PharmTech: Can predictive modeling be used in excipient selection of a solubilizing polymer? If so, can you explain the modeling and
the related variables used in the model?
Morgen (Bend Research): A range of predictive physical models are helpful in selecting solubilizing polymers. Such models are most valuable when
used in conjunction with measured physical properties of the API, the polymer, and/or the formulation. In particular, we often
use models to estimate oral absorption based on key parameters, such as API dose, solubility and supersaturation values, API
partition coefficient into bile salts, and API dissolution and precipitation rates. Several of these parameters depend on
the choice of polymer. In particular, the time profile of the API supersaturation level is often heavily influenced by the
choice of polymer and can dramatically affect absorption and bioavailability. These parameters are typically measured in vitro and then used in predictive models. An important caveat is that it is notoriously difficult to accurately predict certain
in vivo parameters, such as API precipitation rate based on in vitro measurements. Other types of models that are sometimes employed include solubility parameter approaches to predict interactions
between APIs and polymers, including estimates of API solubility in a polymer matrix. Although such predictions can be useful,
it is sometimes faster and more accurate to make the measurements than to do the calculations.
Koblinski (Dow): Predictive modeling can be used to calculate polymer–API interactions that may occur. In addition, tools, such as solubility
parameters, can be used to establish API solubility in the selected polymer. Hansen solubility parameters are often used in
this regard as a screening tool for drug-polymer compatibility, however, they should only be used as a guide in conjunction
with preformulation data.
Asgarzadeh (Evonik): In a conventional empirical formulation development approach, various qualitative and quantitative combinations of drug and
polymers are melted, spray-dried, or film-casted from organic solvents in numerous experiments to identify appropriate solid
solution/dispersion formulations. Such empirical development methodologies are time-consuming and require significant amount
of drug as well as costly analyses. Predictive tools to identify solvents for polymers based upon solubility parameters and
molecular interactions (e.g., hydrogen bonding and ionic interactions) have been used for several decades in paint, polymer,
and organic-chemistry industries. These predictive tools afford faster systematic screening of formulations and processing
conditions at the early stages of solid-dispersion product development, independent of the preparation technology used.
One such tool developed by Evonik is MemFis (Melt-Extrusion Modeling and Formulation-Information System). The MemFis model
uses well-established polymer and organic-chemistry group contribution theories to estimate Hansen solubility parameters of
drug molecules and polymers. In MemFis, the calculations of more than 50 chemical group contributions and the effects of polar
(i.e., dipole moments) as well as 40 different hydrogen-bonding interactions on solubility parameters are considered. MemFis
enables the selection of initial solid-solution/dispersion formulations and melt-extrusion processing conditions without API
consumption. It reduces the number of experiments by bringing quality into formulation development in a systematic rather
than empirical approach targeting appropriate experiments. Other analytical tools, such as Raman mapping, differential scanning
calorimetry, and atomic force microscopy, can allow early detection of solid-solution formation or any recrystallization effects.