RESUMO
Objective: To optimize the formulations of microporosity osmotic pump controlled release tablet of solid dispersion for total flavonoids from Desmodium styracifolium (TFDS) by central composite design-response surface methodology. Methods: The independent variables comprised of the amount of lactose, pore-forming agent, and coating weight gain, and the dependent variables involved the cumulative release after 2, 6, and 12 h, and the linear correlation coefficient of cumulative release curve. Design-expert software was used to fit multivariate linear models and quadratic multinomial models for experimental data. Response surface was delineated according to best-fit mathematic models and the optimum formulation was selected by Numerical Optimization. Results: The correlation coefficients of quadratic multinomial models were better than those of multivariate linear models. There was close agreement between the observed and predicted values for the cumulative release, and bias were all less than 5%. Conclusion: The model established by Design-expert Software is better to predict and could be used to optimize the formulations of microporosity osmotic pump controlled release tablets of solid dispersion for TFDS.