Testing the glycaemic index of foods: in vivo, not in vitro

The glycaemic index (GI) of foods is receiving increasing coverage in medical and nutrition literature and food manufacturers are showing interest in determining and publicising the GI values of their products. Although an extensive international table of GI values for foods was recently published (Foster-Powell et al, 2002), the need for continued testing and product development is evident. In particular, the wider implementation of a low-GI diet will depend on the identification of low-GI foods using standardised in vivo methodology (FAO/WHO expert consultation, 1998). Research evidence to date confirms that the GI values of foods cannot be accurately predicted from their ingredients, nutrient composition or preparation methods (Truswell, 1992).

The in vivo method of measuring GI values for foods is relatively time consuming. Consequently, research has been conducted to determine whether in vitro enzymatic assays that estimate the likely rates of carbohydrate digestion and glucose absorption in the small intestine could be used instead of in vivo methodology as an easier and cheaper method of measuring GI values of foods. Some researchers have reported an association between in vitro measurements of carbohydrate digestion and GI values for foods (Jenkins et al, 1987; Ross et al, 1987; Englyst et al, 1999; Araya et al, 2002), and these results have led to assumptions that the laborious task of in vivo testing is no longer necessary for measuring GI values for foods. However, associations between in vivo and in vitro results have not been consistently found (Urooj & Puttaraj, 2000) and other physiological factors (gastrointestinal function, glucose tolerance, the rate of food consumption) and meal factors (physical form, other nutrients) can confound the relationship (Glycemic index and health, 2001). Therefore, at the present time, we would strongly advise against the use of any in vitro method for producing GI values for food labelling purposes.

Recently, we used standardised in vivo GI methodology to measure the GI values of three carbohydrate-rich processed foods (one breakfast cereal and two snack bars) that are sold in supermarkets throughout Australia and labelled as being low-GI foods on the basis of results obtained from an in vitro test (Brand-Miller & Holt, 2001–2002). However, in vivo, the foods yielded medium to high mean±s.e.m. GI values of 75±3, 68±5 and 65±5, using glucose as the reference food with a fixed GI value of 100. Since foods with a GI value of ≤55 are currently considered to have a low GI (Brand-Miller et al, 2003), these foods are clearly not low-GI foods. Hence, the labelling was not only incorrect but also potentially harmful for someone with diabetes.

It is possible that the rate of carbohydrate digestion in a test tube may mimic the in vivo glycaemic response for some foods, such as whole legumes and grains (Trout et al, 1993). The rate of gastric emptying following food consumption is a major determinant of the rate at which food becomes available for digestion in the small intestine (Mourot et al, 1988). Factors that affect stomach emptying include acidity, osmolality, volume and the concentration of sugars. In vitro methods are unlikely to detect the effect of differences in gastric emptying, nor the profound effect of viscosity on rate of absorption of the products of digestion. Even with a near-perfect study correlation across a wide range of foods, there will always be specific foods for which the in vitro measure will misrepresent the true glycaemic effect of the carbohydrate in the food. However, the in vitro method may still be useful for certain applications, such as product development and research on factors that influence the rate of digestion of carbohydrates in foods.

Access to accurate information about the GI values of foods is important for consumers, health professionals and people with diabetes to assist them in making informed dietary choices. For the time being, we urge food manufacturers to undertake GI testing only with experienced laboratories using the standardised in vivo method. If in vitro methods are employed, the findings should never be used to imply that the GI has been tested because by definition the GI is an in vivo measurement of glycaemic response.


  1. Araya H, Contreras P, Alvina M, Vera G & Pak N (2002): A comparison between an in vitro method to determine carbohydrate digestion rate and the glycemic response in young men. Eur. J. Clin. Nutr. 56, 735–739.

    CAS  Article  Google Scholar 

  2. Brand-Miller J & Holt S (2001–2002): Unpublished observations. Human Nutrition Unit, Sydney University, Australia.

  3. Brand-Miller J, Wolever TMS, Colagiuri S & Foster-Powell K (2003): The New Glucose Revolution. New York: Marlowe & Company.

    Google Scholar 

  4. Englyst K, Englyst H, Hudson G, Cole T & Cummings J (1999): Rapidly available glucose in foods: an in vitro measurement that reflects the glycemic response. Am. J. Clin. Nutr. 69, 448–454.

    CAS  Article  Google Scholar 

  5. FAO/WHO Expert Consultation (1998): Carbohydrates in human nutrition. In FAO WHO Food and Nutrition, Vol. 66. Geneva: Food and Agriculture Organisation, World Health Organisation.

  6. Foster-Powell K, Holt SH & Brand-Miller JC (2002): International table of glycemic index and glycemic load values: 2002. Am. J. Clin. Nutr. 76, 5–56.

    CAS  Article  Google Scholar 

  7. Glycemic index and health: the quality of evidence (2001): Danone Vitapole Nutrition and Health Collection. Montrouge, France: Éditions John Libbey Eurotext.

  8. Jenkins DJ, Jenkins AL, Wolever TM, Collier GR, Rao AV & Thompson LU (1987): Starchy foods and fiber: reduced rate of digestion and improved carbohydrate metabolism. Scand. J. Gastroenterol. 129 (Suppl.), 132-141.

    CAS  Article  Google Scholar 

  9. Mourot J, Thouvenot P, Couet C, Antoine JM, Krobicka A & Debry G (1988): Relationship between the rate of gastric emptying and glucose and insulin responses to starchy foods in young healthy adults. Am. J. Clin. Nutr. 48, 1035–1040.

    CAS  Article  Google Scholar 

  10. Ross S, Brand J, Thorburn A & Truswell A (1987): Glycemic index of processed wheat products. Am. J. Clin. Nutr. 46, 631–635.

    CAS  Article  Google Scholar 

  11. Trout DL, Behall KM & Osilesi O (1993): Prediction of glycaemic index for starchy foods. Am. J. Clin. Nutr. 58, 873–878.

    CAS  Article  Google Scholar 

  12. Truswell AS (1992): Glycaemic index of foods. Eur. J. Clin. Nutr. 46 (Suppl. 2), 91S–101S.

    Google Scholar 

  13. Urooj A & Puttaraj S (2002): Glycaemic responses to cereal-based Indian food preparations in patients with non-insulin-dependent diabetes mellitus and normal subjects. Br. J. Nutr. 83, 483–488.

    Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Brand-Miller, J., Holt, S. Testing the glycaemic index of foods: in vivo, not in vitro. Eur J Clin Nutr 58, 700–701 (2004). https://doi.org/10.1038/sj.ejcn.1601856

Download citation

Further reading


Quick links