We read with interest the revised 2002 Clinical Practice Recommendations as they relate to nutrition therapy for diabetes (1) as well as the associated Technical Review (2). We would strongly endorse the need to individualize this component of treatment because advice is indeed necessary regarding other aspects of lifestyle, oral hypoglycemic agents, and insulin. However, we question some of the recommendations regarding dietary carbohydrates.
The need for evidence-based guidelines in all aspects of medical management is universally recognized. Unfortunately, with regard to nutritional recommendations, there are no randomized-controlled clinical trials with morbidity and mortality as end points. These are regarded as the ultimate type of evidence on which to make recommendations. We therefore have to use less conclusive approaches to study, including several different epidemiological methods and studies of dietary manipulations on surrogate end points known to be related to morbidity and mortality. This inevitably leads to subjective interpretation regarding the quality of studies because it is clearly inappropriate to simply count the numbers of investigations pointing in one direction or another. Furthermore, there is need to determine the emphasis that should be given to one type of evidence compared with another. We suggest that meticulously conducted and controlled human studies of people with diabetes that involve dietary manipulations over a period of weeks or months and that acknowledge clinically relevant end points should provide the most powerful level of evidence, especially when the findings are compatible with epidemiological data. It is also important to consider the manner in which recommendations are likely to be interpreted by health professionals and patients.
With these considerations in mind, we express concern regarding two aspects of the recommendations regarding carbohydrates. The recommendations that the “total amount of carbohydrate in meals or snacks is more important than the source or type” and that “as sucrose does not increase glycemia to a greater extent than isocaloric amounts of starch, sucrose and sucrose-containing foods do not need to be restricted by people with diabetes” (A-level evidence) (1) are in our opinion not backed by convincing evidence and are open to misrepresentation. The first recommendation regarding carbohydrate (also based on A-level evidence) indicates that “foods containing carbohydrate from whole grains, fruits, vegetables and low fat milk should be included in a healthy diet,” but provides no indication that these are the most desirable choices (1). Thus, despite the caveat based on “expert consensus” that “sucrose and sucrose-containing foods should be eaten in the context of a healthy diet,” it appears, according to this set of recommendations, that it is perfectly acceptable for the bulk of dietary carbohydrate to be derived from highly refined (processed) starchy foods, foods rich in sucrose, and other sugars or sucrose. We know of no medium or long-term studies in which such a dietary practice has been shown to be compatible with good glycemic control and optimum levels of risk factors for the complications of diabetes. Indeed, most of the well-controlled studies in which sucrose has been shown to be an acceptable component of the diabetic dietary prescription have included modest intakes of sucrose eaten with meals as part of a high-fiber diet, with the sucrose displacing other fiber-depleted carbohydrate-containing foods (3,4). Such a recommendation also has the potential to increase the energy density of the diet, surely an undesirable step when obesity rates are escalating out of control in all age groups. The latter is especially remarkable in the young age groups, considering that calories from fluids have been shown to satisfy less than solid food (5). A high intake of sugary beverages has been convincingly shown to be related to subsequent risk of obesity in children (6). The potential for misinterpretation has already been clearly demonstrated by a news item in the British Medical Journal (7) that describes the new recommendations under the headline “U.S. relaxes sugar ban for people with diabetes”
Under the heading of B-level evidence is the statement that “there is insufficient evidence of long-term benefit to recommend the use of low-glycemic index diets as a primary strategy in food/meal planning” (1) and that there is no need to recommend that people with diabetes consume a greater amount of fiber than other Americans. There is impressive evidence from carefully controlled studies that diets containing low-glycemic index foods (8) or foods high in fiber (9) are associated with appreciably improved levels of several measures of carbohydrate metabolism and cardiovascular risk factors. These studies confirm a substantial body of earlier research and suggest that these two characteristics of carbohydrate-containing foods may independently influence glycemic control, insulin levels, and lipoprotein-mediated risk of cardiovascular disease (10–12). There is also recent epidemiological evidence that a high intake of dietary fiber improves glycemic control and reduces the risk of ketoacidosis in type 1 diabetes (13). The Food and Agriculture Organization/World Health Organization Expert Consultation on Carbohydrates endorsed the use of glycemic index as a means of determining optimum carbohydrate-containing foods (14).
Thus, we believe that there is a convincing evidence base to advise that although sucrose and other added sugars may be included in moderation in the diets of people with diabetes, the bulk of dietary carbohydrate should be derived from foods with a low glycemic index and/or foods that are rich in soluble fiber. Such a recommendation permits the choice of foods from a wide range of fruits, vegetables, and whole-grain cereals and, although processed starchy foods are not excluded, they are not regarded as equivalent to these food choices. The European recommendations for people with diabetes that include such advice (15) will be updated to include the evidence base from which they were derived.
Address correspondence to Dr. James Mann, Department of Human Nutrition, University of Otago, P.O. Box 56, Dunedin, New Zealand. E-mail: email@example.com.