Saturday, August 28, 2010

Saturated Fat, Glycemic Index and Insulin Sensitivity: More Evidence

Insulin is a hormone that drives glucose and other nutrients from the bloodstream into cells, among other things. A loss of sensitivity to the insulin signal, called insulin resistance, is a core feature of modern metabolic dysfunction and can lead to type II diabetes and other health problems. Insulin resistance affects a large percentage of people in affluent nations, in fact the majority of people in some places. What causes insulin resistance? Researchers have been trying to figure this out for decades.*

Since saturated fat is blamed for everything from cardiovascular disease to diabetes, it's no surprise that a number of controlled trials have asked if saturated fat feeding causes insulin resistance when compared to other fats. From the way the evidence is sometimes portrayed, you might think it does. However, a careful review of the literature reveals that this position is exaggerated, to put it mildly (1).

The glycemic index, a measure of how much a specific carbohydrate food raises blood sugar, is another common concept in the diet-health literature. On the surface, it makes sense: if excess blood sugar is harmful, then foods that increase blood sugar should be harmful. Despite evidence from observational studies, controlled trials as long as 1.5 years have shown that the glycemic index does not influence insulin sensitivity or body fatness (2, 3, 4). The observational studies may be confounded by the fact that white flour and sugar are the two main high-glycemic foods in most Western diets. Most industrially processed carbohydrate foods also have a high glycemic index, but that doesn't imply that their high glycemic index is the reason they're harmful.

All of this is easy for me to accept, because I'm familiar with examples of traditional cultures eating absurd amounts of saturated fat and/or high-glycemic carbohydrate, and not developing metabolic disease (5, 6, 7). I believe the key is that their food is not industrially processed (along with exercise, sunlight exposure, and probably other factors).

A large new study just published in the American Journal of Clinical nutrition has taken the evidence to a new level (8). At 6 months and 720 participants, it was both the largest and one of the longest studies to address the question. Participants were assigned to one of the following diets:
  1. High saturated fat, high glycemic index
  2. High monounsaturated fat, high glycemic index
  3. High monounsaturated fat, low glycemic index
  4. Low fat, high glycemic index
  5. Low fat, low glycemic index
Compliance to the diets was pretty good. From the nature of the study design, I suspect the authors were expecting participants on diet #1 to fare the worst. They were eating a deadly combination of saturated fat and high glycemic carbohydrate! Well to their dismay, there were no differences in insulin sensitivity between groups at 6 months. Blood pressure also didn't differ between groups, although the low-fat groups lost more weight than the monounsaturated fat groups. The investigators didn't attempt to determine whether the weight loss was fat, lean mass or both. The low-fat groups also saw an increase in the microalbumin:creatinine ratio compared to other groups, indicating a possible deterioration of kidney function.

In my opinion, the literature as a whole consistently shows that if saturated fat or high glycemic carbohydrate influence insulin sensitivity, they do so on a very long timescale, as no effect is detectable in controlled trails of fairly long duration. While it is possible that the controlled trials just didn't last long enough to detect an effect, I think it's more likely that both factors are irrelevant.

Fats were provided by the industrial manufacturer Unilever, and were incorporated into margarines, which I'm sure were just lovely to eat. Carbohydrate was also provided, including "bread, pasta, rice, and cereals." In other words, all participants were eating industrial food. I think these types of investigations may be limited by reductionist thinking. I prefer studies like Dr. Staffan Lindeberg's paleolithic diet trials (9, 10, 11). The key difference? They focus mostly on diet quality, not calories or specific nutrients. And they have shown that quality is king!

* Excess body fat is almost certainly a major cause. When fat mass increases beyond a certain point, particularly abdominal fat, the fat tissue typically becomes inflamed. Inflamed fat tissue secretes factors which reduce whole-body insulin sensitivity (12, 13). The big question is: what caused the fat gain?


Jenny said...

Ludicrously stupid in my opinion. Carbohydrates have a very different impact on the health of people with rock solid blood sugars (A1c 4.3-4.9)who can eat all they want without health impacts, than they do on people with higher ones ranging into the pre-diabetic range (usually undiagnosed) who risk glucose toxicity, and they are toxic to many organ systems in anyone diabetes.

