The Digest

Glycemic Index Flawed: All that Fattens is not Told


The glycemic index (GI) was introduced in the early 1980s as a means of comparing the ability of different foods to elevate blood glucose [1]. Intended primarily for use in the management of blood glucose in individuals with diabetes, the index recognises the fact that foods containing the same amount of available carbohydrate do not necessarily bring about similar changes in blood glucose levels. In recent years, however, the index has found more general application, particularly in studies investigating the links between diet and the development of disease.

In this article, Mark Burkitt questions the method by which GI values are calculated which, he argues, gives a false indication of the glycemic impact of many foods. He also suggests that, in ignoring the impact of foods on blood fructose levels, the index is of limited value as a risk factor for type-2 diabetes and cardiovascular disease.  

Is the glycemic index being overextended?

Foods in which the majority of the carbohydrate is present as starch (a polymer of glucose) take a relatively long time to be hydrolysed by the digestive enzymes, whereas simple sugars, such as the disaccharides sucrose (a glucose-fructose dimer), maltose (glucose-glucose) and lactose (galactose-glucose), have a more immediate impact on blood glucose levels. Being a monosaccharide, glucose itself is absorbed without the requirement for hydrolysis. Other factors contributing to the glycemic index of particular foods include the requirement for the cell walls in plant materials to be mechanically ruptured before carbohydrate can be released (whole fruits, for example, tend to have lower GI values than their juices) and the presence of fat, which can act as a physical barrier to the absorption of glucose [1-3].

Although the GI values published for particular foods often show considerable variation – reflecting, for example, regional differences in manufacturing processes, formulations and cooking methods – there are some well known patterns in the values of our most common foods. Root vegetables, for example, have particularly high GI values yet ice cream, which typically contains added sucrose and glucose, has a relatively low value. Given the extent to which the GI has been embraced by both the wider research and ‘healthy eating’ communities, it is perhaps appropriate to consider whether the index – particularly when calculated as originally intended for the management of blood glucose levels in diabetes – is not being overextended ‘beyond its pay grade’.

Granted, we have seen the introduction of the ‘glycemic load’ (GL), the ‘glycemic profile’ and other derivations of the GI, but these are still calculated in a manner which leads one to question the wisdom of their wider application beyond the management of existing diabetes[2,4,5]. For the reasons given below, these indices can be particularly misleading when used to investigate the links between specific foods and the development of diseases within the metabolic syndrome spectrum, including type-2 diabetes and cardiovascular disease [6-8]. Indeed one can go as far as to question the very basis of how the GI is calculated for use in the management of diabetes.

Should the calculation of GI values be modified?

The fundamental problem here lies in the calculation of the portion size of the food being tested for its GI value. This is usually calculated so as to contain 50.00 g of available carbohydrate. Cellulose (a component of fibre) is ignored, but starch and the simple sugars (monosaccharides and disaccharides) are included. In principle, the glycogen in meat should also be included, but in practice meats have such a low impact on blood glucose levels that their GI values are not routinely reported.

The only parameter measured during the GI determination for a particular food is the concentration of glucose in the blood following ingestion (over a two-hour window). Therefore is it a ‘fair test’ to be comparing, for example, foods in which all of the 50.00 g of available carbohydrate is present as glucose, maltose or starch – each of which can undergo 100 % conversion to blood glucose – with those containing, say, 50.00 g of sucrose which, being half fructose by mass, can generate a maximum of only 26.32 g of glucose [9]? In the latter case, the ‘incremental area-under-the-curve’, obtained when blood glucose levels are plotted against time, can never be barely more than half the value of the reference curve, plotted using blood glucose levels following the ingestion of 50.00 g of pure glucose. This means that foods with a high sucrose content – which includes many of those believed to be responsible for the current, unacceptably high occurrence of obesity and metabolic syndrome disorders – are getting off very lightly indeed in the GI test.

High-fructose corn syrup, for example, in which a proportion of the glucose generated by the hydrolysis of corn starch is converted to fructose, must be a godsend for those marketing foods sweetened with this product. Not only is fructose ‘silent’ in the GI test, there is widespread agreement that, due to its ease of conversion to fat in the liver, the monosaccharide is potentially more harmful to health than glucose. This brings us back to the original point that the GI is of real value only to individuals with diabetes, for whom the control of blood glucose is the overriding concern. (Even here, though, the above criticism over how the index is calculated still holds.) Individuals without diabetes who base their food choices on the GI index in the belief that this is conducive to good health are being misled. Even the authors of the popular book ‘The Fast Diet’ extol the virtues of ice cream on the basis on its low GI value, which they quote as 37 (adding, ‘You would bet your house on ice cream being high GI/GL, but not so’)[10]. How many nutritionists would recommend ice cream over root vegetables? (Not many, one would hope.)

How might the glycemic index be improved upon?

There are at least two possible solutions to these shortcomings. One would be to calculate the portion size of the food being tested so as to contain 50 g of glucose ‘equivalents’. Thus glucose itself would count as one glucose equivalent because 1 g provides 1 g of the sugar. Each gram of starch and maltose would also each count as one glucose equivalent because they yield 1 g of glucose upon digestion (ignoring the small contribution of the hydrolysing water to the molecular weight of the released sugar [9]). Each gram of sucrose however would count as 0.5 equivalents because it contains only 0.5 g glucose. Significantly, fructose would count as zero glucose equivalents [11]. Such an approach certainly rebalances the scales. Ice cream, for example, has an estimated GI value of between 63 and 88 when calculated using glucose equivalents [12].

