Interindividual variability and intra-individual reproducibility of glycemic index values for commercial white bread

We sought to assess the intra- and interindividual variability of glycemic index value determinations for white bread using glucose as the reference food.

A total of 23 healthy adults (aged 20-70 years) completed up to three sets of two visits per set. Each pair of visits assessed the glycemic response to 50 g available carbohydrates from commercial white bread and glucose, administered in random order. Glycemic index values were calculated by dividing the 2-h incremental area under the serum glucose response curve after each commercial white bread challenge by the mean area under the curve (AUC) for glucose.

The mean +/- SE ratio of the AUC after white bread intake by the AUC after glucose intake for the first set of determinations was 78 +/- 15 (n = 23; coefficient of variation [CV] 94%). When using glycemic index values calculated with the subset of participants who completed three sets of tests (n = 14), glycemic index values for each of the three sets of determinations were 78 +/- 10, 60 +/- 5, and 75 +/- 10, respectively. CVs were 50, 28, and 50%, respectively. The mean glycemic index value of these three sets was 71 +/- 6, with a CV of 30%. When an ANOVA approach was applied to these data, the interindividual CV was 17.8%, and the intra-individual variation was 42.8%.

These data suggest that in response to a challenge of white bread relative to glucose, within-individual variability is a greater contributor to overall variability than among-individual variability. Further understanding of all the sources of variability would be helpful in better defining the utility of glycemic index values.