HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Extract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Franz, M. J. Articles by Wheeler, M. Search for Related Content PubMed PubMed Citation Articles by Franz, M. J. Articles by Wheeler, M. Social Bookmarking                What's this?

Evidence-Based Nutrition Principles and Recommendations for the Treatment and Prevention of Diabetes and Related Complications

¹ Nutrition Concepts by Franz, Inc., Minneapolis, Minnesota
² Division of Endocrinology and Diabetes, Department of Medicine, University of Minnesota, Minneapolis, Minnesota
³ University of Illinois, Chicago, Illinois
⁴ Department of Medicine, University of Washington, Seattle, Washington
⁵ Research Center, Centre Hospitalier de l’Université de Montréal, Campus Htel-Dieu and Department of Medicine, University of Montreal, Montreal, Ontario, Canada
⁶ Department of Internal Medicine and Center for Human Nutrition, University of Texas Southwestern Medical Center, Dallas, Texas
⁷ Holzmeister Nutrition Consulting, Tempe, Arizona
⁸ Department of Endocrinology, Cleveland Clinic Foundation, Cleveland, Ohio
⁹ Department of Epidemiology and Biostatistics, School of Public Health, University of South Carolina, Columbia, South Carolina
¹⁰ Division of Endocrinology, St. Louis University Medical Center, St. Louis, Missouri
¹¹ Division of Endocrinology, Diabetes, and Clinical Nutrition, Oregon Health and Science University, Portland, Oregon
¹² Diabetes Research and Training Center, Indiana University School of Medicine, Indianapolis, Indiana

	INTRODUCTION

TOP INTRODUCTION GOALS OF MEDICAL NUTRITION... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... DIABETES PREVENTION SUMMARY References

 
Historically, nutrition principles and recommendations for diabetes and related complications have been based on scientific evidence and diabetes knowledge when available and, when evidence was not available, on clinical experience and expert consensus. Often it has been difficult to discern the level of evidence used to construct the nutrition principles and recommendations. Furthermore, in clinical practice, many nutrition recommendations that have no scientific supporting evidence have been and are still being given to individuals with diabetes. To address these problems and to incorporate the research done in the past 8 years, this 2002 technical review provides principles and recommendations classified according to the level of evidence available. It reviews the evidence from randomized, controlled trials; cohort and case-controlled studies; and observational studies, which can also provide valuable evidence (1,2), and takes into account the number of studies that have provided consistent outcomes of support. In this review, nutrition principles are graded into four categories based on the available evidence: those with strong supporting evidence, those with some supporting evidence, those with limited supporting evidence and those based on expert consensus.

Evidence-based nutrition recommendations attempt to translate research data and clinically applicable evidence into nutrition care. However, the best available evidence must still be moderated by individual circumstances and preferences. The goal of evidence-based recommendations is to improve the quality of clinical judgments and facilitate cost-effective care by increasing the awareness of clinicians and patients with diabetes of the evidence supporting nutrition services and the strength of that evidence, both in quality and quantity.

Before 1994, the American Diabetes Association’s (ADA’s) nutrition principles and recommendations attempted to define an "ideal" nutrition prescription that would apply to everyone with diabetes (3,4,5). Although individualization was a major principle of all recommendations, it was usually done within defined limits for recommended energy intake and macronutrient composition. The 1994 nutrition recommendations shifted this focus to one that emphasized effects of nutrition therapy on metabolic control (6,7). The nutrition prescription is determined considering treatment goals and lifestyle changes the diabetic patient is willing and able to make, rather than predetermined energy levels and percentages of carbohydrate, protein, and fat. The goal of nutrition intervention is to assist and facilitate individual lifestyle and behavior changes that will lead to improved metabolic control. This focus continues with the 2002 nutrition principles and recommendations.

Medical nutrition therapy (MNT) is an integral component of diabetes management (8,9) and diabetes self-management education (10). (Medical nutrition therapy is the preferred term and should replace other terms, such as diet, diet therapy, and dietary management.) MNT for diabetes includes the process and the system by which nutrition care is provided for diabetic individuals and the specific lifestyle recommendations for that care. However, recommendations should not only be based on scientific evidence but should also take into consideration lifestyle changes the individual can make and maintain. Cultural and ethnic preferences should be taken into account, and the person with diabetes should be involved in the decision-making process.

Results from the Diabetes Control and Complications Trial (DCCT) and the U.K. Prospective Diabetes Study (UKPDS) convincingly demonstrated the importance of glycemic control in preventing the microvascular complications of diabetes (11,12). In both trials, MNT was important in achieving treatment goals (13,14). MNT in diabetes addresses not only glycemic control but other aspects of metabolic status as well, including dyslipidemia and hypertension—major risk factors for cardiovascular disease. This is important, as macrovascular complications are the major contributors to the morbidity and mortality associated with diabetes (15).

The current nutrition principles and recommendations for diabetes focus on lifestyle goals and strategies for the treatment of diabetes. Now, for the first time, the 2002 recommendations specifically address lifestyle approaches to diabetes prevention; they distinguish MNT for treating and managing diabetes from MNT for preventing or delaying the onset of diabetes, as the two may not necessarily be the same.

Whether for management or prevention of diabetes and its complications, basic to the nutrition recommendations is the underlying concern for optimal nutrition through healthy food choices and an active lifestyle. The ADA supports and incorporates the nutrition recommendations from major organizations, such as the U.S. Department of Agriculture (Dietary Guidelines for Americans) (16), American Heart Association (17), National Cholesterol Education Program (18), American Institute for Cancer Research (19), and Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (20).

Although many studies have focused on the role of single nutrients, food, or food groups in disease prevention or promotion, emerging research suggests there are health benefits from food patterns that include mixtures of food containing multiple nutrients and nonnutrients (21,22,23,24,25,26,27). Although this approach makes it difficult to elucidate mechanisms through which the diet composition affects a particular health outcome, it does represent a practical approach to making realistic nutrition recommendations for improving health.

The health professional with the greatest expertise in providing MNT for diabetes is the registered dietitian (RD) knowledgeable and skilled in diabetes management (8,9,10,28). Outcome studies (13,29,30,31,32,33,34) have demonstrated that MNT provided by RDs results in a 1.0% decrease in HbA_1c in patients with newly diagnosed type 1 diabetes (29), a 2.0% decrease in HbA_1c in patients with newly diagnosed type 2 diabetes (14), and a 1.0% decrease in HbA_1c in patients with an average 4-year duration of type 2 diabetes (30). The effectiveness of dietitian-delivered MNT in improving dyslipidemia has also been demonstrated (35,36,37,38,39,40,41). However, it is essential that all team members involved in diabetes treatment and management be knowledgeable about MNT and supportive of the patient’s need to make lifestyle changes (42,43,44,45).

