Attaining and maintaining an appropriate weight is an increasingly rare accomplishment. It has, in expectation, many health benefits, as well as numerous other practical benefits. (See "The Heavy Non-Health Consequences of Heaviness.") If you share the goal of attaining and maintaining a healthy weight, it is important to focus on the timing of eating, as I argue in "Stop Counting Calories; It's the Clock that Counts." In addition, as I argue In "Forget Calorie Counting; It's the Insulin Index, Stupid," you should focus on the food insulin index in choosing which foods to eat.
"Forget Calorie Counting; It's the Insulin Index, Stupid" gives an organized table of all the data on the food insulin index I have been able to lay my hands on so far. Unfortunately, there are many, many foods and beverages for which the insulin index has not been measured. Sometimes it is possible to guess the insulin index of a food or beverage from the insulin index for similar foods—especially when similar foods have a very high insulin index. But in other cases, not knowing the insulin index for a particular type of food makes it hard to guess whether that type of food is healthy or not. In some fraction of those cases, the glycemic index can be helpful, if used with great caution.
The big advantage of the glycemic index is that it has been measured for a huge range of foods and beverages. At the top of the post is a link to an especially extensive list that I use as my primary data source for this post. This list has several advantages. First, it shows the range of different estimates for different varieties of seemingly similar foods with very different glycemic indexes. Second, it gives standard errors. (Remember that you have to double the standard error on each side of the point index to get a 95% confidence interval.) Third, the article having been peer-reviewed makes me a little more confident that the numbers in the table are of reasonable quality.
I focus here on what I learned from reading this table. Where the message is the same as in "Forget Calorie Counting; It's the Insulin Index, Stupid" (for example "Nuts are great!") I won't say much here. And indeed, the bulk of the table flagged above was saying that foods I categorized as unhealthy in "Forget Calorie Counting; It's the Insulin Index, Stupid" also look unhealthy through the lens of the glycemic index.
Limitations of the Glycemic Index.
The glycemic index measures the effect of a food or beverage on blood sugar. Since a surge in blood sugar reliably generates an insulin response, a high glycemic index almost always indicates a high insulin index. But the reverse is not necessarily true. A food or beverage can have a relatively low glycemic index while having a much higher insulin index. There are three big blind spots of the glycemic index:
Protein often generates an insulin response. So meat and other high-protein foods are not as innocent as they appear from looking at the glycemic index.
Oversimplifying a bit, fructose goes straight to the liver without staying long in the bloodstream. A large amounts of fructose going straight to the liver is not so great. So anything with fructose in it is much less innocent than it might appear from looking at the glycemic index.
Though the science is at a much earlier stage than I would like, some types of nonsugar sweeteners boost insulin. So nonsugar sweeteners are not all as innocent as they appear from looking at the glycemic index.
Although I focus here on what can be learned from the glycemic index, remember that you can experiment with yourself as a guinea pig and quickly learn a lot from your own experience. It doesn't take long for data you collect on yourself to have more statistical power than the typical data set in macroeconomics. And a close look at the table above shows that the precision in the table as indicated by the standard errors has often been achieved from studying a relatively small sample. So don't underestimate what you can learn from your own experience.
The key to experimenting with yourself to guess how big an insulin kick a given food or beverage has is closely connected to why you should care about the insulin index in the first place. The big problem with high-insulin-index foods and beverages is that big spikes in insulin drive down blood sugar and make you hungry again relatively soon. And from my own experience and that of others I know well, I can say that consuming foods and beverages high on the insulin index can make it significantly more difficult to fast. (Though I don't fully understand the mechanism, subjective hunger during fasting, even 24 hours later, seems to be higher if one was eating foods high on the insulin index before beginning the fast.) So the way to get a read on the insulin kick of a given food or beverage is to pay attention to whether you feel hungry again relatively soon afterwards and whether you have extra difficulty fasting after consuming that food or beverage. My wife Gail and I use this approach all the time to get an additional reading on how healthy a given type of food is.
Lessons from the Glycemic Index
In cases where protein, fructose and nonsugar sweeteners are not an issue, my rule of thumb for converting a glycemic index to an insulin index equivalent (as discussed in "Forget Calorie Counting; It's the Insulin Index, Stupid) is to multiply by 1.1. This rule of thumb uses the version of the glycemic index that gives glucose a glycemic index of 100, which is given in the leftmost column of numbers in the table flagged at the top of this post. Confusingly, sometimes a glycemic index with white bread as 100 is used. Making white bread instead of glucose 100 makes the number for the glycemic index 1.4 times as big. This explains some tables in existence that have glycemic indexes that look too high.
