(W1230-07-41) Solubility Measurement and Modeling of Various Drugs in Binary Solvent Mixtures at Different Temperatures

Purpose: The main purpose of this study is to measure the solubility of various drugs in binary solvent mixtures at different temperatures, and to identify correlations of experimental data with calculation data by applying various computational models. Furthermore, types of solubility are divided by pattern of solubility profile.

Methods: The solubility in the binary solvent mixtures was determined through the solid–liquid (S-L_ equilibrium by utilizing the shake-flask method under atmospheric pressure. The concentration of drugs in the solvent mixtures was estimated using a HPLC system. The solid-state of raw and equilibrated samples were characterized by differential scanning calorimetry (DSC) and powder X-ray diffractometry (PXRD) analyses to confirm degradation and/or crystalline transformation during solubility studies. The experimental solubility data for drugs was fitted by computational models, such as the van’t Hoff, modified Apelblat, simplified CNIBS/R-K, Jouyban-Acree, Ma, and Sun models, to confirm correlation between experimental and calculation data.

Results: The various poorly soluble drugs showed different solubility patterns in the binary solvent mixtures. First, the solubility profile of lifitegrast in the PEG 400 + water mixture was the highest at a 1.0 mass ratio of PEG 400 (w₁). Interestingly, the solubility profile in the DEGME + water mixture increased until the mass fraction of DEGME (w₁) increased to 0.9. The 0.9 point showed higher solubility than the 1.0 point due to the synergistic effect between the solvents. The experimental solubility of lifitegrast at 0.9 point at 298.15 K was about 20,800-fold higher than solubility in pure water. Secondly, the solubility profile of rivaroxaban in the dichloromethane + primary alcohol mixtures was represented as concave curves which has maximum values at certain point in all temperature ranges. As increased mass fraction of dichloromethane (w₁), the value of solubility was increased and reached the highest peak at w₁ = 0.8. Its solubility was approximately 11.18 and 85.52 times higher than solubility of pure dichloromethane and pure methanol at 298.15 K, respectively. Similarly, the solubility of telmisartan in dichloromethane + primary alcohol mixtures also showed the highest solubility at the point where the mass fraction of dichloromethane was 0.8. In the binary solvent mixtures, primary alcohol and dichloromethane acts as a hydrogen bond accepting and a hydrogen bond donating solvent, respectively, and these interactions can induce a solvent synergy effect. Certain mass fraction of the binary solvent mixture provides a more polar environment than pure primary alcohol or pure dichloromethane, resulting in high solubility for poorly-water soluble drugs. Finally, in case of resveratrol, the solubility in dichloromethane + primary alcohol mixtures was highest in pure primary alcohol, and the graph slope (mass fraction-solubility) was decreased as the mass fraction of primary alcohol increased. Furthermore, the estimated results obtained using the computational models through the regression analysis have a good correlation with solubility data of various drugs at different compositions of solvent mixtures and temperatures. Therefore, all the models used in this study can be useful for predicting the solubility of various drugs.

Conclusion: Ideally, the solubility profile of the drug in the binary solvent mixture would increase linearly with the addition of solvent 2 to solvent 1, but in practice, as shown in the results of solubility experiments for several drugs, this does not occur. These results of solubility experiments for several drugs with such unpredictable variability can be divided into widely three types. Type (i) indicates that the solubility slightly increases as the solvent 2 is added, but rises rapidly above a certain point, and the data can be used to determine the ratio of antisolvent and solvent, and concentration of the drug solution when manufacturing particles from a top-down method, especially using antisolvent precipitation method. Type (ii) indicates that the solubility rises rapidly as the solvent 2 is added, and then there is no significant change in solubility even after a certain point, which can be used to secure the solvent composition that can dissolve both the drug and the specific excipient in the formulation study. Type (iii) indicates the synergistic effect at a specific ratio of solvent mixture, much greater than solubility in pure solvent 1 and 2, and can be available to identify solvent composition with sufficient solubility. Therefore, these experimental solubilities and computational models for drugs can be useful in process development for the purification and micro/nanoparticles preparation, as well as in future studies on drugs with respect to the formulation design and development of dosage forms.

Acknowledgments: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2020R1A2C4002166).