Sign In

Remember Me

Preserving Insulin Sensitivity

The body adjusts its rate of aging in response to environmental cues.  Most influential is diet.  The story I have told is that

  • The body decides whether food is plentiful (for pregnancy and childrearing) by sensing nutrition and body fat.
  • The body decides either “reproduce and die” or “hang in there and live longer”.
  • The medium for transferring the information is the metabolism of insulin, the pancreatic hormone that controls blood sugar.
  • Insulin resistance (type 2 diabetes) is a primary mode of aging.
  • Sugar and starches quickly become sugar in the blood.  Fat and fiber slow the absorption of sugar.
  • Fasting, exercise and foods with a low glycemic index contribute to better insulin sensitivity and longer life.

Glycemic index (GI) is supposed to be a measure of how rapidly a food is turned into sugar in the blood.  But as I have been reading about GI this week, the story gets muddier and muddier.  Glycemic index as reported is not a good measure of the body’s insulin response.  In fact, insulin response in some studies was found to have no relationship whatever to GI!  Combinations of food turn out to be very important.  Fiber and fats can be helpful, and there is a new category I have learned about recently called “resistant starch” which slows the absorption of other foods that are digested along with it.  Can drinking water with food slow digestion?  (I last blogged about this subject 2 years ago.)

The idea of “negative calorie” foods is very appealing.  Are there things that you can eat that actually make you thinner?  Are there foods that decrease the insulin impact of a meal, or that lead to lower absorption of food that is eaten together at the same meal?


Glycemic Index and Glycemic Load

Glycemic Index was conceived as a characteristic of food ingredients, something that could be reported on a label.  The way it is supposed to work is that low GI foods are better, especially for diabetics, but preserving insulin sensitivity is a concern for all of us as we age.  Glycemic Load (GL) was supposed to be equal to the glycemic index times the portion size.  But instead, the insulin response to food has turned out to be just as complicated as it can be

  • different people respond differently to the same food
  • the same person responds differently at different times
  • eating twice as much of a sugar or starch does not double the glycemic load (as was presumed in the early studies)
  • combinations of food cause a different response from the foods separately

Insulin is a signal to the body that “we have all the sugar we need, thank you.  Take your calories and turn them to fat.”  Insulin causes fat storage.  Insulin is associated with weight gain.  Weight gain and insulin resistance, together and independently, contribute to mortality risk.


Inscrutability of Tables of Glycemic Index and Load

I knew that Glycemic Index (GI) was designed to measure the surge in blood sugar that comes from eating a particular food.  It was my naive expectation that sugar would have the highest GI, then starches, then whole grains, then beans, then nuts.  I thought that greens and fish and meat would have very low GI.  But the GI reference table confounded my expectations on every score.

The glycemic load [GL] is calculated by multiplying the number of grams of carbohydrates in a given food by its glycemic index [GI] (which measures how quickly the food is converted into glucose and released into the bloodstream). The glycemic load is an accurate measure of how much insulin your body will have to produce to neutralize the carbohydrates in a given food. [LEF magazine]

Turns out that this very common-sensical statement is far from true.  The first thing I had to learn is that GI is not standardized by energy value of the food, but by grams of carbohydrate.  Say a peanut is being compared to a potato.  The potato is almost all carb, but the peanut is only ⅕ carb.  So when they compare the glycemic response head-on-head, they compare 50 grams of potato with 250 grams of peanuts.  Of course, this is going to make the peanut’s GI look bad.  The GL is supposed to correct for this, multiplying by a “standard portion”, but the whole idea of a “standard portion” is dubious.

By this same definition, meats and salad oils have an undefined GI.  This is not because they produce no glycemic response, but because they contain no carbs, so the GI measurement protocol is undefined.

Here is a table from Harvard Medical school, the top one that came up in a Google search.  In my experience, it is no more or less unfathomable than others.  For example:

  • Banana cake without sugar has a higher GI than banana cake with sugar.  But it has a lower GL.
  • Cannd chick peas have a GI of 38, while dry cooked chick peas have a GI of 10
  • Unsweetened Apple Juice has a GL twice as high as Coca Cola (30 vs 16) even though Apple Juice has a lower GI (44 vs 63)
  • Whole milk has a higher GI than skim milk (41 vs 32)

Perhaps you can explain the higher GI of whole milk if you assume that cream is finding a pathway to be burned as fuel within two hours.  And the low score for Coca Cola might have to do with burning sugar in response to cafeine.  But there are too many questionable numbers in this table for me to have any faith in it.

