点击显示 收起
Human Nutrition Unit, G 08
University of Sydney
Sydney, NSW 2006
Australia
E-mail: j.brandmiller{at}mmb.usyd.edu.au
Dear Sir:
Das et al (1) are to be congratulated on their 2-phase dietary trial in which they compared the effects of a high- and low-glycemic-load diet in overweight adults in the CALERIE (Comprehensive Assessment of the Long-term Effects of Restricting Intake of Energy) trial. Their study design had many strengths, including the provision of all the foods in the first 6 mo of life, the assessment of body fat (not just weight loss), the use of doubly labeled water methodology to measure total energy expenditure, and the careful control of behavioral factors. Unfortunately, there was one critical element missing in the study. In any study comparing diets with various glycemic loads (GLs), it is necessary, by definition, to demonstrate that the subjects consumed diets that did indeed produce differences in postprandial glycemia (and preferably insulinemia as well). Although Das et al determined the GL based on published glycemic index (GI) values and the nutrient composition of the foods provided, the calculated value is theoretical and not necessarily reliable. This is particularly true of studies in which the GI values have been extrapolated from one manufacturer to another and from one country to another. Thus, in the CALERIE trial, we cannot tell whether the lack of difference in weight loss (or body fat) resulted because postprandial glycemia was not distinctly different or that lowering glycemia is ineffective for weight-loss purposes.
Close inspection of Appendix A in Das et al's article, in which sample foods for the high-GL and low-GL diets are listed, suggests a naïve knowledge of the GI. For example, carrots were placed in the high-GL list, even though several studies have confirmed that carrots have a low GI and that their contribution to total carbohydrate intake is small (2). Similarly, yogurt and oatmeal were placed in the high-GI list, even though these may be low-GI foods. Despite these concerns, some difference in postprandial glycemia would be evident simply because the carbohydrate content of the 2 diets differed (60% compared with 40% for the high- and low-GL diets, respectively).
It is revealing from a mechanistic viewpoint that well-designed, adequately powered studies have detected differences in the rate of weight or fat loss between high- and low-GL diets, at least in the short term (3, 4). In addition, the trend in negative studies almost always favors the low-GI or low-GL diet (5, 6). At 6, 7, 8, and 9 mo, the CALERIE trial also showed a trend to greater weight loss with the lower-GL diet. At 6 mo, those consuming the low-GL diet had lost 23.3% of their body fat compared with only 17.1% of those consuming the high-GL diet. This difference in fat mass change was not statistically significant, but the study was underpowered because there were only 17 subjects in each group. It is also possible that subgroups of people with a high first phase insulin response may respond more favorably to diets with a low GL, as reported earlier in the CALERIE trial (7).
Although 2–4-kg differences in weight loss may be considered modest, we should not expect the differences to be large. Small incremental weight increases over time produce an obese individual. Differences in fat oxidation dictated by differences in postprandial insulin secretion amounting to only 1 g/d (8) will produce a gain in fat mass of 350 g over 1 y and of 7 kg over 20 y. Although studies such as the CALERIE trial are a valuable addition to the literature, future studies should document differences in day-long glycemia and should be adequately powered to detect small, but clinically important, differences in fat mass change.
ACKNOWLEDGMENTS
JCB-M is a coauthor of The Low GI Diet (New York, NY: Marlowe and Co, 2005) and a coauthor of The New Glucose Revolution book series (New York, NY: Marlowe and Co).
REFERENCES