Trout D L, Behall K M, Osilesi O
Carbohydrate Nutrition Laboratory, Beltsville Human Nutrition Research Center, US Department of Agriculture, MD 20705.
Am J Clin Nutr. 1993 Dec;58(6):873-8. doi: 10.1093/ajcn/58.6.873.
Associations between published values for glycemic index (GI) and chemical components were examined by regression analysis in 18 starchy foods. Independent variables were the amounts of protein, fat, phytic acid, and total dietary fiber (TDF) present in food portions containing 50 g available carbohydrate. GI was similarly and significantly correlated (P < 0.05) with TDF, protein, and phytate. When foods were considered as comprising two groups, legumes and others (grains and tubers), food category accounted for as large a fraction of variability in GI as did regression analysis with food components in the undivided sample. Most correlations between GI and components reverted toward zero when calculated only for the legumes or the "others" group. Evidence is cited that the method of preparing foods and the characteristics of starch and starch granules are more important in predicting GI among starchy food than is the content of any of these four food components.
通过回归分析研究了18种淀粉类食物中已发表的血糖生成指数（GI）值与化学成分之间的关联。自变量为含有50克可利用碳水化合物的食物份中蛋白质、脂肪、植酸和总膳食纤维（TDF）的含量。GI与TDF、蛋白质和植酸同样显著相关（P<0.05）。当将食物分为豆类和其他类（谷物和块茎）两组时，食物类别在GI变异性中所占比例与未分组样本中食物成分的回归分析所占比例一样大。仅针对豆类或“其他”组计算时，GI与成分之间的大多数相关性都趋于零。有证据表明，在预测淀粉类食物的GI时，食物的制备方法以及淀粉和淀粉颗粒的特性比这四种食物成分中的任何一种含量都更重要。