How is GI measured?

The GI value of a food is determined by feeding 10 or more healthy people a portion of the food containing 50 grams of digestible (available) carbohydrate and then measuring the effect on their blood glucose levels over the next two hours. For each person, the incremental area under their two-hour blood glucose response (glucose iAUC) for this food is then measured. On another occasion, the same 10 people consume an equal-carbohydrate portion of glucose sugar (the reference food) and their two-hour blood glucose response is also measured. A GI value for the test food is then calculated for each person by dividing their glucose iAUC for the test food by their glucose iAUC for the reference food. The final GI value for the test food is the average GI value for the 10 people.  (Jennie Brand-Miller at Sydney University)

The big problem is in the “portion of food containing 50 grams of digestible carbohydrate”, which can be a very large or a very small portion, depending on the food.  Two additional problems are that different people have very different responses, and also that the body’s response to foods tested in isolation is not a good indication of how the body responds to food combinations typical of a meal.  In this study, GI for different breakfasts was computed by adding up the GIs for individual foods; when the subjects’ insulin response was measured it had no relationship at all to the computed GI—a correlation of zero.


Combining fiber with carbohydrates

Adding fiber to a meal can reduce the subsequent blood glucose and the insulin spike [ref].  Wheat bran has been used to slow the blood sugar uptake for diabetics [ref].  Green leafy vegetables have a similar benefit [ref, ref].

Resistant starch is, by definition, starch that resists stomach enzymes, and is not quickly digested.  It passes through the large intestine, where it is fermented by bacteria that thrive on it.  RS may cause flattulence or indigestion.  Often the symptoms clear up after a few weeks.  Green bananas and raw potatoes are natural sources of RS.  You can also buy it as a flour and mix it into foods.  This study claims that resistant starch slows starch absorption and damps the insulin spike better than fiber.  In this study, RS lowered fasting glucose and improved blood lipid profiles. Glucomannan (konjac), pectin, and guar flour are all reported to have similar effects to RS.

Personally, I have a hunch that fiber (and possibly resistant starch) change the intestinal flora in a way that lowers total food absorption.  You extract fewer calories from the same food.  However, this is almost impossible to measure directly, and to my knowledge the study has not been done.  Less controversially, adding fiber or RS to the diet affect your appetite and how full you feel, and affect the insulin response, which influences whether your body burns the calories or adds to fat stores.  In all events, I think it’s worth the experiment to see if you can lose weight by adding fiber to your meals.

In this context, it is no surprise that long-term studies show that diets rich in green leafy vegetables [ref, ref, ref] and high in fiber [ref,ref, ref, ref] lower the risk of chronic disease.



Drinking water with food, or eating foods with high water content, dilutes the food in your stomach.  I find it reasonable that this alone should slow digestion, delay the absorption of sugar, lowering the effective glycemic load.  Is this a substantial benefit, or is it negligible?  I have been unable to find any data on this question.  There is good evidence that peope tend to eat less calories if they drink more water, especially right before a meal [ref], so water can be a psychological aid to weight loss.  Eating foods with high water content also increases satiety, so that (most people) tend to eat smaller meals, and don’t make up the difference later [ref].  Soups are a weight-loss strategy.


Supplements and drugs that can damp your insulin response:

  • Chromium and magnesium.  These are minerals with no down side.  Recommended for all.
  • Metformin.  This is a classic diabetes drug, so powerful that diabetics taking metformin actually have a lower mortality rate than non-diabetics who don’t take metformin according to a Scottish study last year.
  • Cinnamon and vinegar, green coffee extract and irvingia are among many substances that can be taken 20 minutes before a meal to suppress the insulin spike.


Is Fructose worse than Sugar?

All sugars contribute to insulin resistance and accelerate aging.  Are some sugars worse than others?

Table sugar is sucrose, a 12-Carbon sugar.  It is made of two 6-Carbon sugars, a “dimer” of fructose and glucose.  High-fructose corn syrup (HFCS) has more fructose than glucose (55-65%) and honey is similar.  Agave is yet higher in fructose.

The name “fructose” would suggest that it is the sugar of fruits, but fruits in general don’t tend to have more fructose than glucose.  Apple sugar is about 75% fructose, and melons are 65-70% fructose, but grapes, peaches, and berries are 50%, and bananas actually have less fructose than glucose. [from]

Whether fructose or HFCS is worse than sugar has become a controversial question.  Life Extension magazine and are down on fructose, while andother health advice sources tend to minimize the difference.