Without knowing starting blood sugar status, and breaking out results by that status the results are meaningless.

Plus it takes a good 10 years of exposure to significantly high blood sugars to produce visible damage to health. The impact of those carbs won't show up in just six months.

Finally, all diets studied result in very different test results at 6 months than they do at two or three years.

Who are the imbeciles who come up with these badly designed studies and why aren't they doing the menial jobs their limited intelligence would make more appropriate for them?

Amy Alkon said...

Just to echo what Jenny said with my boyfriend and myself as examples: Although I've cut out carbs (save for the few I get in cheese and green vegetables), when I ate them, I was maybe 8-10 pounds heavier than I am now (I'm now effortlessly slim, and weigh what I did in high school).

My boyfriend, on the other hand, can't eat a cracker without putting on weight. He agreed to an experiment, to eat exactly as I said for a month (only bacon, eggs, hamburger, steak, chicken with the skin on, and coffee and spring water), and lost 17 pounds sans exercise. And his blood pressure went from high to near normal. He now eats dairy and vegetables and drinks some wine, and has lost more weight and maintained that weight loss but without crazy exercise or being hungry.

John said...

Hi Stephan,

I think insulin resistance is a pretty tricky topic. Mice on high fat diets are quite non-responsive compared to low fat mice when injected with much insulin* (and hence don't die--as much). This would imply resistance, but of course, as you know, receiving insulin during times of low blood sugar and insulin sensitivity is a bad idea. I think Peter has a great series of posts on physiological insulin resistance.

Also, it's important to factor in cell "type." Glutamine has been shown to increase fat cell insulin resistance, while perhaps reducing it in muscle cells.* Maybe we want insulin resistant fat cells with muscle cells' sensitivity varying according to blood sugar?

*Sorry, I don't have my computer with my references.

Chris Kresser said...

Yep, this is a pretty complex and controversial issue.

I agree with Jenny that the glycemic control of subjects has to be considered when trying to answer the question "do dietary carbohydrates contribute to insulin resistance?".

The Kitavans, certain Asian populations, and even plenty of Americans and Europeans can eat all the natural carbs they want without much impact. But those who already have dysglycemia may not be able to tolerate even a moderate amount of carbohydrate - regardless of the source.

There are so many variables to consider. I believe stress is one of the most important, and often overlooked. I'm not just talking about psychological and emotional stress, but also physiological stress that occurs with chronic, systemic inflammatory conditions like a leaky gut or autoimmune condition (both incredibly prevalent today). Stress alone is entirely capable of wreaking havoc on glycemic control, and with the recent research linking inflammation to the pathogenesis of type 2 diabetes, it's virtually certain that stress also plays a significant role.

This would explain, at least in part, why certain traditional cultures can eat so much carbohydrate without problems, whereas people living modern lifestyles cannot. It's not the carbohydrates that are causing the blood sugar dysregulation, but the blood sugar dysregulation that is causing the carbohydrate intolerance.

So what causes the blood sugar dysregulation in the first place, aside from stress? I agree with Stephan and many others that modern foods like industrial seed oils, HFCS and refined grains are the most notable dietary contributors. I've also come to believe that environmental toxins play an increasingly big role, via inflammation.

Peter said...

Jenny, LMAO at Susan Jebb being described as an imbecile. Boy, do you know how to insult imbeciles! That's without reading the study.

Stephan, I think insulin resistance is a very slippery concept. As Amy illustrates, it's easy to side step but what about causation or even reversal????? To me reversal means you can eat 2 large pizzas with a 2h BG of 100mmol/l after having been clincally diabetic. I don't think fat loss alone can do this.


Anonymous said...

Like I wrote on Hyperlipid in response to his recent blog post,

"Don't blame the processing when it's the substance itself that is poisonous. We don't hydrogenate animal fat. We render animal fat. It's still "processing" but the end product conserves its healthful properties.

Which brings another point, if even with processing a substance conserves its healthful property, isn't it reasonable to assume that an unhealthful substance also conserves its unhealthful property?

We hydrogenate vegetable oil. Maybe, just maybe it's the vegetable oil that's killing us.

Processed arsenic. It's still arsenic."

