Glycemic Index and Glycemic Load of Popular Weight-Loss Diets

James M. Shikany, DrPH; Sarah E. Thomas, BS; C. Suzanne Henson, MS, RD; David T. Redden, PhD; Douglas C. Heimburger, MD

Abstract

Context: Carbohydrate-restricted diets have been popular in recent years. The theoretical glycemic impact of these diets, compared with other popular weight-loss diets, has not been reported.
Objective: To assess the glycemic index (GI) and glycemic load (GL) of 2 popular, carbohydrate-restricted diets (South Beach and Sugar Busters!) and compare them with a low-fat, high-carbohydrate diet (Ornish) and a moderate-fat, moderate-carbohydrate diet (EatRight).
Design: All available sample menus provided in the book for each diet were extracted and included in the analyses. GI values for all carbohydrate-containing foods were assigned based on a published list. GL values were determined based on these GI values and recommended serving sizes.
Main Outcome Measures: Median daily GI and GL values were calculated for each diet and for each phase of South Beach and each meal pattern of EatRight.
Results: The median daily GLs of South Beach and Sugar Busters! (34 and 48, respectively) were less than one half those of Ornish and EatRight (113 and 104, respectively). Adjusting the diets to 1500 kcal attenuated the differences slightly. The median daily GIs of the diets were in a very narrow range (46-53); however, South Beach (46) was significantly lower than Ornish (53).
Conclusions: The GLs of 2 carbohydrate-restricted diets were significantly lower than those of a low-fat, high-carbohydrate diet and a moderate-fat, moderate-carbohydrate diet. The differences were due primarily to lower carbohydrate content rather than to differences in overall GIs of the diets.

 


Readers are encouraged to respond to George Lundberg, MD, Editor of MedGenMed, for the editor's eye only or for possible publication via email: glundberg@medscape.net

Introduction

The prevalence of overweight and obesity among adults in the United States increased from 47% and 15%, respectively, in the period of 1976-1980 to 65% and 31%, respectively, in the period of 1999-2002.[1] Countless dietary approaches for weight loss have been promulgated over the years. Although first described over 150 years ago, carbohydrate-restricted diets have undergone resurgence in popularity over the past several years. Interest in these diets is demonstrated by the continuing popularity of carbohydrate-restricted diet books in the lay press, with several of these books enjoying long runs on nonfiction bestseller lists.

Several of the most popular carbohydrate-restricted diets (including The South Beach Diet and Sugar Busters!) embrace the concepts of glycemic index (GI) and glycemic load (GL) as adjuncts to weight-loss efforts. GI is a ranking of carbohydrate-containing foods based on postprandial blood glucose response and is a measure of carbohydrate quality.[2] Generally, the lower the GI, the lower the rate of absorption of the carbohydrate and the smaller the rise in postprandial glucose and insulin concentrations.[3] GL is a measure that incorporates both the quality and quantity of dietary carbohydrates.[4] It has been hypothesized that low-GI diets may aid weight-loss efforts by increasing satiety and by promoting fat oxidation at the expense of carbohydrate oxidation.[5] Small clinical studies suggest that low-GI diets may promote weight loss.[6]

A systematic assessment of GI and GL of carbohydrate-restricted diets has not been reported. The purpose of this study was to assess the theoretical glycemic impact of 2 popular carbohydrate-restricted diets and compare them with a low-fat, high-carbohydrate diet and a moderate-fat, moderate-carbohydrate diet.

Materials and Methods

Diets

Two carbohydrate-restricted diets were analyzed: The South Beach Diet[7] and Sugar Busters!.[8] These diets were chosen based on the popularity of their respective books on The New York Times' Best Sellers' list in early 2004. The South Beach Diet book remains on The New York Times' Best Sellers' list as of November 2005.[7] For comparison, a high-carbohydrate, low-fat diet (Ornish)[9] and a moderate-carbohydrate, moderate-fat diet (EatRight)[10] were included. All available sample menus in each diet book were extracted and included in the analyses.

South Beach. The South Beach Diet is a carbohydrate-restricted diet with 3 phases. During Phase 1, all carbohydrates are severely restricted. In Phase 2, the concept of GI is introduced and low-GI carbohydrates, such as apples, high-fiber cereals, multigrain breads, and reduced-fat milk, are gradually reintroduced into the diet. Higher-GI carbohydrates, such as refined-grain breads and potatoes, are proscribed. Phase 3 of the diet is the maintenance phase.

