Response Surface Methodology for the Development of Self-Nanoemulsified Drug Delivery System (SNEDDS) of All-Trans-Retinol Acetate Ehab I. Taha Department of Pharmaceutical Sciences, School of Pharmacy, Texas Tech University Health Science Center, Amarillo, Texas, USA and Department of Pharmaceutical Sciences, Faculty of Pharmacy, Al-Azhar University, Nasr City, Cairo, Egypt Ahmed M. Samy and Alaa A. Kassem Department of Pharmaceutical Sciences, Faculty of Pharmacy, Al-Azhar University, Nasr City, Cairo, Egypt Mansoor A. Khan Department of Pharmaceutical Sciences, School of Pharmacy, Texas Tech University Health Science Center, Amarillo, Texas, USA The purpose was to prepare, characterize, and optimize a self-nanoemulsified drug delivery system (SNEDDS) of a model lipophilic compound, all-trans-retinol acetate. As part of the optimization process, the main effects, interaction effects, and quadratic effects of the formulation ingredients were inves- tigated. Method. A three-factor, three-level Box-Behnken design was used to explore the quadratic response surfaces and construct a second-order polynomial model in the form: Y=A+A 1 X 1 +A 2 X 2 +A 3 X 3 +A 4 X 1 X 2 +A 5 X 2 X 3 +A 6 X 1 X 3 + A 7 X 1 2 +A 8 X 2 2 +A 9 X 3 2 + E. Amount of added oil (X1), surfactant (X2), and cosurfactant (X3) were selected as the factors. Par- ticle size (Y1), turbidity (Y2), and cumulative amount of the active ingredient emulsified after 10 (Y3) and 30 (Y4) min were the observed variables. Response surface plots were used to demonstrate the effect of factors (X1), (X2), and (X3) on the response (Y4). Amount of added soybean oil (X1), Cremophor EL (X2), and Capmul MCM-C8 (X3) showed a significant effect on the emulsification rates, as well as on the physical properties of the resultant emulsion (particle size and turbidity). Observed and predicted values of Y4 obtained from the constructed equa- tions were in close agreement. Response surface methodology was then used to predict the levels of factors X1, X2, and X3 under the constrained variables for an optimum response. Applied constraints were 0<Y1<0.5, 1<Y2<20, 60<Y3<80, and 90 < Y4 < 100. The predicted values were 0.0704 mm for particle size (Y1), 18.95 NTU for turbidity (Y2), 88.88% for drug release after 10 min (Y3), and 110.7% drug release after 30 min (Y4). Two new formulations were prepared according to the predicted levels. The observed and predicted values were in close agreement. INTRODUCTION All-trans-retinol acetate is a fat-soluble vitamin, which is an essential nutrient for humans. It is involved in several important biological functions including vision, growth, reproduction, and the differentiation and mainte- nance of epithelial tissue. [1 – 7] The absorption of vitamin A is governed by factors that determine the absorption of lipids. During digestion, all-trans-retinol ester is hydro- lyzed to all-trans-retinol alcohol (retinol) by pancreatic and intestinal enzymes, and then solubilized by bile acids. Therefore, any disturbance of gastrointestinal enzymes or pancreatic enzymes will lead to disturbance of all-trans- retinol absorption. [8] Self-nanoemulsifing systems are isotropic mixtures of oil, surfactants, and cosurfactants. When these systems come in contact with gastrointestinal fluids, they disperse as very fine droplets in nanometer size range. The droplet size, turbidity, and drug release characteristics depend on formulation variables, such as the nature and concentrations of the oil, surfactant, or cosurfactant. This investigation is aimed at mathematically re- lating the formulation variables with all-trans-retinol ace- tate emulsification. Received 2 April 2004, Accepted 31 December 2004. Address correspondence to Mansoor A. Khan, FDA/CDER/ DPQR, 10903 New Hampshire Ave., Silver Spring, MD 20993, USA; Fax: (301) 796-9816; E-mail: Khanm@cder.fda.gov 363 Pharmaceutical Development and Technology, 10:363–370, 2005 Copyright D 2005 Taylor & Francis Inc. ISSN: 1083-7450 print / 1097-9867 online DOI: 10.1081/PDT-200065675 Order reprints of this article at www.copyright.rightslink.com