Soy foods have low glycemic and insulin response indices in normal weight subjects

1475-2891-5-35 1475-2891 Research Soy foods have low glycemic and insulin response indices in normal weight subjects Blair M Robert drblair@revivalsoy.com Henley EC henleyec@aol.com Tabor Aaron drtabor@revivalsoy.com

Physicians Pharmaceuticals, Inc., 1031 E. Mountain St., Building 302, Kernersville, NC 27284, USA

Nutrition Journal 1475-2891 2006 5 1 35 http://www.nutritionj.com/content/5/1/35 17192192 10.1186/1475-2891-5-35 13 3 2006 27 12 2006 27 12 2006 2006 Blair et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Foods with a low glycemic index (GI) may provide a variety of health benefits. The objective of the present study was to measure the GI and insulin index (II) of select soy foods.

Methods

The study was conducted in two parts with low-carbohydrate products being tested separately. In Experiment 1, subjects averaged 23.2 years of age with BMI = 22.0 kg/m², while subjects in Experiment 2 averaged 23.9 years of age with BMI = 21.6 kg/m². The reference (glucose) and test foods were served in portions containing 10 g of carbohydrates in Experiment 1 (two test foods) and 25 g of carbohydrates in Experiment 2 (four test foods). Subjects consumed the reference food twice and each test food once. For each test, subjects were instructed to consume a fixed portion of the reference food or test food together with 250 g of water within 12 min. Blood samples were collected before each test and at 15, 30, 45, 60, 90, and 120 min after consumption of reference or test foods to quantify glucose and insulin. Two-hour blood glucose and plasma insulin curves were constructed and areas under the curves were calculated. GI and II values for each subject and test food were calculated.

Results

In Experiment 1, both low-carbohydrate soy foods were shown to have significantly (P < 0.05) lower GI and II values than the reference food. In Experiment 2, three of the four test foods had significantly (P < 0.05) lower GI and II values than the reference food.

Conclusion

All but one of the soy foods tested had a low GI, suggesting that soy foods may be an appropriate part of diets intended to improve control of blood glucose and insulin levels.

Background

The glycemic index (GI) was first developed by Jenkins and colleagues 1 as a new method of classifying foods based on the blood glucose response after food consumption. The GI value of a food is a percentage of the 2-hour area under the blood glucose response curve of a reference food, typically glucose 2. Since the GI is determined for a particular quantity of carbohydrates in the food being tested and since the actual amount of carbohydrates consumed in a meal or snack varies greatly, the GI concept was expanded to include the concept of glycemic load (GL). The GL is determined by multiplying the GI of a food by the grams of carbohydrates in a serving. The GL value incorporates the amount of digestible carbohydrates in a serving in order to better gauge the impact of a meal or snack on postprandial glucose response 34.

It has been reported that a high GI diet may have adverse health consequences by increasing the risk for chronic disease 56. Recent evidence suggests that high GI/GL diets may increase the risk for cardiovascular disease (CVD) 789 and type 2 diabetes 341011. A high GI diet may increase the risk for chronic disease through the stimulation of hyperglycemia and hyperinsulinemia 6.

In contrast, a low GI diet has been reported to have health benefits 561213. A low GI diet has been shown to improve glycemic control 14151617, aid in weight loss 1819, and reduce some CVD risk factors 91420212223.

To date, only about 30 – 40 soy foods have been assessed for their GI/GL values 124252627. The objective of the current study was to determine the GI and insulin index (II) values of select soy food products (bars, drinks, pasta, and chips) currently available on the U.S. and international markets.

Methods

The current study was conducted using internationally recognized GI methodology 282930. The experimental procedures used in this study were in accordance with international standards for conducting ethical research with humans and were approved by the Human Research Ethics Committee of Sydney University of Australia where the study was conducted by contract.

A search of the literature using the National Library of Medicine's PubMed search engine showed a paucity of literature on the determination of the glycemic index of soy foods. A search of "soy AND glycemic index" yielded 5 hits; a search of "soy AND glycemic load" yielded 2 hits; a search of "soy AND glycemic" yielded 14 hits and a search of "isoflavone AND glycemic" yielded 5 hits. However, none of the papers found measured the glycemic index or glycemic load of soy foods. A search of the international table of glycemic index and glycemic load values 26 showed that the glycemic index of several soy foods has been measured, but only a few of these have been reported as independent publications.

Study Subjects

For both experiments, 10 healthy, non-smoking subjects, were recruited from the staff and student population of the University of Sydney. Exclusion criteria included being overweight, dieting, impaired glucose tolerance, illness or food allergy, or regular use of prescription medication (other than contraceptive medication). All study participants gave written informed consent before participating in this study.

