Reply to J Brand-Miller

Energy Metabolism Laboratory
USDA Human Nutrition Research Center at Tufts University
711 Washington Street
Boston, MA 02111
E-mail: susan.roberts{at}tufts.edu

Dear Sir:

We were surprised that Brand-Miller questioned whether we chose appropriate foods for diets with different glycemic indexes (GIs) and glycemic loads (GLs) in our recent study (1). In the same protocol as the one she questioned, we previously reported that a subgroup who had high baseline concentrations of insulin secretion did, in fact, lose more weight when randomly assigned to a low-GL diet than did those randomly assigned to a high-GL diet (2). Our group is also known for our finding that single-meal studies support a role for low-GI diets in hunger reduction, and we have specifically published studies in which different forms of oatmeal (the food she questions) were used to create differences in GI (3). These different findings all strongly support our ability to create diets with different GIs and GLs and demonstrate weight-loss benefits in individuals who have risk factors that a low-GI or low-GL diet will help counterbalance.

In this new analysis of the same data from our long-term protocol (1), we found no mean difference in weight loss between individuals randomly assigned to consume a high-GL diet and those randomly assigned to consume a low-GL diet for 6 mo and then further assessed to 12 mo. Another group has also reported an identical finding in a different population (4). Given our track record in designing diets with different GIs and GLs, and the fact that the same result has emerged from a separate study, we feel that there are more logical and useful starting points to explain the results of our new study than a faulty diet design. First, because 2 separate studies have now reported that mean weight loss is the same with high-GL and low-GL diets, but that individuals with high insulin secretion do better with low-GL diets, it seems most likely that this is a real finding that may have important clinical utility for the treatment of obesity in the future.

It is also important to note, however, that both of our diets were high in fiber. To our knowledge there have been no studies of whether low-GI carbohydrates benefit energy regulation in diets that are also high in fiber, and this is an important issue because the effects of dietary fiber and a low GI may not be additive. Also, compared with the single-day studies cited by Brand-Miller, long-term studies may show reduced benefit of approaches to weight control that focus specifically on hunger suppression, because of the increased importance of other challenges to compliance over time such as social pressures to eat out and the desire for favorite foods.

ACKNOWLEDGMENTS

Neither of the authors had a conflict of interest.

REFERENCES