[In vitro regression model of glycemic index for carbohydrate-riched foods].

OBJECTIVE: To Establish the in vitro regression model of glycemic index for carbohydrate-riched food by the in vitro digestibility and composition characteristics. METHODS: Thirty products were commercially available and selected on the basis of their high carbohydrate content. After determined fat, protein and carbohydrate constitutes (sugar, starch), an in vitro method for digestibility characteristics were developed to measure hydrolyzed starch at 20 min, 120 min, 240 min, 16 h, and non-digested resistant starch (RS). In vivo glycemic responses were determined by standardized methods. The relationship between the compositions and GI were also discussed through stepwise regression methods. RESULTS: The observed GI ranged from 26 to 113, and correlated strongly with the digestibility profile of carbohydrates. Significantly positive correlation of S20, S120 (P < 0.05) and negative correlation of RS (P < 0.01) to GI was observed. Easily available glucose (EAG, sum of content of S120 and glucose) and RS were selected as variables to setup regression equation for in vitro GI prediction, GI = 39.65 + 1.008EAG-1.072RS. CONCLUSION: The findings indicate that the digestibility profile of carbohydrates is very essential to explain in vivo glycemic response of carbohydrate rich foods and predict GI value.