Both glucose and fructose go directly into the bloodstream, but glucose is the body’s primary fuel, so the insulin response is determined by the glucose.  In terms of insulin spike, fructose is a lot better than glucose.  Fructose has a much lower GI than glucose.

(from a new study in PNAS )

Despite this, there are two lines of argument against fructose

  • The sweet taste without the insulin was found in this new study to interfere with natural appetite control, increasing desire for more food.  But the difference was not dramatic, as measured either by subjective reports or fMRI.  Satiety and hunger are said to be ruled by leptin and ghrelin, respectively.  Leptin and ghrelin responses to glucose and fructose were not different in this study.
  • In the liver, fructose is converted not to glucose, but to triglycerides, which are stored by the body as fat.  There is broad evidence both from animals and humans that fructose is more fattening than glucose [ref, ref].  This is the basis for the argument that, in the long run, fructose leads to weight gain, insulin resistance, more type 2 diabetes, and thus more accelerated aging than glucose.

The best reference I’ve been able to find on the subject is Basciano et al, 2005.

Everyone agrees that it’s best to minimize both glucose and fructose.  If you prefer to sweeten with glucose instead of fructose, it’s sold as dextrose, not as cheap as cane sugar, but cheaper than honey or agave or maple syrup.


A Calorie is a Calorie is a Calorie – Not!

“The amount you weigh is exactly the difference between the calories you ingest and the calories that you burn exercising.”  We hear this all too often.  Writers who should know better promote it as the “First Law of Thermodynamics.”  This is dangerous nonsense.

In fact, the calorie content of a food is measured simply by burning it and collecting the heat that is released.  But the body’s efficiency in use of foods is a very complex affair, dependent on everything from how well you chew your food to which bacteria reside in your intestine.  Peanuts do not deliver the body as many calories as peanut butter. There are (lucky?) people with very inefficient metabolisms and (unlucky?) people whose bodies are able to extract every last calorie from any meal.  The bacteria that live in our guts digest food for us, but extract a toll in energy that they need for themselves.  Depending on the particular bacteria you have in your intestine, the toll may be only 10% of the food energy coming in, or close to half.

Roughage slows calorie absorption and helps to move food quickly through the intestine, with less total absorption.  Eat as much raw wheat bran as you can stomach, and think of it as a negative-calorie food.  A vegetarian raw foods diet is not for everyone, but if you can live with it, it is a sure way to lose weight.  Raw foods are poorly absorbed, and from our perspective, that’s a good thing.


The Bottom Line

Weight gain or lost is probably a good proxy for how well you are controlling your glycemic load.

Remember that weight is relative to your genetic body type.  A person who has a genetic disposition to being overweight is not at higher risk for diseases of old age unless he actually eats too much.  Conversely, people who are blessed with a “thin metabolism” can be unhealthy if they overeat, even if they don’t look fat.  If you tend to being overweight and have to diet and work out constantly to keep your weight just above “normal”, you are in the optimal position for health and longevity. [Read more]

I suggest that you use your instincts and your experience to lower your glycemic response to food, and not to waste time with the mysteries in the tables of glycemic index.  First cut out sugar, then work on reducing starch.  Refined carbohydrates are worse than whole grains, but in my own diet I’ve eliminated all bread, pasta, rice, grains, and potatoes—so I can attest it’s possible.  Beans and nuts are good protein sources if, like me, you don’t eat meat.  My vegetarian version of the no-starch diet leads to moderate ketosis, less extreme than in a meat-based diet.

Combine fiber into every meal.  Eat large salads and other green vegetables.  Add raw wheat bran and/or resistant starch to your foods.

Everyone should be supplementing with magnesium.  Consider chromium and metformin as well.

Eat soups.  Drink water before and during meals.

Fasting is a powerful way to preserve insulin sensitivity.  Concentrating all your daily eating in an 8-hour window counts as a short, daily fast [read more].  Longer fasts of one to four days have major benefits [read more].

Exercise is the best general tonic for general health.  Exercise before meals in particular is a great habit to culture.  Even a minute or two of vigorous exercise that gets you panting makes a big difference in your glycemic response to a meal.


This article originally appeared here in Josh’s blog Aging Matters. Republished with permission.