Optimization and Evaluation of Time-Dependent Tablets Comprising an Immediate and Sustained Release Profile Using Artificial Neural Network

Abstract

The aim of this work was to optimize time-dependent tablets using artificial neural network (ANN). The time-dependent tablet consisted of a tablet core, which contained sustained release pellets (70% isosorbide-5-mononitrate [5-ISMN]), immediate release granules (30% 5-ISMN), superdisintegrating agent (sodium carboxymethylstarch, CMS-Na), and other excipients, surrounded by a coating layer composed of a water-insoluble ethylcellulose and a water-soluble channeling agent. The chosen independent variables, i.e., X1 coating level of tablets, X2 coating level of pellets, and X3 CMS-Na level, were optimized with a three-factor, three-level Box-Behnken design. Data were analyzed for modeling and optimizing the release profile using ANN. Response surface plots were used to relate the dependent and the independent variables. The optimized values for the factors X1–X3 were 4.1, 14.1, and 29.8%, respectively. Optimized formulations were prepared according to the factor combinations dictated by ANN. In each case, the observed drug release data of the optimized formulations were close to the predicted release pattern. An in vitro model for predicting the effect of food on release behavior of optimized products was used in this study. It was concluded that neural network technique could be particularly suitable in the pharmaceutical technology of time-dependent dosage forms where systems were complex and nonlinear relationships often existed between the independent and the dependent variables.