In Experiment 1 the 10 subjects (two females, eight males) had a mean age of 23.2 years (19.9–25.7 years) and a mean body mass index (BMI) score of 22.0 kg/m²(absolute range = 19.4–25.4 kg/m²). In Experiment 2 the 10 subjects (four females, six males) had a mean age of 23.9 years (20.3–26.9 years) and a mean BMI score of 21.6 kg/m²(absolute range = 19.5–25.8 kg/m²). Four subjects participated in both experiments.

Composition of Test Foods

In Experiment 1, the reference and two low-carbohydrate test foods (products with ≤6 g net carbohydrates (available carbohydrates); net carbohydrates = total carbohydrates – sugar alcohols – fiber) were served to subjects in portions containing 10 g of digestible carbohydrate. In Experiment 2, the reference and four test foods were served to subjects in portions containing 25 g of available carbohydrate. Glucose (Glucodin^®powder, Boots Health Care Company, North Ryde, NSW Australia) dissolved in water was the reference food. Mass and nutrient contents of the reference and test foods (Revival Soy^®from Physicians Pharmaceuticals, Inc., Kernersville, NC, USA) are listed in Tables 1 and 2.

Table 1

Mass and nutrient contents of the test portions of the glucose reference and the two test foods in Experiment 1.

Test Food

Portion Size (g)¹

Energy (kJ [Cal]) {kj/100 g of product}

Protein (g)²{g/100 g of product}

Fat (g) {g/100 g of product}

Available Carbohydrate (g) {g/100 g of product}

Sugars (g) {g/100 g of product}

Fiber (g) {g/100 g of product}

Glucose (Reference Food)

10.0 g glucose 250 g water

160 [38.2]

0.0

10.0

0.0

Chocolate Raspberry Zing™ bar

150.0

2205 [526.7] {1,470}

45.0 {30.4}

12.5 {8.4}

10.0 {7.7}

0.0 {0.0}

< 2.0 {1.3}

Chocolate Daydream™ shake – sucralose

70.0 g powder 500.0 g water

1092 [260.8] {1,551}

40.0 {57.1}

5.0 {7.1}

10.0 {14.3}

2.0 {2.9}

4.0 {5.7}

¹The amount of water listed for some products is the amount of water needed to prepare that product. This volume is in addition to the 250 g of water consumed with every product as part of the test methodology.

²All of the protein in the products above is derived from soy. Each Chocolate Raspberry Zing™ bar contained 21 g soy protein. Each Chocolate Daydream™ shake contained 20 g soy protein.

Table 2

Mass and nutrient contents of the test portions of the glucose reference and the four test foods in Experiment 2.

Test Food

Portion Size (g)

Energy (kJ [Cal]) {kj/100 g of product}

Protein (g)* {g/100 g of product}

Fat (g) {g/100 g of product}

Available Carbohydrate (g) {g/100 g of product}

Sugars (g) {g/100 g of product}

Fiber (g) {g/100 g of product}

Glucose (Reference Food)

25.0 g glucose 250 g water

400 [95.5]

0.0

25.0

0.0

Peanut Butter Chocolate Pal™ bar

48.4

813 [194.2] {1,671}

12.9 {26.7}

4.8 {10.0}

25.0 {51.7}

14.5 {30.0}

0.8 {1.7}

Chocolate Daydream™ shake – fructose

47.1 g powder 305.0 g water

742 [177.2] {1,567}

14.7 {31.3}

1.8 {3.9}

25.0 {53.0}

23.6 {50.0}

1.5 {3.1}

Lightly Salted Sunshine™ soy protein chips

48.1

808 [193.0] {1,672}

13.5 {28.0}

3.8 {8.0}

25.0 {52.0}

0.0 {0.0}

Soy spaghetti

42.4 (dry)

636 [151.9] {1,467}

10.6 {24.6}

1.1 {2.6}

25.0 {57.9}

1.5 {3.5}

0.8 {1.8}

²The protein in all of the products except the spaghetti is from soy. For the spaghetti, one serving provided 14 grams of protein, 8 grams from soy and 6 grams from semolina. Each Chocolate Daydream™ shake contained 20 g soy protein. Each Peanut Butter Chocolate Pal™ bar contained 16 g soy protein. The soy protein chips contained 7 g soy protein per serving

Each portion of the reference food was prepared the day before required by dissolving the glucose in 250 g of water and storing overnight at 4°C. The individual portions of the test foods were prepared the day before required, except for the soy spaghetti. The individual portions of the uncooked soy spaghetti were weighed the day before required. On the testing day, each portion of dry spaghetti was cooked for 4 minutes in boiling water and drained. The reference and test foods were served with 250 g of plain water. The subjects consumed all the food and fluid served to them at a comfortable pace within 12 minutes.