	GOALS OF MEDICAL NUTRITION THERAPY FOR DIABETES

TOP INTRODUCTION GOALS OF MEDICAL NUTRITION... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... DIABETES PREVENTION SUMMARY References

 
Goals of MNT that apply to all persons with diabetes are as follows:

1. To attain and maintain optimal metabolic outcomes, including
   
2.  a. blood glucose levels in the normal range or as close to normal as is safely possible to prevent or reduce the risk for complications of diabetes
   
3.  b. a lipid and lipoprotein profile that reduces the risk for macrovascular disease
   
4.  c. blood pressure levels that reduce the risk for vascular disease
   
5. To prevent and treat the chronic complications of diabetes; modify nutrient intake and lifestyle as appropriate for prevention and treatment of obesity, dyslipidemia, cardiovascular disease, hypertension, and nephropathy
   
6. To improve health through healthy food choices and physical activity
   
7. To address individual nutritional needs, taking into consideration personal and cultural preferences and lifestyle while respecting the individual’s wishes and willingness to change
   

The goals of MNT that apply to specific situations include the following:

1. For youth with type 1 diabetes, to provide adequate energy to ensure normal growth and development, and to integrate insulin regimens into usual eating and physical activity habits
   
2. For youth with type 2 diabetes, to facilitate changes in eating and physical activity habits that reduce insulin resistance and improve metabolic status
   
3. For pregnant or lactating women, to provide adequate energy and nutrients needed for optimal outcomes
   
4. For older adults, to provide for the nutritional and psychosocial needs of an aging individual
   
5. For individuals being treated with insulin or insulin secretagogues, to provide self-management education for treatment (and prevention) of hypoglycemia, acute illnesses, and exercise-related blood glucose problems
   
6. For individuals at risk for diabetes, to decrease the risk by encouraging physical activity and promoting food choices that facilitate moderate weight loss or at least prevent weight gain
   

	MEDICAL NUTRITION THERAPY FOR TYPE 1 AND TYPE 2 DIABETES

TOP INTRODUCTION GOALS OF MEDICAL NUTRITION... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... MEDICAL NUTRITION THERAPY FOR... DIABETES PREVENTION SUMMARY References

 
Carbohydrate and diabetes
When referring to common food carbohydrates, the following terms are preferred: sugars, starch, and fiber (Table 1). This classification is based on the recommendations of the Food and Agriculture Organization of the United Nations and the World Health Organization in which carbohydrates are classified according to their degree of polymerization and are initially divided into three principal groups—sugars, oligosaccharides, and polysaccharides (46). Terms such as simple sugars, complex carbohydrates, and fast-acting carbohydrates are not well defined; use of these terms should be abandoned.

Carbohydrate and type 1 diabetes
In trying to establish nutrition recommendations for the prescription of carbohydrate for patients with type 1 diabetes, there are several problems: the paucity of studies, the small number of subjects studied, and the lack of long-term studies. Although several studies in subjects with type 1 diabetes have been done comparing differing amounts or percentages of calories from carbohydrate (59,60,61), there is no body of evidence to suggest changing the 1994 recommendation that 60–70% of total energy be distributed between carbohydrate and monounsaturated fat based on nutrition assessment and treatment goals (6). A basis for choosing between carbohydrate and monounsaturated fat, other than ethnic or cultural preferences, is not readily available. In some studies (59,60), a high-carbohydrate diet compared to a high-monounsaturated fat diet resulted in a higher glycemic profile, with no differences in the lipid profile; however, insulin may not have been adjusted appropriately to cover the amount of carbohydrate ingested. In another study (61) comparing a high-carbohydrate diet to a high-monounsaturated fat diet, there were no differences in glycemia, but there was an increase in postprandial triglycerides with the high-monounsaturated fat diet.

More information is available regarding the effects of different types of carbohydrate on postprandial glycemia. In type 1 patients with diabetes, the ingestion of a variety of starches or sucrose, both acutely (62,63,64,65,66,67,68) and for up to 6 weeks (69,70,71,72), was shown to produce no significant differences in glycemic response if the total amount of carbohydrate is similar. Studies in controlled settings (62,63,64,65,66,67,68,69) and in free-living subjects (70,71,72) have demonstrated similar results.

Studies show a strong relationship between the premeal insulin dosage and the postprandial response to the carbohydrate content of the meal (73,74,75,76). In individuals receiving intensive insulin therapy, the total amount of carbohydrate in the meal did not influence glycemic response if the premeal insulin was adjusted for the carbohydrate content of the meal (73). The premeal insulin dosage required was not affected by the glycemic index, fiber, fat, or caloric content of the meal. Furthermore, wide variations in meal carbohydrate content did not modify the basal (long-acting) insulin requirement. The concept of total meal carbohydrate determining the premeal insulin dosage is further supported by the DCCT, in which it was shown that individuals who adjusted their premeal insulin dosages based on the carbohydrate content of meals had 0.5% (P < 0.03) lower HbA_1c levels than those who did not adjust premeal insulin (13).

For individuals receiving fixed dosages of short- and intermediate-acting insulin, day-to-day consistency in the amount and source of carbohydrate has been associated with lower HbA_1c levels (77). Day-to-day variations in energy and protein or fat intakes were not significantly related to HbA_1c.

Glycemic index.
The usefulness of low-glycemic index diets in individuals with type 1 diabetes is controversial. Five studies (n = 48; range 12 days to 6 weeks) (78,79,80,81,82) have compared low-glycemic index diets to high-glycemic index diets for longer than 1 day. The results from these studies did not provide convincing evidence of benefit. Four studies (78,80,81,82) measured HbA_1c, and none reported differences in HbA_1c between low- and high-glycemic index diets. Four studies (78,79,80,81) measured glycated albumin or fructosamine; three of those (78,79,80) reported decreases in glycated albumin or serum fructosamine after the incorporation of low-glycemic index food in the diet, and one reported no differences in serum fructosamine (81). Three studies (78,79,80) measured fasting plasma glucose (FPG) concentrations, and none reported differences in FPG between low- and high-glycemic index diets. Insulin requirements were measured in four studies (79,80,81,82); one (79) reported lower insulin requirements from low- compared to high-glycemic index diets, whereas three (80,81,82) reported no differences in insulin dosages. Therefore, although the use of low-glycemic index food may reduce postprandial glucose levels, there is not sufficient evidence of long-term benefit to recommend general use of low-glycemic index diets in individuals with type 1 diabetes.

In a cross-sectional study of 2,810 people with type 1 diabetes from the EURODIAB IDDM Complications Study (83), the glycemic index calculated from 3-day food records was examined for its relation to HbA_1c and serum lipid concentrations. HbA_1c levels were lower in the lowest glycemic index quartile compared with the highest quartile. Of the serum lipids, only HDL cholesterol was independently related to the glycemic index. Interestingly, the consumption of bread and pasta had the biggest effect on the overall glycemic index.