Sugar Busters!. The Sugar Busters! diet is a 14-day dietary plan based on the restriction of dietary sugars rather than all carbohydrates. The diet emphasizes GI and restricts high-GI carbohydrates. The plan also discusses GL and advises against the overconsumption of even low-GI carbohydrates because of deleterious effects on dietary GL.

Ornish. In this diet plan, dietary fat is restricted to 10% of total energy intake. In contrast to the carbohydrate-restricted diets, the Ornish diet includes 70% of total energy from carbohydrates. So-called complex carbohydrates and high-fiber foods are promoted. Sugars are restricted, although not completely eliminated from the diet.

EatRight. The EatRight Weight Management Program is a 12-week lifestyle-oriented, weight-control program that was developed at the University of Alabama at Birmingham. The program is based on the concept of "time-calorie displacement," which emphasizes the ingestion of large quantities of high-bulk, low-energy-density foods (vegetables, fruits, high-fiber grains, and cereals) and moderation in high-energy-density foods (meats, cheeses, sugars, and fats). Women are placed on a 1200-kcal/day regimen (Meal Pattern 1) and men on a 1500-kcal/day regimen (Meal Pattern 2). Meal Pattern 3 is an 1800-kcal regimen that is recommended for some persons who have reached their weight goals.

Nutritional Analysis

The nutrient composition of each diet was determined by computerized nutritional analysis, with the Nutrition Data System for Research (NDS-R) software, version 4.06 (Nutrition Coordinating Center, University of Minnesota, Minneapolis). Nutrient values used in the calculations came from the NDS-R software or were provided by the manufacturer.

Determination of Dietary GI and GL

Dietary GI and GL values were determined for each meal and day for each of the 4 diets. Calculations were also done for each phase or meal pattern of a diet, where applicable. To determine the GI and GL values for a given meal, a list of all carbohydrate-containing foods in the meal was generated by NDS-R, along with the amount of total carbohydrate and dietary fiber in the serving size of each food specified by the respective diet. Available carbohydrate was used in the GL calculations and was calculated by subtracting the amount of dietary fiber from the amount of total carbohydrate. Each food was assigned a glucose-based GI value from a published table of GI values.[11] The GL value of an individual food was calculated by multiplying the GI value of the food by the grams of available carbohydrate per serving and dividing by 100.[11] The GLs of all carbohydrate-containing foods in the meal were summed, resulting in the GL for the meal. The GL for a day was obtained by summing the meal GLs for that day.

The GI of a meal was calculated by multiplying the meal GL by 100 and dividing by the total grams of available carbohydrate in the meal, resulting in a weighted median GI, with weights based on the proportion of available carbohydrate content of the meal contributed by each food.[12] With the same methods, the GI for a day's intake was also determined. These methods assume that the overall glycemic effect of the diet is proportional to the GI and the amount of carbohydrate that each food contributes to the diet. These assumptions have been demonstrated to be correct in a mixed diet of low- and high-GI foods, which was shown to have a glycemic response similar to that predicted based on the GI values of the component foods.[13]

GI values for foods with no published values were imputed based on published GI values of similar foods. Foods containing little carbohydrate (eg, meats, poultry, fish, cheese, eggs, and salad vegetables) were not included in the GL and GI analyses due to their minimal glycemic effect. When there were multiple GI values listed for a given food in the GI table, the mean GI value was assigned. For mixed dishes with no published GI value, the GI was calculated as the weighted average of the GIs of the component carbohydrate ingredients with the methods described above.

Statistical Analysis

Descriptive statistics (medians and ranges) were used to summarize the diets' nutrient contents. Weighted descriptive statistics, with "glycemic" carbohydrates as a weighting variable, were calculated for the GI. Due to small sample sizes, nonparametric tests (Kruskal-Wallis) were used to compare nutrient information among the diets. For significant results, follow-up procedures with Mann-Whitney tests were conducted with a Bonferroni correction to control for multiple comparisons.

Results

The number of days of sample menus provided for each diet varied from 7 (Ornish and EatRight) to 14 (South Beach and Sugar Busters!). Overall, the compositions of both carbohydrate-restricted diets were very similar ( ). When compared with Ornish and EatRight (Meal Pattern 3), the carbohydrate-restricted diets were lower in energy, carbohydrates, and dietary fiber. Both South Beach (Phase 3) and Sugar Busters! were significantly higher in fat (both in absolute terms and as a percentage of total energy) and cholesterol compared with Ornish, but not compared with EatRight.