Experimental Procedures

The experimental methods used in the current study have been previously described 31 and are briefly outlined here. In both experiments, study subjects consumed the reference food on two separate occasions and each of the test foods on one occasion only after a 10-hour overnight fast. The reference food was consumed on the first and last test sessions, and test foods were consumed in random order in between. Each test session was completed on a separate morning with at least a day between subsequent sessions.

On each test day a baseline, finger-prick blood sample was obtained for blood glucose and plasma insulin determinations using an automatic, non-reusable lancet device (Safe-T-Pro^®, Boehringer Mannheim GmbH, Mannheim, Germany). Following consumption of the reference or test food, additional blood samples were collected at 15, 30, 45, 60, 90 and 120 minutes. Blood glucose concentrations were measured immediately after the blood samples were collected. Blood samples collected for plasma insulin determination were centrifuged for 30 seconds immediately after collection and the plasma layer from each sample was transferred into a labeled, uncoated microcentrifuge tube and stored at -20°C until analyzed.

Blood Glucose and Glycemic Index Determinations

The glucose concentration in each of the whole capillary blood samples was analyzed in duplicate using a glucose dehydrogenase/mutarotase enzymatic reaction using a HemoCue^®beta-glucose photometric analyzer (HemoCue AB, Ängelholm, Sweden). Duplicate readings were accepted if the two separate measurements for each time point were within 0.3 mmol/L of each other. If the readings were not within 0.3 mmol/L of each other, then an additional 2 blood glucose sample readings were taken from the subject within approximately 40 – 60 seconds after the initial readings. The two or three similar (i.e. within 0.3 mmol/L) readings were then averaged together to obtain the blood glucose response for that time point. A two-hour blood glucose response curve was constructed and the incremental area under the glucose response curve (IAUC) was calculated. The GI value for each test food was calculated for each subject by dividing the two-hour blood glucose IAUC value for the test foods by their average two-hour blood glucose IAUC value for the reference food and multiplying by 100 to obtain a percentage score. The final reported GI value for each test food is the mean GI value for that food in the group of 10 subjects.

Plasma Insulin and Insulin Index Determinations

The concentration of insulin in each plasma sample was analyzed using a solid-phase antibody-coated tube radioimmunoassay kit (Diagnostic Products Corporation, Los Angeles, CA, USA). A two-hour plasma insulin response curve was constructed from the data and the IAUC of the insulin response curve was calculated. The II value for each test food was calculated for each subject by dividing their plasma insulin IAUC value for the test foods by their mean plasma insulin IAUC value for the reference food and multiplying by 100 to obtain a percentage score.

Statistical Analyses

Sample size calculations (90% power, level of significance = 0.05) using data from published GI studies indicated that a minimum of eight study subjects would be needed to detect significant differences among the GI values of the reference and test foods. Analysis of variance (ANOVA) and the Fisher PLSD test for multiple comparisons were used to determine significant differences between the test foods' mean GI and II values. Statistical analyses were conducted using Statview Student™ software (version 4, Abacus Concepts Inc., Berkley, CA, USA). Significance was assumed at P < 0.05.

Results

Experiment 1: Low-Carbohydrate Soy Foods

The mean blood glucose response curves for the reference and two test foods are shown in Figure 1. The reference food produced a much larger rise in blood glucose during the first 30 minutes and a greater overall glycemic response than the two test foods. The two test foods produced slightly different glucose response curves with the Chocolate Raspberry Zing™ bar producing a higher glycemic response than the Chocolate Daydream™ sucralose shake. However, both foods produced very low glycemic response curves.

Figure 1

Blood glucose response curves of low-carbohydrate soy products in Experiment 1

Blood glucose response curves of low-carbohydrate soy products in Experiment 1. The mean blood glucose response curves for the equal-carbohydrate portions of the reference food (glucose) and the two soy-based, low-carbohydrate food products tested in Experiment 1. Data are expressed as the change in blood glucose concentration from the fasting baseline concentration. Bars for each data point represent standard error of the means (SEM).

The mean plasma insulin response curves for the reference and two test foods are shown in Figure 2. The plasma insulin responses observed for the reference food and the test foods showed a similar profile to their concurrent blood glucose responses. The reference food produced the highest peak plasma insulin concentration and the largest overall plasma insulin response, followed by the Chocolate Raspberry Zing™ bar and the Chocolate Daydream™ sucralose shake, respectively.