There have been other attempts to improve upon the GI. Monro and Shaw, for example, have introduced the term ‘glycemic glucose equivalence’. This concept holds that the glycemic impact of a food should be stated as the weight of glucose that brings about the same increase in blood glucose. Thus the glycemic glucose equivalent of an apple is 7.9 in that it brings about the same increase in blood glucose as the ingestion of 7.9 g of glucose [4]. Although such modifications to the calculation of the GI provide a more realistic reflection of the ability of foods to increase blood glucose levels, they still fail to address the fact that the GI tells us nothing about how a given food affects blood fructose levels. In studies where the index is being used to assess the wider links between diet and metabolic syndrome diseases – including the recent analysis of data from three prospective cohort studies [7] – it would be more meaningful to either combine the blood levels of both glucose and fructose into a single index or, better still, devise a separate index to reflect the impact of foods on blood fructose levels. Until this is done, we are in danger of being misled by the glycemic index: epidemiologists, health gurus and the general public take note.

References and notes

  1. Jenkins DJA, Wolever TMS, Taylor RH, Barker H, Fielden H, Baldwin JM et al. Glycemic index of foods: a physiological basis for carbohydrate exchange. Am J Clin Nutr 1981; 34: 362-6

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

  3. University of Sydney. Online glycemic index database.

  4. Monro JA, Shaw M. Glycemic impact, glycemic glucose equivalents, glycemic index, and glycemic load: definitions, distinctions and implications. Am J Clin Nutr 2008; 87 (suppl): 237S-43S

  5. Chulp M, Peterson, K, Zapletalová J, Kudlová P, Sečkař P. Extended pranial glycemic profiles of foods as assessed using continuous monitoring enhance the power of the 120-minute glycemic index. J Diabetes Sci Technol 2010; 4:615-24

  6. Kristo AS, Matthan NR, Lichenstein, AH. Effect of diets differing in glycemic index and glycemic load on cardiovascular risk factors: review of randomized controlled-feeding trials. Nutrients 2013; 5:1071-80

  7. Muraki I, Imamura F, Manson JE, Hu FB, Willett, WC, van Dam RM, Sun Q. Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies. BMJ 2013; 347:f5001 doi: 10.1136/bmj.f5001

  8. Schwingshackl L, Hoffmann G. Long-term effects of low glycemic index/load vs. high glycemic index/load diets on parameters of obesity and obsesity-associated risks: a systematic review and meta-analysis. Nutr Metab Cardiovasc Dis 2013; 23: 699-706

  9. The hydrolysis of 50.00 g of sucrose generates 26.32 g of glucose and 26.32 g of fructose. (The increase in mass reflects the water added ‘across’ the 1,2‑glycosidic bonds in sucrose during hydrolysis.)

  10. Mosley M, Spencer M. The Fast Diet. The simple secret of intermittent tasting: lose weight, stay healthy, live longer. Short Books, 2013

  11. Although, in theory, fructose can be converted to glucose through phosphorylation and entry into the gluconeogenesis pathway, this is unlikely to be of significance in the fed state, particularly when accompanied by dietary glucose (e.g. when fructose is derived from sucrose) which, through the actions of insulin, would ensure that any glucose 6-phosphate derived from fructose is directed towards glycogen synthesis.

  12. Burkitt MJ. Healthy Eating Through Informed Choice. Westcott Research and Consulting, 2013. e-Book [ASIN: B00DFMA668]. Paperback to be published on 28 July 2014 (Matador)(ISBN-10: 178306479X; ISBN-13: 978-1783064793).


  1. It is argued that the glycemic index can be misleading when applied to foods containing fructose, either as the monosaccharide or as a component of the disaccharide sucrose (table sugar). This is because fructose is counted as ‘available carbohydrate’ when determining the portion size of the food being tested – even though the sugar has no impact on blood glucose levels.

  2. In order to make appropriate comparisons between foods containing fructose with those containing glucose (either as the monosaccharide or in the form of maltose or starch), it is suggested that the portion size of the food being tested should be modified so as to reflect its content of carbohydrate only which can contribute directly to blood glucose levels.

  3. It is suggested that the glycemic index is of only limited value as a risk factor for type-2 diabetes and cardiovascular disease. This is because the index gives no indication of the capacity of foods to elevate blood fructose which, due in part to its ease of conversion to fat in the liver, can be regarded to present a greater risk to health than glucose. This omission of fructose may explain the inconsistent findings of studies in which the GI has been evaluated as a risk factor for cardiovascular disease and type-2 diabetes.


The assertion of a link between particular dietary components and metabolic syndrome disorders occurs regularly in the media. This is often preceded by the publication of a scientific study, which is then disseminated to the wider public.

Several studies have employed the glycemic index as a risk factor for metabolic diseases. The index has also found widespread popularity as a means of selecting foods by those wishing to counteract obesity and metabolic diseases. 

I am concerned that the GI index is being used with little thought given as to what it is actually measuring. Not only does the index distort the impact of foods containing sucrose on blood glucose, it ignores the important role of fructose in the development of metabolic diseases.

Mark Burkitt

Westcott Research and Consulting [Home Page]

Article published 22 April 2014

For more articles, go to: The Digest

If you would like to learn more about the topics discussed in this article, it is recommended you read Dr Burkitt’s book, Healthy Eating Through Informed Choice, where you will find, for example, descriptions of the various types of fats and oils (saturates, monounsaturates, polyunsaturates, trans fats, omega-3 oils etc) and explanations of how they affect health.

Whilst the book is written in non-technical language and is intended primarily for readers with absolutely no background in science, it is hoped that trained scientists and health professionals will also find the material to be of interest – the book is extensive in its scope and challenges some of the conventional views on the role of nutrition in human disease.