Journal of Nutrition OpenSOurce Diets- www.ResearchDiets.com

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Venn, B. J.
Right arrow Articles by Green, T. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Venn, B. J.
Right arrow Articles by Green, T. J.
© 2006 American Society for Nutrition J. Nutr. 136:1377-1381, May 2006

Methodology and Mathematical Modeling

The Glycemic Load Estimated from the Glycemic Index Does Not Differ Greatly from That Measured Using a Standard Curve in Healthy Volunteers1

Bernard J. Venn*,2, Alison J. Wallace{dagger}, John A. Monro{dagger}, Tracy Perry*, Rachel Brown*, Chris Frampton** and Tim J. Green*

* Department of Human Nutrition, University of Otago, Dunedin, New Zealand; {dagger} New Zealand Institute for Crop and Food Research Limited, Christchurch and Palmerston North, New Zealand; and ** Christchurch School of Medicine, University of Otago, Christchurch, New Zealand

2 To whom correspondence should be addressed. Email: bernard.venn{at}stonebow.otago.ac.nz.


   ABSTRACT
TOP
 ABSTRACT
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 LITERATURE CITED
 
Glycemic load (GL) is calculated indirectly as glycemic index (GI) times the weight of available carbohydrate. Alternatively, GL may be measured directly using a standard glucose curve. The purpose of this study was to test the agreement between GL values obtained using direct and indirect methods of measurement in 20 healthy volunteers. A standard curve in which glucose dose was plotted against blood glucose incremental area under the curve (iAUC) was generated using beverages containing 0, 12.5, 25, 50, and 75 g glucose. The GI and available carbohydrate content of 5 foods were measured. The foods (white bread, fruit bread, granola bar, instant potato, and chickpeas) were consumed in 3 portion sizes, yielding 15 food/portion size combinations. GL was determined directly by relating the iAUC of a test food to the glucose standard curve. For 12 of 15 food/portion size combinations, GL determined using GI x available carbohydrate did not differ from GL measured from the standard curve (P > 0.05). For 3 of the test products (100 g white bread, and 100- and 150-g granola bars), GI x available carbohydrate was higher than the direct measure. Benefits of the direct measure are that the method does not require testing for available carbohydrate and it allows portion sizes to be tested. For practical purposes, GI x available carbohydrate provided a good estimate of GL, at least under circumstances in which available carbohydrate was measured, and GI and GL were tested in the same group of people.

KEY WORDS: • glycemic load • glycemic index • available carbohydrate • standard curve

Blood glucose response to a food is related to the glycemic load (GL)3 of the food consumed. GL is usually calculated as the glycemic index (GI) multiplied by the weight of the available carbohydrate in the food. In epidemiologic studies, GL was applied to diets to examine the relation between the type and amount of carbohydrate and health. Although not definitive, the findings of a number of investigators suggest that consuming a high GL diet is an independent risk factor for type 2 diabetes (1–3), coronary heart disease (4), and certain types of cancer (5–11). Further, GL is used on food labels and in GI/GL reference tables (12). In these applications, the GL of a food or a diet has not been measured directly and despite its widespread use, the method of determining GL indirectly from GI has not been fully validated. Some work was carried out comparing the glycemic responses among foods that were served at portion sizes calculated to have the same GL (13,14). The glycemic responses to the foods were not always equivalent. One explanation for this discrepancy may be that GI is calculated using an amount of food containing 50 g of available carbohydrate. When the amount of food consumed is either larger or smaller, then linearity is assumed between the amount of food and the glycemic response. If postprandial glycemia does not respond linearly with varying amounts of food, it suggests that testing the GI of a food to calculate GL will not always provide a reliable estimate of GL.

An alternative way to determine GL is to construct a standard glucose reference curve and impute GL values for a test food (Fig. 1). Note that GL was used specifically to describe GI x available carbohydrate, but the concept applies to the use of a standard curve because each food unit of GL has a glycemic effect equivalent to 1 g of the reference food (15). In this paper, the term GL is used to describe the numerical values derived from either GI x available carbohydrate or to values imputed from a standard curve. It is also important to validate GL across varying amounts of food consumed because in practice, GL is applied to people's diets. The aims of this study were to test how GL determined indirectly using GI x available carbohydrate compared with GL measured directly from a standard glucose curve and to extend the validation of GL to include foods with varying portion sizes.