The effects on lipids after low- compared to high-glycemic index diets appear to be minimal. Two studies (79,80) measured cholesterol concentrations and three studies (78,79,80) measured HDL cholesterol concentrations, but none reported differences in the low- compared to the high-glycemic index diets. One study (80) reported lower triglyceride levels, but one (78) did not.

Although it is clear that carbohydrates do have differing glycemic responses, the data reveal no clear trend in outcome benefits. If there are long-term effects on glycemia and lipids, these effects appear to be modest. Moreover, the number of studies is limited, and the design and implementation of several of these studies is subject to criticism.

Fiber.
Early short-term studies using large amounts of fiber (>30 g/day) in small numbers of suboptimally controlled type 1 subjects (84,85,86,87) suggested a positive effect of fiber on glycemia. However, in a study of type 1 diabetes subjects on intensive insulin therapy, 56 g of fiber had no beneficial effect on glycemic control (81). A recent study (88) randomized subjects being treated with two or more injections of insulin per day and HbA_1c levels of 7–10% to either a high-fiber (50 g/day), low-glycemic index diet or a low-fiber (15 g/day), high-glycemic index diet for 24 weeks. The high-fiber diet significantly reduced mean daily blood glucose concentration (P < 0.05), the number of hypoglycemic events (P < 0.01), and, in the subgroup of patients compliant to diet, HbA_1c (P < 0.05), but had no beneficial effect on cholesterol, HDL cholesterol, or triglyceride concentrations. Conversely, a cross-sectional analysis of dietary fiber in type 1 diabetes patients enrolled in the EURODIAB IDDM Complications Study revealed that a higher intake of total fiber (grams per day) was independently associated with higher levels of HDL cholesterol in both men and women, and lower LDL cholesterol levels in men but not women (89). No substantial differences were observed between soluble and insoluble fiber intakes. Mean total fiber intake was 18.5 g/day in men and 16.2 g/day in women.

The Dietary Guidelines for Americans (16) recommends that all Americans choose a variety of fiber-containing food, such as whole grains, fruits, and vegetables, because they provide vitamins, minerals, fiber, and other substances important for good health. This is an appropriate recommendation for people with type 1 diabetes as well.

There is strong evidence for the following statements:

• Studies in healthy subjects support the importance of including food containing carbohydrate from whole grains, fruits, vegetables, and low-fat milk in the diet.
  
• With regard to the glycemic effects of carbohydrates, the total amount of carbohydrate in meals and snacks is more important than the source or type.
  
• Individuals receiving intensive insulin therapy should adjust their premeal insulin dosages based on the carbohydrate content of meals.
  

There is some evidence for the following statements:

• Individuals receiving fixed daily insulin dosages should try to be consistent in day-to-day carbohydrate intake.
  
• Although the use of low-glycemic index food may reduce postprandial hyperglycemia, there is not sufficient evidence of long-term benefit to recommend use of low-glycemic index diets as a primary strategy in food/meal planning for individuals with type 1 diabetes.
  
• As for the general public, consumption of fiber is to be encouraged; however, there is no reason to recommend that people with type 1 diabetes consume a greater amount of fiber than other Americans.
  
• Percentages of carbohydrate should be based on individual nutrition assessment.
  

The following statement is based on expert consensus:

• Carbohydrate and monounsaturated fat together should provide 60–70% of energy intake.
  

Carbohydrate and type 2 diabetes
As is the case for type 1 diabetes, there is no body of evidence relating to people with type 2 diabetes to suggest changing the 1994 recommendation that 60–70% of total energy be divided between carbohydrate and monounsaturated fat. In weight-maintaining diets for type 2 patients with diabetes, replacing carbohydrate with monounsaturated fat reduces postprandial glycemia and triglyceridemia (90,91), but there is concern that increased fat intake in ad libitum diets may promote weight gain and potentially contribute to insulin resistance (92,93,94,95,96,97,98,99,100). Thus the contributions of carbohydrate and monounsaturated fat to energy intake should be individualized based on nutrition assessment, metabolic profiles, and weight and treatment goals.

In individuals with type 2 diabetes, postprandial glucose levels and insulin responses to a variety of starches and sucrose are similar if the amount of carbohydrate is constant (69,71,101,102,103,104,105,106). This has been demonstrated in both controlled (69,101,102,103,104) and in free-living subjects (71,105,106). When studied, the effects of starches and sucrose on plasma lipids were similar and no adverse effects were observed (103,104,105,106).

Glycemic index.
There have been nine studies (80,82,107,108,109,110,111,112,113) involving type 2 diabetes subjects (n = 129) that have compared low-glycemic index and high-glycemic index diets for longer than 1 day. One study (107) reported lower HbA_1c levels in low- compared to high-glycemic index diets, whereas four studies (80,82,108,109) reported no differences in HbA_1c levels. Three studies (110,111,112) reported significantly lower fructosamine levels in low- compared to high-glycemic index diets, whereas three other studies (108,109,113) reported no significant differences in fructosamine. No differences in fasting plasma glucose concentrations were reported in eight studies (80,107,108,109,110,111,112,113), and no differences in insulin levels were found in two studies (107,109).

In studies that also assessed the effects of low- and high-glycemic index diets on plasma lipids, there were no consistent results. One study (112) reported positive differences on cholesterol levels, whereas four (80,107,108,113) reported no differences. One study reported positive differences in HDL cholesterol levels (109), whereas five (80,108,110,112,113) reported no differences. Four studies (108,109,110,112) reported no differences in LDL cholesterol levels. One study (80) reported positive differences in triglyceride levels; five studies (107,108,110,111,112) reported no differences.

Although studies in type 2 diabetes subjects have not consistently reported a relation between glycemic index and insulin and lipid levels, studies in other populations have reported an association between either lower glycemic index diets or lower glycemic loads with lipids, in particular HDL cholesterol, and insulin levels. In a cross-sectional study of middle-aged adults, the glycemic index of the diet was the only dietary variable significantly related to serum HDL cholesterol concentration (114), and a recent analysis (115) of the Third National Health and Nutrition Examination Survey (NHANES III) reported a change in HDL concentration of –2.3 mg/dl per 15-unit increase in glycemic index. In a study of 32 patients with advanced coronary heart disease, 4 weeks of a low-glycemic index diet improved glucose tolerance and insulin sensitivity compared to a high-glycemic index diet over the same period (116). The same group reported that a low- compared to a high-glycemic index diet improved adipocyte insulin sensitivity in women at risk for coronary heart disease (117).