Table 1.  Energy, Macronutrient, Fiber, and Cholesterol Content of the Diets*

Dietary Variable South Beach Sugar Busters! Ornish EatRight
Phase 1 Phase 2 Phase 3 Meal Pattern 1 Meal Pattern 2 Meal Pattern 3
Energy (kcal/d) 1264 (1064-1735) 1303 (829-2158) 1250 (878-2893) 1226 (966-2042)† 1726 (1176-2089) 1045 (979-1243) 1533 (1317-1733) 1913 (1693-2053)†
Carbohydrate (g/d) 69 (41-95) 128 (68-220) 123 (74-301)†,‡ 132 (109-227)§,║ 320 (219-421)†,§ 166 (139-181) 224 (192-234) 267 (235-311)‡,║
Carbohydrate (% of energy) 21.0 (14.7-31.2) 37.8 (26.7-45.4) 40.1 (23.3-53.0)†,‡ 45.7 (24.5-61.1)§ 80.3 (74.2-81.4)†,§,║ 58.9 (45.7-67.2) 58.4 (48.0-69.1) 56.5 (49.1-64.1)‡,║
Fat (g/d) 67.9 (41.9-104.8) 71.9 (28.1-116.7) 48.1 (25.5-105.2)† 38.3 (17.4-151.8) 19.6 (13.7-24.7)†,‡,§ 24.6 (17.7-46.3) 34.3 (24.9-65.2) 46.8 (35.0-71.0)§
Fat (% of energy) 48.5 (35.4-58.2) 41.2 (24.4-48.2) 36.6 (20.4-52.1)†,‡ 31.9 (14.8-66.9)§ 10.7 (8.6-12.5)†,§,║ 19.0 (16.3-34.3) 20.8 (16.8-34.4) 20.5 (18.6-33.4)‡,║
Protein (g/d) 109.6 (76.4-129.6) 85.7 (52.2-140.4) 85.8 (62.9-187.0) 76.4 (62.3-93.1)† 65.6 (56.1-84.2) 64.3 (53.2-87.1) 86.8 (63.4-130.6) 111.7 (88.3-158.4)†,‡
Protein (% of energy) 33.7 (22.3-40.9) 27.8 (19.5-34.6) 26.3 (19.4-33.2)† 25.4 (14.9-31.0) 16.1 (13.7-19.1)†,‡,§ 22.3 (18.1-28.0) 24.6 (18.1-30.1) 22.8 (18.2-30.9)§
Dietary fiber (g/d) 16.2 (9.5-26.8) 24.6 (13.1-49.8) 18.5 (12.2-36.8)†,‡ 19.6 (14.8-33.2)§,║ 50.5 (34.1-74.1)†,§ 24.2 (18.1-33.6) 34.3 (22.8-45.5) 42.8 (24.2-46.8)‡,║
Cholesterol (mg/d) 285 (193-858) 307 (109-663) 207 (114-536)† 187 (112-682) 14 (3-52)†,‡,§ 113 (78-509) 167 (98-531) 216 (154-537)§

 

*All values are medians; ranges in parentheses

†,‡,§,║ Values with the same superscript within a horizontal row are significantly different from one another (overall comparison used P < .05; Bonferroni follow-up comparisons used P < .008 = .05/6; Phase 3 was used for South Beach and Meal Pattern 3 for EatRight).

Analyses of median GIs and GLs of the diets are reported in . With Phase 3 of South Beach and Meal Pattern 3 of EatRight for comparison, the lowest overall median daily GLs were seen in the 2 carbohydrate-restricted diets, South Beach and Sugar Busters!. The GLs of these 2 diets were similar in magnitude (34 and 48, respectively) and were less than half that of both Ornish and EatRight (P < .001 for differences), which were also similar in magnitude (113 and 104, respectively). Within the carbohydrate-restricted diet with multiple phases (South Beach), there were increases in GL with progression of the phases: 3, 32, and 34 for Phases 1, 2, and 3, respectively; however, only the differences between Phases 1 and 2 and Phases 1 and 3 were statistically significant. This pattern was also seen in the comparison diet with multiple meal patterns -- EatRight, with GLs of 60, 82, and 104 for Meal Patterns 1, 2, and 3, respectively.

