Glycaemic Index Effects on Fuel Partitioning in Humans – Research Review

Title and Abstract

Diaz EO et. al. Glycaemic index effects on fuel partitioning in humans. Obes Rev. (2006) 7:219-26.

The purpose of this review was to examine the role of glycaemic index in fuel partitioning and body composition with emphasis on fat oxidation/storage in humans. This relationship is based on the hypothesis postulating that a higher serum glucose and insulin response induced by high-glycaemic carbohydrates promotes lower fat oxidation and higher fat storage in comparison with low-glycaemic carbohydrates. Thus, high-glycaemic index meals could contribute to the maintenance of excess weight in obese individuals and/or predispose obesity-prone subjects to weight gain. Several studies comparing the effects of meals with contrasting glycaemic carbohydrates for hours, days or weeks have failed to demonstrate any differential effect on fuel partitioning when either substrate oxidation or body composition measurements were performed. Apparently, the glycaemic index-induced serum insulin differences are not sufficient in magnitude and/or duration to modify fuel oxidation.


The glycemic index (GI) of foods is yet another place where endless argument and debate exists in the world of nutrition, especially as it applies to body composition.

In the early days of nutrition, as many may recall, carbohydrates were rather simplistically divided into simple and complex sources with the even simpler belief that ‘simple = bad’ and ‘complex = good’.  While this was applied to general health and such, one of the major applications and concerns over carbohydrate intake had to do with diabetic meal planning.

When it became clear that simple vs. complex was insufficient, researchers went looking for more accurate methods of measuring the differences between carbohydrates.  Sometime in the 80’s, the GI was born.

Conceptually, the GI refers to the blood glucose response to a given carbohydrate food.  A little more technically, the GI of a food relates to the area under the curve (AUC for nerdy types) of blood glucose versus time after the ingestion of a fixed amount of a test food.

Researchers would first test a fixed amount (currently 50 grams digestible carbohydrate) of some standard food, they originally used pure glucose but switched to white bread years later.  The blood glucose response to that standard food was defined as having a GI value of 100.  I want to make it clear that this value has no inherent meaning, it was simply a defined value.

Then other foods were tested, again 50 grams of digestible carbohydrate (perhaps baked potato or cereal) were given by itself after an overnight fast and the blood glucose response was measured. The GI of that food was then defined relative to the 100 value of the test standard.  So a GI of 80 meant that the test food had 80% of the blood glucose response of the test food; a GI of 120 means that it had 120% the blood glucose response of the test food.  Again, keep in mind that these values don’t really ‘mean’ anything, they are just relative value.

In any case, from the standpoint of diabetic meal planning, the GI seemed important as it would let diabetics decide which foods would have the best effect on blood glucose levels without causing problems.  Of course, for a variety of reasons, the GI concept was also adopted by athletes and the physique obsessed.

I’d note that there is much more to the GI than I have space to go into here, I’ll be writing a full article on it soon enough.  Sufficed to say that GI becomes much more complicated when you start mixing foods together, or the person isn’t fasted (e.g. you’ve eaten a meal).  Even the aerobic training status of a person modifies the GI as I detail in the research review The Influence of the Subject’s Training State on the Glycemic Index.

In any event, the big argument over the GI of foods at least with regards to body composition usually involves the insulin response and potential impact on things like fat mass and fuel utilization.  It was usually inferred that a higher GI value (remember, larger and/or longer blood glucose response) meant a bigger insulin response and for the physique obsesses, insulin equals badness.

I’d note that things aren’t this simple and at least one study suggests that foods with a lower GI may have a lower GI because of a LARGER initial insulin response as detailed in Different Glycemic Indexes of Breakfast Cereals Are Not Due to Glucose Entry into Blood but to Glucose Removal by Tissue.


The Paper

But I’m getting off topic.  What today’s paper looks at is the idea of whether differences in the insulin response (from foods differing in GI) actually have meaningful differences in terms of their effect  on insulin, fuel utilization or body composition.

Because that’s the real issue: there’s no debate that foods differing in GI generate different blood glucose responses, there is indication that this impacts on the insulin response.  But the bottom line question is whether those differences in hormonal response actually meaningfully affect anything.

In looking at the topic, the researchers examined a variety of different data sets including more acute studies along with those looking at actual changes in body composition.

