Broadening the scope of amorphous solid dispersions: alternative manufacturing and formulation approaches

Because of the different challenges still faced in solid dispersion research today, an alternative manufacturing approach was investigated in the form of fluid bed coating, which enables efficient drying and the omission of additional processing steps which can destabilize solid dispersions. Also, alternative formulation approaches are exploited by incorporating controlled release polymers into solid dispersions, either as a rate controlling membrane or as (part of) the solid dispersion carrier.

A first study describes the search for an mDSC sample preparation method for the analysis of INDO-PVP glass solutions coated onto inert sucrose beads using fluid bed coating. The spherical shape of these beads compromises the contact area with the bottom of a DSC sample pan. Grinding the coated beads and separating the resulting particles into different particle size ranges, resulted in a visible glass transition signal. This glass transition, however, shifted and broadened in particle samples with increasing size range. This phenomenon was confirmed in glass solutions prepared as isolated films by rotary evaporation, with different API’s and from different solvent systems. Resulting from TGA analysis, where a difference was made between sub-Tg and above-Tg residual solvent evaporation, it could be concluded that the observed Tg shift and broadening could be ascribed to the differences in residual solvent mass loss from the bulk of the particles and from the surface. Since particles with a smaller size range exhibit a higher surface to mass ratio, they possess more solvent poor surface, as compared to particles with a larger size range, which possess more solvent rich bulk. These findings were confirmed by a correlation between the deviation from the Gordon-Taylor derived Tg and solvent mass loss from the Tg on.

In order to control the release of INDO from coated glass solutions, an additional rate controlling coating was applied on top of the glass solution. This can be done consecutively using fluid bed coating. The resulting multilayer coated beads require a combination of surface and bulk characterization to understand the phase behaviour. Different rate controlling membranes were applied on INDO-PVP glass solutions comprising two possible controlled release polymers, different amounts of pore former can be added and the rate controlling membrane can be applied from an ethanol solution or an aqueous dispersion. The investigation of these different formulations on the phase behaviour of the drug delivery system is described in a second investigation. Surface and cross-sectional topography was investigated by SEM. Chemical composition and distribution analysis of these surfaces and cross-sections was performed using ToF-SIMS. Polymer miscibility was assessed with mDSC and crystallinity with XRPD. Topography differences observed on the surface or in the cross-sections of the coated beads can be ascribed to polymer miscibility differences or coating from a solution or a dispersion. PVP presence at the surface of pure ERL or EC coatings is the result of a coating contamination. The distributional changes of PVP, when incorporated as a pore former can also be explained by polymer miscibility differences. Limited INDO and PVP migration into the rate controlling membrane can be evidenced from cross-sectional ToF-SIMS analysis. Lastly, XRPD analysis shows that INDO remains amorphous after application of a rate controlling membrane, even if it is coated from an aqueous dispersion.

The influence of the above described formulation changes and rate controlling membrane thickness on the release of INDO was investigated. In addition to this, the role of a charge interaction between drug and controlled release polymer on the release of the former was investigated as well. Diffusion experiments showed a clear influence of the controlled release polymer used, pore former concentration and coating from a solution or suspension on the permeability of rate controlling membranes. These findings could be readily translated to their influence on drug release, pinpointing diffusion through the rate controlling membrane as the rate limiting step for drug release and showing the potential of these diffusion experiments for screening purposes of rate controlling membranes. A charge interaction between INDO and ERL was confirmed by ss-NMR but no clear influence of this interaction on the drug release was observed. The diffusion and release differences through ERL and EC coatings are mainly the result of the higher hydrophilicity of the former.

Finally, the use of ERL as a solid dispersion carrier is investigated, either alone or in combination with the hydrophilic polymer PVP. The solid dispersions are produced by spray drying and analysed with respect to their phase behaviour and in vitro drug dissolution. After in vitro dissolution, precipitates are collected and analysed again with mDSC. ERL solid dispersions with INDO and NAP showed extended supersaturated drug concentrations, when compared to the hydrophilic polymer PVP. Combinations of PVP and ERL as a carrier combined this extended supersaturation with higher drug concentrations compared to ERL alone. Phase behaviour analysis showed that ERL can form glass solutions and, in the case of INDO, one phase systems are found after 24h dissolution as well. Low drug loadings in combination with ERL as a carrier resulted in slow diffusion out of the carrier making this approach unfavourable. Oversaturated INDO and NAP solutions formed stable nanocrystals in presence of ERL. This formation can be explained by a dynamic interplay of dissolution, sorption and desorption. High sorption levels are necessary for this nanocrystal formation, and a charge interaction between INDO/NAP and ERL provide the necessary driving force for sorption.