Figure 2

Plasma insulin response curves of low-carbohydrate soy products in Experiment 1

Plasma insulin response curves of low-carbohydrate soy products in Experiment 1.The mean plasma insulin response curves for the equal-carbohydrate portions of the reference food (glucose) and the two soy-based, low-carbohydrate food products tested in Experiment 1. Data are expressed as the change in plasma insulin concentration from the fasting baseline concentration. Bars for each data point represent standard error of the means (SEM).

Experiment 2: Other Soy Foods

The mean blood glucose response curves for the reference and the four test foods are shown in Figure 3. Similar to the results observed in Experiment 1, the reference food produced a large rise in blood glucose during the first 30 minutes and the greatest overall glycemic response. The four test foods varied in their peak blood glucose concentrations and their overall glycemic responses. Among the test foods, the baked soy protein chips produced the largest glycemic response followed by the Peanut Butter Chocolate Pal™ bar, the soy spaghetti, and the Chocolate Daydream™ fructose shake.

Figure 3

Blood glucose response curves of soy products in Experiment 2

Blood glucose response curves of soy products in Experiment 2.The mean blood glucose response curves for the equal-carbohydrate portions of the reference food (glucose) and the four soy-based food products tested in Experiment 2. Data are expressed as the change in blood glucose concentration from the fasting baseline concentration. Bars for each data point represent standard error of the means (SEM).

The mean plasma insulin response curves for the reference and the four test products are shown in Figure 4. The foods' average plasma insulin responses were similar to their respective mean plasma glucose responses. The reference food produced the largest plasma insulin response, followed by the four test foods in the same order as their glycemic responses.

Figure 4

Plasma insulin response curves of soy products in Experiment 2

Plasma insulin response curves of soy products in Experiment 2.The mean plasma insulin response curves for the equal-carbohydrate portions of the reference food (glucose) and the four soy-based food products tested in Experiment 2. Data are expressed as the change in plasma insulin concentration from the fasting baseline concentration. Bars for each data point represent standard error of the means (SEM).

Glycemic and Insulin Indices

The mean GI value of the glucose reference was significantly greater (P < 0.001) than the mean GI values of each of the test foods with the exception of the baked soy protein chips (Figure 5). The mean GI value of the soy protein chips was not different from the glucose reference, but was significantly greater (P < 0.001) than the mean GI values for the other five test foods. Despite a high GI value, the soy protein chips had only a medium GL value due to the small serving size and relatively low carbohydrate level (Table 3).

Figure 5

Glycemic and insulin index values of tested soy products

Glycemic and insulin index values of tested soy products. The mean (± SEM) glycemic index and insulin index for the reference food (glucose) and the six tested soy-based food products. The dark bars represent the glycemic index values and the light bars represent the insulin index values. For the GI values, columns with different superscripts (a, b) are significantly (P < 0.001) different. Columns representing the II values with different superscripts (w, x, y, z) are significantly different (P < 0.001).

Table 3

Glycemic index and glycemic load¹values for the six tested soy food products

Glycemic Index

Glycemic Load

Test Food

Carbohydrate Tested (g)

Value ± SEM²

Category³

Carbohydrates/Serving (g)

Value

Category⁴

Chocolate Daydream™ shake – sucralose

10.0

25.00 ± 4.28

Low

1.25

Low

Chocolate Daydream™ shake – fructose

25.0

32.73 ± 4.41

Low

11.13

Medium

Soy spaghetti

25.0

47.03 ± 7.48

Low

15.52

Medium

Chocolate Raspberry Zing™ bar

10.0

47.42 ± 4.55

Low

2.85

Low

Peanut Butter Chocolate Pal™ bar

25.0

51.82 ± 3.86

Low

16.06

Medium

Lightly Salted Sunshine™ soy protein chips

25.0

86.79 ± 7.86

High

11.28

Medium

¹Glycemic Load = (GI × net carbohydrates)/100

²SEM = Standard Error of the Means

³Glycemic Index Category: Low = ≤ 55; Medium = 56 – 69; High = > 70 [54]

⁴Glycemic Load Category: Low = ≤ 10; Medium = 11 – 19; High = > 20 [54]

The mean II value of the glucose reference was significantly greater (P < 0.001) than the mean II values of each of the six test foods (Figure 5). The mean II of the soy protein chips was significantly higher (P < 0.001) than the mean II values for the soy spaghetti, Chocolate Daydream™ fructose shake, Chocolate Raspberry Zing™ bar, and the Chocolate Daydream™ sucralose shake. The mean II values for the Peanut Butter Chocolate Pal™ bar and the Chocolate Raspberry Zing™ bar were both significantly greater (P < 0.001) than the mean II values for the Chocolate Daydream™ fructose and Chocolate Daydream™ sucralose shakes.