Table 2  Glycemic Index and Glycemic Load* of the Diets

Diet Days of Menus Glycemic Index† Glycemic Load Presented Glycemic Load Adjusted§
South Beach, Phase 1 14 26 (4-38) 3 (1-6) 4 (1-7)
South Beach, Phase 2 14 42 (35-61) 32 (14-47) 36 (16-54)
South Beach, Phase 3 14 46 (38-52) 34 (12-77) 41 (14-92)
Sugar Busters! 14 50 (38-68) 48 (39-70) 59 (48-86)
Ornish 7 53 (48-55) 113 (72-146) 98 (63-127)
EatRight, Meal Pattern 1 7 46 (44-54) 60 (48-71) 86 (69-102)
EatRight, Meal Pattern 2 7 48 (44-55) 82 (67-103) 80 (66-101)
EatRight, Meal Pattern 3 7 48 (45-56) 104 (83-117) 82 (65-92)

 

*Values are medians; ranges in parentheses

†Values are weighted medians; ranges in parentheses

Values for the diet as presented

§Values for the diet adjusted to 1500 kcal/day

A comparison of GLs among the diets was also carried out after standardizing the diets to 1500 kcal to adjust for the effects of differences in energy content on GL ( and Figure 1). Adjusting the diets had the effect of attenuating the differences in GL among the diets to some degree; however, the GLs of the 2 carbohydrate-restricted diets (41 and 59 for South Beach and Sugar Busters!, respectively) remained lower than the GLs of the 2 comparison diets (98 and 82 for Ornish and EatRight, respectively). However, the difference between Sugar Busters! and EatRight was no longer statistically significant. Within South Beach, adjustment had little effect on GL in any phase, reflecting the nearly constant energy levels (albeit different median GIs) of the 3 phases. The converse was true for EatRight -- adjustment had the effect of equalizing the GLs of the different meal patterns due to the different energy levels but nearly equal median GIs of the 3 meal patterns.

Table 2  Glycemic Index and Glycemic Load* of the Diets

Diet Days of Menus Glycemic Index† Glycemic Load Presented Glycemic Load Adjusted§
South Beach, Phase 1 14 26 (4-38) 3 (1-6) 4 (1-7)
South Beach, Phase 2 14 42 (35-61) 32 (14-47) 36 (16-54)
South Beach, Phase 3 14 46 (38-52) 34 (12-77) 41 (14-92)
Sugar Busters! 14 50 (38-68) 48 (39-70) 59 (48-86)
Ornish 7 53 (48-55) 113 (72-146) 98 (63-127)
EatRight, Meal Pattern 1 7 46 (44-54) 60 (48-71) 86 (69-102)
EatRight, Meal Pattern 2 7 48 (44-55) 82 (67-103) 80 (66-101)
EatRight, Meal Pattern 3 7 48 (45-56) 104 (83-117) 82 (65-92)

 

*Values are medians; ranges in parentheses

†Values are weighted medians; ranges in parentheses

Values for the diet as presented

§Values for the diet adjusted to 1500 kcal/day

Figrue 1.

 

Comparison of median daily glycemic load across the diets (adjusted to 1500 kcal). Phase 3 was used for South Beach and Meal Pattern 3 for EatRight. Median values of diets with the same letters were significantly different (overall comparison used P < .05; Bonferroni follow-up comparisons used P < .008 = .05/6).

There was no clear pattern of dietary GI among the diets. Although one of the carbohydrate-restricted diets (South Beach) had the lowest median GI (46), the GI of EatRight (48) was lower than that of the other carbohydrate-restricted diet, Sugar Busters! (50) ( and Figure 2). The only statistically significant difference in GIs was between South Beach (46) and Ornish (53) (P = .001). In South Beach, there were increases in GI with progression of the phases: 26, 42, and 46 for Phases 1, 2, and 3, respectively; however, only the differences between Phases 1 and 2 and Phases 1 and 3 were statistically significant. Median GI was relatively constant among the meal patterns of EatRight.

Table 2  Glycemic Index and Glycemic Load* of the Diets

Diet Days of Menus Glycemic Index† Glycemic Load Presented Glycemic Load Adjusted§
South Beach, Phase 1 14 26 (4-38) 3 (1-6) 4 (1-7)
South Beach, Phase 2 14 42 (35-61) 32 (14-47) 36 (16-54)
South Beach, Phase 3 14 46 (38-52) 34 (12-77) 41 (14-92)
Sugar Busters! 14 50 (38-68) 48 (39-70) 59 (48-86)
Ornish 7 53 (48-55) 113 (72-146) 98 (63-127)
EatRight, Meal Pattern 1 7 46 (44-54) 60 (48-71) 86 (69-102)
EatRight, Meal Pattern 2 7 48 (44-55) 82 (67-103) 80 (66-101)
EatRight, Meal Pattern 3 7 48 (45-56) 104 (83-117) 82 (65-92)

 

*Values are medians; ranges in parentheses

†Values are weighted medians; ranges in parentheses

Values for the diet as presented

§Values for the diet adjusted to 1500 kcal/day

Figure 2.

