We discuss in detail methods for calculating drug-polymer interactions in the important case of formulating poorly soluble drugs as amorphous solid dispersions. Amorphous drugs yield an apparent solubility advantage but are thermodynamically unstable relative to its crystalline counter-part and are at risk of crystallization. Stabilization of amorphous drugs in a polymer matrix is a common formulation strategy. Experiments to understand the physical stability of drug-polymer solid dispersions are time consuming, expensive, yet yield minimal molecular-level understanding of the system. Computational methods that speed up and steer experimental screening in a rational way can be a great and practical tool. As example and proof of concept we demonstrate the methods using two generic drugs (Nifedipine, and Indomethacin) and a few common polymers (PVP, PVAc, and PVP-VA). We tested various approaches: (1) calculating pair energies of interaction between isolated molecules; (2) Molecular Dynamics (MD) methods using atom detailed force fields; (3) rapid screening protocols derived from engineering thermodynamics (quasi-chemical COSMO approach). We compared the theoretical findings with Flory-Huggins parameters extracted from measured melting-point depression curves. Some methods fail: simply calculating pair energies in vacuum may be rapid, but inaccurate. Likewise MD using Dreiding force fields was not sufficiently accurate. However, we do find quite reasonable numbers from the MD approach using a modified OPLS force field, and also the engineering thermodynamics method. In all, the conclusion is that by combination of proper protocols, it is indeed possible to find valuable windows of stability, by computational design only.

Computational design of polymer matrices for formulation of poorly soluble drugs

246th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.