Glycemic Index Values Vary from One Test to the Next

BOSTON, Sept. 27 -- In blood glucose responses to a white bread challenge, individual patients have significant variations from test to test, said investigators here.

There are also significant differences in blood glucose responses to a white bread challenge between different patients, reported Alice H. Lichtenstein, D.Sc., of Tufts University, and colleagues, in the October issue of Diabetes Care. They said the findings call into question the reliability of glycemic index values.

When healthy adults were given 50 g of a commercial white bread, their between-individual variability in blood glucose responses over three test runs was about 17%, and the within-individual variability was about 43% reported.

"Using glucose as the control food, previous studies indicate that white bread has a glycemic index of about 70," Dr. Lichtenstein said. "In our study the combined average was 71, virtually identical to the published value. However, quite strikingly, individual values ranged from 44 to 132. What is critical is to determine why there is such a wide range of responses among individuals."

The glycemic index value of a food is calculated by monitoring the incremental area under the curve (AUC) of two-hour blood glucose responses to 50 g of carbohydrate in a test food (e.g., pasta) and a reference standard food, consisting of either 50 g of glucose or white bread containing 50 g of carbohydrate. The result is expressed as a percentage of the test food in relation to the standard food.

"Implicit in the recommendation to formalize glycemic index as a dietary guidance tool is the assumption that the glycemic response an individual has to a given food is similar among individuals regardless of metabolic and physiological factors," the authors wrote.

To see whether this was true, they recruited 23 healthy adults from the ages of 20 to 70, and had each volunteer complete up to three tests sets with two visits per set.

On each pair of visits, the volunteers were administered, in random order, 50 g of available carbohydrates from either commercial white bread or glucose.

The authors then calculated the glycemic index values by dividing the two-hour incremental AUC for serum glucose response each white bread challenge by the mean AUC for glucose.

They found that for the first set of determinations among all 23 participants, the mean ratio of the AUC after white bread intake to the AUC after glucose intake was 78 + 15, with a coefficient of variation of 94%.

Glycemic index values calculated from the 14 participants who completed all three test set were 78 +10, 60 + 5, and 75 + 10 for runs one, two and three, respectively The co-efficient of variation for the first set of tests was 50%, for the second was 28%, and for the third was 50%.

For the three test sets combined, the mean glycemic index value was 71 + 6, with a coefficient of variation of 30% .

When the authors conducted an analysis of variation, they found that the inter-individual (between-subject) variability 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," the investigators wrote. "Further understanding of all the sources of variability would be helpful in better defining the utility of glycemic index values."

Dr. Lichtenstein noted that many different factors may affect the glycemic index value of a given food.

"For example," she said, "a piece of white bread may have a high glycemic index but, if a person eats a slice of turkey and cheese with that bread, the effect of the multiple foods may result in a different glycemic index than if that person had eaten the white bread alone."

"Since most food is consumed as combinations during meals and snacks, there is a need to assess the significance of using glycemic index values determined on individual foods for food mixtures," she continued. "Similarly, it is important to know whether the food consumed prior to a meal or snack alters subsequent glycemic response."