Published: January 24, 2023
In this blog post, we will continue our discussion of all things blood sugar regulation, and talk about the glycemic index and glycemic load, two concepts that are critical to understanding how the food we eat affects our blood sugar levels.
After reading this post, you will be able to use the glycemic index of foods as one tool to keep your blood sugar levels stable, and specifically to avoid major blood sugar spikes.
To make it easier for you to concretely use the information covered, I have created a free poster that lists the glycemic index and glycemic load data of many common foods.
The terms glycemic index and glycemic load both relate to how much a food, a meal, or a diet affects blood sugar levels. Think of them as scoring systems: the higher their number, the more the food (or the meal or diet) will increase blood sugar levels.
What Is The Glycemic Index, and How Is It Calculated?
According to the current protocol for scientifically measuring the glycemic index of foods, participants must each come into a clinic on several occasions, always in the fasting state. Let’s just assume that a participant comes in on 15 different occasions throughout a month.
Every time they come in, investigators prepare a ‘meal’ for them consisting of just one carbohydrate-rich food. Each meal is calculated and prepared to contain exactly 50g of available carbohydrates.
Available carbohydrates are those carbohydrates that are not fiber.
On at least three of these visits, spaced apart a bit, each participant is given the reference food. That reference food is a beverage containing 50g of pure glucose in a glass of water.
Right before drinking the beverage, the researchers draw blood, and then they do this again repeatedly over 2 hours, with blood draws at least at 0, 15, 30, 45, 60, 90, and 120 min.
Blood sugar levels typically rise quickly and quite a bit even in healthy people after drinking a beverage with 50g of pure glucose. So, let’s imagine the blood sugar response of our participant to drinking 50g of glucose looks like this here.
Investigators then calculate the area-under-the-glucose-curve above the baseline glucose concentration. We also call this the incremental area-under-the-curve. The figure above shows what this may look like.
This is then repeated two more times with the glucose beverage, and then researchers average the three areas-under-the-curve. That average is then set to 100, because by convention 50g of pure glucose has a glycemic index of 100.
On the other occasions the participant comes in, they are given 50g of available carbohydrates from other foods. For example, they may be given white rice, brown rice, bananas, white bread, whole grain bread, oatmeal made from quick oats, oatmeal made from steelcut oats, potatoes, corn, kidney beans, ice cream, and quinoa.
Again, participants eat 50g of available carbohydrates from each of these foods, the researchers draw blood repeatedly and calculate the area-under-the-curve for each.
Let’s look at a concrete example. Let’s assume the figure below shows the glucose curve of a participant after eating 50g of available carbohydrates from brown rice in red. And let’s assume that this red area-under-the-curve was 66% of the one in blue, the blue being the average of three separate tests with the reference food, i.e., 50g of pure glucose. Then the glycemic index of brown rice for this one person would be 66.
Now, the glycemic index of a food is not determined from the data of one person. So, worldwide, this exact experiment may have been repeated in dozens or hundreds of people, and then the glycemic index of brown rice is set as the average of all of the individual tests.
One thing that always leads to confusion is that everyone has different glucose responses in general, so how can we come up with one glycemic index for a food? I think this is an important point, so let me explain.
In general, by the most recent standard, the glycemic index is based on the glucose responses in healthy people who don’t have diabetes or pre-diabetes, so this already reduces differences between people. But even if person A has a substantially smaller glucose response to, say, brown rice than person B, it is likely that person A also has a substantially smaller response to the glucose beverage than person B, because person A likely is simply more glucose tolerant. The likely result is that the relationship between the red area relative to the blue area may be similar in both people.
Differences between people still remain, so if the overall glycemic index for brown rice is 66, it may be 62 for one person and 73 for another. Therefore, you can think about the glycemic index as the average glucose response you can expect from a food relative to 50g of pure glucose.
The Glycemic Index of Common Foods
Let’s take a look at the glycemic index of a few foods.
Let me say right away though that the glycemic index should not be understood as a ranking system of the health effects of a food. It’s just a ranking system of the glycemic effects of a food, that is, how much a food raises blood sugar levels. In other words, it gives us information about one aspect of a food. It doesn’t mean that a food with a higher glycemic index is automatically less healthy.
Now, I realize that this may be confusing to some of you, and you may wonder why we are even talking about this if it isn’t a ranking of the overall healthfulness of a food. The reason is that – in my opinion – there cannot be an overall ranking of the healthfulness of foods. Every food, or – let me correct that – every whole food that has not been subjected to ultraprocessing has some benefits and some potential downsides. So what I am hoping is that we all start to move away from this idea that we should look for a food that is perfect in all regards.
Back to the glycemic index.
