(M1330-08-54) Critical Insights into High Drug Load Amorphous Formulations: Challenges and Advanced Techniques

Purpose: In the current pharmaceutical paradigm, most new drugs fall into the BCS Class II category, which have low water solubility. Amorphous solid dispersion (ASD) is a viable strategy to overcome dissolution challenges. However, the drug load in the polymer matrix is limited to retain the amorphous state. Exceeding this limit (LoC) can decrease solubility and dissolution rates. While spray drying with copovidone is robust for ASDs, characterization at the drug’s LoC is limited. This study explores advanced techniques to understand formulation risks at the LoC.

Methods: ASD batches with 20% w/w drug loading were prepared at two scales using spray drying (10 g, ASD20-S; 50 g, ASD20-L). Danazol (Coral drug Pvt Ltd) and copovidone (Kollidon® VA 64; BASF) were dissolved in ethanol (10% w/v) and spray dried with a Buchi B-290 mini spray dryer under nitrogen. Spray drying parameters: inlet air temperature 68 °C, aspirator 100%, and pump feed rate 2-3 mL/min; outlet air temperature was ~41 °C. The ASDs were vacuum dried at 40 °C for 12 h. Glass transition temperature (Tg) was measured via DSC (TA DSC Q2000). Water and volatiles content were measured by Karl Fischer titration (Mettler Toledo DL39) and TGA (TA Instruments TG55). For dissolution, ASDs in “000” capsules (80 mg drug) were tested in USP Apparatus II (1000 mL FaSSIF v2, 100 rpm, 37 °C), with drug release measured by UV-Vis. ASDs were compressed into pellets for Raman mapping (HORIBA XploRA Confocal Raman Microscope) over 100x100 μm areas, using 3- or 5-μm steps. XRPD patterns were collected on a PANalytical Empyrean diffractometer with Cu Kα radiation (2-40° 2θ range), with samples between 3-μm-thick films analyzed in transmission geometry.

Results: Ethanol was selected as a solvent for the spray drying process due to the solubility of copovidone (up to 40% solids content) and danazol; and to attain a suitable viscosity for the spray solution ( < 300 mPas). The large batch of ASDs showed only 30% drug release, while the small batch achieved 70%, despite identical composition and process parameters. During the large batch spray drying, a pressure drop caused by filter clogging was observed, prompting further investigation into the properties of both ASD batches. Water content of ASD20-S (~3.31%) was slightly higher than water content of ASD20-L (~1.7 %). Volatiles content (TGA) data are similar for all batches, with 2.85% weight loss for ASD20-S and 1.62% weight loss for ASD20-L. XRPD results indicate all samples are amorphous, with some variations in shape and position of the halos. DSC analysis showed all ASDs batches had a Tg around 102 °C, higher than amorphous danazol (73 °C) [1] and similar to copovidone. No danazol melting (Tm) appeared in the thermogram. Earlier work [2] demonstrated that, although DSC and XRPD found no significant differences in ASDs at various drug loadings, Raman mapping revealed wavelength shifts as drug load increased. Lower loading yielded greater amorphous character; higher loading displayed more crystalline features. Raman spectra compared the two ASD batches (Figures 2a and 2b). The C=O stretching region from both batches is compared to a co-povidone reference in Figure 2a. The large batch C=O stretch matches co-povidone at 1670 cm-1, while the small batch C=O stretch shifts to 1665 cm-1. This shift suggests polymer H-bonding in the small batch, possibly due to the drug or higher water content [3]. The C=C stretch and aromatic ring mode region (1560-1650 cm-1) for both batches is compared to a danazol reference in Figure 2b. Both batches show broadening of the major peak relative to crystalline danazol, indicating the API is amorphous in both dispersions. To further interrogate the C=O stretch for danazol, intensity profiles were generated from the Raman maps of both batches. Intensity at 1598 cm-1 (crystalline danazol) was divided by intensity at 1602 cm-1 (amorphous danazol). Figure 3a shows the small batch (left) with mostly blue pixels, while the large batch (right) has more red pixels, suggesting micro-domains of crystalline danazol in the large batch. Extracted spectra from red and blue pixels revealed slight shifts, and subtracting these produced a different spectrum matching crystalline danazol. Though the overall dispersion is amorphous, the presence of these crystalline regions could account for the differences in dissolution observed. Even though the overall dispersion itself is amorphous (represented by the peak width), there are some regions that still retain a detectable amount of crystalline character of the danazol. These micro-domains of crystalline character could explain the differences observed with the dissolution results.

Conclusion: This study highlights the risks of developing ASD formulations with high drug loads at the limit of congruency. Slight variations in process parameters during scale-up can negatively impact results. Critical quality differences between batches should be evaluated with advanced tools like Raman mapping. To improve process success and formulation robustness, target drug loading below the LoC and optimize composition within a broader design space.

References: [1] Claudia Borredon, Luis A. Miccio, Anh D. Phan, Gustavo A. Schwartz, Estimating glass transition temperature and related dynamics of molecular glass formers combining artificial neural networks and disordered systems theory, Journal of Non-Crystalline Solids: X, Volume 15, 2022, 100106, ISSN 2590-1591,
[2] K. Chen, N. Swarnakar, S. Byrn, D. Engers, D. Purcell, P. Smith, S. Tiwari, “Copovidone-Based Spray Dried Amorphous Solid Dispersions: Investigation of Drug Loadings and the Limit of Conguency”, presented at AAPS PharmSci 360 (2024)
[3] Chen, Y., Wang, S., Wang, S. et al. Initial Drug Dissolution from Amorphous Solid Dispersions Controlled by Polymer Dissolution and Drug-Polymer Interaction. Pharm Res 33, 2445–2458 (2016).

Acknowledgements: No public funds supported this work. The authors have no conflicts of interest to disclose.

Figure 1 Dissolution data of 20 DL Danazol ASDs



Figure 3 (a) Comparison of the crystalline character for the small batch (left) vs a large batch (right). Note: the map on the left used a step size of 3 microns whereas the map on the right used a step size of 5 microns. (b) Comparison of spectra extracted from the large batch map (blue, red), a spectral subtraction (pink), and a reference spectrum of danazol (green)