Drug−Polymer−Water Interaction and Its Implication for the Dissolution Performance of Amorphous Solid Dispersions Yuejie Chen, † Chengyu Liu, † Zhen Chen, † Ching Su, ‡ Michael Hageman, ‡ Munir Hussain, § Roy Haskell, ∥ Kevin Stefanski, ‡ and Feng Qian* ,† † Department of Pharmacology and Pharmaceutical Sciences, School of Medicine, and Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, Tsinghua University, Beijing 100084, China ‡ Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Company, Lawrenceville, New Jersey 08648, United States § Drug Product Science and Technology, Bristol-Myers Squibb Company, New Brunswick, New Jersey 08903, United States ∥ Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Company, Wallingford, Connecticut 06492, United States ABSTRACT: The in vitro dissolution mechanism of an amorphous solid dispersion (ASD) remains elusive and highly individualized, yet rational design of ASDs with optimal performance and prediction of their in vitro/in vivo perform- ance are very much desirable in the pharmaceutical industry. To this end, we carried out comprehensive investigation of various ASD systems of griseofulvin, felodipine, and ketoconazole, in PVP-VA or HPMC-AS at different drug loading. Physiochemical properties and processes related to drug−polymer−water interaction, including the drug crystallization tendency in aqueous medium, drug−polymer interaction before and after moisture exposure, supersaturation of drug in the presence of polymer, polymer dissolution kinetics, etc., were characterized and correlated with the dissolution performance of ASDs at different dose and different drug/polymer ratio. It was observed that ketoconazole/HPMC-AS ASD outperformed all other ASDs in various dissolution conditions, which was attributed to the drug’s low crystallization tendency, the strong ketoconazole/HPMC-AS interaction and the robustness of this interaction against water disruption, the dissolution rate and the availability of HPMC-AS in solution, and the ability of HPMC-AS in maintaining ketoconazole supersaturation. It was demonstrated that all these properties have implications for the dissolution performance of various ASD systems, and further quantification of them could be used as potential predictors for in vitro dissolution of ASDs. For all ASDs investigated, HPMC- AS systems performed better than, or at least comparably with, their PVP-VA counterparts, regardless of the drug loading or dose. This observation cannot be solely attributed to the ability of HPMC-AS in maintaining drug supersaturation. We also conclude that, for fast crystallizers without strong drug−polymer interaction, the only feasible option to improve dissolution might be to lower the dose and the drug loading in the ASD. In this study, we implemented an ASD/water Flory−Huggins parameter plot, which might assist in revealing the physical nature of the drug−polymer interaction. We also introduced supersaturation parameter and dissolution performance parameter as two quantitative measurements to compare the abilities of polymers in maintaining drug supersaturation, and the dissolution performance of various solid dispersions, respectively. KEYWORDS: amorphous solid dispersion, drug−polymer interaction, Flory−Huggins interaction parameter, dynamic vapor sorption, dissolution, FT-IR, NMR, crystallization ■ INTRODUCTION A successfully designed amorphous solid dispersion system (ASD) must possess two key characteristics: (1) physical stability during downstream processing and storage and (2) optimal dissolution performance upon in vivo dosing to achieve its desired bioavailability enhancement. In the past, there have been extensive studies focusing on the first area, and topics such as the thermodynamic aspects of amorphous small mole- cule, 1−10 the molecular interaction between drug and polymer, 11−14 crystallization of amorphous drugs in the presence of polymers, 15−17 the phase behavior of the binary drug−polymer systems, 18−22 etc. have been widely investigated and reviewed. Satisfactory physical stability is only halfway toward the ultimate success of these thermodynamically metastable systems. We have demonstrated earlier that a physically stable ASD could have disappointing in vivo bioavailability perform- ance, and ASDs of the same drug, same drug loading, yet different polymer carriers, could perform very differently in vivo despite their equally adequate physical stability in the dry state. 23 The poor in vivo performance predictability of the ASDs has become increasingly recognized as a touchy issue during Received: October 2, 2014 Revised: December 18, 2014 Accepted: December 23, 2014 Published: December 23, 2014 Article pubs.acs.org/molecularpharmaceutics © 2014 American Chemical Society 576 DOI: 10.1021/mp500660m Mol. Pharmaceutics 2015, 12, 576−589