Do you truly believe it's the processing and not the substance itself that kills us? We've been processing our food for a long time yet only recently has it started killing us like that. Maybe it's what we eat, not how we prepare it.

If you feed arsenic to everybody, then it doesn't matter if one eats refined arsenic while the other eats whole arsenic, does it?

Then there's the time element. If you test over a period that is too short, then the result will show no effect. Taubes noted that it takes about 20 years for diabetes to develop in a population. Then that's the time scale we must test the hypothesis on.

But then if the intake is high enough, it could take as little as a month. That's where processing comes in. When it's processed, especially refined, then the intake can be as high as we want.

It's about the substance itself, then the quantity, then the time scale.

qualia said...

what, i thought it's pretty much clear by now that it's the fructose load in a hyper-caloric context?

Tucker Goodrich said...

Stephan, I'm going to guess that part of your issue with this study is with the dietary variables. I'm guessing the saturated fat they used likely wasn't coconut.

If you were going to design this study, how would you set it up? Let's assume you could change any of the variables, including duration.

Richard said...

It strikes me, reading the research review and the comments, that clearly poisonous foods, such as wheat and PUFA, ought to be avoided by sensible people. The problem or damage caused is not something that can be quickly determined, on a six month basis, precisely BECAUSE the problems take years to develop. That is, of course the changes over a short period are inconsequential: that is exactly the problem. The body does not know what is killing it, because there is no immediate reaction or sensation of illness.

In turn, this is precisely BECAUSE people are designed, just about like dogs, to eat almost anything. On a short term basis people can live on wheat or turnips or fruit. That is one reason why people have been so successful as a species.

To do a short term test merely proves that people are omnivores. But we knew that.

Really, the US and many other countries are in a massive, although uncontrolled, experiment with the effects of PUFA and wheat. What we do know, on a long term basis: it makes people sick. people get heart disease, cancers, Alzheimers, and diabetes of course.

The whole point is that these diseases come about because of chronic abuse of the body's ability to cope with a less than optimal diet.

My thought, on all of this, is that aside from wheat and PUFA, people can probably tolerate a wide range in their diet, so long as the basis of the nutrition is saturated fat and not flour, which is simply too accessible as nutrition.

Scott Pierce said...

It seems the tide is changing regarding glycemic index and people are beginning to speak in terms of "carbohydrate load" or total carbs consumed over I guess a 24 hour time period. But what is this based on? I have come to the point of just trying to avoid wheat, sugar and vegetable oil (and living notably healthier for it) and leaving it at that until a more concrete, verifiable explanation comes along.

Innovator said...

May I suggest Clinical Ramifications of Malabsorption of Fructose and Other Short-chain Carbohydrates ["FODMAPs"].

IANAN, but may I ask the Nutritionists here, is a blindingly simple Occam's Razor at work in the conclusion:

"After years of unsatisfactory dietary manipulation, it is evident that a low FODMAP diet is successful in providing relief from functional gut symptoms, whether in the setting of a FGD or IBD. More research is required into the FODMAP content of foods around the world, as differences in food composition are known to vary. Refinement in the dietary advice given will improve the diet, especially with regard to completion of the analysis of all foods for their FODMAP content. Further studies are required to correct the limitations associated with current food composition tables, to refine this dietary approach, and to better explore its application in patients with irritable bowel syndrome and other disorders such as inflammatory bowel and celiac disease."

In short, if this Occam's Razor is right, then metabolic syndrome arises from FODMAPs, and in turn, the many human diseases so far associated with FODMAPs?

To be clear, if this Occam's Razor is right, then our other debates -- whole high-carb/low carb, saturated/unsaturated simply misses the point. It is FODMAPs are the cause of everything.

Please don't flame me, IANAN and I'm just asking!!!

Ned Kock said...

Great post Stephan.

What do you think as the theory of lipotoxicity as a possible explanation for insulin resistance? Here is a post with some links:

Consistent with your last point, this theory (lipotoxicity) puts forth abnormal fat metabolism as a possible precondition for type 2 diabetes. It contradicts the "tired pancreas" theory that we hear so much about.

It does seem to me that the glycemic index is not a very good variable in a study like this. The glycemic load would probably be more appropriate. And, indeed, they should have controlled for insulin resistance status.

Helen said...

