Issue 20 / 30 May 2016

THE glycaemic index (GI) of foods is useful for people with diabetes, especially when choosing a single food, but it’s not necessarily a marker of healthy foods. Lists of GI values can also be confusing.

Healthy low GI choices include legumes, pasta, oats, foods with whole pieces of grain and most fruits. Less healthy choices include pure fructose, ice cream, potato crisps, many cakes and chocolate cookies!

The GI of potatoes illustrates the confusion over values, ranging from a low 24 for “potato, type not specified” to 49 for “potatoes boiled” and 54 to 87 for the standard Sebago potato.

Mixed meals also offer problems as the fat, protein and dietary fibre content of the meal affect the rate of digestion. And giving foods a GI value based on consuming a portion of food that provides 50 g of carbohydrate may not represent the usual amount of the food likely to be consumed.

Glycaemic responses also vary between individuals, and even for the same person at different times.


It’s commonly claimed that low GI foods increase satiety and thus reduce subsequent food intake. Not at breakfast, according to a recent meta-analysis. And not in children. Nor if a combination of foods is consumed – as occurs in most meals.

Weight loss or weight maintenance?

For weight loss, studies testing diets with different GI values provide scant evidence, at least over the long term. Several well conducted studies show no difference with high or low GI diets for weight loss or weight maintenance.

Positive results for low GI diets have usually been from short-term studies, or have used healthier low GI foods that can be recommended for their dietary fibre and good nutritional value.

The multicentre European Diet, Obesity, and Genes (Diogenes) trial is often cited as evidence that low GI diets help with weight maintenance.

It’s worth reading the actual underwhelming results before accepting some of these claims. Participants first lost weight on an 8-week diet based on a commercial meal substitute plus vegetables.

With an average body mass index greater than 30, they were then assigned either to a control diet or one of four diets that were high (25% energy) or low (13% energy) protein with either high or low GI foods.

After 6 months and eight counselling sessions with a dietitian, the participants’ diet diaries claim significant reductions in their energy intake, yet no further weight loss. How accurate were these records?

In the Diogenes trial, the difference in GI between the high and low GI groups was just 4.7 GI units. That’s well below what has been shown to be inter-individual and intra-individual differences. From the huge variations in dietary intake within each diet group, we can only conclude that this was yet another study showing the difficulty of sticking to any new diet long term. Those assigned to differing levels of protein intake or GI values migrated back to the centre position.

Short-term studies avoid these realities, but they don’t solve the problem of our poor eating habits.

The “elephant in the room” for the increasing national girth is our intake of “discretionary” foods – products with high levels of saturated (or trans) fats, sugars, salt and low levels of essential nutrients.

These currently contribute 35–41% of Australians’ energy intake and eat up almost 60% of the average food budget.

Dietary guidelines already recognise legumes, minimally processed wholegrain products, fruit, vegetables and nuts for their nutritional attributes. They also have a low GI. But judging foods by their GI also gives a tick (or a big G) to foods that do not fit with dietary guidelines.

The most obvious example is packets of sugar displaying the big G of approval – 99.4% sugar but with the grains sprayed with molasses to reduce the GI.

Other foods approved on the basis of their low GI include some ice creams, whose major ingredients (after milk) are maltodextrin (a refined starch) and fructose; an instant pudding; powdered chocolate or malt drinks with more than 45% sugar; a breakfast cereal with 26% sugar and meal replacement drink products with 30% sugar (fructose is the second most prominent ingredient).

The GI is useful for people with diabetes, especially when they are eating a single food. But having a low GI does not make potato crisps healthy. Nor does it mean that cakes or biscuits made with pure fructose or low GI sugar are somehow less damaging to teeth or waistlines.

Dr Rosemary Stanton, OAM, is a leading Australian nutritionist.

4 thoughts on “Limits to the glycaemic index

  1. Robin Willcourt says:

    The GI has never been a good measure of carbs and instead should have been replacedd by the glycemic LOAD which is a real world practical measurement of carb intake.

  2. Guy Hibbins says:

    I think that the problem here is the whole concept of ranking foods as healthy or less healthy based upon comparison with other foods in the same class.  What if the entire class of food is unhealthy?  Cane sugar, being sucrose, has a lower GI value than glucose, but this does not make it a health food.  If we had a similar rating system for tobacco products, then some would score more stars than others based on their tar and nicotine content,  Such a rating system often leaves out more than it reveals.

    See, for example, the article by Marion Nestle, author of Food Politics at

  3. Dr Rosemary Stanton says:

    Dr Willcourt

    Glycaemic Load may give more information that the GI, but it still doesn’t look at the total worth of the food. I think we need to stop judging any food on the basis of its macronutrients – it’s just too gross. And that applies whether we are talking about fats or carbohydrates – and even proteins. People shop for foods, not individual nutrients. And how do you estimate GL accurately when it depends on GI which has so many complications due to accompanying foods and individual variation.

  4. Arne Astrup says:

    As a scientist I love criticism – that’s what brings science ahead. By contrast, I have issues with Dr Stanton’s brief review of the Diogenes study. This study was funded by the European Commission research programme with 15 million Euro, and conducted by a consortium of more than 50 of Europe’s leading scientists in nutrition and health, and the results was published in leading peer-reviewed scientific journals such as New England Journal of Medicine, Circulation and Pediatrics. The study clearly showed that a lowering of the glycemic index by just 4-5 units, and increasing protein slightly, was effective in preventing weight regain over 6 months among overweight people who have just lost 11 kg. Moreover, it also produced a significant improvements in inflammation and in diabetes and cardiovascular risk factors, and reduced the prevalence of overweight and obesity among their children. This was achieved without putting any restriction on amount of calories – simply relying on the greater satiety produced by protein and low GI meals. It is correct that we could not find any changes in self-reported calorie intake  – as expected.  All nutrition scientists know that one cannot measure caloric intake accurately by food records.  What matters is the improvement in weight control and risk factors.  I hope Dr Stanton will read the publication once more.

Leave a Reply

Your email address will not be published.