Discussion

The results of this study demonstrate that 5 of the 6 soy food products tested have a low GI value (GI ≤ 55). Of the 6 products tested, only the baked soy protein chips had a high GI value (GI > 70). However, when the amount of available carbohydrates in one serving of the soy protein chips was used to calculate a GL value, the soy protein chips had a medium GL value. The other products tested also were either low or medium GL foods.

An increasing body of evidence suggests that the GI and/or GL values of foods impact human health (see 56123233 for reviews). Low-GI diets have been shown to improve glycemic control in diabetic 1415 and non-diabetic 1617 subjects. Diet impacts the incidence of type 2 diabetes and the regulation of dietary carbohydrate has taken on a prominent role in dietary control of this chronic disease. Two recent meta-analyses 1334 reported that consumption of low-GI foods rather than high-GI foods appears to modestly improve glycemic control by reducing plasma cholesterol, fructosamine, and hemoglobin A_1C(Hb_A1c) levels.

A number of studies suggest that high GI/GL diets may increase CVD risk 789, while several others indicate that low GI diets may reduce some CVD risk factors 9141520212223. Meta-analysis results indicated that low GI diets significantly reduced total cholesterol (average reduction = 0.17 mmol/L; P = 0.03) and HbA_1c(average reduction after 12 weeks = 0.45%; P = 0.02) compared to high GI diets 35. These data suggest that dietary GI may improve some, but not all markers of cardiovascular disease risk.

Few studies have examined the effect of low GI diets on weight loss; however, there is some evidence that low GI diets may be beneficial 33. In obese women, an energy-restricted, low GI diet significantly increased weight loss compared to an energy-restricted, high GI diet 18. Spieth and co-workers 19 demonstrated that an ad libitum low GI diet significantly (P < 0.05) reduced BMI to a greater extent than did an energy-restricted, low fat diet. A recent study demonstrated that dietary GI was inversely associated with thigh intramuscular fat while GL was inversely associated with visceral abdominal fat in men 36. Despite the potential benefits of low-GI diets on weight loss, a number of studies report no effect on weight loss 3738.

Soy protein is a high quality protein that has been extensively studied. The quality of soy protein has been assessed through several metabolic studies of nitrogen balance 394041, which have demonstrated that soy protein supports nitrogen balance on par with beef and milk proteins. One recent study reported that amino acids from soy protein appear in the serum sooner, but that this may lead to a more rapid breakdown of the amino acids in the liver 42.

Dietary soy consumption has been shown to have beneficial effects on several aspects of human health, including the diseases potentially influenced by dietary GI levels 434445. Soy consumption has been reported to modestly improve plasma lipid profiles 4647, improve bone health 48, help reduce menopausal symptoms 49, and slightly reduce the risk of breast 50 and prostate cancers 51. The health benefits of dietary soy have been attributed to its isoflavones as well as to the biological actions of its constituent proteins. However, an additional means of providing health benefits may be through the low GI of soy and soy foods.

The international table of GI and GL values 26 reports the GI/GL values of a number of soy foods. These values range from a low GI of 14 for soybeans canned in brine to a high GI of 115 for a tofu-based frozen dessert 24. The Revival Soy^®products tested in this study fell within this range and with the exception of the baked soy protein chips were all within the low GI category. Only a few other studies have reported on the GI of soy-based foods. Packer et al. 25 indicated that gluten-free, soy-based, bread had a high GI value. In contrast, the addition of soy foods has been shown to lower the GI value of mixed meals 27. Similarly, the replacement of unrefined wheat flour with soy flour lowers the GI value of parantha, an Indian snack food 52. Similar to the Revival Soy^®bars tested in the current study, other snack bars containing soy have been shown to have low to medium GI values 53. Previous data report that the GI values of spaghetti ranges from 27 – 68. The Revival Soy^®thin spaghetti had a GI value = 47, similar to other spaghetti products. Overall, these studies indicate that soy-based foods generally have a low to medium GI value and would be suitable for individuals concerned with regulating blood glucose and insulin levels.

The ingredients and form of a food product affect its GI value. For example, while soybeans have a low GI value, the use of high GI ingredients in soy foods can increase the GI value of the final product. This was likely the case with the baked soy protein chips. The baked soy protein chips contain potato starch and potatoes have a high GI value 33. Additionally, the baked soy protein chips have a puffed physical form, which may lead to high GI values. Similar to planning diets with types of fat and protein in mind, types of carbohydrates should also be considered since carbohydrate types influence the GI. The substitution of high GI ingredients with low GI ingredients in food products like the baked soy protein chips may help keep the final GI value down.

