Purpose: Amorphous solid dispersions (ASDs), formed by molecularly mixing a poorly water-soluble drug with a water-soluble polymer, are widely used to improve the solubility and bioavailability of poorly-soluble drugs. A successful ASD formulation should remain stable throughout its shelf-life, and possess optimal dissolution performance in the human gastrointestinal tract to achieve high bioavailability. In the past, extensive efforts have been directed towards maintaining the solid-state stability of ASDs, including inhibiting drug crystallization, minimizing phase separation, and increasing drug-polymer interactions; whereas there has been less effort to achieve optimal dissolution performance. The impact of ASD microstructure, such as phase separation, on dissolution kinetics remain largely unexplored. Therefore, the purpose of this study was to build a link between solid-state microstructure and solution-state performance of ASDs, and thus enabling rational design and process control of ASD formulations.
Methods: Lopinavir (LPV) was selected as a slow-crystalizing model compound. Hydroxypropylmethylcellulose (HPMC) E5 was used as a model polymer to form ASDs.
ASDs were prepared at different drug loadings (DLs) with selected solvent combinations. A solvent mixture of 1:1 methanol : dichloromethane (v:v) was used to dissolve the drug and the polymer to achieve a total solid content of 4mg/mL. ASD films were prepared by rotary evaporation directly from scintillation vials. To create ASD microstructures, water was introduced during either ASD preparation or storage. Briefly, 4% (v:v, water:solvent, %) water, unless specified elsewhere, was added to the solvent to produce phase-separated ASDs; for storage samples, initially miscible ASD films were stored at 97% RH for 7 days. All films were dried under vacuum overnight and stored in dry conditions.
ASD dissolution was carried out at lopinavir concentrations of 17ug/mL (amorphous “solubility”) and 30ug/mL at all DLs. Sample aliquots were taken at selected time points, filtered, and then analyzed with high performance liquid chromatography (HPLC). The surface morphology of films was determined by atomic force microscopy (AFM) and transmission electron microscopy (TEM). Glass transition temperatures (Tgs) of local nanostructures on the ASD films were determined by AFM-thermal analysis (TA). Homogeneous samples at seven DLs were measured to contruct a calibration curve, and the Gordon-Taylor equation was used to estimate drug content based on Tg.
Results: ASD phase separation greatly altered the dissolution kinetics. At both concentrations tested, improved dissolution was observed for ASDs with 33% to 85% DL with microstructures created through both routes of phase separation, compared to the initially miscible ASDs. At 15%DL, slightly lower drug release was observed in phase-separated ASDs than for miscible ASDs. The extent of increase in dissolution rates was higher in phase-separated samples prepared with water in solvents than those prepared by storage at 97% RH.
Further, the amount of water added to the solvent and drug loading were also found to affect ASD dissolution. An optimal water-to-solvent ratio was found to be 4% to 8% for 33% lopinavir ASDs, and 6% to 20% for ASDs containing 50% drug. Nevertheless, the extent of increase in drug release was much less in 50% lopinavir ASDs than that in 33% DL ASDs.
AFM-TA measurements revealed the formation of continuous phases with different Tgs. The Tgs of pure lopinavir and pure HPMC were 94±6 and 209±11°C. For 33% DL ASD systems, a continuous phase with Tg values of 168±6°C were formed in storage samples, whereas Tgs of 207±5°C were determined in the continuous phase in samples prepared with 4% H2O added in solvent, corresponding to drug concentrations of 20-30% and 0-5%, respectively. The Tgs of discrete phases in these two samples were 91.9±0.2 and 92±5°C, corresponding to 95-100% lopinavir. This is consistent with the lower dissolution rate observed in storage samples. For 33%DL samples prepared with water in solvent, when the water-to-solvent ratio increased to 70% and 100%, Tgs obtained in the continuous phases were 120±10°C and 124±7°C, corresponding to drug contents of 50-80%. The discrete domains in these samples were found to be drug-rich, with Tgs ranging from 97 to 105°C. This suggests the surface of these two phases were both drug-rich, resulting in slow drug release observed.
For 33%DL ASD samples prepared with 8% to 40% water-to-solvent ratios, the Tgs of the continuous phases were found to be similar. Nevertheless, increased water content resulted in the formation of larger discrete domains, suggesting particle size of the discrete phase also played an important role in dissolution.
Conclusion: ASD microstructure formation can affect dissolution performance. The altered dissolution kinetics of phase-separated ASDs may be explained by the creation of multiple phases: polymer-rich regions readily undergo fast dissolution, whereas drug-rich regions dissolve slowly. The composition of polymer-rich regions, the size and location of the drug-rich phases, all affected ASD dissolution rates. Creating microstructures in ASDs formed with slowly crystalizing compounds can be used as a way to improve the dissolution kinetics of high drug-loading ASDs. These results also contributed to the understanding of ASD dissolution mechanisms.
Lynne Taylor– Professor, Department of Industrial and Physical Pharmacy, College of Pharmacy, Purdue University, West Lafayette, Indiana