HEALTH PROFESSIONALS

The Glycemic Index:
Clinical and Public Health Significance

Written by G. Harvey Anderson, PhD, Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, with support from the Nutrition Professionals of the Canadian Sugar Institute.

Carbohydrate is the most important source of energy in the majority of human diets and has a wide range of physical, chemical, and physiological characteristics. Historically, carbohydrate has been defined based on its chemical properties, but more recently a physiological classification of carbohydrate has been developed.

The Glycemic Index (GI) is intended to quantitatively rank carbohydrate based on its impact on blood glucose levels (1). It is a measurement of blood glucose response (over two hours) after ingesting a test food that typically contains 50 grams of glycemic carbohydrate. This value is then expressed as a percentage of the response to 50 grams of glycemic carbohydrate from either white bread or glucose when consumed by the same subject.

Generally, rapidly digested carbohydrate has the highest response while carbohydrate that is more slowly digested produces a much smaller area under the 2 hour blood glucose curve (AUC) (Figure 1). The goal is to characterize carbohydrates in foods by quantifying this response.

Glycemic Load (GL) describes the effect of both the amount of carbohydrate consumed and the GI value of a carbohydrate on blood glucose response. The glycemic load is calculated by multiplying the amount of carbohydrate in a serving of the food by the GI. For instance, watermelon has a high GI, but because it contains very little carbohydrate, the effect of a serving on blood glucose is low (2). Table 1defines GI and GL and provides an example of how the numbers are derived.

Since 1981, the GI of many carbohydrate-containing foods has been calculated and these values are listed to rank foods based on this clinical test (3). In general, foods made from grain flour such as snack foods, breads and ready to eat cereals have a high GI. Whole grains, pasta and fruits tend to have a moderate GI, whereas legumes and dairy products have the smallest glycemic response.

Although GI values are widely available and used in evaluating food choices, there are conflicting views on the reliability and application of this measure in clinical and public health settings. Proponents of the GI system consider it to be an important strategy to reduce fluctuations in blood glucose and insulin, improve glucose and lipid metabolism in diabetes, lower blood triglycerides in those who have elevated levels, and aid in body weight management and athletic performance. The GI has also been popularized in a variety of diet books and GI symbols have emerged to promote “healthy eating” in countries such as Australia, Sweden, and South Africa.

Those who are skeptical of the GI and GL have questioned its usefulness for reasons including:

1. the failure of the index to consider the insulin response (4);
2. the high intra- and inter-subject variation in the glucose response to a food (5,6);
3. the lack of discrimination between foods found as part of a mixed meal (7); and
4. the varied methodologies used by different laboratories (2).

Blood Glucose versus Insulin Response

Although a linear relationship is typically observed between insulin and blood glucose in foods that are primarily carbohydrate, some foods do not display this same association (4). Milk products are a good example, because even though lactose is a low glycemic sugar, milk proteins stimulate insulin release (8).

Intra- and Inter-Individual Variability

The high variability in GI values has been attributed to differences in the chemical and physical properties of the food (7,8), the biological variation between individuals (5,8), day-to-day variability within individuals (6), and the methodology used when testing foods (9). The method by which food is prepared (cooking time, method of cooking, and temperature of food) may result in a different GI value for a similar food (10). The physical state of foods (e.g., raw or cooked, whole or ground) has a large influence on the GI rating, particularly in starchy foods. For instance, the GI of freshly cooked potatoes increases when they are mashed versus whole, but if allowed to cool prior to eating they have a low GI due to the formation of resistant starch (3). Similarly, the geographic region where the food is consumed can impact the glycemic response because of differences in regional varieties and cooking methods - rice is a good example (3). Individual variability in the rate of eating and the extent to which food is chewed affects digestion and absorption and influences the GI (11,12). These factors vary both between individuals and within the same individual and may contribute to the lack of reproducibility in GI values as observed in the same individual on different days (6).

Single Foods versus Mixed Meals

When a food is eaten as part of a mixed meal, the GI of the meal may not be accurately predicted (7). One technique that has been used to evaluate a mixed meal is to rank the meals by the GI of the food that provides the most carbohydrate. Alternatively, the GIs of individual components of a meal have been summed to predict the overall GI of the meal. However, combining foods of varying composition has different effects on the overall glycemic response. For instance, adding protein and fat to a carbohydrate containing meal reduces the glycemic response (13,14).

Variability in Estimates

The GI values from different laboratories can vary for the same food. Figure 2 illustrates the variability of GI reported for certain foods due to different analytical methods (2) and to other sources of variation as described above (15) (Figure 2).

FIGURE 2. Ranges of GI values for a given food as part of an interlaboratory study by Wolever et al. (2003). With the exception of the bread, foods were purchased centrally and distributed to participating laboratories. 8-12 subjects were studied in 7 laboratories around the world using a method recommended by the FAO/WHO (15).

Clearly the debate on the value of the GI concept in a clinical and public health setting will and must continue. The wide variation in individual responses to the same carbohydrate food as well as the poor reproducibility of the GI of a given food makes it extremely difficult to provide reliable advice to guide food choices in the clinical or public health context (6). Although several countries currently allow voluntary labelling of a food product with its GI, the unresolved issues surrounding the measurement, the lack of reproducibility, and the benefits of the GI in dietary guidance suggests that it is premature to support the labelling of foods with the GI.