I'm still trying to sort out all of this stuff regarding my own glucose control situation and this information is quite helpful.

There's a lot of discussion on this loose circle of blogs (Stephan's, Peter's, Jenny's, and a lot of the paleo/primal blogs) about various fats and carbohydrates, their forms, quantities, and effects, but what about that other macronutrient, protein? Could protein source, quality, and quantity play a role?

Innovator said...

Tried to post before but left the Wikipedia summary out. Could a low FODMAP diet be the simple answer?

Chris Kresser said...

I think inflammation plays a significant role, and although it has been a focus in the diabetes literature for 10 years, it's still not widely recognized as a contributing factor in mainstream medicine.

Several studies have found that inflammation causes insulin resistance and dysregulation of lipid and glucose metabolism. This happens in both obese and nonobese people. (

This theory is quite compatible with the theory of lipotoxicity that Ned wrote about and mentioned in his comment above. It also provides additional insight on why n-6 PUFA, refined carbohydrates and excess fructose are all associated with rising rates in metabolic syndrome and type 2 diabetes.

Also, recent research has shown that altered gut microbiota and bacteria LPS are major contributing factors to both obesity and diabetes. This is yet another mechanism implicating the modern lifestyle's role in the pathogenesis of these conditions. Stress, antibiotic overuse, poor diet, lack of breastfeeding during infancy and medications like birth control all have a significant adverse effect on the gut microbiota, and thus could predispose to type 2 diabetes and metabolic syndrome.

guyberliner said...

One thing I'm wondering about, Stephan: A lot of people who subscribe to Cordain's "Paleo Diet" avoid dairy products, which are said to cause "insulin spikes" (due to long chain fatty acids in milk??). I know you've said that an "insulin spike" does not imply anything to the effect of "insulin resistance." Still, I would find more elaboration enlightening. I mean, why shouldn't we expect that routine exposure to something that causes rapid elevation of insulin to have some adverse effect?

Stancel said...

Google "insulin spike" and you will see it is merely a trendy bodybuilding buzzword.

Diabetics know full well how VITAL insulin is, because they fail to produce it at optimal levels and the body suffers from high blood sugar.

Insulin is our body's natural way of keeping blood sugar levels down. But low blood sugar is not a good thing either. Diabetics can experience traumatic episodes of both high and low blood sugar, because their ability to produce glucagon is impaired.

Insulin and glucagon work together to keep blood sugar at a normal level.

Unknown said...

Helen wrote,

"There's a lot of discussion on this loose circle of blogs (Stephan's, Peter's, Jenny's, and a lot of the paleo/primal blogs) about various fats and carbohydrates, their forms, quantities, and effects, but what about that other macronutrient, protein? Could protein source, quality, and quantity play a role?"

Helen, this is a great question and I would strongly suspect that there is some benefit to getting your protein from high-quality animal sources such as meat, eggs, and dairy.

Innovator said...

Stephan, you provide references for the following:

(2, 3, 4) "controlled trials as long as 1.5 years have shown that the glycemic index does not influence insulin sensitivity or body fat gain."

(5, 6, 7) "examples of traditional cultures eating absurd amounts of saturated fat and/or high-glycemic carbohydrate, and not developing metabolic disease."

(8) "neither saturated fat nor high glycemic carbohydrate influence insulin sensitivity in humans, at least on the timescale of most controlled trials."

(9, 10, 11) "studies like Dr. Staffan Lindeberg's paleolithic diet trials . The key difference? They focus mostly on diet quality, not calories or specific nutrients. And they have shown that quality is king!"

(12, 13) "When fat mass increases beyond a certain point, particularly abdominal fat, the fat tissue typically becomes inflamed. Inflamed fat tissue secretes factors which reduce whole-body insulin sensitivity. The big question is: what caused the fat gain?"

However, you've also noted studies showing that fructose reduces insulin sensitivity in humans, along with many other harmful effects. Cite

It would seem that all these references are at least consistent with the (very simple) hypothesis that fat gain is caused by: 1) the proportion of a) glucose to fructose in sugar and many fruits and juices and b) glucose to fructans in wheat; along with 2) the limit on the amount of fructose/fructans consumed per sitting (25g to 50g)? Cite

Innovator said...