We have glycemic index data only for foods that contain a meaningful amount of carbohydrates. Meat or eggs or fish, for example, have so few carbohydrates – basically, none – that one cannot practically determine their glycemic index. We are not going to mention these foods here, but that is only because they don’t really have a meaningful acute impact on blood sugar levels. So you are not making a mistake if you think of them as having a glycemic index of zero, or close to zero.
That’s very different for grains. You’ll have some grains with a very high glycemic index, such as White Jasmin Rice at 89, instant oatmeal at 82, cornflakes at 80, and white wheat bread at 73. With these, you get a blood sugar response almost as high as with pure glucose water. Not good.
The grains category provides some better options though. For example, the glycemic index of brown rice or parboiled white rice is about a quarter lower than that of white Jasmin rice. Oatmeal made from rolled or steel-cut oats is also a lot lower than instant oatmeal. And sourdough rye bread with a GI of 54 compares favorably with white wheat bread at 73.
Another category in which we find many foods with a high glycemic index is tubers. Baked white or sweet potatoes, for example, come in at 88, mashed potatoes at 77, boiled white potatoes at 73, and French fries at 65. Lower glycemic index options in this category include boiled sweet potatoes with a glycemic index of 46, and carrots at 32.
A third category in which we find high-glycemic index foods are sweets and snack foods. Marshmallows at 65, popcorn at 60, and potato chips at 55 are some examples.
Among beverages, an obvious example for a high-glycemic index food are soda and fruit juices, but particularly beer.
Looking at fruit and berries, some also have a high glycemic index. Dried fruit such as raisins, for example, at 59, pineapple at 58, blueberries at 53, and grapes at 52. This sounds high, but before you stop eating fruit, wait until we cover the glycemic load in a bit . Also, there are lower-GI options, such as strawberries at 40 and pears at 29.
Consistently relatively low glycemic index values can be found among dairy foods, and among legumes and nuts. Low-fat milk has a glycemic index of 27, full-fat milk of 37, plain yogurt of 26, and even flavored and sweetened commercial yogurt is not all that high at 31 for low-fat and 35 for full-fat. Even full-fat ice cream has a glycemic index of only 38. This is a good reminder of what I talked about earlier: just because a food has a fairly low glycemic index does not mean that it is a health food!
Then there’s a category in which I have grouped legumes and nuts. Kidney beans have a glycemic index of 40, chickpeas of 37, and lentils come in at 42. Mixed nuts have a glycemic index of 24.
All of these values, and more, are shown on our poster, available for download here.
The Glycemic Load, and How It’s Calculated
I personally find the glycemic index useful, but it does have one important limitation. The glycemic index is always based on 50g of available carbohydrates from each food. And that’s nice in that it makes it easy to get a sense of how much each food raises blood sugar levels compared to others, assuming that the exact same amount of available carbohydrates are eaten.
But that exactly is the limitation: foods differ a lot in their carbohydrate content, and one would need to eat only small amounts of certain foods but unrealistically huge amounts of others to get to 50g of available carbohydrates. For example, table sugar is a pure carbohydrate, and so we would need to eat only 50g of that. At the same time, we would need to eat about 175g (~6 oz) of white rice, or 431g of apple (that’s more than two large apples), or more than 1kg (35 oz) of full-fat yogurt, to get to 50g of available carbohydrates. That is because these foods contain a lot less carbohydrates than table sugar. Whereas table sugar is 100% carbs, full-fat yogurt contains less than 5% carbs.
You can probably readily appreciate that many people will regularly eat, for example, 175g of white rice in a meal, but few people will regularly eat two large apples in one sitting, and I don’t think many people have ever had more than 1,000g of full-fat yogurt in a single meal. That would be about five servings of yogurt.
To take the differences in carbohydrate content of foods into consideration, researchers have developed the glycemic load. In simple terms, the glycemic load is the product of the glycemic index of a food and the grams of available carbohydrates in that food.
To illustrate how this is done, and why the glycemic load is useful, let’s calculate the glycemic load of a typical serving of a few different foods.
Let’s look at table sugar first. It has a glycemic index of 66. By the way, if you are wondering why it isn’t 100, that’s because 100 relates to pure glucose, but table sugar is sucrose, which is 50% glucose and 50% fructose. A typical serving of table sugar is one teaspoon, about 5g, and these contain 5g of available carbohydrates. Using the most commonly used formula, shown below, the glycemic load of one teaspoon of table sugar is 3.
For white rice, the glycemic index is 60. A typical serving size is about 160g, which contains about 45g of available carbohydrates. This leads to a glycemic load of 27.
Here now you can see how the glycemic load is useful. The glycemic load of white rice of 27 is nine times as much as that of a tablespoon of table sugar. This means that in an average person, you can expect a similar glucose response to a serving of white Basmati rice as to 9 teaspoons of table sugar.