The glycemic load, defined as the product of the glycemic index value of a food and its carbohydrate content, has been reported to be positively associated with the risk of developing type 2 diabetes in men and women (118,119) and coronary heart disease in women (120). In a cross-sectional study of healthy postmenopausal women, dietary glycemic load was inversely related to plasma HDL cholesterol and positively related to fasting triglycerides (121). In the analysis of the NHANES III results, a high glycemic load was associated with a lower concentration of plasma HDL cholesterol (115).

Fiber.
Early studies of the effects of fiber on glycemia showed promising results, but may have suffered from methodological errors (i.e., poor control of confounding variables such as weight loss, differences in energy consumed, different food sources with potential differences in starch digestibility, and differences in dietary fat content) (122). In a study in which dietary variables were controlled for, increasing the fiber content of the diet from 11 to 27 g/1,000 kcal did not lead to improvements in glycemia, insulinemia, or lipemia (123).

In contrast, a diet supplemented with large amounts of water-soluble, gel-forming fiber, such as guar gum, reduced postprandial glycemia (124). In support of this finding, another study comparing a diet containing 24 g fiber per day (high usual intake) to a diet containing 50 g fiber per day found that the intake of food high in dietary fiber improved glycemic control, reduced hyperinsulinemia, and decreased plasma lipids (125). It thus appears that ingestion of large amounts of fiber is necessary to confer metabolic benefit. It is not clear whether the palatability and gastrointestinal side effects of fiber in this amount would be acceptable to most people.

A meta-analysis of 67 controlled clinical trials indicated that diets high in soluble fiber decrease total and LDL cholesterol, but had a small HDL-lowering effect and did not affect triglyceride concentrations (126). Patients with hypercholesterolemia were not more responsive to dietary fiber than healthy individuals. However, the authors concluded that the effect of soluble fiber within practical ranges on cholesterol was modest (daily intake of 3 g soluble fiber, e.g., 3 apples or 3 bowls [29-g servings] oatmeal can decrease total cholesterol by 5 mg/dl, an 2% reduction), and on risk of heart disease may be only small.

Newer fiber supplements such as psyllium (127) and ß-glucan (128,129) have mixed short-term effects on glycemia and lipemia and require further study.

There is strong evidence for the following statements:

• Studies in healthy subjects and those at risk for type 2 diabetes support the importance of including food containing carbohydrate from whole grains, fruits, vegetables, and low-fat milk in the diet.
  
• With regard to the glycemic effect of carbohydrates, the total amount of carbohydrate in meals or snacks is more important than the source or type.
  

There is some evidence for the following statements:

• Although the use of low-glycemic index food may reduce postprandial hyperglycemia, there is not sufficient evidence of long-term benefit to recommend general use of low-glycemic index diets in type 2 diabetes patients.
  
• As for the general population, consumption of fiber is to be encouraged. Although large amounts of dietary fiber (50 g per day) may have beneficial effects on glycemia, insulinemia, and lipemia, it is not known if such high levels of fiber intake can be maintained long-term.
  

The following statement is based on expert consensus:

• Carbohydrate and monounsaturated fat should together provide 60–70% of energy intake. However, the individual’s metabolic profile and need for weight loss should be considered when determining the monounsaturated fat content of the diet. Increasing fat intake may result in increased energy intake.
  

Nutritive sweeteners
Sucrose.
Sucrose is a common, naturally occurring disaccharide composed of a glucose and a fructose molecule. Average per capita consumption of sucrose and other sugars in the U.S. is estimated to be 94 g/day, accounting for 22% of energy intake (130). Historically, the most widely held belief about nutrition and diabetes was that added sugars should be avoided and naturally occurring sugars restricted. This belief was based on the assumption that sucrose and other sugars were more rapidly digested and absorbed then starch-containing food and thereby aggravated hyperglycemia. However, scientific evidence does not support this assumption.

The available evidence from clinical studies demonstrates that dietary sucrose does not increase glycemia more than isocaloric amounts of starch (67,69,101,103,104,106,131,132). Thus the intake of sucrose and sucrose-containing food in diabetic individuals need not be restricted because of a concern about aggravating hyperglycemia. If sucrose is part of the food/meal plan, it should be substituted for other carbohydrate sources or, if added, should be adequately covered with insulin or other glucose-lowering medication. In addition, the intake of other nutrients (such as fat) often ingested with sucrose-containing food should be taken into account. In one study, when individuals with type 2 diabetes included sucrose in their daily meal plan, no negative impact on dietary habits or metabolic control was observed (133).

Fructose.
Fructose is a common, naturally occurring monosaccharide that accounts for 9% of average energy intake in the U.S. (134). Fructose is somewhat sweeter than sucrose. It has been reported that 33% of dietary fructose comes from fruits, vegetables, and other natural sources in the diet and 67% comes from food and beverages to which fructose has been added (135).

In several studies in diabetic subjects, fructose produced a reduction in postprandial glycemia when it replaced sucrose or starch as a carbohydrate source (69,106,136,137). Thus fructose might be a good sweetening agent in the diabetic diet. However, this potential benefit is tempered by the concern that fructose may have adverse effects on plasma lipids. Consumption of large amounts of fructose (15–20% of daily energy intake [90^th percentile of usual intake]) has been shown to increase fasting total and LDL cholesterol in subjects with diabetes (137) and fasting total and LDL cholesterol and triglycerides in nondiabetic subjects (138,139,140,141).

Sugar alcohols (polyols).
Sugar alcohols are classified as hydrogenated monosaccharides (e.g., sorbitol, mannitol, xylitol), hydrogenated disaccharides (e.g., isomalt, maltitol, lactitol), and mixtures of hydrogenated mono- (sorbitol), di- (maltitol), and oligosaccharides (e.g., hydrogenated starch hydrolysates) (46). They are used in food as sweeteners and bulking agents. Sugar alcohols have been designated by the U.S. Food and Drug Administration (FDA) as safe for use as food additives or as Generally Recognized as Safe (GRAS) by affirmation petitions accepted for filing by the FDA (Table 2). The FDA has not indicated a need to designate an acceptable daily intake. Food prepared with sugar alcohols may claim on the label that there is an association between sugar alcohols and reduced risk of dental caries. Because sugar alcohols are only partially absorbed from the small intestine, the claim of reduced energy values per gram is allowed. However, if certain polyols are used in food, a warning concerning excess consumption and laxative effects of polyols is required on the food label (Table 2).

There is strong evidence for the following statements:

• Sucrose does not increase glycemia to a greater extent than isocaloric amounts of starch.
  
• Sucrose and sucrose-containing food do not need to be restricted by people with diabetes based on a concern about aggravating hyperglycemia. However, if sucrose is included in the food/meal plan, it should be substituted for other carbohydrate sources or, if added, be adequately covered with insulin or other glucose-lowering medication.
  

There is some evidence for the following statements:

• Fructose reduces postprandial glycemia when it replaces sucrose or starch in the diabetic diet.
  