In addition to the effects of form and content of foods on the GI value, consumption of other foods with low GI foods can affect the overall GI value of meal. Sugiyama et al. 34 demonstrated that adding soybean products (miso, natto, and ground soybean) lowered the GI of white rice by 20 – 40%. However, further studies are required before any conclusions can be drawn.

Conclusion

With the apparent resurgence of interest in low-GI diets for weight loss and health benefits, it is important that information on the GI value of foods is available. Therefore, we conducted the current study to determine the GI value of a small variety of commercially available soy foods. The results of the current study demonstrate that soy food products generally have low GI values and low to medium GL values. Improvements in ingredient selection and usage may further improve glycemic responses to soy foods. The low GI of soy foods appears to be an additional benefit of soy for human health and suggests that soy foods are an appropriate part of diet plans intended to improve control over blood glucose and insulin levels.

Competing interests

Robert M. Blair is the Research Manager of Physicians Pharmaceuticals, Inc. E. C. Henley is a consultant to Physicians Pharmaceuticals, Inc. Aaron Tabor is the CEO and Medical Research Director of Physicians Pharmaceuticals, Inc.

Authors' contributions

RMB prepared the manuscript. ECH and AT participated in the coordination and design of the study and assisted with manuscript preparation.

Acknowledgements

This study was funded by Physicians Pharmaceuticals, Inc. and conducted via contract research at the University of Sydney, Australia.

