Insulin Index

J S Coleman

Bionomic Nutrition Forum, 2000

This article has been written to clear up some of the dietary inaccuracies and myths surrounding the role of different types of food, and how they affect insulin secretion. Some dietary authorities and fad diet groups are now proposing dietary changes based on the glycemic (i.e. glucose in blood) index (GI) method. At the moment the GI method has yet to be shown to be anything other than a crude yardstick, and so it had not gained broad scientific acceptance. The GI method was developed to rank foods according to the extent to which they increase blood glucose concentrations, this being a useful guide to help those people with diabetes choose foods with lower glycemic responses. Insulin promotes the uptake of glucose from general circulation, into the cells for use in energy production.

Some of the over-simplistic concepts being circulated run along these lines:

  • foods high in carbohydrate have higher glycemic indexes than protein rich foods...
    and therefore high protein foods (meat/fish) are safer
  • fruits are high in sugar...
    and therefore have higher glycemic indexes
  • a higher glycemic index means the body must produce more insulin...
    and therefore low glycemic index foods are safer

A better attempt at understanding how diet affects insulin levels has been proposed by Susanne HA Holt, Janette C Brand Miller and Peter Petocz in their paper entitled "An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods" (Am. J. Clin. Nutr., Nov 1997, Vol.66, Iss.5, p.1264-76). The authors point out that their results are "preliminary", and it must also be noted that only a few foods (38) have been studied. Even so, their food choice method is more realistic, and their method more thorough than the GI method. Their conclusions challenge some previous beliefs based on GI findings.

In this paper, the researchers identify a number of problems with the GI method. The most obvious problems are that GI uses a 50g carbohydrate serving of foods, which is not representative of how people really eat, and also that although protein rich foods produce a low blood glucose response, it does not follow that there is a correspondingly low insulin response. In short, the GI method is inaccurate, incomplete and unrealistic, although perhaps better than nothing. The researchers state that the GI concept does not consider concurrent insulin responses, and that little research reports both a GI value and accompanying insulin responses. Real diets do not consist of meals where single food items are eaten to 50g carbohydrate levels, in addition when foods with different qualities are mixed; the insulin response can be unpredictable. The GI method is not an accurate predictor of insulin response, and the new paper proposes a method for obtaining a more realistic assessment of dietary factors to insulin response, based on a more realistic isoenergetic basis.

There are a number of factors other than carbohydrate content that mediate in stimulation of insulin secretion, for example it is stated that protein-rich foods or the addition of protein to a carbohydrate-rich meal can stimulate a modest rise in insulin secretion, without increasing blood glucose concentration. Similarly adding fat to a carbohydrate-rich meal also increases insulin secretion even though plasma glucose response is reduced. Several insulinotropic factors have been found to potentiate the stimulatory effect of glucose and mediate posprandial insulin secretion. These factors include fructose, some amino acids and fatty acids, and gastrointestinal hormones. So protein and fat rich foods also induce substantial insulin secretion despite producing relatively small blood glucose responses.

The GI is a ratio of the measure of blood glucose levels found after eating a 50g portion of white bread (or sugar), to a 50g carbohydrate portion of the test food. White bread is often taken as the reference food, and given a score of 100%, so a food that produces half of the blood glucose response over the same test period would be given a score of 50%. In contrast, the Insulin Score is a ratio based on insulin levels found over 2 hours after consuming a 1000kJ meal of the test food, to a 1000kJ meal of white bread. The equation is similar to that developed for the GI value. The glycemic score measures blood glucose levels in a similar fashion.

Although personal variations in response to identical meals occurred in the study, the researches found a stable correspondence between foods and insulin and glucose scores across the group. On average the snack foods produced the highest food group IS, followed by bakery products, carbohydrate-rich foods, fruit, protein rich foods and then breakfast cereals respectively (see figure). The researchers found significant variations in foods of the same food group, so food group alone is not a good predictor of insulin or glucose scores. Furthermore, at the food group level, variations are not as dramatic as between specific foods, so that generalisation about food groups and insulin or glucose scores are inaccurate. The graph to the right adapted from the study results, shows the mean glucose and insulin scores of the food groups.

The researchers did find that jellybeans (made of sugar and animal protein) produced the highest mean IS, whereas peanuts (an oily legume) had the lowest IS. The reference food, white bread, consistently had the highest glucose and insulin responses, and had a higher insulin score than most of the other foods. On average fish produced twice as much insulin secretion as did the equivalent portion of eggs. Amongst the few fruits examined, oranges and apples produced significantly lower scores than grapes and bananas, despite similar carbohydrate content. Potatoes had significantly higher scores than all of the other carbohydrate-rich foods. White and brown rice have similar scores, as do white and brown pasta. Despite containing similar amounts of carbohydrate, jellybeans induced double the insulin secretion as any of the four fruits. These findings are presented in the figure below, showing both scores for all 38 foods, in their food groups.