• Consumption of fructose in large amounts may have adverse effects on plasma lipids.
  
• The use of sugar alcohols as sweetening agents appears to be safe.
  
• Sugar alcohols may cause diarrhea, especially in children.
  

There is limited evidence for the following statement:

• The use of added fructose as a sweetening agent is not recommended.
  

The following statements are based on expert consensus:

• Sucrose and sucrose-containing food should be eaten in the context of a healthy diet, and the intake of other nutrients ingested with sucrose, such as fat, should be taken into account.
  
• There is no reason to recommend that diabetic individuals avoid naturally occurring fructose in fruits, vegetables, and other food.
  
• It is unlikely that sugar alcohols in the amounts likely to be ingested in individual food servings or meals will contribute to a significant reduction in total energy or carbohydrate intake, although no studies have been conducted to support this.
  

Resistant starch.
Resistant starch (nondigestible oligosaccharides and the starch amylose) (Table 1) is not digested and therefore not absorbed as glucose in the small intestine. It is, however, almost completely fermented in the colon and produces about 2 kcal/g of energy (46). It is estimated that resistant starch and unabsorbed starch represent 2–5% (usually <10 g/day) of the total starch ingested in the average Western diet (150). Legumes are the major food source of resistant starch in the diet, containing 2–3 g resistant starch per 100 g cooked legumes. Uncooked cornstarch contains about 6 g resistant starch per 100 g dry weight (151). It has been suggested that ingestion of resistant starch produces a lesser increase in postprandial glucose than digestible starch and correspondingly lower insulin levels. As a result, it has been proposed that food containing naturally occurring resistant starch (cornstarch) or food modified to contain more resistant starch (high amylose cornstarch) may modify postprandial glycemic response, prevent hypoglycemia, and reduce hyperglycemia; these effects may explain differences in the glycemic indexes of some food.

There have been several one-meal (152,153,154) and second-meal studies (155,156,157) in nondiabetic subjects, comparing subjects’ physical response to food high in resistant starch and their response to food with an equivalent amount of digestible starch. All studies found some reduction in postprandial glucose and insulin responses to the first meal, but observed mixed results after the second meal. Long-term studies have not consistently confirmed these results (155,158,159,160,161).

Published studies involving people with diabetes have focused on uncooked cornstarch and its potential to prevent nighttime hypoglycemia (162,163,164,165). In uncontrolled studies, evening cornstarch in specific dosages or dosages based on g/kg body weight resulted in less hypoglycemia around 0200 h in all groups (166,167). Longer term studies of evening cornstarch snacks in adults with type 1 diabetes reported less hypoglycemia at 0300 h (168). In subjects with type 2 diabetes, evening cornstarch snacks increased nocturnal glucose and insulin (165). It has not been established that bedtime cornstarch snacks are more effective in preventing nocturnal hypoglycemia than other types of carbohydrate.

The is limited evidence for the following statement:

• Resistant starches have no established benefit for people with diabetes.
  

Nonnutritive sweeteners.
There are presently four nonnutritive sweeteners (also referred to as high-intensity sugar alternatives, low calorie, or alternative sweeteners) approved for use in the U.S.: saccharin, aspartame, acesulfame potassium (acesulfame K), and sucralose. Saccharin, originally linked to human cancer based on a study in rats more than two decades ago, has now been dropped from the FDA list of cancer-causing chemicals (169).

The newest product approved by the FDA is sucralose (made from sucrose through a multistep process in which three hydrogen-oxygen groups are replaced with three chlorine atoms). Sucralose has been shown to have no effect on glucose homeostasis in diabetic subjects (170,171). FDA approval is being sought for three other nonnutritive sweeteners: alitame (formed from the amino acids aspartic acid and alanine), cyclamates (removed from the market in 1970), and neotame (similar to aspartame but 30–60 times sweeter and will not require special labeling for phenylketonuria) (172). A recent trend in the food industry is to blend high-intensity sweeteners. This decreases the total amount of individual sweeteners used and may improve taste.

Nonnutritive sweeteners approved by the FDA must undergo rigorous scrutiny and are not allowed on the market unless they are demonstrated to be safe for the public, including people with diabetes, to consume. For all food additives, including nonnutritive sweeteners, the FDA determines an acceptable daily intake (ADI), defined as the amount of a food additive that can be safely consumed on a daily basis over a person’s lifetime without risk. Actual intake is much less than the ADI. Although the daily ADI for aspartame is 50 mg/kg body wt, the range of actual daily aspartame intake at the 90^th percentile is 2–3 mg/kg body wt (173). Table 3 lists ADIs of nonnutritive sweeteners (174).

There is strong evidence for the following statement:

• Nonnutritive sweeteners are safe for people with diabetes when consumed within the ADI levels established by the FDA.
  

The following statement is based on expert consensus:

• It is unknown if use of nonnutritive sweeteners improves long-term glycemic control or assists in weight loss.
  

Protein and diabetes
In the U.S., protein intake accounts for 15–20% of average adult energy intake, a statistic that has varied little from 1909 to the present. Protein intake is also fairly consistent across all ages, from infancy to older age (176,177), and appears to be similar in individuals with diabetes. Protein intake in subjects with type 2 diabetes in the U.K. Prospective Diabetes Study was 21% of daily energy (32); protein intake in children with type 1 diabetes has been reported to be 17% of daily energy (178).

Protein needs.
It has been assumed that in people with diabetes, abnormalities of protein metabolism are less affected by insulin deficiency and insulin resistance than abnormalities of glucose metabolism (179). However, moderate hyperglycemia may contribute to increased turnover of protein in type 2 diabetic subjects. During moderate hyperglycemia, obese subjects with type 2 diabetes, compared to nondiabetic obese subjects, had an increase in whole-body nitrogen flux and a higher rate of protein synthesis and breakdown (180,181).

A high-quality protein (95 g protein/day), very-low-energy diet capable of maintaining nitrogen balance in obese subjects without diabetes did not prevent negative nitrogen balance in diabetic subjects, despite weight loss and improved glycemic control (181). This increased protein turnover was restored to normal only with oral glucose-lowering agents or exogenous insulin sufficient to achieve euglycemia and with increased protein intake (182,183). These study results suggest that people with type 2 diabetes have an increased need for protein during moderate hyperglycemia and an altered adaptive mechanism for protein sparing during weight loss. Thus with energy restriction, the protein requirements of people with diabetes may be greater than the recommended dietary allowance (RDA) of 0.8 g protein/kg body wt, although not greater than usual intake, which is 1.0 g protein/kg body wt or 100 g protein/day (32). However, individuals consuming very low energy intakes may have a deficiency in protein intake and require an assessment of protein adequacy.

