Glycemic Index:

Excerpt From David Mendoza

The glycemic index is about the quality of the carbohydrates, not the quantity. Obviously, quantity matters too—and that is the reason for the glycemic load values—but the measurement of the glycemic index of a food is not related to portion size. It remains the same whether you eat 10 grams of it or 1000 grams. That's because to make a fair comparison tests of the glycemic indexes of food usually use 50 grams of available carbohydrate in each food. You can eat twice as many carbohydrates in a food that, for example, has a glycemic index of 50 than one that has a glycemic index of 100 and have the same blood glucose response.

Basically, test foods are fed to various people, some with diabetes, others without, in portions that contain 50 grams of available carbohydrates. The 50 gram carbohydrate portion is specified in Dr. Wolever's methodology paper (see bibliography below) as 50 grams of available carbohydrates. "That means it excludes the fiber," Professor Brand-Miller writes me. "We have always used a 50 gram available carbohydrate portion and often relied on manufacturers to give us the composition data. I am aware of only one instance where we been given incorrect information and therefore inadvertently included the fiber in the 50 gram carbohydrate portion…but there may be some papers from developing countries where the data is not reliable."

For example, to test boiled spaghetti, the scientists give their subjects 200 grams of spaghetti, which according to standard food composition tables provide 50 grams of available carbohydrate. The scientists compare this response with the volunteer's response to a reference food, which may be either glucose or white bread. Both for the test and for the reference foods the volunteer's response over the next two or three hours is calculated. Rather than measuring a single point, they make the more precise measurement of the area under the curve. Then, they repeat the whole process on different days to reduce the effect of day-to-day variations.

Next, the area under the response curve for the test food is expressed as a percent of the mean value for the reference food for the same subject. Finally, these percentages from each subject are averaged together to obtain the GI for that food. For more information, see Wolever, Thomas M.S. et al. "The Glycemic Index: Methodology and Clinical Implications," listed in the bibliography below.

The glycemic index should not be your only criterion when selecting what to eat. The total amount of carbohydrate, the amount and type of fat, and the fiber and salt content are also important dietary considerations. The glycemic index is most useful when deciding which high-carbohydrate foods to eat. But don't let the glycemic index lull you into eating more carbohydrates than your body can handle, particularly if you have diabetes. The number of grams of carbohydrate we consume is awfully important. Make sure you know the carbohydrate content of the foods you eat—study the nutritional information on the package.

But first you need to decide the composition of your diet in terms of carbohydrate, fat, and protein. Almost all the experts agree that we should minimize our intake of saturated and trans fat and eat a lot more fiber than we do. Some other fats, particularly those from cold-water fish, seem to be beneficial. Beyond that, the battle rages between those who would have us eat more protein and those who say that carbohydrates should provide most of our calories. I'm no expert and am genuinely puzzled myself, although I have begun to cut back on my carbohydrates and eat more protein. Generally, foods high in fat and protein have lower glycemic indexes than foods high in carbohydrate. In a real sense, the glycemic index is not applicable to high-fat and/or high-protein foods.

The problem is that even among the complex carbohydrates not all are created equal. Some break down quickly during digestion and can raise blood glucose to dangerous levels. These are the foods that have higher glycemic indexes. Other carbohydrates break down more slowly, releasing glucose gradually into our blood streams and are said to have lower glycemic indexes.

Before the development of the glycemic index, scientists assumed that our bodies absorbed and digested simple sugars quickly, producing rapid increases in our blood glucose levels. This was the basis of the advice to avoid sugar, a proscription recently relaxed by the American Diabetes Association and others.

Contrariwise, the experts thought that our bodies absorbed starches such as rice and potatoes slowly, causing only small rises in blood glucose. Clinical trials of the glycemic index have also proven that assumption to be false.

Factors such as variety, cooking, and processing may effect a food's GI. Foods particularly sensitive to these factors include bananas, rice, and potatoes (for a fuller discussion of the GI of rice and potatoes see the section below). A 1992 study by Hermansen et al. reported that the GI for under-ripe bananas was 43 and that for over-ripe bananas was 74. In under-ripe bananas the starch constitutes 80-90 percent of the carbohydrate content, which as the banana ripens changes to free sugars. Particle size is also an important factor, according to a 1988 study by Heaton et al. The researchers found that the GI of wheat, maize, and oats increased from whole grains (lowest GI), cracked grains, coarse flour, to fine flour (highest GI).

In addition, the glucose response to a particular food may be somewhat individual. So it is probably a good idea to carefully watch your own blood glucose level after eating foods you have questions about and determine if they have high or low GI for you.

So, the idea of glycemic index is a very useful one. But if you find a specific food produces an unexpected result, either high or low, take note of it and incorporate that into your meal planning.

Also note that the numbers vary from study to study. This may be due to variations in the individuals in a particular study, other foods consumed at the same time, or different methods of preparation, since your body can absorb some foods better when they are well cooked.

Most, but not all, of the foods tested are high in carbohydrates. Some may wonder at the gaps—why other high-carbohydrate low-calorie foods like celery (or tomatoes or similar foods) have never been tested. The problem is a technical one for the testers, because they would be so hard put to get anyone to volunteer to eat 50 grams of carbohydrate from celery—it's just too much celery to think about! Essentially, from a glycemic index standpoint, celery and foods like it can be considered as free foods. I now have a list of the common vegetables and fruits that are free foods on-line at David Mendosa

Some people wonder if the glycemic index can predict the effect of a mix meal containing foods with very different indexes. Studies have shown that it does that job very well, too.

More than fifteen studies have looked at the glycemic index of mixed meals. Twelve of them showed an excellent correlation between what was expected and what was actually found.

You can quite readily predict the glycemic index of a mixed meal. Simply multiply the percent of total carbohydrate of each of the foods by its glycemic index and add up the results to get the glycemic index of the meal as a whole. Professor Brand-Miller has an example on page 78 of her book The New Glucose Revolution.

The three studies that did not show the expected correlation came from a group of researchers who were not using standardized methodology for working out the glycemic index from the area under the curve. In addition, their meals were high in fat, which tends to reduce the impact of any one carbohydrate food.