Variability of the glycemic response to single food products in healthy subjects.

BACKGROUND: Many studies on the health effects of the glycemic index (GI) are confounded by differences in the intakes of other macronutrients and fibre. Little data exist about the within- and between-subject variability of the GI. OBJECTIVE: Our objectives were therefore (i) to calculate the GI of eight commonly used food products with similar macronutrient and fibre composition, but with different sources of carbohydrates, (ii) to examine the inter- and intra-individual variability of the incremental area under the curve (iAUC) after consuming the reference solution, and (iii) to compare the effect of three different methods on 2-h postprandial blood glucose responses. DESIGN: Four groups of 10 healthy subjects consumed in random order the increased (iGI) and decreased GI (dGI) variants and twice a glucose solution. All products consisted of 25g available carbohydrates (CHO). For the fruit drink, glucose values were simultaneously analyzed using venous and capillary blood samples, and by using a continuous glucose monitoring system (CGMS). RESULTS: The GIs for increased and decreased variants were (mean+/-standard error of the mean (SEM)) 69+/-15 and 40+/-4 for bread, 86+/-14 and 48+/-8 for a fruit drink, 51+/-12 and 20+/-4 for cake, and 63+/-17 and 37+/-10 for a cookie. The inter- and intra-individual coefficient of variation (CV) of the iAUCs of the reference solution was large and varied respectively between 13 and 38%, and between 33 and 80%. CONCLUSIONS: These data suggest that the GI is difficult to use at the individual level.