Glycemic index of foods: a physiological basis for carbohydrate exchange Jenkins DJ Wolever TM Taylor RH Barker H Fielden H Baldwin JM Bowling AC Newman HC Jenkins AL Am J Clin Nutr 1981 34 362 366 6259925 The glycemic index: methodology and clinical implications Wolever TMS Jenkins DJA Jenkins AL Josse RG Am J Clin Nutr 1991 54 846 854 1951155 Dietary fiber, glycemic load, and risk of NIDDM in men Salmeron J Ascherio A Rimm EB Colditz GA Spiegelman D Jenkins DJ Stampfer MJ Wing AL Willett WC Diabetes Care 1997 20 545 50 9096978 Dietary fiber, glycemic load, and risk of non-insulin-dependent diabetes mellitus in women Salmeron J Manson JE Stampfer MJ Colditz GA Wing AL Willett WC JAMA 1997 277 472 477 10.1001/jama.277.6.472 9020271 Glycemic index in chronic disease: a review Augustin LS Franceschi S Jenkins DJA Kendall CWC La Vecchia C Eur J Clin Nutr 2002 56 1049 1071 10.1038/sj.ejcn.1601454 12428171 The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease Ludwig DS JAMA 2002 287 2414 2423 10.1001/jama.287.18.2414 11988062 A prospective study of dietary glycemic load, carbohydrate intake, and risk of coronary heart disease in US women Liu S Willett WC Stampfer MJ Hu FB Franz M Sampson L Hennekens CH Manson JE Am J Clin Nutr 2000 71 1455 1461 10837285 Primary prevention of coronary heart disease in women through diet and lifestyle Stampfer MJ Hu FB Manson JE Rimm EB Willett WC New Engl J Med 2000 343 16 22 10.1056/NEJM200007063430103 10882764 Correlation between dietary glycemic index and cardiovascular disease risk factors among Japanese women Amano Y Kawakubo K Lee JS Tang AC Sugiyama M Mori K Eur J Clin Nutr 2004 58 1472 1478 10.1038/sj.ejcn.1601992 15127092 Glycemic index and dietary fiber and the risk of type 2 diabetes Hodge AM English DR O'Dea K Giles GG Diabetes Care 2004 27 2701 2706 15505008 Glycemic index, glycemic load, and dietary fiber intake and incidence of type 2 diabetes in younger and middle-aged women Schulze MB Liu S Rimm EB Manson JE Willett WC Hu FB Am J Clin Nutr 2004 80 348 356 15277155 Glycemic index and load – dynamic dietary guidelines in the context of diseases Colombani PC Physiol Behav 2004 83 603 610 10.1016/j.physbeh.2004.07.029 15621065 Meta-analysis of the health effects of using the glycaemic index in meal-planning Opperman AM Venter CS Oosthuizen W Thompson RL Vorster HH Br J Nutr 2004 92 367 381 10.1079/BJN20041203 15469640 Improved glycemic control and lipid profile and normalized fibrinolytic activity on a low-glycemic index diet in type 2 diabetic patients Jarvi AE Karlstrom BE Granfeldt YE Bjorck IE Asp NG Vessby BO Diabetes Care 1999 22 10 18 10333897 Long-term dietary treatment with increased amounts of fiber-rich low-glycemic index natural foods improves blood glucose control and reduces the number of hypoglycemic control in children with type 1 diabetes Giacco R Parillo M Rivellese AA Lasorella G Giacco A D'Episcopo L Riccardi G Diabetes Care 2000 23 1461 1466 11023137 Metabolic effects of a low-glycemic-index diet Jenkins DJ Wolever TM Collier GR Ocana A Rao AV Buckley G Lam Y Mayer A Thompson LU Am J Clin Nutr 1987 46 968 975 2825505 The effect of low-glycemic carbohydrate on insulin and glucose response in vivo and in vitro in patients with coronary heart disease Frost G Keogh B Smith D Akinsanya K Leeds S Metabolism 1996 45 669 672 10.1016/S0026-0495(96)90129-8 8637438 Effects of a low-insulin-response, energy-restricted diet on weight loss and plasma insulin concentrations in hyperinsulinemic obese females Slabber M Barnard HC Kuyl JM Dannhauser A Schall R Am J Clin Nutr 1994 60 48 53 8017337 A low-glycemic index diet in the treatment of pediatric obesity Spieth LE Harnish JD Lenders CM Raezer LB Pereira MA Hangen SJ Ludwig DS Arch Pediatr Adolesc Med 2000 154 947 951 10980801 Low-glycemic index diet in hyperlipidemia: use of traditional starchy foods Jenkins DJ Wolever TM Kalmusky J Guidici S Giordano C Patten R Wong GS Bird JN Hall M Buckley G Am J Clin Nutr 1987 46 66 71 3300252 Diets high and low in glycemic index versus high monounsaturated fat diets: effects on glucose and lipid metabolism in NIDDM Luscombe ND Noakes M Clifton PM Eur J Clin Nutr 1999 53 473 478 10.1038/sj.ejcn.1600779 10403584 Glycemic index and serum high-density lipoprotein cholesterol concentration among US adults Ford ES Liu S Arch Intern Med 2001 161 572 576 10.1001/archinte.161.4.572 11252117 Relation between a diet with a high glycemic load and plasma concentrations of high-sensitivity C-reactive protein in middle-aged women Liu S Manson JE Buring JE Stampfer MJ Willett WC Ridker PM Am J Clin Nutr 2002 75 492 498 11864854 Frozen desserts and glycaemic response in well-controlled NIDDM patients Bukar J Mezitis NHE Saitas V Pi-Sunyer FX Diabetes Care 1990 13 382 385 2180658 The glycaemic index of a range of gluten-free foods Packer SC Dornhurst A Frost GS Diabet Med 2000 17 657 660 10.1046/j.1464-5491.2000.00356.x 11051285 International table of glycemic index and glycemic load values: 