Protein degradation and conversion of endogenous and exogenous protein to glucose in type 1 diabetes depends on the state of insulinization and corresponding glycemic control. Insulin deficiency increases whole-body protein synthesis, protein breakdown, oxidation of essential amino acids (184), and gluconeogenesis (185). Conversion of excess dietary protein or endogenous protein to glucose may occur and, in turn, adversely influence glycemia.

Short-term kinetic studies have demonstrated increased protein catabolism in type 1 diabetic subjects treated with conventional insulin therapy (186,187,188). In one study, to protect against increased protein catabolism, type 1 subjects required near-normal glycemia and an adequate protein intake (188). Because most adults eat at least 50% more protein than required, people with diabetes appear to be protected against protein malnutrition when consuming a usual diet.

Protein and development of nephropathy.
An association between dietary protein intake and the development of renal disease has been suggested. In seven studies, dietary protein intake was reported to be similar in patients with diabetes with and without nephropathy (189,190,191,192,193,194,195). In all studies, protein intake was in the range of usual intake and rarely exceeded 20% of the energy intake. In a cross-sectional study of 2,500 type 1 diabetic subjects, those who reported protein consumption <20% of total energy had average albumin excretion rates <20 mcg/min (196). Those in whom protein intake was 20% of daily energy (22% of patients) had average albumin excretion rates >20 mcg/min (in the range of microalbuminuria). In individuals with macroalbuminuria, 32% consumed >20% of total energy from protein versus 23% for those with microalbuminuria and 20% for those with normal albumin excretion. This suggests that a high-protein intake may have a detrimental effect on renal function.

The long-term effects of consuming >20% of energy as protein on the development of nephropathy has not been determined. However, intake of protein in the usual range does not appear to be associated with the development of diabetic nephropathy.

Glucose responses to protein.
A number of studies in healthy, normal-weight subjects (197) and subjects with controlled type 2 diabetes (blood glucose <200 mg/dl) (198,199,200) have demonstrated that ingested protein does not increase plasma glucose concentration. Gannon et al. (200) reported that during the 8-h period after subjects with type 2 diabetes ingested 50 g protein in the form of very lean beef, 20–23 g of protein were deaminated, which in theory could yield 11–13 g glucose. However, the amount of glucose appearing in the circulation was only 2 g, confirming that ingested protein does not result in a significant increase in glucose concentration. This raises the question that if gluconeogenesis from protein is occurring, why does the glucose produced not appear in the general circulation after the ingestion of protein?

In type 2 diabetic subjects, the peak plasma glucose response to carbohydrate is similar to the response to carbohydrate plus protein (197,198,199), suggesting that protein does not slow the postprandial absorption of carbohydrate.

In individuals capable of secreting insulin, protein ingestion is just as potent as glucose ingestion in stimulating insulin secretion (197,198,199,200). The net effect on glucose output by the liver depends on the ratio of insulin to glucagon. In type 1 or type 2 diabetic subjects, the glucagon response to protein is considerably greater than in nondiabetic subjects (201).

In one study of subjects with well-controlled type 1 diabetes, the addition of protein to a meal did not slow the absorption of carbohydrate or change either the postprandial peak glucose response to the meal or glucose levels at 5 h (202). Furthermore, in type 1 diabetic subjects, the rate of restoration to euglycemia after hypoglycemia did not differ when treatment was given with carbohydrate or carbohydrate plus protein (203). Glucose levels, the time to peak plasma glucose levels, and subsequent rate of glucose fall were similar after both treatments.

Protein’s effect on satiety and/or energy balance.
It has been claimed that high-protein, low-carbohydrate diets produce weight loss and improvements in glycemia. It should be noted that most of these diets tend to be high in fat. In one study of high-protein diets, there was a significant decrease in weight and plasma triglycerides at 12 weeks (204). However, plasma LDL cholesterol levels were increased. There is no research available to document that high-protein diets maintain long-term weight reduction any better than traditional weight-loss diets and that they are safe for long-term use.

The Continuing Survey of Food Intake by Individuals 1994–1996 (177) was used to examine the relationship among prototype popular diets (205). In a comparison of low-carbohydrate diets (30% of energy from carbohydrate [high protein diets]) and high-carbohydrate diets (>55% of energy from carbohydrate), diet quality was lower and total and saturated fat intake was higher on the lower carbohydrate diet, whereas energy intake and BMI were approximately similar between the two.

The effects of dietary protein on the regulation of energy intake and satiety have not been adequately studied (206,207). Short-term meal studies suggest that protein does exert a positive effect on satiety (208,209,210,211). However, results from one study demonstrated that although hunger was suppressed to a greater extent after a high-protein than a high-fat or high-carbohydrate breakfast, the changes in hunger were not of sufficient magnitude to change ad libitum lunchtime energy intake 5 h later or energy intake for the rest of the day, which were similar after all three breakfast types (210).

There is strong evidence for the following statement:

• In individuals with controlled type 2 diabetes, ingested protein does not increase plasma glucose concentrations, although ingested protein is just as potent a stimulant of insulin secretion as carbohydrate.
  

There is some evidence for the following statements:

• For diabetic individuals, there is no evidence to suggest that usual protein intake (15–20% of total daily energy) should be modified if renal function is normal.
  
• For diabetic individuals, especially those with less-than-optimal glycemic control, the protein requirement may be greater than the RDA, but not greater than usual intake.
  
• Contrary to advice often given to patients with diabetes, the available evidence suggests the following: 1) dietary protein does not slow the absorption of carbohydrate and 2) dietary protein and carbohydrate do not raise plasma glucose later than carbohydrate alone and thus do not prevent late-onset hypoglycemia.
  

There is limited evidence for the following statement:

• It may be prudent to avoid protein intake >20% of total daily energy.
  

The following statement is based on expert consensus:

• The long-term effects of diets high in protein and low in carbohydrate are unknown. Although such diets may produce short-term weight loss and improved glycemia, it has not been established that weight loss is maintained. The long-term effect of such diets on plasma LDL cholesterol is also a concern.
  

Dietary fat and diabetes
Saturated fats and dietary cholesterol.
The primary goal regarding dietary fat in patients with diabetes is to decrease intake of saturated fat and cholesterol (212,213,214). Saturated fat is the principal dietary determinant of LDL cholesterol (213). Compared to nondiabetic subjects, diabetic subjects have an increased risk of coronary heart disease with higher intakes of dietary cholesterol (215). The ADA (8,212) and the National Cholesterol Education Program’s Adult Treatment Panel III (18) have recommended that the serum LDL cholesterol goal be <100 mg/dl. To assist in achieving this goal, it is recommended that food with a high content of saturated fatty acids and cholesterol be limited (17,18).