Unfortunately, while scientific consensus and a regulatory framework for allowing the labelling of GI on foods in Canada is not present, industry and lobbying pressures are being applied to achieve political action. Thus it is time to have an open and critical review of the state of the science and the application of the GI in Canada.

While the benefits of GI in general dietary guidance may be uncertain, dietitians can be confident that promoting foods and eating patterns that provide lower blood glucose response is not likely to cause adverse effects. Fortunately, most individuals appear to naturally consume a moderate glycemic diet by consuming foods that range from low to high GI (22,23). However, it is important to note that a food which elicits a low glucose response is not necessarily a healthier choice because it may be high in fat or low in essential nutrients. For the present time, directing consumers to the Nutrition Facts table on foods, providing education on interpreting the nutritional information, and recommending an eating pattern based on Canada’s Food Guide should remain the focus of nutritional guidance.

References

Carbohydrate Terminology

Written by the Nutrition Professionals of the Canadian Sugar Institute with support from Alison Stephen, PhD, Population Nutrition Research, MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, UK

Dietary carbohydrate has a wide range of chemical, physical, and physiological properties. Carbohydrate is classified according to chemical composition, but these groupings are not always helpful when describing physiological and nutritional functions. As a result, a number of terms have emerged to group carbohydrate based on physiological properties and to help focus on specific health benefits. A recent review by Cummings and Stephen (2007) addresses the challenges in reconciling the various chemical and physiological terms and proposes some terms to be more useful than others.

The three major chemical classifications of carbohydrate comprise sugars (mono- and disaccharides), oligosaccharides (composed of 3-9 monosaccharides), and polysaccharides (composed of >9 monosaccharides). This classification system is fundamental for the measurement of carbohydrate and for the assessment of dietary intake; however it is not always useful in describing nutritional benefits given the overlapping physiological effects of major chemical groupings. As a result, a number of terms or categories have evolved that attempt to group carbohydrate by physiological effect or health benefit, such as glycemic carbohydrate, dietary fibre, prebiotics and resistant starch (see definitions below).

Grouping carbohydrate by physiological properties or nutritional characteristics is more difficult than classifying by chemical composition because the physiological effects of a particular carbohydrate can vary between individuals. For example, lactose is poorly digested and absorbed in the small intestine by many individuals, whereas others are able to break down lactose and utilize its constituent sugars, glucose and galactose. Furthermore, the physiological or nutritional classification of carbohydrate requires ongoing revision due to evolving scientific evidence in the area of food science and metabolism.

Cummings and Stephen (2007) reviewed the wide range of chemical and physiological terms used to describe carbohydrate. Terms were considered useful if they were: 1) measureable by a laboratory analyst; 2) understandable to the consumer; 3) indicative of the properties of the carbohydrates rather than the food itself. Based on their review of the categories used to classify carbohydrate, Cummings and Stephen (2007) present a number of conclusions. A number of the conclusions relevant to the above discussion are summarized below.

1. Dietary carbohydrate should be classified according to chemical composition.
2. The physiological and health effects of carbohydrate are dependent not only on their chemical form but are also dependent on their physical properties.
3. Many terms exist to describe sugars in the diet. The most useful are total sugars and mono-and disaccharides. The use of other terms creates difficulties for the laboratory analyst, confusion for the consumer, and suggests properties of foods that may not be related to sugars themselves, but to the food matrix.
4. Because neither chemical nor physical descriptions of carbohydrate reflect its physiological properties and health benefits, a number of terms to describe carbohydrate based on their physiology have been created. Of these, prebiotic, resistant starch, dietary fibre, and glycemic carbohydrate are useful.
5. The term dietary fibre should not be used to describe physiological or health properties that vary considerably with type, but rather should reflect the health benefits of a diet rich in fruits, vegetables, and whole grains.
6. The distinction between soluble and insoluble forms of fibre is inappropriate considering that this separation is pH dependent and does not reflect the physiological properties of whole foods in the gut. Further investigation is required to examine the effect of foods containing different types of fibre on glycemic control and lipid levels and to determine the specific properties of the carbohydrate that induce improvements in these health biomarkers.

It is clear that dichotomy exists – the chemical classification of carbohydrate is insufficient to describe nutritional or physiological properties and physiological terms are not necessarily useful for measurement or labelling purposes. Furthermore, the use of a number of these chemical and physiological terms is not suggested by Cummings and Stephens (2007) because they may cause consumer confusion by describing the food matrix, rather than the carbohydrate.

References

Scientific Advisory Council

G. Harvey Anderson, Ph.D.
University of Toronto

N. Theresa Glanville, PhD, PDt
Mount St. Vincent University

David D. Kitts, PhD
University of British Columbia

Rena Mendelson, D.Sc., RD
Ryerson University

Robert Ross, PhD
Queen’s University

Huguette Turgeon-O'Brien, Ph.D., Dt.P.
Laval University

This publication may be reproduced provided the source is acknowledged.

Publié en français sous le titre: «Glucides-Info»

Back to top »