Solubilities of crystalline drugs in polymers: an improved analytical method and comparison of solubilities of indomethacin and nifedipine in PVP, PVP/VA, and PVAc.

A previous method for measuring solubilities of crystalline drugs in polymers has been improved to enable longer equilibration and used to survey the solubilities of indomethacin (IMC) and nifedipine (NIF) in two homo-polymers [polyvinyl pyrrolidone (PVP) and polyvinyl acetate (PVAc)] and their co-polymer (PVP/VA). These data are important for understanding the stability of amorphous drug-polymer dispersions, a strategy actively explored for delivering poorly soluble drugs. Measuring solubilities in polymers is difficult because their high viscosities impede the attainment of solubility equilibrium. In this method, a drug-polymer mixture prepared by cryo-milling is annealed at different temperatures and analyzed by differential scanning calorimetry to determine whether undissolved crystals remain and thus the upper and lower bounds of the equilibrium solution temperature. The new annealing method yielded results consistent with those obtained with the previous scanning method at relatively high temperatures, but revised slightly the previous results at lower temperatures. It also lowered the temperature of measurement closer to the glass transition temperature. For D-mannitol and IMC dissolving in PVP, the polymer's molecular weight has little effect on the weight-based solubility. For IMC and NIF, the dissolving powers of the polymers follow the order PVP > PVP/VA > PVAc. In each polymer studied, NIF is less soluble than IMC. The activities of IMC and NIF dissolved in various polymers are reasonably well fitted to the Flory-Huggins model, yielding the relevant drug-polymer interaction parameters. The new annealing method yields more accurate data than the previous scanning method when solubility equilibrium is slow to achieve. In practice, these two methods can be combined for efficiency. The measured solubilities are not readily anticipated, which underscores the importance of accurate experimental data for developing predictive models.