Figure 1
View larger version (9K):
[in this window]
[in a new window]
  		FIGURE 1  A participant's standard glucose curve. Values are means and SE, n = 3 tests at glucose doses of 12.5, 25, and 75 g, and 4 tests at 50 g. The dashed lines demonstrate how GL was imputed from the standard curve.

 

   SUBJECTS AND METHODS
TOP
 ABSTRACT
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 LITERATURE CITED
 
Volunteers (n = 20; 11 women and 9 men) were recruited. The Human Ethics Committee of the University of Otago approved the study and all participants gave informed consent. None of the participants had impaired glucose tolerance according to the WHO classification (fasting glucose <7 mmol/L, 2-h blood glucose concentration after a 50 g glucose load <7.8 mmol/L). The mean (SD) age and BMI of the group was 23.3 (3.5) y and 23.4 (3.3) kg/m2, respectively. Each person completed a series of 33 tests over a 4-mo period between September and December 2004. The tests were carried out following a predetermined balanced randomization of products whose sequence differed for each person.

Participants were instructed to consume an evening meal containing a large proportion of carbohydrate-rich food the night before each test. It was suggested that the meal be based on rice, bread, or potato. Participants were asked not to drink alcohol the evening before the test. They were also instructed not to consume food or beverages (other than water) after 2200 h the night before the tests. On the morning of the test, the participants were asked to refrain from physical exercise and to report to the clinic in a fasting state. Capillary blood samples were taken using a lancet. A drop of blood was collected into a HemoCue cuvette, and blood glucose concentration measured using a Hemocue Glucose 201 Analyzer. Each morning the instrument was checked using its own internal system to confirm that it was functioning correctly.

The mean of 2 fasting blood glucose concentrations determined 5 min apart was used as a baseline measure. The time of the second fasting sample was recorded as the start time of the test. Participants were given a test product in accordance with the randomization sequence. The product was consumed at an even pace over a period of 15 min and capillary blood samples were taken at 15, 30, 45, 60, 90, and 120 min after the start time. If the blood glucose concentration had not returned to within 0.2 mmol/L of the baseline concentration after 2 h, further blood samples were taken at 150 and 180 min after the start time. Participants were asked to remain seated for the duration of the tests. Blood glucose response was measured over 2–3 h after consumption of a reference or a test food. The incremental area under the curve (iAUC) was calculated using the method described by the FAO/WHO (16). The reference food was comprised of beverages containing 0, 12.5, 25, 50, and 75 g anhydrous glucose. A "standard" glucose iAUC dose-response curve was generated for each person, with tests at glucose doses of 0, 12.5, 25, and 75 g carried out in triplicate and the 50-g glucose test replicated on 4 occasions.

The test foods were commercially available white bread, fruit bread (13% dried fruit), granola bar, instant mashed potatoes, and chickpeas. The 5 foods were each tested at 3 serving sizes corresponding to a common standard measure, 100 g of the food, and an amount containing 50 g available carbohydrate except for the chickpeas for which an amount containing 25 g available carbohydrate was used. The white bread serving size, corresponding to 50 g available carbohydrate, was tested in triplicate. The GI of the food for each individual was calculated by dividing the iAUC of a 50-g carbohydrate portion by the mean iAUC of the four 50-g glucose beverages and multiplying by 100, except for chickpeas for which the iAUC of a 25-g carbohydrate portion was divided by the iAUC of the mean of the three 25-g glucose beverages.

Available carbohydrate for each of the 5 foods was determined by measuring the total starch and total sugar content of the food. Total starch was determined according to AOAC International (17). Total sugars were determined as the sum of sucrose, lactose, maltose, glucose, and fructose, each measured by a GLC method.

    Data analysis. GL was calculated using the following 2 methods: 1) multiplying the GI of each food by the proportion of available carbohydrate in the portion consumed; 2) equating the iAUC for each food, at each serving size, against each participant's standard glucose response curve, referred to herein as the direct measure. Linear, log-linear, and log-log relations between glucose doses and blood glucose responses were tested. The log-log model consistently produced higher r2 values. Thus, data used to generate the glucose standard curve and the test food iAUC responses were log transformed. For clarity, the data shown in the figures were not log transformed. Comparisons between GL derived from GI x available carbohydrate and the direct measure were made using paired t tests. Values in the text are means ± SD. Differences were considered significant at P < 0.05.


