Predicting process design spaces for spray drying amorphous solid dispersions

Amorphous solid dispersions (ASDs) are commonly manufactured using spray-drying processes. The product quality can be decisively influenced by the choice of process parameters. Following the quality-by-design approach, the identification of the spray-drying process design space is thus an integral task in drug product development. Aiming a solvent-free and homogeneous ASD, API crystallization and amorphous phase separation needs to be avoided during drying. This publication provides a predictive approach for determining spray-drying process conditions via considering thermodynamic driving forces for solvent drying as well as ASD-specific API/polymer/solvent interactions and glass transitions. The ternary API/polymer/solvent phase behavior was calculated using the Perturbed-Chain Statistical Associating Theory and combined with mass and energy balances to find appropriate spray-drying conditions. A process design space was identified for the ASDs of ritonavir and naproxen with either poly(vinylpyrrolidone) or poly(vinylpyrrolidone-co-vinylacetate) spray dried from the solvents acetone, dichloromethane, or ethanol. More on Design space for spray drying of amorphous solid dispersions (full article, open access)

Stefanie Dohrn, Pranay Rawal, Christian Luebbert, Kristin Lehmkemper, Samuel O. Kyeremateng, Matthias Degenhardt, Gabriele Sadowski, Predicting process design spaces for spray drying amorphous solid dispersions,
International Journal of Pharmaceutics: X, 2021,100072, ISSN 2590-1567, https://doi.org/10.1016/j.ijpx.2021.100072.

Materials

The polymers PVP with a molar mass of 1,220,000 g mol⁻¹ (Kollidon® K90), and the copolymer PVPVA64 (Kollidon® VA64) with a molar mass of 65,000 g mol⁻¹ were purchased from BASF (Ludwigshafen, Germany). Naproxen was purchased from Sigma-Aldrich (Steinheim, Germany), ritonavir was obtained from AbbVie Deutschland GmbH & Co. KG (Ludwigshafen, Germany).

Conclusion

This work predicted design spaces for ASD spray-drying process encompassing solvent-drying performance, spray-dryer outlet temperatures, residual-solvent content in the final ASDs, risk of API crystallization in the ASD, and regions of glass transition. It was shown that different drying inlet parameters have a decisive influence on the stability of the obtained ASD after drying. By investigation of h-Xdiagrams, it became visually clear that solvent drying differs due to the different solvent-load capacities of the drying gas, nitrogen. Due to the highest volatility of DCM compared to acetone or ethanol (DCM has the lowest boiling temperature and the lowest evaporation enthalpy), the highest load X was predicted to be achievable with DCM as solvent for the drying process. However, h-X diagrams are only of limited use in designing the ASD spray-drying process, since they consider only the nitrogen/solvent systems and do not account for non-ideality in the vapor phase nor for the API and the polymer in the liquid phase. The influence of solvents, polymers, and API load on the drying performance and on ASD stability was therefore accounted for using PC-SAFT when predicting the process design spaces in this work. It was shown that different drying conditions decisively influence the drying performance and the ASD stability, resulting in different outlet temperatures and residual-solvent contents in the ASD. The influence of the API load on the spray-drying outlet temperature and residual-solvent content was investigated for RIT/PVP ASDs with DCM as solvent and was found to be small. However, by adjusting the nitrogen inlet temperature while keeping the kind of solvent and API/polymer/solvent feed constant, it is possible to obtain ASDs above or below glass transition, metastable API-supersaturated or thermodynamically-stable ASDs. To avoid API crystallization and to generate ASDs below the glass transition, the required inlet temperature (at constant feed rate) significantly depends on the ASD composition. The solvent influence was predicted for NAP/PVP and RIT/PVPVA64 ASDs spray dried using the solvents DCM, ethanol, and acetone. It was found that the solvent-drying performance strongly differs for the same ASD, while additionally the solvent strongly influences the potential risk of API crystallization as well as the glass transition at the end of the drying process. Thermodynamically-stable NAP/PVP ASDs were predicted to be obtained in the whole process design spaces using DCM or ethanol. Nevertheless, due to the more-moderate drying conditions required for a certain residual-solvent content, DCM seems to be the more appropriate solvent compared to ethanol for these ASDs. In case of RIT/PVPVA64 ASDs, it was found that acetone seems to be the more appropriate solvent compared to DCM. Although the drying performance of acetone was predicted being slightly poorer compared to DCM, the risk of RIT crystallization can be reduced using acetone. The polymer influence on the drying performance was investigated for NAP/PVPVA64 and NAP/PVP ASDs spray dried with DCM. It was found that the polymer mainly influences the NAP solubility, while in these systems the solvent-sorption behavior and glass transition was only slightly affected by the polymer. Due to the lower NAP solubility in PVPVA64 compared to PVP, the spray-dryer inlet temperature needs to be higher when using PVPVA64 for obtaining a thermodynamically-stable ASD.

This work thus presents an approach for predicting process design spaces for spray-drying ASDs allowing for the best-achievable product quality without carrying out spray-drying experiments. It could be shown that the process design space of ASDs does not only depend on mass and energy balances, but also on the intermolecular interactions in the API/polymer/solvent system. Knowing the thermodynamic phase behavior comprising API solubilities, solvent sorption and glass transitions combined with spray-dryer mass and energy balances enable identification of spray-drying conditions at which ASDs do not crystallize, contain low amounts of residual solvent, and/or lie below glass transition. This approach therefore is a useful tool for choosing appropriate solvent candidates and process conditions for the ASD spray drying with minimal experimental effort. These process conditions also affect other product properties, like size and shape of the particles and therewith also their dissolution rate. The proposed approach can therefore also be used to support a Design of Experiments in process development.