This glycaemic response to carbohydrate can be measured by sampling blood after a portion of food is eaten. To get good data you need to start with a group of volunteers. Tests are done first thing in the morning before anything is eaten so that the normal blood glucose levels should have become stable at the “fasting level”. The tests involve taking a drop of blood, usually from a finger prick, and measuring the blood glucose level.
Next a food sample is given and finger prick samples are taken every 15 minutes at the start and 30 minutes later on for a period of 2 hours after the sample is eaten. Figure 1 shows how the blood glucose level varies after a 50g glucose glucose drink is given. The dotted line is the fasting level. This is a typical curve for a non diabetic. The glucose that has been eaten finds its way into the bloodstream fairly rapidly, but the blood glucose level is rapidly brought back to normal as insulin, produced in
the pancreas, is injected into the bloodstream where it reacts with the glucose to convert it to glycogen in the liver for future use or storage as fat. In the case of diabetes the rise in blood glucose would be greater and the level would remain elevated for several hours in the absence of externally injected insulin.
Similar responses are obtained by eating portions of any food containing carbohydrate. In figure 2 the glucose response is shown for a 50g portion of cornflakes (in blue) and also a 50g portion of Lizi’s granola (in pink). In order to quantify the curves obtained the “Glycaemic Response” is measured which is the area under the curve and above fasting level. To compare the amount of glucose that enters the bloodstream directly from different foods, the glycaemic response needs to be referred to the standard response obtained from ingesting a reference amount of glucose.
One way of comparing different foods is the Glycaemic Index (GI). The GI of a food is simply the ratio of the glycaemic response of a portion of food containing 50g of carbohydate to that of the response to 50g of glucose and expressed as a percentage. Thus GI is scale on which the GI of glucose is 100% and other foods rate less. Some numbers: cornflakes 84, white bread 70, table sugar (sucrose) 55, baked beans 48, fructose
(fruit sugar) 19. GI is a good way of sorting out the “good carbs” ie those that have low GI from the “bad carbs” the high GI ones. But, in order to get an indication of how a given portion of food affects your blood sugar you also need to know how much carbohydrate is in the food. For example, a 100g portion of watermelon (GI 70) hardly affects your blood sugar at all since the carbohydrate quantity is so low, but a 100g portion of spaghetti, even with a low GI of 41, can have a significant affect on blood sugar. The glycaemic impact of any portion of food is thus a combination of the GI of that food and the amount of carbohydrate it contains. This is where GL comes in.
GL, the Glycaemic Load of a food is a measure of the glycaemic impact that the food has on your bloodstream. It depends directly on serving size; double the serving and you’ll get double the glycaemic load. GL can be calculated from GI by multiplying GI by the weight of carbohydrate in the serving and dividing by 100. Typically, if GL is less than 10 it will have little impact on blood glucose, more than 50 and you are quite seriously challenged. The GL of a serving of food can also be measured directly by finding the “Glucose Equivalent” of that serving of food. This is the amount of glucose you would need to eat to give you the same glycamic response. So GL comes with units and they are “grams glucose equivalent”.
We prefer to measure the GL of servings of our granola, because calculating GL from GI where the GI is small can lead to significant errors, not least because carbohydrate is the one food component most subject to errors. But the main reason is the fact that GI is a non-linear scale but the GL calculation makes an assumption that the scale is linear. This is a complex argument which you can get more details from in the attached research papers.
To measure GL we use a “bracketing” technique which is explained in figure 3. By making a guess at the GL of the food serving you can give servings of glucose which you know to be above the glucose equivalent and another serving which you know to be below. If these servings are not too different the glycaemic response may be assumed to be linear over the range and the glucose equivalent can be obtained by linear extrapolation.
This is how we measure the GL of our products. The tests are done in university laboratories using groups of up to 12 volunteers so that results can be averaged across the sample. This is an important consideration which consumers who wish to use GI or GL need to be aware of. The numbers obtained are population averages, the response of an individual on a particular day can be significantly different. But over the long term these measures give a good way of quantifying the glycaemic effect of foods.
The paper by Venn et al gives more details on how the glucose equivalent is measured. Click Here
The paper by Shaw et al is an unpublished comment on the Venn paper which shows that a more accurate calculation of GL can be obtained using a non-linear glycaemic response. Click Here
The Monro and Shaw paper elaborates this point Click Here