   RESULTS
TOP
 ABSTRACT
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 LITERATURE CITED
 
The 20 participants completed all 33 tests. A list of the test foods, serving sizes, available carbohydrate contents, and GI are shown in Table 1. The GI was calculated using the food serving that provided 50 g of available carbohydrate, except chickpeas for which a serving containing 25 g available carbohydrate was used. Two additional GI were determined for granola bars using food servings that contained ~25 and 75 g available carbohydrate using glucose as the reference food at doses of 25 and 75 g, respectively. The granola bar GI values calculated on the basis of consuming 1 (56 ± 21), 2 (53 ± 22), or 3 (52 ± 19) bars did not differ (P > 0.05). For drinks containing ≤ 50 g of glucose and for most of the test foods, the postprandial rise in blood glucose concentrations returned to baseline within 2 h. For the 118-g white bread serving, it took up to 3 h for blood glucose concentrations to return to baseline. Truncation of the iAUC at 2 h resulted in a GI for white bread of 79 ± 20, compared with 84 ± 21 if the 3-h data were used.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Test foods, serving amounts, available carbohydrate contents, and GI values

 
We determined the GL values for the various foods and serving sizes according to GI x available carbohydrate and by the direct measure (Table 2). The 2 methods gave values that did not differ (P > 0.05) in the majority of cases. For most foods, the absolute difference in GL between the methods was <3 GL units (g), with a median proportional difference of 10%. Using the direct measure, increasing the portion size of a food tended to result in an increase in GL proportional to the portion. Granola bars comprised a notable exception: tripling the intake from 1 to 3 bars produced a 120% increase in GL.


View this table:
[in this window]
[in a new window]
  		TABLE 2 Comparison of GL determined by GI x available carbohydrate and direct measure

 
Each food was tested at 3 different portion sizes making it possible to fit iAUC dose-response curves to each food (for an example, see Fig. 2). Using equations developed from the iAUC dose-response curves, the GL was calculated for white bread, fruit bread, granola bar, and potato in increments of 1 g up to 50 g of available carbohydrate, and up to 25 g available carbohydrate for chickpeas, and plotted against GI x available carbohydrate (Fig. 3). The resulting plots are approximately linear, suggesting a direct relation between GI x available carbohydrate and the direct measure. However, at higher intakes of granola for instance, the curve tends to flatten off, exaggerating the overestimation of GI x available carbohydrate.


Figure 2
View larger version (10K):
[in this window]
[in a new window]
  		FIGURE 2  Mean standard glucose curve and mean blood glucose dose-response curve to 1, 2, and 3 granola bars represented by 25, 50, and 75 g available carbohydrate, respectively. Values are means and SE, n = 20.

 

Figure 3
View larger version (13K):
[in this window]
[in a new window]
  		FIGURE 3  Comparison of mean GL (dashed line, n = 20) obtained using the direct measure and calculated from GI x available carbohydrate in increments of 1 g up to 50 g for white bread (A), fruit bread (B), potato (C), granola bar (D), and up to 25 g for chickpeas (E). The solid straight line is the line of equivalence.

 

   DISCUSSION
TOP
 ABSTRACT
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
LITERATURE CITED
 
The most widely used means of determining GL is by multiplying a food's GI by the amount of available carbohydrate in the food consumed. This study tested whether this indirect method provided a good estimate of GL by comparing the results to that food's GL measured directly from a glucose standard curve. Our findings support the use of the indirect method. However, GI x available carbohydrate tended to overestimate GL, an effect that may be more pronounced as the portion size of some foods increases.

A tendency for GI x available carbohydrate to overestimate GL is a novel finding given that Brouns et al. (18) reasoned that GI values may be underestimated. The rationale was that some carbohydrate found to be digestible by in vitro analysis may not be available in vivo, leading to an overestimation of available carbohydrate, a smaller portion of food being used for the GI test, and hence an underestimation of GI. However, the calculation of GL using GI x available carbohydrate may tend to compensate for the imprecision in the estimate of GI because an "underestimated" GI would be multiplied by an "overestimated" weight of available carbohydrate. This suggests that GL calculated from GI x available carbohydrate is a relatively robust measure that accommodates some discrepancy between in vitro release and in vivo release of available carbohydrate.