Combined with my rule of thumb of avoiding any food that has an insulin index of 50 or above, the rule of thumb of multiplying a glycemic index by 1.1 to get an insulin-index equivalent means my recommendation is to avoid anything with a glycemic index above 45. But I give a little extra leeway to whole fruit, since I find myself unable to completely shake the conventional wisdom that fruit is healthy. (I have no trouble in completely dismissing the conventional wisdom that fruit juice is healthy. It isn't. Fruit juice has way too much sugar that hits your system like a ton of bricks.)
I should say that the advice of avoiding anything with an insulin index 50 or above or a glycemic index above 45 is for anyone who has ever had a problem with their weight. If you have been thin all your life so far, my advice is less rigorous: those who have always been thin might well be able to stay that way simply by strictly avoiding anything with added sugar or sugars, avoiding sweet beverages and processed food, and keeping their eating window each day down to 10 or 12 hours. I base this advice on my view of the epidemiology of obesity: the two things that have approximately the right timing to explain the secular rise of obesity are the ready availability of sugar, the rise of processed foods and beverages and the trend toward eating around the clock every waking moment.
Now, to what I learned from studying glycemic index tables:
A. Bean and Lentils May Be OK. The insulin index data on beans and lentils are quite sparse. The glycemic index data on beans and lentils are reassuring that they may be OK. But you should make sure to pay attention to data from your own experience as your own personal guinea pig for two reasons:
the glycemic indexes for different types of beans and lentils vary quite a bit;
it is possible that the protein in beans and lentils adds to their insulin kick.
Nevertheless, I think it is quite reasonable to experiment and see if you can eat beans and lentils without getting too hungry later on. It should go without saying that you should avoid processed foods made from beans and lentils that have added sugar. Unfortunately, that will include most canned soup as well as many varieties of canned beans. On the likely dangers of sugar, see "Diseases of Civilization", "Sugar as a Slow Poison", "The Case Against Sugar: Stephan Guyenet vs. Gary Taubes", The Case Against the Case Against Sugar: Seth Yoder vs. Gary Taubes", and "Gary Taubes Makes His Case to Nick Gillespie: How Big Sugar and a Misguided Government Wrecked the American Diet." On processed food more generally, see "The Problem with Processed Food."
I should also warn that beans and lentils are likely to be in the category I labeled "Portion Sizes Should Be Kept Small Except on Special Occasions" in "Forget Calorie Counting; It's the Insulin Index, Stupid." In other words, beans and lentils are not freebies. My prediction is that most people can eat moderate amounts of them without getting rebound hunger, but will get rebound hunger if they eat large amounts of beans and lentils.
B. Nonstarchy Vegetables are Indeed Good. In "Why a Low-Insulin-Index Diet Isn't Exactly a 'Lowcarb' Diet" I mirror Wikipedia's list of nonstarchy vegetables. Below I give that list again, plus a few additions, with glycemic index numbers beside them. Strangely, the table flagged above reports almost no glycemic indexes for nonstarchy vegetables, so these numbers are from other sources, particularly the charts I found by searching for Google Images related to the glycemic index. I mirror a few of those images after the list. When they disagreed, I chose the higher of the numbers. In the few cases when something had a high glycemic index I removed it from this list and put it on the list of (presumptively) starchy vegetables further below.
Bamboo shoots (raw) 20
Beet greens 20
Bok choy 10
Celery (raw) 0
Collard greens 20
Green beans <15
Kale 15 (but other sources suggest kale is a bit lower than lettuce)
Okra (raw) 15
Summer squash <15 (Note that other kinds of squash that are eaten without the rind have a much higher glycemic index.)
C. Starchy Vegetables are Bad, with the Exception of Carrots, Chickpeas and True Yams. The list flagged at the top was better at reporting glycemic indexes for starchy vegetables. I have spoken against potatoes before. There are so many entries for varieties of potatoes in the table flagged at the top I will let you do the "find" command within that pdf yourself to see them all. Here are some other starchy vegetables:
Beets (called "beetroot" in the table) 64
Carrots (raw) 16
Carrots (boiled) 41 (I averaged the 32 and 49 readings to get this number. The introduction to the table flagged at the top questions the 92 number, which comes from an early study that may not have been up to later standards.)
Chickpeas (Garbanzo beans) 28
Fava beans (called "broad beans" in the table) 79
Sweet potato 61
Carrots: The glycemic index of carrots depends a lot on whether they are raw or cooked.
Chickpeas: I am puzzled that hummus, which is ground-up chickpeas, would have a lower glycemic index than intact chickpeas. One possible reason is that different types of chickpeas have different glycemic indexes (the table flagged at the top has 10, 31, 33 and 36, averaged to 28) and hummus is often made from a type with an especially low glycemic index. But I don't know.
Yams: In the United States and Canada, the word "yam" is often used to describe a sweet potato. But sweet potatoes are not true yams. Take a look at the Wikipedia article on yams.
