Solubility of Benzodiazepines in Polyethylene Glycol 200 + Water Mixtures at 303.2 K Abolghasem Jouyban,* ,† Javad Shokri, ‡ Mohammad Barzegar-Jalali, ‡ Davoud Hassanzadeh, § William E. Acree, Jr., | Taravat Ghafourian, ⊥ and Ali Nokhodchi ⊥ Faculty of Pharmacy and Drug Applied Research Center, Biotechnology Research Center, and Research Center for Pharmaceutical Nanotechnology, Tabriz University (Medical Sciences), Tabriz 51664, Iran, Department of Chemistry, University of North Texas, Denton, Texas 76203-5070, and Medway School of Pharmacy, Universities of Kent and Greenwich, Kent ME4 4TB, United Kingdom Experimental solubilities of chlordiazepoxide, diazepam, and lorazepam in polyethylene glycol 200 (PEG 200) + water mixtures at 303.2 K were reported. The solubility of each drug increased exponentially with the addition of PEG 200 and reached the maximum value in neat PEG 200. The Jouyban-Acree model was used to mathematically describe the experimental data, and the solubilities were predicted using a previously trained version of the Jouyban-Acree model for PEG + water mixtures and the solubility data in monosolvents. The overall mean relative deviations of the models were 3.7 % and 18.3 %, respectively, for the fitted model and the trained version. Introduction Polyethylene glycols (PEGs) are water-soluble polymers with a general formula of HO-(CH 2 CH 2 O) n -H and have been used in many pharmaceutical formulations including oral, parenteral, topical, ophthalmic, and rectal preparations. 1,2 Generally, they are regarded as nontoxic and nonirritant materials. PEGs are also utilized in the industry to precipitate/crystallize proteins, and stable formulation of protein powders has been developed employing PEG-induced precipitation. 3 PEGs are used in pharmaceutical formulations by acting as a cosolvent (lower molecular weight PEGs) to increase the aqueous solubility of poorly water-soluble drugs in liquid pharmaceutical formulations and/or as dissolution rate enhancers (higher molecular weight PEGs) in solid dispersions. 4 PEGs are stable and low toxic pharmaceutical excipients which are used in many commercially available oral and parenteral pharmaceutical formulations of poorly soluble drugs. 5,6 In addition to the solubilization power of polyethylene glycol 200 (PEG 200), it possesses stabilization ability on the enzymes and facilitates the biotransformation of low aqueous soluble substrates in aqueous solutions. 7 The solubility of drugs in PEG 200 + water mixtures is essential information in drug development studies. The composition of PEG 200 in pharmaceutical preparations should be kept as low as possible (usually a volume fraction less than 50 %), and the often used method to optimize the solvent composition of the mixtures for dissolving a desired amount of a drug in a given volume of the solution is the trial-and-error approach which is both time- consuming and expensive. Moreover, in the early stages of drug discovery processes, the scarcity of the drug and/or drug candidate is another limiting factor. To address this issue, numerous mathematical models have been presented for predicting the solubility of drugs in water-cosolvent mixtures. These models and their advantages and limitations were recently reviewed. 8 Of the models developed in recent years, the Jouyban-Acree model is perhaps one of the more versatile models. The model provides very accurate mathematical descriptions for how the solute solubility varies with both temperature and solvent composition. The model is log x m,T Sat )  1 log x 1,T Sat +  2 log x 2,T Sat +  1  2 T/K ∑ i)0 2 J i ( 1 -  2 ) i (1) where x m,T Sat is the solute (mole fraction) solubility in the binary solvent mixtures at temperature T;  1 and  2 are the volume fractions of the solvents 1 (PEG 200) and 2 (water) in the absence of the solute; x 1,T Sat and x 2,T Sat denote the mole fraction solubility of the solute in solvents 1 and 2, respectively; and J i are the constants of the model computed by a regression analysis. 8 Knowledge of these model constants, which requires a number of solubility data in water-cosolvent mixtures for the training process, is a limitation for the model when the solubility predictions are the goal of the computations in early drug discovery studies. This limitation could be overcome using a trained version of the model for a given water -cosolvent mixture. The trained version of the Jouyban-Acree model for prediction of drug solubility in PEG 400 + water mixtures at temperature (T) is 9 log x m,T Sat )  1 log x 1,T Sat +  2 log x 2,T Sat + 394.82 1  2 T/K - 355.28 1  2 ( 1 -  2 ) T/K + 388.89 1  2 ( 1 -  2 ) 2 T/K (2) which was trained using 80 data sets of Rytting et al. 10 and produced reasonable prediction error (≈40 %). 9 It should be noted that we considered that the addition of the solvent 1 * To whom correspondence should be addressed. E-mail: ajouyban@ hotmail.com. Fax: +98 411 3363231. † Faculty of Pharmacy and Drug Applied Research Center, Tabriz University (Medical Sciences). ‡ Biotechnology Research Center, Tabriz University (Medical Sciences). § Research Center for Pharmaceutical Nanotechnology, Tabriz University (Medical Sciences). | University of North Texas. ⊥ Medway School of Pharmacy, Universities of Kent and Greenwich. J. Chem. Eng. Data 2010, 55, 519–522 519 10.1021/je900330p CCC: $40.75  2010 American Chemical Society Published on Web 06/15/2009