In a meta-analysis (216) of 37 dietary intervention studies in free-living subjects, plasma total cholesterol decreased from baseline by 24 mg/dl (10%), LDL cholesterol by 19 mg/dl (12%), and triglycerides by 15 mg/dl (8%) in Step I (10% saturated fat and 300 mg cholesterol) interventions (P < 0.01 for all). In Step II interventions (7% saturated fat and 200 mg cholesterol), total cholesterol decreased by from baseline by 32 mg/dl (13%), LDL cholesterol by 25 mg/dl (16%), and triglycerides by 17 mg/dl (8%) (P < 0.01 for all). HDL cholesterol decreased by 7% (P = 0.05) in response to Step II but not Step I dietary interventions. Positive correlations between changes in dietary total and saturated fatty acids and changes in total, LDL, and HDL cholesterol were observed. Adding exercise resulted in greater decreases in total and LDL cholesterol and triglycerides and prevented the decrease in HDL cholesterol associated with low-fat diets. However, studies in diabetic subjects demonstrating the effects of specific percentages of saturated fatty acids (e.g., 10 vs. 7% of energy) and specific amounts of dietary cholesterol (e.g., 300 vs. 200 mg) are not available. Therefore, the goal for patients with diabetes remains the same as for the general population: to reduce saturated fat intake to <10% of energy intake. Some individuals (i.e., those with LDL cholesterol 100 mg/dl) may benefit by reducing saturated fat to <7% of energy intake. The goal for dietary cholesterol intake is <300 mg/day and for individuals with LDL cholesterol 100 mg/dl, <200 mg/day.

For patients with diabetes, the debate has focused not on the extent to which saturated fatty acids and cholesterol intake should be limited, but rather on what is the best alternative energy source. Plasma cholesterol reductions of 9–29% have been reported in four studies in which saturated fat was replaced with carbohydrate in diabetic diets (217,218,219,220). Two of these studies also measured plasma LDL and HDL cholesterol and reported that substituting a low-fat (<30% of total daily calories), high-carbohydrate diet for a high-saturated fat diet resulted in reductions in LDL, but not HDL, cholesterol. (217,218) Glycemic control was improved or unchanged as a result of restricting dietary saturated fat and replacing it with carbohydrate (217,219,221,222,223,224). See Table 4 for classification of fatty acids (225).

Polyunsaturated fats.
Only a few studies have evaluated the effects of polyunsaturated fat on plasma lipid levels and glycemic control in subjects with diabetes. In one study of type 2 subjects with diabetes, a diet high in total and polyunsaturated fat resulted in lower plasma total and LDL cholesterol than a diet high in total and saturated fat, but produced no difference in other plasma lipid levels (231). Another study in type 2 diabetic subjects compared a diet high in polyunsaturated fat with one high in monounsaturated fat and reported higher plasma total and LDL cholesterol, fasting glucose, and insulin levels with the polyunsaturated fat diet (232).

N-3 polyunsaturated fat (omega-3 fatty acids).
Food sources of n-3 polyunsaturated fatty acids include fish, especially fatty fish, as well as plant sources such as flaxseed and flaxseed oil, canola oil, soybean oil, and nuts. N-3 fatty acid supplements have been shown to reduce plasma triglyceride levels, especially in hypertriglyceridemic individuals (233), and to have beneficial effects on platelet aggregation and thrombogenicity (234). Increasing the intake of n-3 polyunsaturated fatty acids has been shown to be beneficial in subjects with diabetes (235,236,237). Although fish oil supplementation may be beneficial in lowering plasma triglyceride levels in type 2 diabetic subjects, the accompanying rise in plasma LDL cholesterol is of concern (238,239). Therefore, if n-3 fatty acid supplements are used, the effects on plasma LDL cholesterol should be monitored. Glucose metabolism is not likely to be adversely affected with the use of n-3 supplements (236,237,240). N-3 supplements may be most beneficial in the treatment of severe hypertriglyceridemia (236,241). Although studies of the effects of n-3 fatty acids in patients with diabetes have primarily used supplements, there is evidence from the general population that food containing n-3 fatty acids, specifically eicosapentaenoic acid and docosahexaenoic acid, has cardioprotective benefits (242,243,244,245,246).

Transunsaturated fatty acids.
Trans fats—unsaturated fatty acids formed when vegetable oils are processed and made more solid (hydrogenation)—are found in some margarines and in food prepared or fried in hydrogenated vegetable oils. Trans fats also occur naturally in small amounts in meats and dairy products. The mean intake of trans fatty acids in the U.S. has been estimated at 2.6% of total caloric intake and 7.4% of total fat intake (247). When studied independently of other fatty acids, the effect of trans fatty acids is similar to that of saturated fats in raising plasma LDL cholesterol. Trans fatty acids also lower plasma HDL cholesterol (248,249,250). Studies in nondiabetic subjects support limiting the intake of trans fatty acids.

Stanols/sterols.
Plant stanols are found in very small amounts in food from plants such as corn and soy and in other vegetable or plant oils. Within plant tissue they are derived from plant phytosterols. They differ from plant sterols in that their ring structure is saturated. Plant stanols are esterified to other vegetable oil fatty acids to enhance their lipid solubility and make them easier to use as an ingredient in food. Plant sterol and stanol esters block the intestinal absorption of dietary and biliary cholesterol (251) by competing with cholesterol for entry into the mixed micelles that must form during digestion for dietary fatty acids, cholesterol, and fat-soluble vitamins to be absorbed. Plant stanols/sterols in the amount of 2 g/day have been shown to lower serum total cholesterol by up to 10% and LDL cholesterol by up to 14% (251,252,253,254,255).

Low-fat diets.
There are potential benefits from low-fat diets. Low-fat diets are usually associated with modest loss of weight, which can be maintained as long as the diet is continued (230,256) and if combined with aerobic exercise (216,257,258). In studies evaluating the effect of ad libitum energy intake as a function of dietary fat content, low-fat, high-carbohydrate intake is associated with a transient decrease in caloric intake and modest weight loss, leading to a new equilibrium body weight (221,259,260,261,262,263,264,265,266,267,268). With this modest weight loss, a decrease in total cholesterol and plasma triglycerides and an increase in HDL cholesterol occur. Consistent with this, low-fat, high-carbohydrate diets over long periods of time have been shown to not increase plasma triglycerides and, when reported, have led to modest weight loss (230,269,270). Although the significance of the effects of reduced dietary fat intake are controversial (256,271,272,273,274,275,276,277,278,279), reduced-fat diets have been shown to maintain weight loss better than other types of reduced energy diets (216,257,280,281,282,283).

In type 2 diabetic subjects, restrained eating behaviors combined with dietary fat restriction have been shown to have beneficial effects on glycemia, plasma lipids, and/or weight (284,285,286). A higher intake of total dietary fat is associated with higher levels of plasma LDL cholesterol, and the adverse effect of a higher carbohydrate intake on triglycerides has been found in individuals who have undiagnosed diabetes or have gained weight during the previous year (287).