For clarity, the direct action could be (hypothetically) that fat gain is caused by the malabsorption of fructose and fructans in the intestine, leading to gut permeability, rogue fructose and unsaturated fatty acids in the blood stream, and the growth of adipose tissue, aka the "fat organ".

As further narrative, the "fat organ" could be genetic selected to survive winter, increasing in size just when pre-winter fructose is available in large quantities.

Helen said...

J.A. Deep -

I think it's important to keep in mind there can be many routes to fat gain. I gained the most weight (i.e., I'd never had a weight problem before, but recently gained 20 pounds in a short period), while consuming the least fructose in my life (no sweeteners, rarely any fruit), avoiding seed oils, and avoiding refined grains, and indeed, most grains. I was also exercising moderately to vigorously 20-30 minutes daily, and was not a couch potato - or desk-worker - the rest of the time.

I still don't know why I gained the weight, but I've now lost 10 of those pounds while adding back to my diet 1-2 pieces of fruit a day as well as legumes, which I'd also been minimizing in an effort to follow a more paleo diet.

Anonymous said...

REally interesting thank you!!!

Adolfo David said...

The key is not glycemic index, but glycemic load, 'refined' carbs and total carbs consumed.

Alex said...
This comment has been removed by the author.
Innovator said...


Yes, I agree with you. I think the conventional theory that a sufficiently hypercaloric diet can cause fat gain still has some validity, even if fructose/fructans are appropriately limited.

What I'm trying to understand is the sudden epidemic of fat gain in affluent societies. In the simple hypothesis, the epidemic is caused by eating the many industrial foods that grossly exceed the "natural" fructose/fructans content of traditional or "paleo" diets, which is:

a) no worse than a 45/55 proportion of fructose/fructans to glucose and

b) a total fructose/fructans load or limit per sitting (25 g to 50g) of fructose/fructans.

What makes the epidemic worse is that fructose and fructans are hidden in a great many industrial foods, and there are no labelling requirements! For a more complete list of "forbidden fruit", see the FODMAP study.

I'd also be interested in knowing more information on the foods you ate that resulted in the unexpected fat gain. Perhaps you could email me with more information at johnadeep at gmail dot com? Thanks for your reply!

Stephan Guyenet said...

I don't think fructose polymers should be in the same category as sugar and other sources of free fructose, so I don't like the FODMAP concept.

Malabsorption of fructose monomers and fructose polymers is different. Monomeric fructose that isn't absorbed gets fermented in the small intestine, where it shouldn't. Polymeric fructose (inulin, FOS) isn't supposed to be absorbed, so using the term "malabsorption" is misleading. Polymeric fructose is a fermentable fiber that gets broken down by bacteria in the large intestine, where fermentation is supposed to occur. Inulin and FOS don't contribute a meaningful amount of absorbed fructose because the colon (where they are broken down) doesn't absorb fructose. It gets eaten by bacteria.

If you eat too much, it gives you gas. But that doesn't mean it's harmful in moderation.

Stephan Guyenet said...

Hi guyberliner,

I'm not aware of any evidence that post-meal insulin spikes are harmful. To the contrary, traditional cultures eating very large amounts of carbohydrate are lean, insulin sensitive and have low levels of atherosclerosis, CHD and diabetes. That doesn't prove that eating 70% carb is the ideal diet, but it does show that under good conditions, the human metabolism can tolerate it without obvious ill effects.

I think there's a lot of confusion between fasting insulin and post-meal insulin. High fasting insulin is usually bad, and can indicate insulin resistance, while the significance of eating foods that increase post-meal insulin to long-term health is unclear as far as I know.

Chris Kresser said...

Also, while the FODMAP concept may alleviate symptoms people experience d/t fructose malabsorption, it doesn't address the root of the problem (why people can't absorb fructose properly in the first place).

I always prefer to identify the underlying mechanism and address it at that level with the hope of restoring normal function.

Innovator said...


Yes, fructans can be eaten in moderation (7.5-10g?), but the modern industrial diet of wheat -- in cereal, bread and pasta -- may cause symptoms of IBS, and restricting fructan intake in a diet may reduce symptoms in a variety of gastrointestinal disorders. (Id.)

