Textural Profiling and Statistical Optimization of Crosslinked Calcium-Alginate-Pectinate-Cellulose Acetophthalate Gelisphere Matrices VINESS PILLAY, MICHAEL P. DANCKWERTS University of the Witwatersrand, Department of Pharmacy and Pharmacology, Medical School, 7 York Road, Parktown, 2193, Johannesburg, Gauteng, South Africa Received 17 March 2002; revised 3 June 2002; accepted 25 June 2002 ABSTRACT: A2 5 factorial design was employed to statistically evaluate the textural properties of a crosslinked calcium-alginate-pectinate-cellulose acetophthalate geli- sphere system. In accordance with the factorial matrix, gelispheres were formulated by titrating a combination polymeric solution comprised of sodium alginate, pectin and/or cellulose acetophthalate into an inducer solution (crosslinking agent) consisting of calcium and/or acetate ions. A Texture Analyzer was used to profile the gelisphere matrices for their resilience in the unhydrated and hydrated states, the fracture energy involved in matrix rupture, and the matrix hardness achieved with different levels of crosslinking. Significantly different textural properties were found among the cross- linked formulations. In particular, the unhydrated matrix resilience was selected as a parameter for optimization of the gelisphere formulation because of its large impact on drug release modulation, matrix integrity, and sensitivity to the crosslinking process. Resilience increased with increasing polymer concentration, irrespective of the polymer combination. Furthermore, resilience was not significantly influenced by the concentra- tion of the crosslinking agents, but rather by the application of a higher polymer concentration in the crosslinking reaction; again irrespective of the polymer combination. In addition to the use of a factorial design, artificial neural modeling was employed to predict the textural properties based on the factorial matrix as a statistically suitable data source. Neural networks appeared to be a strong competitor of factorial regression for the prediction of textural data. ß 2002 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 91:2559–2570, 2002 Keywords: crosslinked gelispheres; factorial design; statistical analysis; artificial neural networks; matrix resilience; fracture energy; matrix hardness INTRODUCTION In recent years, much attention has been directed to the application of hydrophilic swellable single and multiple-unit drug delivery systems particu- larly for the attainment of site-specific oral target- ing of bioactive materials. 1–5 In the present study, we evaluated a new crosslinked multiple-unit system for which the term ‘‘gelispheres’’ has been coined (i.e., a spherical crosslinked hydrophilic polymeric entity capable of extensive gelation and swelling in simulated biological fluids). Hydrophilic swellable polymers in their native state usually demonstrate rapid gelation, first- order dissolution, and unsynchronized erosion/ relaxation in relation to the timescale required for modulated drug delivery. 6 These properties may be significantly adjusted and controlled through the process of crosslinking, which essentially changes the extent of polymer entanglement and hence the rate of disentanglement. In addition, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 12, DECEMBER 2002 2559 Correspondence to: Viness Pillay (Telephone: 27-11-7172274; Fax: 27-11-6424355; E-mail: pillayvi@therapy.wits.ac.za) Journal of Pharmaceutical Sciences, Vol. 91, 2559–2570 (2002) ß 2002 Wiley-Liss, Inc. and the American Pharmaceutical Association