Response to the Diabetes Nutrition Study Group of the European Association for the Study of Diabetes

We welcome the European perspective, although we do not completely share it. We agree with the statement of Mann et al. (1) that “meticulously conducted and controlled human studies of people with diabetes that involve dietary manipulations over a period of weeks or months… provide the most powerful evidence.” However, we take issue with several concerns they express. First, we take issue with the statement that it is important to consider how recommendations are likely to be interpreted by health professionals and patients. Second, we continue to believe that the total amount of carbohydrate is more important than the source or type. Third, we continue to believe that sucrose does not need to be restricted, relative to other carbohydrates, because of concern about aggravating hyperglycemia.

It was an initial determination of the American Diabetes Association Task Force that it was our task to write, as accurately as possible, evidence-based nutrition principles and recommendations. The implementation of these principles and recommendations was to be determined by health professionals in their individualized nutrition counseling with patients. Furthermore, we concluded that patients have the right to read and know accurate nutrition information. With this information, it is then their right to make decisions about their own food choices. Too often in the past, health professionals have taken a parental approach, such as “do this because it is good for you,” or a “food police” approach, such as “don’t eat sugar.” These approaches have not led to successful outcomes. Table 1 outlines this well.

With respect to amount and source of carbohydrate, we stand behind our original recommendation. However, as stated in our response to Wolever (2), our recommendation about the amount of carbohydrate might be more clear if it were changed to say the total amount of “available” carbohydrate is more important than the source or the type. In type 1 diabetic subjects, the amount of carbohydrate in test meals influenced the amount of insulin necessary to control glycemia, whereas glycemic index, fiber content, and caloric content did not (3).

With regard to the concern about sucrose, it is clearly stated in our introduction that “basic to the nutrition recommendations is the underlying concern for optimal nutrition through healthy food choices and an active lifestyle” (4). The section on sucrose, as noted by Mann et al. (1), also states that “sucrose and sucrose-containing foods should be eaten in the context of a healthy diet” (4,5). Mann et al. states that they know of no medium or long-term studies where the practice of focusing on total carbohydrate was shown to be compatible with good glycemic control. Please note that, in the 20 studies quoted, when total carbohydrate came from a variety of starches or starches plus sucrose, the sucrose intake represented approximate usual intake, and only Peterson et al. (6) made an attempt to use sucrose with fiber-containing foods. In most of the studies, rigorous control of the nutrients under study was established by providing meals to subjects. One of the studies provided 23% and another 30% of energy from sucrose. Two of the studies lasted 28 days. If the total carbohydrate intake was kept similar, the responses were also similar. Was the European perspective that both sucrose and starch should be restricted in the diabetic diet because both aggravate hyperglycemia generated with these studies in mind? If so, does this not affirm the concept that the total amount of carbohydrate is more important than the source or type?

The headline “U.S. relaxes sugar ban for people with diabetes,” which appeared in the British Medical Journal (7), surprised us. The relaxation of the restriction on sucrose was nothing new, having been recommended in 1994 (8).

With regard to the statement by Mann et al. (1) that a “high intake of sugary beverages has been convincingly shown to be related to subsequent risk of obesity in children,” we would call attention to another study (9) in which added sugars were found to be relatively unimportant when it came to overall diet quality in individuals between 2 and 19 years of age.

With regard to the glycemic index, the study by Jarvi et al. (10) did find benefit, but as noted in the previous reply to the letter by Irwin (11), other studies (4) have not confirmed long-term benefit from low-glycemic index diets. One study is not “impressive evidence.” The same applies to fiber. Whereas some intervention studies have reported benefit (12,13), others have not (14–16). Moreover, the study by Chandalia et al. (13), which compared 24 g fiber with 50 g fiber, would support our statement that it “appears that ingestion of large amounts of fiber is necessary to confer metabolic benefits. It is unclear whether the palatability and gastrointestinal side effects of fiber in this amount would be acceptable to most people” (5,6). The control arm of the study used 24 g dietary fiber and had no beneficial effects on glucose, lipid, or insulin levels. This amount of fiber is clearly at the upper end of usual intake for most Americans and would, by itself, require major lifestyle changes for most Americans to achieve. The 50-g dietary fiber diet included two servings of oatmeal (15 g carbohydrate/serving), six slices of whole wheat bread, six to seven servings of fruit (15 g carbohydrate/serving), and three servings of vegetables (15 g carbohydrate/serving). For many individuals, this type of food plan would require very dramatic changes in eating habits.

In conclusion, we stand behind our original recommendations, as we believe they are evidence based.