The short-, mid- and long-term studies typically examined things like blood glucose, insulin, blood fatty acid levels, carb and fat oxidation and/or energy expenditure over periods ranging from 6-24 hours (or longer) after the ingestion of foods or meals differing in GI.  I’m not going to detail each and every one but, with one or two exceptions, the majority simply found no significant difference in things like fatty acid suppression or fuel oxidation despite significant differences in blood glucose and insulin response.  Even longer term intervention studies of 30 days to 10 weeks found no significant impact on weight or body composition for diets designed with different GI levels.

So in terms of data directly examining the topic, the researchers comment that:

High fasting serum insulin concentration or high first-phase serum insulin response to intravenous glucose has been proposed as a risk factor for weight gain.This may have led Ludwig to state that ‘functional hyperinsulinemia associated with high-GI diets ma promote weight gain by preferentially directing nutrients away from oxidation in muscle and towards storage in fat’.  Evidence for this hypothesis is still lacking since no effects of GI on fuel partitioning have been demonstrated to date.

Of course, there are studies suggesting that lower GI diets generate more weight loss than higher GI but there are often subtle confounds including the fact that typically GI is not the only difference between diets.  Often, with changes in the GI come differences in fiber intake, energy density of the diet, at least one study I can think of changed protein intake between groups.  So concluding that the GI per se is having an impact is incorrect.

It’s worth mentioning that low GI foods are often claimed to better control appetite than higher GI foods.  And about half of the studies examining this do find this effect, with the other half finding no real effect.  As I discuss in Is a Calorie a Calorie, this is another confound, if eating lower GI foods causes someone to eat less total food, they will tend to lose weight but it’s not due to the GI of the foods per se.  As well, if high GI foods make people eat more, they will tend to gain fat, as a function of eating more.

But this is far different than claiming that high GI foods will make someone gain fat (and/or lose muscle) at an identical caloric intake, an argument that does not seem to be supported by the above studies looking at fuel utilization directly.

I should note that there is at least some indication of an interaction between high and low GI diets and insulin sensitivity, as I discuss in Insulin Sensitivity and Fat Loss, at least one study has shown that people with insulin resistance lose more weight with lower GI diets while those with higher insulin sensitivity actually do better with higher GI diets.

Wrapping up the paper, the researchers examine the impact of insulin on fuel utilization in general terms mentioning that both the magnitude and duration of insulin response has the potential to affect fuel and fat utilization.  Without detailing all of the information, they conclude

Taking into account all of the above arguments, we speculate that under postprandial conditions, GI-induced serum insulin differences are not sufficient in magnitude and/or duration to modify fat oxidation.

Given that even tiny increases in insulin pretty much shut off fat oxidation, this actually isn’t surprising.  As I discussed in The Stubborn Fat Solution, even fasting levels of insulin inhibit fat cell lipolysis by 50% from maximal rates and almost any increase in insulin is sufficient to shut off lipolysis completely.

As this research review points out, it simply doesn’t appear that some vs. more insulin has any major impact on this.  I’d note, mind you, that fat cell metabolism can also be impacted by eating even if insulin doesn’t increase; oral ingestion of pure dietary fat also shuts down lipolysis but that’s beyond the scope of this article.


Summing Up

Ok, what does this all mean and what am I saying?  First let me clarify what I am not saying.  I don’t want folks to read this as a suggestion to go scarf down as much high GI, refined stuff as they can put down their gullets.  That would be asinine although I’m sure someone will manage to read this article as advocating exactly that.

Even if there is no significant impact on acute fuel utilization, fat oxidation or storage in the short-term for higher GI vs. lower GI foods, that doesn’t suggest that eating nothing but high GI foods is the way to automatically go.

As I noted above, for many people lower GI foods tend to control hunger better and, in general, lower GI foods are typically less refined, contain more fiber and nutrients, etc.  Even if there are no significant differences in how they impact on fuel utilization, health should always be a consideration.  There are other issues such as the glycemic load (a topic I’ll discuss in some detail later) and overall health as well.

But from the standpoint of fuel utilization, fat oxidation and the rest, there appears to be no meaningful differences in the impact of higher vs. lower GI foods in humans (the study also examined animal data where things, as usual, are different but simply don’t apply to non-rats).

I find that many people become nearly clinically insane over the issue of GI, it becomes a level of absolute dietary extremism that is simply not necessary.   For these folks, anything without a super low GI is a devil food and will cause one’s muscles to instantly fall off and be replaced by body fat.

