Screening of Bioequivalent Extended-Release Formulations for Metformin by Principal Component Analysis and Convolution-Based IVIVC Approach

Bioequivalence (BE) is usually hard to achieve for extended-release (ER) dosage form products due to not only its complicated formulation but also to the BCS classification of the investigated drugs. Considering the difficulties in establishing full-scale IVIVC and limited in vivo pharmacokinetics data in the early stage of formulation development, we have selected BCS III drug metformin as a model drug to demonstrate a novel approach for the selection of BE formulations. Firstly, dissolution tests in both standard and biorelevant media were performed followed by identification of the most similar formulation WM to the reference product (GXR) based on principal component analysis (PCA) of the dissolution data. Then, we developed an IVIVC model using the reported GXR pharmacokinetics profiles via a convolution-based approach. Based on our established IVIVC and in vitro dissolution profiles of generic metformin ER products, we were able to predict their in vivo pharmacokinetic profiles and quantitatively compare the differences in AUC and C_max to ensure the correct selection of BE product. Finally, the selection of WM as the BE formulation of GXR was confirmed with a pilot BE study in healthy volunteers under fasting state. Moreover, the in vivo data from the fed state study were further integrated into our IVIVC model to identify FeSSIF-V2 as the biorelevant media for WM. Our novel integrative approach of PCA with a convolution-based IVIVC was successfully adopted for the screening of the BE metformin ER formulation and such an approach could be further utilized for the effective selection of BE formulation for other drugs/formulations with complex in vivo absorption processes.