overview with simple and practical approach for developing in vitro–in vivo correlation

“…In this context, implementation of quality by design (QbD) and identifying design space has been expected to provide confirmation of quality. Shifting out of design space may be treated as change process and would normally require a regulatory approval (Suarez‐Sharp, Li, Duan, Shah, & Seo, ; Yu et al, ).…”

In Vitro-In Vivo Correlation (IVIVC): From Current Achievements Towards the Future

“…Although complex systems are being developed and regulators allow nonconventional modeling techniques to be used, provided they meet specified criteria for predictability, conventional IVIVC is mainly established and presented in NDAs (6,68). All the newer methods for achieving IVIVC are still in early development phases and are too new to be routinely used for regulatory purposes.…”

“…According to survey conducted by Fotaki et al, the majority of IVIVC types of modeling approaches are using deconvolution and the simple linear models to achieve Level-A correlation (71%) (5). Moreover, as Suarez-Sharp et al pointed out, most new drug applications (NDAs) use a two-step model-independent approach to establish IVIVC (6). During recent years, a strong increase of scientific publications with the focus on IVIVC has been observed.…”