The supposition that different foods with the same GL produce similar blood glucose responses was tested previously (13,14). In both studies, there was good agreement among some, but not all of the foods tested. In a study by Liu et al. (13), the postprandial blood glucose iAUC after rice consumption was approximately half the iAUC obtained with other foods estimated to have the same GL. The authors suspected that the published GI value for rice used in the calculation of portion size was higher than the GI of the rice used in their study, leading to the consumption of a smaller portion size. Brand-Miller et al. (14) found that lentils induced a far smaller glycemic response than other foods served at an equivalent GL. Possible reasons given by the authors were that the starch in legumes might be so slowly absorbed and digested that GL is independent of portion size, the large volume of the meal delayed gastric emptying, or enzyme inhibitors interfered with digestion and absorption. This implies that GL estimated from GI x available carbohydrate may not be applicable to all foods in all portion sizes. However, the chickpeas used in our study gave the expected dose response based on the direct measure, with a doubling in portion size resulting in a doubling of GL (Table 2). The largest portion we used was 136 g chickpeas, similar in weight to the 138 g of lentils used in the study by Brand-Miller et al. (14). Thus, our data show that GL is responsive to the portion size of legumes, at least for chickpeas in amounts of up to 136 g.

There is debate concerning the best way to measure the glycemic potency of foods; some investigators argue for a whole-food approach rather than a GI approach (19). There are scientific and practical merits and limitations to both approaches. A benefit of using GI x available carbohydrate is its practicality. The GI of the food is tested at 1 serving size corresponding to 50 g available carbohydrate, and the result is used to calculate GL at any serving size of the food. This approach assumes that GL changes linearly with varying portion size because GI is assumed to be a food constant, i.e., the ratio of foodiAUC relative to glucoseiAUC is fixed. In our study, the similar values obtained for GI using 1, 2, or 3 granola bars support this contention. The assumption that GL in turn increases linearly with portion size is generally supported by the linear relation shown between GI x available carbohydrate and GL measured from a glucose standard curve (Fig. 3).

One area of potential imprecision in the determination of GI has been the measurement of available carbohydrate. Several major food databases report total carbohydrate values obtained by difference. The preferred approach is by direct analysis of the starch and sugar components of foods (20). Foster-Powell et al. (12) stressed that accurate laboratory measurements of the available carbohydrate content of foods is an essential preliminary step in GI testing, and that failure to measure accurately may be one of the reasons for variation in GI among apparently similar foods, even among foods listed in the International Tables of GI. Nevertheless, the use of GI x available carbohydrate to estimate GL is popular because a large number of GI values have been published, making the determination of GL possible over a wide range of foods. In the absence of a direct measure, investigators often rely on published GI data to make their assessment of dietary GL (1,2). Among the criticisms of the GI approach is the fact that food is eaten along a continuum of intakes, not simply in amounts containing 50 g of available carbohydrate (21). An advantage of using a direct measure of GL is that food may be tested in amounts corresponding to the manufacturers' portion sizes or to amounts typical of foods consumed. Further, use of a direct measure to determine GL does not require analysis of the food's available carbohydrate content because the in vivo glycemic response to food is simply equated to that of the glucose standard curve. The use of a direct measure thus eliminates issues arising from the measurement of a food's carbohydrate content as a potential source of error.

Although the results of this study show that GI x available carbohydrate may provide a reasonable estimate of a food's GL, at least under circumstances in which the available carbohydrate content of a food has been accurately determined, and the GI and GL determinations were made in the same subjects, still more precision may be gained from a direct measure. A practical limitation of the direct measure is that generating a glucose standard curve requires more tests to be performed, and foods may also have to be tested at various portion sizes, thereby increasing the cost of testing.


   FOOTNOTES
 
1 Jointly funded by the New Zealand Institute for Crop and Food Research Limited and the University of Otago. Back

3 Abbreviations used: iAUC, incremental area under the curve; GI, glycemic index; GL, glycemic load. Back

Manuscript received 15 November 2005. Initial review completed 4 January 2006. Revision accepted 1 March 2006.


   LITERATURE CITED
TOP
 ABSTRACT
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 LITERATURE CITED
 

1. Salmeron J, Ascherio A, Rimm EB, Colditz GA, Spiegelman D, Jenkins DJ, Stampfer MJ, Wing AL, Willett WC. Dietary fiber, glycemic load, and risk of NIDDM in men. Diabetes Care. 1997;20:545–50.[Abstract]

2. Salmeron J, Manson JE, Stampfer MJ, Colditz GA, Wing AL, Willett WC. Dietary fiber, glycemic load, and risk of non-insulin-dependent diabetes mellitus in women. JAMA. 1997;277:472–7.[Abstract]

3. Schulze MB, Liu S, Rimm EB, Manson JE, Willett WC, Hu FB. Glycemic index, glycemic load, and dietary fiber intake and incidence of type 2 diabetes in younger and middle-aged women. Am J Clin Nutr. 2004;80:348–56.[Abstract/Free Full Text]