 

Comparison of median daily glycemic index across the diets. Phase 3 was used for South Beach and Meal Pattern 3 for EatRight. Median values of diets with the same letters were significantly different (overall comparison used P < .05; Bonferroni follow-up comparisons used P < .008 = .05/6).

Discussion

The GLs of the 2 carbohydrate-restricted diets were less than half that of the 2 comparison diets. Although the median GIs of 2 of the diets were significantly different from each other, there was no clear pattern among the diets' GIs, and the range among the median GIs (46- 53) was very narrow and most likely of little clinical significance. Carbohydrate-restricted diets would be expected to have low GLs due to their reduced carbohydrate content. The contribution of GI to GL is generally less powerful than that of carbohydrate content. Although a reduced carbohydrate content contributed to the low GLs of both South Beach and Sugar Busters!, one of these diets (South Beach) also had a slightly lower median GI than the other diets, which contributed to its low GL value.

Studies have suggested that low-GI diets may play a role in weight loss. A low-GI diet resulted in greater weight loss compared with a high-GI diet in obese women in a 12-week clinical study.[6] Low-GI carbohydrates may help maintain weight and prevent obesity by increasing satiety.[5] Because low-GI foods are digested and absorbed more slowly, nutrient receptors in the gastrointestinal tract are stimulated for a longer period of time, resulting in prolonged feedback to the satiety center in the brain.[14] This results in reduced energy intake at a given meal and at subsequent meals.[15] High-GI diets, compared with isoenergetic and nutrient-controlled, low-GI diets, result in higher average 24-hour blood glucose and insulin concentrations.[16] The postprandial rise in glucose and insulin concentrations in response to high-GI foods increases carbohydrate oxidation through the rapid activation of key rate-limiting enzymes.[17] Concomitant reduction in the rate of fat oxidation has been linked with greater weight gain in several prospective studies.[18,19]

The effectiveness of reducing dietary GI and GL as adjuncts to weight-loss efforts requires further study. Clinical trials of the effects of lowering dietary GI and GL on weight loss have not been conducted. In addition, the concept of GI has been criticized on the grounds that it is impractical and that long-term clinical benefits have not been demonstrated.[20] It should also be noted that a recent study suggested that the most important predictor of success of weight-loss diets was adherence, not the specific dietary prescription.[21]

Despite the questionable clinical utility of GI and GL in weight-loss diets, the concept of GI was included in the diet books for both South Beach and Sugar Busters!. This is contrasted with another well-known, carbohydrate-restricted diet -- the Atkins diet.[22] In Atkins, although extremely low-carbohydrate intake was recommended, there was no reference to GI. Atkins was not included in our analyses because the Atkins book included only 1 day of sample menus for phases beyond induction.

Limitations of this study should be noted. First, although all menus in each diet book were analyzed, these were sample menus, and actual foods consumed by persons following the diets may vary to some extent. Data regarding the estimated GI and GL values of the diets in actual practice would be of interest to determine the degree to which they differ from the theoretical values presented here. Second, the GI and GL values presented are theoretical in nature and do not consider other macronutrients, including fat and protein, and other dietary constituents, such as phytates and lectins, that may influence the GI of foods.[23,24] There is no way to adequately control for the effects of other dietary components in GI/GL calculations. Third, it is possible that the exclusion of foods with very low carbohydrate contents could have resulted in calculated GLs that were slightly lower than what would be expected based on the total carbohydrate content. Because these foods comprised only a small fraction of the total carbohydrate content of each diet, it is unlikely that excluding these foods materially altered the absolute GL values or the differences in GLs among the diets. Finally, there were a few foods for which no published GI values exist. In these cases, GI values were imputed, which could have introduced bias.

In summary, 2 carbohydrate-restricted diets had significantly lower median GLs than a high-carbohydrate, low-fat diet and a moderate-carbohydrate, moderate-fat diet, primarily due to their lower carbohydrate contents. By contrast, the range of GIs among the diets was very narrow and probably has little clinical significance. The GIs of the diets that we analyzed that emphasized the concept of GI were not substantially different from those of the diets that did not emphasize it. It may therefore be impossible to isolate an independent effect of GI on weight-loss outcomes. The impact of GI and GL on efforts to prevent and treat obesity remains to be determined.