And yes, fructans cannot be converted to fructose in the small intestine, but there is much we do not know about where fructans end up -- in the large intestine! In one example, the fatty acids created from fructans by the gut bacteria in the large intestine do eventually find their way into our bloodstream, and the fatty acids released in this way may not trigger an appetite suppressing response. Cite

Indeed, whatever we have so far learned about the ten trillion strong population of ancient creatures in our gut only shows how little we know, or may ever know.

What I don't see any disagreement over here is the correlation between fat gain and a diet of a) a proportion of fructose/fructan to glucose greater than 1/2; and/or b) greater than the 25g limit per sitting of fructose/fructans.

Dr. Art Ayers said...

Thanks for another entertaining post.

BTW, I didn't know that you have published so many articles on another health and diet site:

They are exactly the same as on Whole Health Source, without your name attached, and with ads. For example, search the site for "saturated fats".

Billy Oblivion said...

So that study did:

1. High saturated fat, high glycemic index
2. High monounsaturated fat, high glycemic index
3. High monounsaturated fat, low glycemic index
4. Low fat, high glycemic index
5. Low fat, low glycemic index

What about High Saturated Fat, Low Glycemic Index? Hmmm?

Unknown said...

I enjoy reading your posts. Excellent resource for those interested in various health issues. Keep up the good work

Juliana Lee

karl said...

A lot of people are gluten sensitive - I ran a test from

and found I was. I suspect that there is something about gluten that causes systemic inflammation

Todd Hargrove said...


Check out the cover story for the recent PCC newsletter: sat fat healthy, corn oil bad:

Stephan Guyenet said...

Whoa, I haven't seen that yet. Awesome.

robrob said...

good points stephan, I agree with you sugar is not the problem it is the malnorishment that tends to come with it due ot dieting for weight loss and not eating enough good stuff.

it is also known that eating saturated fat with carb lowers the gi of that carb even more slowing absorption by coating the carbs, carbs that are absorbed to fast in high amounts cause problems when your malnorished from poor diet or dieting (hypocal) diets. we know that sugars like glucose, fructose are very reactive to oxygen right? this means if the body doesn't handle sugar well you get inflmmation, cellular damage from sugar sticking to cells and oxidizing them.

in fact body builders will inject insulin to help store more protein in muscles as well as up their glycogen stores, insulin is a anabolic steriod you know. I don't think cells become insulin resistant, they become glucose resistance because they do not have the minerals and vita ect to handle the oxidative nature of the glucose. but only my opinion.

in fact I would guess that insulin being a protein molecule gets coated and oxidized too doesn't it if sugar sticks to it?

if you have all the nutrients you need you can handle glucose (rocket fuel)and if you eat enough saturated fats (natural not manmade from veggie oil sat fat) less fat cells are needed to provide the sat fats you need and in theory should reduce in number due to apoptosis in fact you need more fat cells if your low fatting and have low mineral consumption like calcium and magnesium or fat soluable vitamines. our bodies hoard what we dont get enough of. can't imagine how anyone can absorb cal vita d etc without saturated fat intake?

it will hoard fat soluable vita in fat cells, it will hoard fat if you don't eat enough of it converting all those carbs to fat.

your body can become overwhlemed if your eating alot of carbs in protecting the body from its oxidative nature. I think if you stick to whole carbs with all their phytonutrients and natural fats (organic is best?) you should be okay. but not being an expert (except at losing weight permanently over a 35 year period) I can only guess, but frankly I am sick of all the false info/hype perpetuated that cause such problems with knowing what ot eat.

Santhy George said...

I grew up on refined sunflower oil. Last year we did some tests, and my fasting insulin was 4, my husband's was 11. Vitamin D was rock bottom for both. We switched to coconut oil for cooking and tried to get more sunlight. Now eight months later, our Vitamin D has gone up, but my insulin is 9 and my husband's 19! The only things we have changed are adding more healthy fats, switching to raw milk from pasteurised and cutting down slightly on carbs and eat outs. I'm shocked and disheartened, not to mention confused, any insight would be greatly appreciated.

Unknown said...

I think that That Sugar Film answers many of the questions brought up here over the years. Very controversial topic. I am in search of the meaning of GI and find the numbers all over the place. I am pretty sure that GI is totally pointless.