Survey of Crystallographic Data and Thermodynamic Stabilities of Pharmaceutical Solvates: A Step toward Predicting the Formation of Drug Solvent Adducts

Solid form screens of active pharmaceutical ingredients are a key component of drug development. Beyond forming multiple neat crystal structures, also known as polymorphs, drugs can also crystallize with a variety of solvent molecules to form solvates. They can display physiochemical properties different from the solid form selected for development. Hence, using computational crystal structure prediction (CSP) methods to predict the formation of solvates would be valuable to the pharmaceutical industry. However, CSP for solvate prediction faces a range of obstacles. This work addresses two important challenges: the determination of a computationally feasible search space for CSP on pharmaceutical solvates and the extent to which the formation of these adducts is driven by thermodynamics. First, a survey of ∼12,000 solvates of druglike molecules extracted from the Cambridge Structural Database revealed that a majority of them have stoichiometries and numbers of constituents in the asymmetric unit that are routinely accessible by CSP. Then, quantum mechanical calculations performed on a subset of solvates of drug molecules demonstrated that their formation is generally driven by thermodynamics: ∼80% are more stable than their constituents by an average of 6.6 kJ·mol–1. The implications on the potential of CSP on pharmaceutical solvates are discussed.