Objective: One of the methods for reducing glycative stress is to suppress postprandial hyperglycemia (PPHG). The purpose of this study is to establish a non-invasive and easy-to-implement means for suppressing PPHG. Based on the results of the past intake tests of various foods, a model formula for predicting the degree of PPHG from food contents was created.
Methods: A model formula was created to predict the indices for PPHG,
i.e. iAUC (incremental area under the curve), ΔCmax (maximum blood glucose concentration), based on iAUC (mg/dL·min) or ΔCmax when ingested a standard food (
i.e., cocked rice, udon, and bread) and the nutritional component of the test food. The past results of the model food intake test in our laboratory were used to create the predictive model formula. We applied 18 kinds of food to the formula and verified the degree of coincidence with the actual postprandial glucose change. Then, the mean absolute relative difference (MARD) between the predicted value and the measured value was calculated for each food (n = 18) and for each subject (n =159) in the 18 tests. In a subclass analysis, subjects were divided into three groups: top 25% (n = 42, iAUC; 7,379.9 ± 146.5), middle (n = 75, iAUC; 5,302.7 ± 73.5), and bottom 25% (n = 42, iAUC; 3,243.9 ± 61.5), based on iAUC at standard food intake. Pearson's correlation analysis was used to test the correlation between predicted and measured values, and Turkey's HSD test was used to analyze MARD.
Results: In the simulation of the food intake test (18 types), a highly positive correlation of r = 0.7 was observed between the predicted and measured value, and the average MARD was less than 15%. A subclass analysis showed the MARD in the top 25% group were lower than those in the bottom 25% group (p < 0.05).
Conclusion: A high correlation was found between the predicted value from the model formula and the measured value. Among them, the accuracy of prediction tended to be higher as the data of the subjects whose blood glucose was more likely to rise.
View full abstract