References

  1. National Center for Health Statistics. Health, United States, 2004 With Chartbook on Trends in the Health of Americans. Hyattsville, Md: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2004. DHHS Publication No. 2004-1232.

  2. Jenkins DJ, Wolever TM, Taylor RH, et al. Glycemic index of foods: a physiological basis for carbohydrate exchange. Am J Clin Nutr. 1981;34:362-366. Abstract

  3. Wolever TM. The glycaemic index. World Rev Nutr Diet. 1990;62:120-185. Abstract

  4. Liu S, Manson JE, Stampfer MJ, et al. Dietary glycemic load assessed by food-frequency questionnaire in relation to plasma high-density-lipoprotein cholesterol and fasting plasma triacylglycerols in postmenopausal women. Am J Clin Nutr. 2001;73:560-566. Abstract

  5. Brand-Miller JC, Holt SH, Pawlak DB, et al. Glycemic index and obesity. Am J Clin Nutr. 2002;76(suppl):281S-285S.

  6. Slabber M, Barnard HC, Kuyl JM, et al. Effects of a low-insulin-response, energy-restricted diet on weight loss and plasma insulin concentrations in hyperinsulinemic obese females. Am J Clin Nutr. 1994;60:48-53. Abstract

  7. Agatson A. The South Beach Diet: The Delicious, Doctor-Designed, Foolproof Plan for Fast and Healthy Weight Loss. New York: Rodale Inc.; 2003.

  8. Steward HL, Bethea MC, Andrews SS, Balart LA. Sugar Busters! Cut Sugar to Trim Fat. New York: Ballantine Books; 1998.

  9. Ornish D. Eat More, Weigh Less: Dr. Dean Ornish's Life Choice Program for Losing Weight Safely While Eating Abundantly. New York: Quill; 2001.

  10. Weinsier RL, Wilson NP, Morgan SL, et al. EatRight Lose Weight: 7 Simple Steps. Birmingham, Ala: Oxmoor House; 1997 (plus accompanying workbook).

  11. Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index and glycemic load values: 2002. Am J Clin Nutr. 2002;76:5-56. Abstract

  12. Wolever TM, Nguyen PM, Chiasson JL, et al. Determinants of diet glycemic index calculated retrospectively from diet records of 342 individuals with non-insulin-dependent diabetes mellitus. Am J Clin Nutr. 1994;59:1265-1269. Abstract

  13. Collier GR, Wolever TM, Wong GS, Josse RG. Prediction of glycemic response to mixed meals in noninsulin-dependent diabetic subjects. Am J Clin Nutr. 1986;44:349-352. Abstract

  14. Lavin JH, Wittert GA, Andrews J, et al. Interaction of insulin, glucagon-like peptide 1, gastric inhibitory polypeptide, and appetite in response to intraduodenal carbohydrate. Am J Clin Nutr. 1998;68:591-598. Abstract

  15. Ludwig DS. Dietary glycemic index and obesity. J Nutr. 2000;130(suppl):280S-283S.

  16. Jenkins DJ, Wolever TM, Collier GR, et al. Metabolic effects of a low-glycemic-index diet. Am J Clin Nutr. 1987;46:968-975. Abstract

  17. Wolfe RR. Metabolic interactions between glucose and fatty acids in humans. Am J Clin Nutr. 1998;67(suppl):519S-526S.

  18. Zurlo F, Lilloja S, Esposito-Del Puente A, et al. Low ration of fat to carbohydrate oxidation as predictor of weight gain: study of 24-h RQ. Am J Physiol. 1990;259:E650-E657. Abstract

  19. Weyer C, Pratley R, Salbe A, et al. Energy expenditure, fat oxidation, and body weight regulation: a study of metabolic adaptation to long-term weight change. J Clin Endocrinol Metab. 2000;85:1087-1094. Abstract

  20. Pi-Sunyer X. Glycemic index and disease. Am J Clin Nutr. 2002;76:290S-298S. Abstract

  21. Dansinger ML, Gleason JA, Griffith JL, et al. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial. JAMA. 2005;293:43-53. Abstract

  22. Atkins RC. Dr. Atkins' New Diet Revolution. New York: Avon Books; 2002.

  23. Jenkins DJ, Wolever TM, Taylor RH, et al. Glycemic index of foods: a physiological basis for carbohydrate exchange. Am J Clin Nutr. 1981;34:362-366. Abstract

  24. Jenkins DJ, Wolever TM, Jenkins AL. Starchy foods and glycemic index. Diabetes Care. 1988;11:149-159. Abstract