And as with so many other topics, that’s just not the case.  Small differences in GI, especially within the context of mixed meals and lean individuals who are training regularly appear to have no significant impact on overall fuel utilization, fat oxidation, or anything else.




12 thoughts on “Glycaemic Index Effects on Fuel Partitioning in Humans – Research Review

  1. Lyle

    What about the effect on insulin blood glucose fuel partioning and body composition in response to a zero carbohydrate or low carbohydrate meal?


  2. That would be a different article then, wouldn’t it? Since it has nothing to do with what I was talking about today.

  3. Perhaps GI issues produce more confounded (and consequently less readily identifiable) results on the particular metrics the paper (and you) have chosen; but there are other effects worth attending to. See e.g.

    “Meta-analysis of the health effects of using the glycaemic index in meal-planning”
    British Journal of Nutrition (2004), 92, 367–381

    While in general I laud your inclination to seek the mean wrt nutritional hysteria, more and more the posts here seem shying away from taking any kind of substantive stand; in other words, you’re right, your head won’t explode if you eat a candy bar. But compelling reasons exist to prefer low- to high-GI all else being equal, and reading this post you wouldn’t walk away with that conclusion. That’s too bad.

  4. This isn’t exactly the same question, but it sounds like glycemic index is related to partitioning in a UD2 style refeed, at least after the first few hours. Except here it’s more likely to be about glycogen storage vs use as fuel, rather than fat gain.

    It reminds me of the vinegar glycogen storage thing – the same kind of effect I guess. Some diabetic people take vinegar with starches because it supposedly slows gastric emptying (lowering glycemic index).

    Ever think about taking tryptophan + complex carbs + vinegar before bed for a 24hr refeed? Apparently it’s the only amino acid that slows things down, plus it puts you to sleep.

  5. Another neuance that just cuts me up is the glycemic load deal. I laugh my head off when i see someone scarfing down 800 calories of whole wheat pasta with no problem but would crap their pants at the thought of eating a small serving of white rice or a few grapes or pineapple. We all remember the guys who wouldn’t eat a banana to save their life back in the day.

    I am also surprised to see the variance of a food’s G.I. rating from one study to another. They mostly corellate modestly, but i remember a few foods changing pretty significantly.

  6. Working in the food industry, I’ve had the opportunity to talk to several diabetics who use the idea of GI load when eating. Some people report that at a regular meal they will simply eat the protein first — this allows them to then consume some carbs with very little insulin response.

    If you watch a group of people at a meal (say dinner), inevitably they will consume a bunch of carbs (liquid or otherwise) before they jump into the protein.

    Lyle, Thanks for the great articles

  7. Shane: as I clearly wrote:

    “As I noted above, for many people lower GI foods tend to control hunger better and, in general, lower GI foods are typically less refined, contain more fiber and nutrients, etc. Even if there are no significant differences in how they impact on fuel utilization, health should always be a consideration.

    Even if there is no significant impact on acute fuel utilization, fat oxidation or storage in the short-term for higher GI vs. lower GI foods, that doesn’t suggest that eating nothing but high GI foods is the way to automatically go.”

    Almost as if what you wrote was something I made sure to explicitly address.

  8. As a Type 2 with virtually no Phase 1 insulin response but still quite effective levels of Phase 2, GI serves to indicate things I really need to avoid. I can only glance briefly at a banana the other side of the room. Yet other Type 2s, presumably with a good Phase 1 and a reduced Phase 2, can get away with them, and other high GI foods in sufficiently small quantities, whereas a load of low GI carbs will be more deleterious for them.

    Another potential factor might be digestive enzymes, a small but significant number of diabetics including some Type 1s, have an extreme glycemic response to wheat which doesn’t occur with other grains (nondiabetics too.)

    Personally I do best by avoiding most overt carbs altogether, but on the occasions when I need to eat them adding fats and fibrous veggies and washing them down with red wine serves to reduce the GI enough that I can get away with more than I usually do. Then a fast walk while my insulin output is at its highest will help burn off the excess glucose, that’s another potential confounding factor – changing the rate of glucose disposal may appear to alter the GI

    It’s an indicator and no more IMO, individual glycemic responses and the resulting glucose and insulin curves can vary significantly as anyone with a BG meter can attest – even time of day and changed insulin resistance levels can have an effect.

  9. Do say high or low GI foods like Whole Wheat vs White Pancakes have a difference on a structured refeed day?

Leave a Comment