From the above data, we can conclude at least, that some fruits do not produce insulin responses much greater than protein rich foods such as beef or fish. Perhaps surprisingly, the insulin scores for cheese, beef and fish are greater than those for starchy foods such as porridge. This will lay rest to the claim that protein rich foods are somehow insulin safe when compared to carbohydrate rich foods. Each food must be evaluated individually, and more realistically, each meal.

Overall, although GS is a good predictor of IS, the researchers found that the nutrient levels analysed, only explain 33% of the variation of the insulin response for the foods studied. It seems then, that the individual properties of a food, other than those studied here, account for two-thirds of the remaining insulin response.

In the authors discussion they conclude that important Western staples, bread and potato were among the most insulinogenic foods. Highly refined bakery and confection also induce substantially more insulin secretion per kilojoule or per gram of food than did other test foods. If any of these high carbohydrate foods were eaten with either fat or protein rich foods, say bread and cheese or meat, or pizza, then the scores would be far higher. The authors also note, as above, that some protein-rich foods induce as much insulin secretion as some carbohydrate-rich foods. Fibre was not found to predict the magnitude of insulin response. Their conclusion is that the findings imply that typical Western diets are likely to be significantly more insulinogenic than traditional whole food diets. The research method is not ideal, because some of the serving sizes, for apples, oranges, fish and potatoes were felt to be unrealistic, presumably due to excess. However, the method is still superior to the crude 50g carbohydrate portions found in GI study meals. The researchers have found that increased insulin secretion did not account for the low glycemic responses produced by low-GI foods such as pasta and porridge. These findings challenge the scientific basis of carbohydrate exchange tables, which are now clearly making invalid assumptions.

Other important factors such as the rate of gastric emptying, rate of starch digestion, the amount of rapidly available glucose and resistant starch, the degree of osmolarity and the viscosity of the gut's content, must be significant factors in influencing the degree of postprandial insulin secretion. As a cautionary note, the researchers suggest that additional studies are needed to determine whether the IS concept is useful, and reproducible, and more importantly whether it is predictable in a mixed-meal context. When these questions are answered the role of IS in the treatment of diabetes, hyperlipidemia and overweight will be better known. Until then, we can at least dispatch with the some of the urban diet myths that were presented at the top of this article. We can conclude that protein-rich foods are not necessarily some insulinogenic panacea, and that fruits are not some kind of sweet bogeyman. We also find that food refining and mixing is potentially problematic.

Another related set of pop myths, being circulated again by similar nutri-babble factions, concern the claim that fruits contain so much sugar that they are "addictive", and also that sugar is itself addictive. While it is true that the taste of sugars on the tongue does promote release of satiety chemicals in the brain within seconds (an adaptive feeding reflex common in mammals - perhaps more so in frugivorous primates?), a thorough examination of all Medline papers revealed no relevant papers concerning fruit or sugar consumption and addiction. Intriguingly, although the reporters of these anecdotes identify "cravings" for fruit as suggestive of "addictive" properties, reported cravings for animal foods are somehow seen as adaptive survival reflexes. Should we be concerned about fruit addiction stories? In another light scan of the entire Medline record at Healthgate, I found over 3,500 articles concerning the healthful effects of fruit consumption, but only a handful concerning a few special problems induced by fruit eating in some cases, such as workers at citrus farms suffering tooth erosion. Obviously if fruit really is addictive, then we would expect to find fruit sales soaring above those of cheese and beef, with cattle farmers queuing to buy fruit trees and proffer from the new market in addiction - but the cold facts of the matter say otherwise.

While not wishing to add to the surfeit of dietary anecdotes by peddling more patent absurdities, observations of cycles of bingeing and withdrawal while adding more sweet fruits to the diet have been found. Eventually though, over a period of 1 to 2 years, in all cases I have seen, these "cravings" remedy themselves, and one adopts a more varied diet. Perhaps these "cravings" are in fact some part of a healing and then balancing process, as the body realigns feeding behaviour to maintain homeostasis. As they often say, "more research is needed".

Our Privacy Policy, Conditions of Web site use, copyright and disclaimer

© 2002, portions © Jupiter Web Services Ltd 2002
Website Designed & Maintained By Jupiter Web Services Limited