Testing and labelling GI

The GI ranks food on a scale from zero to 100, based on how quickly it raises your blood glucose levels. A low value is less than 55, medium is 56 to 69, and high is greater than or equal to 70 and above.

The GI rating of a food is assessed in laboratory trials involving real people (usually in a batch of 10) and real food. First, the subjects must fast overnight, before being fed 50 grams of pure glucose.

Blood samples are taken over a two-hour period to measure what happens to their blood glucose levels. The response is given a rating of 100.

Then portions of the trial food, containing 50g of carbohydrate, are fed to the subjects. If the food is sugar, which is all carbohydrate, it weighs just 50g. For other foods the portion sizes will be bigger.

Blood samples are taken and the blood glucose responses measured against the glucose standard. If a subject's response is exactly half that of glucose, the GI is said to be 50. The average GI of all subjects becomes the published GI.

What's in a logo?

There's a GI logo you may start noticing on some foods here, developed by the University of Sydney, Diabetes Australia and other groups. Companies apply to use the mark and pay for a licence, much like the Heart Foundation tick.

We advise caution. First, the GI mark doesn't mean low GI. Any food (whether it gets a low, medium or high GI rating) can be eligible for the mark, provided it has had its GI accurately tested, contains at least 10g of carbohydrate, and meets certain nutritional criteria for energy, total and saturated fat, fibre and sodium.

If you see the GI logo, check the number on the back.

Second, the guidelines do not limit the amount of refined sugar in a food, so it's possible a pavlova (provided it meets the other nutrition criteria) could display the logo. Low GI perhaps, but definitely no better than a treat food!

Third, some foods without the logo will have a lower GI than others that have it.