Application Of Experimental Design And Canonical Analysis Through The Eigen Value Problem To Drug Release Of Microencapsulated Microparticles

• By: KHAMANGA, Sandile (Rhodes University, South Africa)
• Co-author(s): Prof Sandile Khamanga (Rhodes University, Makhanda (Grahamstown), South Africa)
  Prof Roderick B. Walker (Rhodes University, Makhanda (Grahamstown), South Africa)
  
• Abstract:

  Background
  Fitting data to a second order polynomial for a system that contains some curvature is not well accommodated or interpreted. Canonical analysis can solve the problems by finding local optima for classical minimization and saddle point problems. In vitro dissolution is one of the most important elements of drug product development. Several models can be used to describe drug dissolution profiles where f (t) is a function of t (time) that is related to the amount of drug dissolved from a pharmaceutical dosage form. The quantitative interpretation of values generated in dissolution studies is facilitated using generic equations that mathematically translate dissolution curves as a function of certain parameters related to the dosage forms being tested. In some cases, the equations can be deduced by theoretical analysis of processes that a dosage form is subject to.
  Purpose
  The response surface methodology and canonical analysis through the Eigen value problem were employed to find the most suitable conditions for drug release of captopril (CPT) from microencapsulated microparticles.
  Methods
  In this study the modeling of percent drug release and optimization of percent released was undertaken using canonical, mathematical and Lagrange methods or Lagrange multipliers and Eigenvalue problems. A total of 30 sets of experiments were developed to obtain second-order polynomial equations in terms of independent variables. The variables were concentration of Eudragit® RS, Methocel® K15M, Methocel® K100M and homogenizing speed. The experiments were carried out according to a central composite design. An empiric quadratic equation that correlated drug release and the independent variables was proposed.
  Results
  The optimal conditions determined by the canonical analysis of the fitted model were X1= 0.43 g, X2= 2.04 g, X3 = 1.02 g and X4 = 3000 rpm. Subsequent formulation, carried out under optimal conditions, confirmed the release predicted by the adjusted model. The results of curve-fitting studies reveal that CPT release from the microparticles could be described by the Makoid-Banakar, Korsmeyer-Peppas, Kopcha and Higuchi models. These coefficients of determination were higher than 0.900 for all analyses and the corresponding sum squared regression values were lower than for the other models used.
  Conclusion
  Central composite rotatable design and canonical analysis of the response surfaces proved to be useful tools for determining the release of CPT from the microparticles. Canonical analysis seems to be a proper choice for finding local optima for classical minimization and saddle point problems.