2002 Foster-Powell K Holt SHA Brand-Miller JC Am J Clin Nutr 2002 76 5 56 12081815 Glycemic index of single and mixed meal foods among common Japanese foods Sugiyama M Tang AC Wakaki Y Koyama W Eur J Clin Nutr 2003 57 743 752 10.1038/sj.ejcn.1601606 12792658 Carbohydrates in Human Nutrition Joint FAO/WHO Report FAO Food and Nutrition, Paper #66. Rome: FAO 1998 Glycaemic index methodology Brouns F Bjorck I Frayn KN Gibbs AL Lang V Slama G Wolever TMS Nutr Res Rev 2005 18 145 171 10.1079/NRR2005100 Determination of the glycaemic index of foods: interlaboratory study Wolever TM Vorster HH Bjorck I Brand-Miller J Brighenti F Mann JI Ramdath DD Granfeldt Y Holt S Perry TL Venter C Xiaomei W Eur J Clin Nutr 2003 57 475 482 10.1038/sj.ejcn.1601551 12627186 Physiological validation of the concept of glyemic load in lean young adults Brand-Miller JC Thomas M Swan V Ahmad ZI Petocz P Colagiuri S J Nutr 2003 133 2728 2732 12949357 Glycemic index, postprandial glycemia and cardiovascular disease Dickinson S Brand-Miller J Curr Opin Lipidol 2005 16 69 75 10.1097/00041433-200502000-00012 15650566 Glycemic index and obesity Brand-Miller JC Holt SHA Pawlak DB McMillan J Am J Clin Nutr 2002 76 281S 285S 12081852 Low-glycemic index diets in the management of diabetes: a meta-analysis of randomized controlled trials Brand-Miller J Hayne S Petocz P Colagiuri S Diabetes Care 2003 26 2261 2267 12882846 Low glycaemic index diets for coronary heart disease Kelly S Frost G Whittaker V Summerbell C The Cochrane Database of Systematic Reviews 2004 4 Art. No.: CD004467.pub2. DOI: 10.1002/14651858.CD004467.pub2 15495112 Dietary glycemic index and load, measures of glucose metabolism, and body fat distribution in older adults Sahyoun NR Anderson AL Kanaya AM Koh-Banerjee P Kritchevsky SB de Rekeneire N Tylavsky FA Schwartz AV Lee JS Harris TB Am J Clin Nutr 2005 82 547 552 16155266 Beneficial effect of low-glycemic index diet in overweight NIDDM subjects Wolever TMS Jenkins DJA Vuksan V Jenkins AL Wong GS Joss RG Diabetes Care 1992 15 562 564 1499480 Reduced glycemic index and glycemic load diets do not increase the effects of energy restriction on weight loss and insulin sensitivity in obese men and women Raatz SK Torkelson CJ Redmon JB Reck KP Kwong CA Swanson JE Liu C Thomas W Bantle JP J Nutr 2005 135 2387 2391 16177201 Evaluation of the protein quality of an isolated soy protein in young men: relative nitrogen requirements and effects of methionine supplementation Young VR Puig M Queiroz E Scrimshaw NS Rand WM Am J Clin Nutr 1984 39 16 24 6537870 Gastro-jejunal digestion of soya-bean-milk protein in humans Baglieri A Mahe S Zidi S Huneau JF Thuillier F Marteau P Tome D Br J Nutr 1994 72 519 532 10.1079/BJN19940056 7986784 Nutritional value of [15N]-soy protein isolate assessed from ileal digestibility and postprandial protein utilization in humans Mariotti F Mahe S Benamouzig R Luengo C Dare S Gaudichon C Tome D J Nutr 1999 129 1992 1997 10539774 Postprandial kinetics of dietary amino acids are the main determinant of their metabolism after soy or milk protein ingestions in humans Bos C Metges CC Gaudichon C Petzke KJ Pueyo ME Morens C Everwand J Benamouzig R Tome D J Nutr 2003 133 1308 1315 12730415 Beneficial role of dietary phytoestrogens in obesity and diabetes Bhathena SJ Velasquez MT Am J Clin Nutr 2002 76 1191 1201 12450882 The cardiovascular effects of soy products Hasler CM J Cardiovasc Nurs 2002 16 50 63 12597262 The role of isoflavones in cancer chemoprevention Sarkar FH Li Y Front Biosci 2004 9 2714 2724 15353308 Meta-analysis of the effects of soy protein containing isoflavones on the lipid profile Zhan S Ho SC Amer J Clin Nutr 2005 81 397 408 15699227 Phytoestrogens and lipoproteins in women Merz CNB Johnson BD Braunstein GD Pepine CJ Reis SE Paul-Labrador M Hale G Sharaf BL Bittner V Sopko G Kelsey SF for the Women's Ischemia Syndrome Evaluation Study Group J Clin Endocrinol Metab 2006 91 2209 2213 10.1210/jc.2005-1853 16595605 Skeletal benefits of soy isoflavones: a review of the clinical trial and epidemiologic data Messina M Ho S Alekel DL Curr Opin Clin Nutr Metab Care 2004 7 649 658 10.1097/00075197-200411000-00010 15534433 Isoflavone therapy for menopausal flushes: a systematic review and meta-analysis Howes LG Howes JB Knight DC Maturitas 2006 55 203 211 10.1016/j.maturitas.2006.03.008 16675169 Meta-analysis of soy intake and breast cancer risk Trock BJ Hilakivi-Clarke L Clarke R J Natl Cancer Inst 2006 98 459 471 16595782 Meta-analysis of soy food and risk of prostate cancer in men Yan L Spitznagel EL Int J Cancer 2005 117 667 669 10.1002/ijc.21266 15945102 Impact of incorporating soya fibre and processed soyabean flours on the glycaemic index of parantha Jain M Verma N Asia Pac J Clin Nutr 2004 13 S97 15294614 Glycemic index of popular nutrition and energy bars Wolever TMS Radmard R J Amer Dietetic Assoc 2002 102 A78 10.1016/S0002-8223(02)90428-1 Sydney University Glycemic Index Research Service http://www.glycemicindex.com