One of the reasons why nutritionists and dietitians are not that excited about refined grains, such as white rice or white flour bread is that they are not all that different from eating pure sugar. It’s just starch, which are chains of pure glucose, and they are easily digested, so they enter the blood very quickly after eating them.
Apples have a glycemic index of 44, which seems pretty high. However, a serving size for an apple is about 150g, which contains only 15g of available carbohydrates, leading to a glycemic load number of 7. Similarly, we can calculate the glycemic load for full-fat yogurt, and arrive at 3.
So what you can readily see is that you could eat a lot of apples and yogurt and still have a lower blood sugar response than you’d get from a single serving of white rice.
As it happens, right now, I have plain yogurt with a small apple and a handful of nuts for breakfast most days. And check out the changes in my sugar levels in the figure below. On that day, I had breakfast pretty late, at 11:00 AM. Before the meal, my sugar level was about 105mg/dL, because I had already had a few cups of tea with milk. After the meal, it rose very gradually to about 115 mg/dL, and then returned to around 90 mg/dL.
Now, let me clarify that the glycemic index and the glycemic load can not just be applied to individual foods, but even to meals or entire diets. This application of the glycemic index and glycemic load may be irrelevant to you, but I wanted to mention it because it’s often used that way in the scientific literature to assess associations between the dietary glycemic index and glycemic load with long-term health outcomes.
To calculate the glycemic load of a meal or entire diet, one can simply add up all of the individual glycemic loads of all foods eaten. Each individual glycemic load of each food is calculated by multiplying the glycemic index of the food with the grams of available carbohydrates of that food, divided by 100 (as shown in the example above for table sugar, Basmati rice, apples, and yogurt).
Summary and Conclusion
To wrap it up, let me give you the main points:
First, the glycemic index is a measure of how much a food raises blood sugar levels compared to the same amount of available carbohydrates from pure glucose. In a way, you can think about it as a percentage, because the glycemic response to glucose is always set to 100 and all other foods are compared to that. Therefore a glycemic index of, say, 80, means that a food raises blood sugar levels 20% less than pure glucose. A glycemic index of 40 means 60% less than pure glucose.
Second, the glycemic load is based on the glycemic index, but also considers how many available carbohydrates are eaten of a food. It can be used, for example, as a measure of how much a typical serving of a food affects blood sugar levels. Consuming one serving of a food or meal with a glycemic load of 10 would be expected to raise blood sugar about twice as much as one serving of a food or meal with a glycemic load of 5.
Now, why is it that foods have such vastly different glycemic indices, even if they contain similar amounts of carbohydrates? And, is there any evidence that eating a low-glycemic index diet or low-glycemic load diet matters for our long-term health? These two questions will be answered in the next blog post.
Until then, please feel free to post a comment below if you have any questions or suggestions.
Atkinson FS, Brand-Miller JC, Foster-Powell K, Buyken AE, and Goletzke J. International tables of glycemic index and glycemic load values in 2021: a systematic review. American Journal of Clinical Nutrition 2021; 114: 1625-32.
Would you kindly do an article explaining how fats and proteins differentially lower glycemic response of carbs and get into the science of it? I would also love to see an article on brain fog/cognitive dysfunction and why it happens after meals in some people and whether or not there is any truth to type 3 diabetes?
Thank you for the excellent suggestions.
Hi, thank you very much for your content.
1. Do you have a Patreon so that the tip can automatically be recurring instead of one-time only?
2. Could you create a poster version that includes the typical serving size of a meal? It will be helpful to have it handy when one is assessing the glycemic load
Thank you. That is very kind of you to consider leaving a tip. Currently, I don’t have patreon account, though. I am hoping to build something in the long term that actually also brings some additional value to people (rather than just a tip for me …;-), such as an online community with live calls that would also allow members to communicate with each other and be a place to store guides on certain topics, recipes etc. That’ll take some time to build, though, and I’ll first need to figure out if there is enough interest in that.
Regarding your second question, I had initially done that, but it became really ugly and confusing, particularly because I’d have to provide serving sizes in g or mL as well as (for my mostly American audience) ounzes and fluid ounces. It was a mess. I felt that the GL values were still useful, because I thought people had a rough idea of what a typical serving of these foods look like.
Here is an idea, though: I could add serving sizes on a second page of the poster, to be printed on the back!? That may be easy, and quick to do. Haven’t heard from anyone else that they miss the portion sizes, but I do agree with you it would be useful.
How can I get a copy of the Glycemic Index poster that Dr Katz recommends in his videos?
– LINK TO DOWNLOAD GLYCEMIC INDEX / GLYCEMIC LOAD POSTER –
Thank you for your interest, Ross.