Use and Abuse of the Glycemic Index

The promise and pitfalls of the glycemic index


The glycemic index is a very good tool, and its invention redounds to the credit of my friend and colleague, David Jenkins, of the University of Toronto.

That said, the handsaw is also a very good tool; I have no idea who originally invented it, but kudos to him or her.

Of course, a saw is of no help for opening a jar, typing a letter, or tying shoes, and would make a truly bad choice for removing lint from a dress, combing hair, or brushing teeth. The handsaw, and all variations on its theme, is an excellent tool for what it was designed to do but not necessarily for anything else. It can be used well or badly. So, too, can the glycemic index.

In simple terms, the glycemic index measures how much a food raises blood sugar levels. The word "glycemic" refers to sugar in the blood. The term "index" is used when scores are established by comparing one measurement to another. Weight is not an index because pounds or kilograms are not measured relative to anything else. The body mass index (BMI) is an index because weight is measured relative to height.

In the case of the glycemic index, which uses a scale from 0 to 100, the reference standard is generally pure glucose (sugar), set at 100. In some cases, white bread has been used as the reference standard. Either way, how much a food affects blood sugar is compared to the reference standard, generating a higher or lower number accordingly.

As with all good measures and most good tools, a certain complexity underlies the apparent simplicity of the glycemic index. The measure actually represents that area under the blood glucose curve, plotted for a period of 2 hours following ingestion of the test food. If you ever took calculus, you may recall that calculating the area under a curve was the principal focus of this advanced branch of mathematics, developed by Sir Isaac Newton. Fortunately for the rest of us, no knowledge of calculus is required to understand the glycemic index GI: High numbers mean foods raise blood sugar more, and low numbers mean they raise it less.

The comparison to weight actually harbors a precautionary message about the limits of the glycemic index. Consider, for instance, that one person weighs 170 pounds, and another weighs 120 pounds. Who is heavier?

The answer is obvious: The person who weighs 170 pounds. But that obvious answer might be totally misleading. What if the person who weighs 170 pounds is 6 foot 5, and the person who weighs 120 pounds is 4 foot 5? In all the ways that matter, the shorter person is now the "heavier" of the two.

The glycemic index is subject to this same limitation, because it compares comparable portions of "carbohydrate" in foods. Traditionally, comparisons are based on a 50-gram portion.

See the problem? Carrots, famously, have a rather high glycemic index. But it takes 4 cups of carrots to amount to 50 grams of carbs. In contrast, a cup and a half of vanilla ice cream has 50 grams of carbs. A small portion of ice cream is compared to a very large portion of carrots. Cotton candy provides more than 50 grams of carbs in less than 2 ounces.

So the glycemic index, while useful when comparing similar foods, may be rather misleading when comparing dissimilar foods—just as weight can be misleading when comparing people of very dissimilar heights.

Now, let's consider how this can be corrected. In the case of weight, the standard solution is to use the BMI. Again, there is some devilry in the details of the calculation (e.g., weight in kilograms divided by height, in meters, squared), but the concept is simple: weight, adjusted for height. A person who is 6 foot 5 and weighs 170 pounds has a low-normal BMI of 20.2. A person who is 4 foot 5 and weighs 120 pounds is obese, with a BMI of 30.

The corresponding measure for the effects of food on blood sugar is the glycemic load. The glycemic load adjusts the glycemic index for the amount of food being eaten. So whereas carrots do have a relatively high glycemic index (47), they have a very low glycemic load (2). The idea behind the glycemic load is to compare the effects on blood sugar of comparable and realistic amounts of different foods.

By that measure, most foods widely recognized as nutritious get a predictably low score. But the glycemic index is still better known, and has been adopted into a wide array of popular diets. This has all too often played out like using a saw as a lint brush: badly.

Carrots, apples, chickpeas, walnuts, black beans, and strawberries all have a much higher glycemic index than either heavy cream, or diet soda. If that leads you to believe these last two are the better choices for your health, you can enjoy them for lunch while enjoying the view from the bridge I would like to sell you.

Heavy cream is all fat, most of it saturated. It has a glycemic index (and glycemic load, for that matter) of 0, since it contains no carbs. The same is true for pure trans fat. And for diet soda, which in my opinion is a chemistry experiment in a cup. But since it's an artificially sweetened cup of chemistry, its GI is 0.

On the glycemic-index scale, pure, cold-pressed, extra virgin olive oil and pure trans fat would be identical. So, too, would diet soda and water. And so would unsalted peanuts, lightly salted peanuts, and heavily salted peanuts. Even when it is used well, the glycemic index is only one measure among many of the overall nutritional quality of foods. Whenever we focus exclusively on just one aspect of nutrition, we risk missing the forest for the trees.

In the case of foods, the forest extends not only to the full nutritional composition of the food, but also to the overall pattern of the diet. Studies show, for instance, that foods we eat at one time can influence the glycemic effects of foods we eat at another. Soluble fiber blunts blood sugar responses, so a breakfast of soluble fiber-rich oat cereal, for example, can lower the glycemic response to whatever is eaten at lunch.

Finally, while the glycemic load of the overall diet certainly is important for both health and weight, as studies by David Ludwig, a pediatric endocrinologist and child obesity expert, at Harvard, among others, show convincingly—there are different ways to lower your load, so to speak, and some are better than others. A 2006 study in the Archives of Internal Medicine specifically compared a low-carb, low-glycemic diet to a high-carb, mostly plant-based, low-glycemic diet, and the high-carb, low-glycemic diet had better overall effects on a variety of metabolic and cardiac risk measures.

The glycemic index might convince you, as it has convinced many diet authors, that you should avoid carrots, or all fruits. But let's be blunt here: You find me the person who can blame obesity or diabetes on an excess of carrots or apples, and I will give up my day job and become a hula dancer!

The glycemic index, or preferably glycemic load, can help guide you to better choices within a category where nutrition varies widely, such as breads, cereals, sauces, and dressings. Better still, focus on the overall nutritional quality of foods, and choose wholesome foods, mostly plants, direct from nature as much as possible—and the glycemic load will tend to take care of itself. Health-promoting diets, from Mediterranean to vegan, generally feature a low-glycemic load among their virtues.

Like a saw, or hammer, or screw driver, the glycemic index is only good for doing what it was designed to do. It measures the effect of a fixed dose of carbs on blood sugar responses, and that's all it does. It is oblivious to protein quality, so cannot give eggs credit for a perfect array of essential amino acids. It is oblivious to fat quality, so cannot reward wild salmon for its high concentration of omega-3s.

Like any tool, it's as good as your use of it.

Hungry for more? Write to with your questions, concerns, and feedback.

David L. Katz, MD, MPH, FACPM, FACP, is a specialist in internal medicine and preventive medicine, with particular expertise in nutrition, weight management, and chronic-disease prevention. He is the founding director of Yale University's Prevention Research Center, and principal inventor of the NuVal nutrition guidance system. Katz was named editor-in-chief of Childhood Obesity in 2011, and is president-elect of the American College of Lifestyle Medicine.