D. Other Than Dates, Figs and Raisins, Dried Fruit May Be a Little Healthier than Fresh Fruit. This makes some sense to me: raisins, figs and dates are soft and so seem like they should be easily digestible. Most other dried fruit is tough and chewy, and may be digested more slowly than the corresponding fresh fruit. Here is the data I base this idea on—apples, apricots and plums vs. dates, figs and grapes—some from the table flagged at the top, some from the table mirrored just below the table. I am predicting that dried peaches will be more like dried apricots, apples and plums than like dried dates, figs or grapes. (Remember that dried fruit purchased in the store often has sugar added. That is a different matter. I am talking here about dried fruit with no sugar added. Also, remember that the quantity of dried fruit you are eating needs to be judged by the fresh fruit equivalent.)
Apple (raw) 38 (average of several types)
Apple (dried) 29
Apricots (raw) 57
Apricots (dried) 31
Prunes (dried plums) 15
Dates (dried) 103
Figs (raw) 35
Figs (dried) 67
Raisins (dried grapes) 64
For reference, here is the table of glycemic indexes and a few insulin indexes for fruit that I have in ""Forget Calorie Counting; It's the Insulin Index, Stupid":
E. Underripe Bananas May Be Healthier. I love bananas. But I don't eat them anymore because of their high insulin index. The glycemic index table flagged at the top suggests that underripe bananas have a lower glycemic index:
Banana overripe (Denmark) 52
Banana, overripe, yellow flecked with brown (48)
Banana, ripe, all yellow (USA) 51
Banana, slightly underripe, yellow with green sections (USA) 42
Banana, underripe (Denmark) 30
I suspect a similar principle applies to other unripe fruit—for example the unripe mango that is an important ingredient in Southeast Asian cooking.
F. Couscous and Popcorn Are Bad
G. Pleasant Surprises
I don't trust the low glycemic indexes on a few types of cakes and cookies. Even if they were accurate for a particular type, it would be hard to know that what is still being made is the same type. But there are some pleasant surprises for less processed food that might be valid. Here are the pleasant surprises of relatively unprocessed food with a reasonably low glycemic index and no protein or fructose issue that I found in the table flagged at the top. I include some things where different types tested out differently, and it might be hard to find exactly the right type. I give a range of values. The experimental approach with oneself as guinea pig that I talked about above could be used to try to figure out if a given type was OK.
Coarse Barley Bread 27, 34, 40
Pumpernickel bread 41, 41, 46, 55
Pearl barley 22, 22, 25, 27, 29
27%+ amylose rice grown in Bangladesh 35, 32, 27, 33
Mango 41, 51, 60, 51
Egg fettucine 32, 47, 40
Mung bean noodles 26, 39
Cheela (thin savory Bengali pancake made from legume flour batter) 42, 36, 45, 38
Laddu (sphere-shaped sweets made of popped amaranth, foxtail millet, legume powder, fenugreek seeds, ghee and sugar or sweet sauce) 24, 29 (given the added sugar, this seems a bit too good to be true, but it might be worth the risk)
A big reason I made the effort to study glycemic index tables was to identify a wider variety of food that you might be able to eat without getting an insulin backlash. As I emphasized above, you can quickly experiment to see if you do get hungry again soon after eating a particular food item, or if you have a hard time fasting after eating one of these types of food. So you will quickly be able to confirm or disconfirm the prediction implicitly made by the glycemic index measurements above. I hope the discussion above helps you find some additional types of food that are both tasty and work well with your weight loss or weight maintenance program. And I hope it helps you notice some types of food that mess up your weight loss or weight maintenance program so you can avoid those foods. Remember this rule of thumb: anything with a glycemic index of 27 or above could easily cause you trouble if you eat a lot of it. But a type of food with a glycemic index of 45 or so, might well be OK to eat in moderation. Try and see.
Don’t miss my other posts on diet and health:
I. The Basics
II. Sugar as a Slow Poison
III. Anti-Cancer Eating
IV. Eating Tips
V. Calories In/Calories Out
VI. Other Health Issues
VIII. Debates about Particular Foods and about Exercise
Julia Belluz and Javier Zarracina: Why You'll Be Disappointed If You Are Exercising to Lose Weight, Explained with 60+ Studies (my retitling of the article this links to)
IX. Gary Taubes
X. Twitter Discussions
XI. On My Interest in Diet and Health
See the last section of "Five Books That Have Changed My Life" and the podcast "Miles Kimball Explains to Tracy Alloway and Joe Weisenthal Why Losing Weight Is Like Defeating Inflation." If you want to know how I got interested in diet and health and fighting obesity and a little more about my own experience with weight gain and weight loss, see “Diana Kimball: Listening Creates Possibilities” and my post "A Barycentric Autobiography. I defend the ability of economists like me to make a contribution to understanding diet and health in “On the Epistemology of Diet and Health: Miles Refuses to `Stay in His Lane’.”