The research on the insulin index of foods is intriguing but limited. Only 16 peer-reviewed articles in MEDLINE even mention the term "insulin index," and only one of them actually reports the results of food comparisons. By comparison, 244 peer-reviewed articles mention the glycemic index.
They…found that glycemic and insulin scores were highly correlated.
That study is "An Insulin Index of Foods: The Insulin Demand Generated by 1000-kJ Portions of Common Foods" in the American Journal of Clinical Nutrition 1997, Vol. 66: pages 1264-1276 by Susanne HA Holt, Janette C. Brand Miller, and Peter Petocz. The three co-authors were then associated with the University of Sydney in Australia. Susanne Holt was then a graduate student working under the supervision of Janette Brand Miller, and Peter Petocz provided statistical support. Subsequently, Ms. Holt—now Dr. Susanna Holt—obtained her Ph.D. degree and is directs the Glycemic Index Research Service (SUGiRS) in the University of Sydney's department of biochemistry. Ms. Brand Miller—now Professor Jennie Brand-Miller—directs glycemic index research at the University of Sydney's department of biochemistry.
They tested only 38 foods and found that glycemic and insulin scores were highly correlated. Their most interesting finding was that "protein-rich foods and bakery products (rich in fat and refined carbohydrate) elicited insulin responses that were disproportionately higher than their glycemic responses."
One J.S. Coleman finds the insulin index to be superior to the glycemic index. Comparing the insulin index study cited above with glycemic index studies, that article (which is no longer online) states that "their food choice method is more realistic, and their method more thorough than the GI method."
What that author apparently fails to realize was that the leading researchers of the glycemic index and the insulin index are the same people. Here is what Professor Brand-Miller has to say about the insulin index in the latest version of her best-selling book, The New Glucose Revolution (New York: Marlowe and Company, 2003, pages 57-58:
While it's clear that the insulin demand exerted by foods is important for long-term health, it doesn't necessarily follow that we need an insulin index of foods instead of a glycemic index. When both have been tested together, the glycemic index is extremely good at predicting the food's insulin index. In other words, a low-GI food has a low insulin index value and a high-GI food has a high insulin index value. Furthermore, the level of glucose in the blood is directly related to adverse reactions such as protein glycosylation (linkages between glucose and protein) and oxidative molecules.
There are some instances, however, where a food has a low glycemic value but a high insulin index value. This applies to dairy foods and to some highly palatable energy-dense "indulgence foods." Some foods (such as meat, fish, and eggs) that contain no carbohydrate, just protein and fat (and essentially have a GI value of zero), still stimulate significant rises in blood insulin.
At the present time, we don't know how to interpret this type of response (low glycemia, high insulinemia) for long-term health. It may be a good outcome because the rise in insulin has contributed to the low level of glycemia. On the other hand, it may be not-so-good, because the increased demand for insulin contributes to beta-cell "exhaustion" and the development of type 2 diabetes. Until studies are carried out to answer these types of questions, the glycemic index remains a proven tool for predicting the effects of food on health.
The following table shows how the glycemic scores and insulin index of these 38 foods compare. Note that here the glycemic scores are based on white bread set to equal 100, although the now more common glycemic index sets glucose to be to equal 100.
Especially note that glycemic scores differ in other ways from the glycemic index. "It's important to discriminate between glycemic index values—for 50 gram-carbohydrate portions of foods—and glycemic scores—for 1000 kJ portions of foods," the lead author of the study, Dr. Susanna Holt, writes me.
"In the insulin index study, we measured glycemic scores and insulin scores for 1000 kJ portions of foods. They are not GI values. In a healthy person that has fasted for more than 10-12 hours overnight, cheese and steak can cause a small rise in blood glucose in the second hour of our 2 hour test periods due to gluconeogenesis. Also the normal fluctuations in blood glucose around the fasting value that our experiments start from produce some area above the fasting blood glucose level , which is used to calculate both GI and glycemic score values."
|Food||Glycemic Score||Insulin Score|
|Grain [rye] bread||60||56|
|SNACKS AND CONFECTIONARY|
First published here: July 13, 2003
Go back to Home Page