Mixture experimental design in the development of a mucoadhesive gel formulation.

The objective of the present study is to apply response surface methodology to the design and analysis of composite experiments containing independent covariate(s). The approach is illustrated here by the study of viscosity characteristics of a polymeric mucoadhesive formulation in multicomponent solvent vehicles. The nonaqueous formulation will produce a gel network with significant rheological change when in contact with body fluids. The process of water inclusion will induce not only solvent compositional change of the mixture but also concomitant dilution of the polymer concentration. To study the viscosity change over the solvent compositions and polymeric concentrations of interest, an experimental design is utilized consisting of a 10-point simplex-centroid lattice augmented with three interior points at each polymeric concentration. The contour patterns are compared with the experimental data using the variance and lack of fit, starting with the Scheffe linear model and building up to the full cubic model including the covariate terms. The fitted model provides information needed to predict optimum formulations, i.e., initial viscosity of less than 100 cP, but yielding rheological profiles commensurate with high degrees of substantivity when diluted with water. For illustrative purposes, the Carbopol resins neutralized with a 1:1 molar equivalent ratio of triethanolamine in three primary solvents, propylene glycol, glycerol formal, and water, were chosen for this study.