4. Liu S, Willett WC, Stampfer MJ, Hu FB, Franz M, Sampson L, Hennekens CH, Manson JE. A prospective study of dietary glycemic load, carbohydrate intake, and risk of coronary heart disease in US women. Am J Clin Nutr. 2000;71:1455–61.[Abstract/Free Full Text]

5. Franceschi S, Dal Maso L, Augustin L, Negri E, Parpinel M, Boyle P, Jenkins DJ, La Vecchia C. Dietary glycemic load and colorectal cancer risk. Ann Oncol. 2001;12:173–8.[Abstract/Free Full Text]

6. Augustin LS, Polesel J, Bosetti C, Kendall CW, La Vecchia C, Parpinel M, Conti E, Montella M, Franceschi S, et al. Dietary glycemic index, glycemic load and ovarian cancer risk: a case-control study in Italy. Ann Oncol. 2003;14:78–84.[Abstract/Free Full Text]

7. Augustin LS, Gallus S, Bosetti C, Levi F, Negri E, Franceschi S, Dal Maso L, Jenkins DJ, Kendall DJ, et al. Glycemic index and glycemic load in endometrial cancer. Int J Cancer. 2003;105:404–7.[Medline]

8. Augustin LS, Galeone C, Dal Maso L, Pelucchi C, Ramazzotti V, Jenkins DJ, Montella M, Talamini R, Negri E, et al. Glycemic index, glycemic load and risk of prostate cancer. Int J Cancer. 2004;112:446–50.[Medline]

9. Augustin LS, Gallus S, Negri E, La Vecchia C. Glycemic index, glycemic load and risk of gastric cancer. Ann Oncol. 2004;15:581–4.[Abstract/Free Full Text]

10. Higginbotham S, Zhang ZF, Lee IM, Cook NR, Giovannucci E, Buring JE, Liu S. Dietary glycemic load and risk of colorectal cancer in the Women's Health Study. J Natl Cancer Inst. 2004;96:229–33.[Abstract/Free Full Text]

11. Lajous M, Willett W, Lazcano-Ponce E, Sanchez-Zamorano LM, Hernandez-Avila M, Romieu I. Glycemic load, glycemic index, and the risk of breast cancer among Mexican women. Cancer Causes Control. 2005;16:1165–9.[Medline]

12. 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/Free Full Text]

13. Liu P, Perry T, Monro JA. Glycaemic glucose equivalent: validation as a predictor of the relative glycaemic effect of foods. Eur J Clin Nutr. 2003;57:1141–9.[Medline]

14. Brand-Miller JC, Thomas M, Swan V, Ahmad ZI, Petocz P, Colagiuri S. Physiological validation of the concept of glycemic load in lean young adults. J Nutr. 2003;133:2728–32.[Abstract/Free Full Text]

15. Liu S, Manson JE, Stampfer MJ, Holmes MD, Hu FB, Hankinson SE, Willett WC. 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;73:560–6.[Abstract/Free Full Text]

16. FAO/WHO. Carbohydrates in human nutrition. FAO food and nutrition paper no. 66. 1998; Rome:FAO.

17. Association of Official Analytical Chemists International. Method 996.11. Starch (total) in cereal products, amyloglucosidase-{alpha}-amylase method (total starch assay). 15th edition. Official methods of analysis. 1998; AOAC.

18. Brouns F, Bjorck I, Frayn KN, Gibbs AL, Lang V, Slama G, Wolever TMS. Glycaemic index methodology. Nutr Res Rev. 2005;18:145–71.

19. Monro J. Redefining the glycemic index for dietary management of postprandial glycemia. J Nutr. 2003;133:4256–8.[Abstract/Free Full Text]

20. Gibson RS. Principles of nutritional assessment. 2nd edition. Oxford University Press; 2005.

21. Pi-Sunyer FX. Glycemic index and disease. Am J Clin Nutr. 2002;76:290S–8.[Abstract/Free Full Text]




This article has been cited by other articles:


Home page
 Am. J. Clin. Nutr.Home page
J. A Monro and M. Shaw
Glycemic impact, glycemic glucose equivalents, glycemic index, and glycemic load: definitions, distinctions, and implications
Am. J. Clinical Nutrition, January 1, 2008; 87(1): 237S - 243S.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Venn, B. J.
Right arrow Articles by Green, T. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Venn, B. J.
Right arrow Articles by Green, T. J.

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]