Liu et al (2003) recently conducted a large and well-designed study, the purpose of which was to validate the concept of glycaemic glucose equivalent (GGE) as a predictor of the relative glycaemic effect of foods. However, the conclusion that GGE is a valid predictor of glycaemic response is not supported by the data. The blood glucose responses of subjects with diabetes (n=12) and without diabetes (n=11) were measured after they consumed five different carbohydrate foods at two levels of GGE (10 and 20) and one food at two different levels of GGE (24 and 48). The accuracy of prediction was judged using three criteria. Different foods fed at the same level of GGE should elicit similar glycaemic responses in the same group of subjects, and doubling GGE should double the glycaemic response. The final criterion was related to a quantity known as relative glycaemic effect (RGE), which is the incremental area under the curve (IAUC) divided by the glycaemic sensitivity factor of each subject (defined as IAUC/GGE). According to the authors: ‘…if GGE accurately predicted glycaemic responses, the numbers for GGE intake and RGE should be the same.’ The authors conclude that ‘Relative glycaemic effects were accurately predicted by GGE intake after adjusting for individual glycaemic sensitivity.’ and ‘GGE content predicted glcyaemic impact of foods over a practical range of carbohydrate intakes…’.
If GGE is valid, foods fed at the same level of GGE should elicit similar glycaemic responses. This study provides four separate tests (two groups of subjects, two levels of GGE) of this hypothesis. In three of these four comparisons, there were large (2–3-fold) and statistically significant differences in glycaemic response among the five test foods: diabetic subjects, GGE=10, range of IAUC 79–202 (P<0.05); diabetic subjects, GGE=20, range of IAUC 181–469 (P<0.05); normal subjects, GGE=10, range of IAUC 52–106 (P<0.05); normal subjects, GGE=20, range of IAUC 157–216 (NS). Thus, five foods fed at the same level of GGE produced significantly different glycaemic responses in three of four cases.
This is a preview of subscription content, access via your institution
