(M1030-07-39) Days or Centuries? Predicting the Shelf-Life of ASDs

Purpose: Amorphous solid dispersions (ASDs) are a state-of-the-art enabling formulation technique for poorly soluble small-molecule drugs. The bioavailability is significantly enhanced by the drug dissolution in a polymer carrier. The drugs’ recrystallization in the polymer is highly unwanted as it marks the end of shelf life of such a formulation. The current work presents a novel fast-track prediction method that combines thermodynamic factors (water absorption, fundamental crystallization driving forces), kinetic factors (glass transition, diffusivity in the ASD), drug-specific crystallization properties (nucleation, crystal growth, glass-forming ability), thermal history (spray drying, melt extrusion, aging) and their mutual impact to predict the shelf life of an ASD at any drug load/ temperature/humidity. This method is developed to predict ab-initio the crystallization onset time and thus the shelf life of metastable ASDs prior to storage tests and hereby dramatically reduces the risk of late-stage product failures due to stability issues.

Methods: The crystallization onset time is predicted combining of few key experiments with a strong in-silico prediction approach. The crystallization velocity of the pure drug is characterized via one simple and fast differential scanning calorimetry or dynamic vapor sorption measurement enduring two days at maximum. This serves as input for the in-silico shelf life prediction tool that considers the different aspects influencing the shelf-life of ASDs. Thermodynamic properties were predicted using the PC-SAFT model^1-4. The glass transition temperatures at dry and humid conditions were predicted with the Gordon-Taylor/Kwei equation and the viscoelastic properties of the polymer matrix as a function of the glass-transition temperature (Tg) with an Arrhenius expression (below Tg) or Williams-Landell-Ferry equation (above Tg)⁵. The nucleation and crystal growth properties were approximated with the Johnson-Mehl-Avrami-Kolmogorow equation. All these contributions to the shelf life of ASDs were linked with an expression derived from the classical nucleation theory, allowing a robust extrapolation to any storage condition of interest. Validation ASDs were manufactured with spray drying, stored for three years at different ICH-compliant storage conditions and characterized in recurring X-ray diffraction measurements.

Results: FEL crystallization onset times observed during storage varied between 9 days (90wt.% FEL/PVPVA64 ASD at 40 °C / 75% RH) and 880 days (60w% FEL/PVPVA64 ASD at 25 °C / 0% RH). This was found in excellent agreement with the shelf life predictions in the entire drug load/ storage temperature/ relative humidity range (Figure 1). Several investigated ASDs remained amorphous at 25 °C / 0% RH, although they were expected to crystallize as well according to the PC-SAFT solubility prediction of FEL in PVPVPA64. As can be seen in Table 1, their predicted crystallization onset time is indeed significantly greater than the duration of this study. The model enabled extrapolating the shelf life (1) of other FEL/PVPVA64 ASDs and (2) at other storage conditions to define critical formulation windows ensuring a desired shelf life (Figure 2). Remarkably, at 40°C/75% RH, the longest predicted shelf life is expected at medium drug loads and not -as intuitively expected- at low drug loads (Figure 2). This high stability can be explained by an optimal ratio of low supersaturation in the polymer and a low impact of humidity due to moderate hydrophilicity of the ASD at that drug load (compare predicted values for water content, viscosity, Tg and supersaturation ratio in Table 1). At low drug loads, the amount of absorbed water increases and thus induces a stronger destabilization. The shelf life of the same ASDs (20w% FEL) stored at 25°C/0% RH tremendously increases, e.g. from 0.02 years (9 days) at 40 °C / 75% RH to 172 years at 25 °C / 0% RH. This remarkable stability of metastable ASDs was also observed by other authors⁶, though they could not estimate the shelf life in their ASDs prior to the experiments.

Conclusion: A novel in-silico model has been successfully applied for predicting the shelf life of metastable ASDs. The shelf life predictions were validated by three-years-enduring stability tests. This approach serves as robust tool for predicting the shelf life of metastable ASDs in the entire temperature/drug load/-humidity space and allows a revolutionary risk assessment of storage test results at early stage of formulation development and even before conducting the tests. This work proofs that the results obtained at so-called ‘accelerated’, harsh storage conditions within short time do not at all reflect the shelf life at more moderate storage conditions, as thermodynamic and kinetic aspects differ completely. The developed model handles these complex dependencies and can be customized to company-specific manufacturing formulations and conditions. It is highly flexible with respect to potential data input and provides a revolutionary approach for ASD shelf-life optimization in the future.
>

References: 1 Luebbert C, Sadowski G. Moisture-induced phase separation and recrystallization in amorphous solid dispersions. Int. J. Pharm. 2017: 532 (1), 635–646.
2 Lehmkemper K, Kyeremateng SO, Heinzerling O, Degenhardt M, Sadowski G. Long-Term Physical Stability of PVP- and PVPVA-Amorphous Solid Dispersions. Mol Pharm. 2017: 14(1):157-171.
3 Prudic A, Ji Y, Luebbert C, Sadowski G. Influence of humidity on the phase behavior of API/polymer formulations. Eur J Pharm Biopharm. 2015: 94:352-62.
4 Wolbert F, Fahrig, IK, Gottschalk T; Luebbert C; Thommes M, Sadowski G. Factors Influencing the Crystallization-Onset Time of Metastable ASDs. Pharmaceutics 2022: 14 (2):269.
5 Wolbert F, Stecker J, Luebbert C, Sadowski G. Viscosity of ASDs at humid conditions. European J. Pharm. Biopharm. 2020: 154, 387–396.
6 Theil F, Anantharaman S, Kyeremateng S, van Lishaut O, Dreis-Kühne H, Rosenberg J, Mägerlein M, Woehrle, GH. Frozen in Time. Kinetically Stabilized Amorphous Solid Dispersions of Nifedipine Stable after a Quarter Century of Storage. Mol. Pharmaceutics 2017: 14 (1), 183–192.

Acknowledgements: The authors declare no conflicts of interest.


Figure 1: Predicted-vs. true plot of observed and computed FEL/PVPVA64 ASD crystallization onset times at the storage conditions 25 °C / 0% RH (green), 25 °C / 60% RH (blue) and 40 °C / 75% RH (orange).


Figure 2: Predicted shelf life of FEL/PVPVA64 ASDs at the storage conditions 25°C/0% RH and 40°C/75%RH as function of FEL load.


Table 1: Shelf-life prediction relevant parameters of all FEL/PVPVA64 ASDs investigated in this work. Italic numbers indicate predicted values of ASDs that did not crystallize during the investigated time period of 880 days. All blue values are predictions obtained from the model.