Fat replacers/substitutes.
Dietary fat intake can be decreased by reducing the amount of high-fat food in the diet. Another option is to provide lower fat or fat-free versions of food and beverages. This can be accomplished by removing some fat or by using fat replacers (ingredients that mimic the properties of fat with significantly fewer calories than fat) in food formulations. The FDA has approved the majority of the fat replacers as GRAS because the substances’ ingredients have a long history of safe use in food. A few replacers (notably olestra) have been approved as food additives; approval of these requires both demonstration of safety and premarket approval (288,289,290). Although olestra has no effect on water-soluble nutrients, it can lead to a loss of fat-soluble vitamins.

Two recent studies involving diabetic subjects and food made with fat replacers have been reported (291,292). One of these, a short-term study (292), provided correctly labeled regular or fat-free food to free-living subjects with and without diabetes. Use of fat substitutes/replacers in reasonable amounts (five low-fat or no-fat products per day) produced a small decrease in dietary fat, saturated fat, and cholesterol intake with little or no decrease in total energy intake or weight. When fat replacers are used in larger amounts (293,294), there can be a significant decrease in energy intake. Long-term studies are needed to assess the effect of food containing fat replacers/substitutes on the macronutrient content of the diets patients with diabetes and their utility in achieving treatment goals.

Studies in nondiabetic subjects provide strong evidence for the following statements:

• In all, <10% of energy intake should be derived from saturated fats. Some individuals (i.e., those with LDL cholesterol 100 mg/dl) may benefit from lowering saturated fat intake to <7% of energy intake.
  
• Dietary cholesterol intake should be <300 mg/day. Some individuals (i.e., those with LDL cholesterol 100 mg/dl) may benefit from lowering dietary cholesterol to <200 mg/day.
  
• Intake of transunsaturated fatty acids should be minimized.
  
• Current fat replacers/substitutes approved by the FDA are safe for use in food.
  

Studies in diabetic subjects provide strong evidence for the following statement:

• To lower plasma LDL cholesterol, energy derived from saturated fat can be reduced if concurrent weight loss is desirable or replaced with carbohydrate or monounsaturated fat if weight loss is not a goal.
  

There is some evidence for the following statements:

• Polyunsaturated fat intake should be 10% of energy intake.
  
• In weight-maintaining diets, when monounsaturated fat replaces carbohydrate, it may beneficially affect postprandial glycemia and plasma triglycerides but not necessarily fasting plasma glucose or HbA_1c.
  
• Incorporation of two to three servings of plant stanols/sterols (2 g) food per day, substituted for similar food, will lower total and LDL cholesterol.
  
• Reduced-fat diets when maintained long term contribute to modest loss of weight and improvement in dyslipidemia.
  

There is limited evidence for the following statement:

• Two or more servings of fish per week provide dietary n-3 polyunsaturated fat and can be recommended.
  

The following statements are based on expert consensus:

• Monounsaturated fat and carbohydrate together should provide 60–70% of energy intake. However, increasing fat intake may result in increased energy intake.
  
• Fat intake should be individualized and designed to fit ethnic and cultural backgrounds.
  
• Use of low fat food and fat replacers/substitutes by patients with diabetes may reduce total fat and energy intake and thereby facilitate weight loss.
  

Energy balance and obesity
Many individuals with type 2 diabetes are overweight, with 36% having a BMI 30 kg/m², which would classify them as obese (295). The prevalence of obesity is higher in women and members of minority populations with type 2 diabetes (295). As body adiposity increases, so does insulin resistance (296,297,298). Obesity may also aggravate hyperlipidemia and hypertension in type 2 patients with diabetes (299).

Because of the effects of obesity on insulin resistance, weight loss is an important therapeutic objective for obese individuals with type 2 diabetes. Short-term studies lasting 6 months or less have demonstrated that weight loss in type 2 diabetic subjects is associated with decreased insulin resistance, improved measures of glycemia, reduced serum lipids, and reduced blood pressure (300,301,302). Long-term data assessing the extent to which these improvements can be maintained in people with type 2 diabetes are not available.

Data from the general public suggest that long-term maintenance of weight loss is challenging. In two observational studies on weight maintenance after weight loss in nondiabetic subjects, one study (303) reported that only 6% in the final study group maintained a 5% weight loss over 9–15 years, while in a random telephone survey (304), 21% of 228 overweight subjects reported that they had intentionally lost weight and maintained a weight loss of 10% for at least 5 years. However, long-term data assessing the extent to which weight loss is maintained in patients with diabetes are not available. In studies of weight loss in type 2 diabetic subjects (305,306,307), the most successful long-term weight loss from diet was reported in the Diabetes Treatment Study (307), with weight loss of 9 kg maintained over the 6-year study period. To accomplish this required long-term access to therapeutic contact.

The reason that long-term weight loss is difficult for most people to accomplish is probably because energy intake and energy expenditure, and thereby body weight, are controlled and regulated by the central nervous system (308,309,310). Although our understanding of central nervous system regulation of energy balance is incomplete, it is thought that the hypothalamus may be the center of control. Neuropeptide Y, leptin, insulin, and a variety of other neural, endocrine, and gastrointestinal signals also appear to be involved. Individual characteristics of central nervous system control of energy balance may be genetically determined. For example, in a study of Danish adoptees, there was a strong relation between BMI of the adoptees and their biological parents, and no relation whatsoever between the BMI of the adoptees and their adoptive parents (311). These study results suggest that genetic factors have an important role in determining body weight. Other data support this conclusion (312,313). Furthermore, environmental factors often make losing weight difficult for those genetically predisposed to obesity.

The National Weight Control Registry has enrolled over 3,000 subjects successful at long-term maintenance of weight loss (314). A group of 800 people who lost an average of 30 kg and maintained a minimum weight loss of 13.6 kg (30 lb) for 5 years were identified from the registry (315). Slightly more than half lost weight through formal programs, and the remainder lost weight with a program of their own. Average energy consumption was 1,400 kcal/day, with 24% of energy derived from fat. Average energy expenditure through added physical activity was 2,800 kcal/week. Importantly, nearly 77% of this sample of people who were successful in achieving and maintaining weight loss reported a triggering event that preceded the weight loss. The most common triggering events were acute medical conditions and emotional problems. Thus a new diagnosis of type 2 diabetes could trigger lifestyle changes that result in reduced fat and energy intake, increased physical activity, and associated weight loss.

Structured programs to produce lifestyle change.
The recently completed Diabetes Prevention Program (DPP) demonstrated long-term benefit in people with glucose intolerance from structured, intensive lifestyle programs (316,317). In the DPP, participants randomly assigned to an intensive lifestyle intervention that included a low-fat diet, increased physical activity, edu