Monday, August 22, 2011

Refined carbohydrate-rich foods, palatability, glycemic load, and the Paleo movement

A great deal of discussion has been going on recently revolving around the so-called “carbohydrate hypothesis of obesity”. I will use the acronym CHO to refer to this hypothesis. This acronym is often used to refer to carbohydrates in nutrition research; I hope this will not cause confusion.

The CHO could be summarized as this: a person consumes foods with “easily digestible” carbohydrates, those carbohydrates raise insulin levels abnormally, the abnormally high insulin levels drive too much fat into body fat cells and keep it there, this causes hunger as not enough fat is released from fat cells for use as energy, this hunger drives the consumption of more foods with “easily digestible” carbohydrates, and so on.

It is posited as a feedback-loop process that causes serious problems over a period of years. The term “easily digestible” is within quotes for emphasis. If it is taken to mean “refined”, which is still a bit vague, there is a good amount of epidemiological evidence in support of the CHO. If it is taken to mean simply “easily digestible”, as in potatoes and rice (which is technically a refined food, but a rather benign one), there is a lot of evidence against it. Even from an unbiased (hopefully) look at county-level data in the China Study.

Another hypothesis that has been around for a long time and that has been revived recently, which we could call the “palatability hypothesis”, is a competing hypothesis. It is an interesting and intriguing hypothesis, at least at first glance. There seems to be some truth to this hypothesis. The idea here is that we have not evolved mechanisms to deal with highly palatable foods, and thus end up overeating them.  Therefore we should go in the opposite direction, and place emphasis on foods that are not very palatable to reach our optimal weight. You might think that to test this hypothesis it would be enough to find out if this diet works: “Eat something … if it tastes good, spit it out!”

But it is not so simple. To test this palatability hypothesis one could try to measure the palatability of foods, and see if it is correlated with consumption. The problem is that the formulations I have seen of the palatability hypothesis treat the palatability construct as static, when in fact it is dynamic – very dynamic. The perception of the reward associated with a specific food changes depending on a number of factors.

For example, we cannot assign a palatability score to a food without considering the particular state in which the individual who eats the food is. That state is defined by a number of factors, including physiological and psychological ones, which vary a lot across individuals and even across different points in time for the same individual. For someone who is hungry after a 20 h fast, for instance, the perceived reward associated with a food will go up significantly compared to the same person in the fed state.

Regarding the CHO, it seems very clear that refined carbohydrate-rich foods in general, particularly the highly modified ones, disrupt normal biological mechanisms that regulate hunger. Perceived food reward, or palatability, is a function of hunger. Abnormal glucose and insulin responses appear to be at the core of this phenomenon. There are undoubtedly many other factors at play as well. But, as you can see, there is a major overlap between the CHO and the palatability hypothesis. Refined carbohydrate-rich foods generally have higher palatability than natural foods in general. Humans are good engineers.

One meme that seems to be forming recently on the Internetz is that the CHO is incompatible with data from healthy isolated groups that consume a lot of carbohydrates, which are sometimes presented as alternative models of life in the Paleolithic. But in fact among influential proponents of the CHO are the intellectual founders of the Paleolithic dieting movement. Including folks who studied native diets high in carbohydrates, and found their users to be very healthy (e.g., the Kitavans). One thing that these intellectual founders did though was to clearly frame the CHO in terms of refined carbohydrate-rich foods.

Natural carbohydrate-rich foods are clearly distinguished from refined ones based on one key attribute; not the only one, but a very important one nonetheless. That attribute is their glycemic load (GL). I am using the term “natural” here as roughly synonymous with “unrefined” or “whole”. Although they are often confused, the GL is not the same as the glycemic index (GI). The GI is a measure of the effect of carbohydrate intake on blood sugar levels. Glucose is the reference; it has a GI of 100.

The GL provides a better way of predicting total blood sugar response, in terms of “area under the curve”, based on both the type and quantity of carbohydrate in a specific food. Area under the curve is ultimately what really matters; a pointed but brief spike may not have much of a metabolic effect. Insulin response is highly correlated with blood sugar response in terms of area under the curve. The GL is calculated through the following formula:

GL = (GI x the amount of available carbohydrate in grams) / 100

The GL of a food is also dynamic, but its range of variation is small enough in normoglycemic individuals so that it can be treated as a relatively static number. (Still, the reference are normoglycemic individuals.) One of the main differences between refined and natural carbohydrate-rich foods is the much higher GL of industrial carbohydrate-rich foods, and this is not affected by slight variations in GL and GI depending on an individual’s state. The table below illustrates this difference.

Looking back at the environment of our evolutionary adaptation (EEA), which was not static either, this situation becomes analogous to that of vitamin D deficiency today. A few minutes of sun exposure stimulate the production of 10,000 IU of vitamin D, whereas food fortification in the standard American diet normally provides less than 500 IU. The difference is large. So is the difference in GL of natural and refined carbohydrate-rich foods.

And what are the immediate consequences of that difference in GL values? They are abnormally elevated blood sugar and insulin levels after meals containing refined carbohydrate-rich foods. (Incidentally, the GL  happens to be relatively low for the rice preparations consumed by Asian populations who seem to do well on rice-based diets.)  Abnormal levels of other hormones, in a chronic fashion, come later, after many years consuming those foods. These hormones include adiponectin, leptin, and tumor necrosis factor. The authors of the article from which the table above was taken note that:

Within the past 20 y, substantial evidence has accumulated showing that long term consumption of high glycemic load carbohydrates can adversely affect metabolism and health. Specifically, chronic hyperglycemia and hyperinsulinemia induced by high glycemic load carbohydrates may elicit a number of hormonal and physiologic changes that promote insulin resistance. Chronic hyperinsulinemia represents the primary metabolic defect in the metabolic syndrome.

Who are the authors of this article? They are Loren Cordain, S. Boyd Eaton, Anthony Sebastian, Neil Mann, Staffan Lindeberg, Bruce A. Watkins, James H O’Keefe, and Janette Brand-Miller. The paper is titled “Origins and evolution of the Western diet: Health implications for the 21st century”. A full-text PDF is available here. For most of these authors, this article is their most widely cited publication so far, and it is piling up citations as I write. This means that not only members of the general public have been reading it, but that professional researchers have been reading it as well, and citing it in their own research publications.

In summary, the CHO and the palatability hypothesis overlap, and the overlap is not trivial. But the palatability hypothesis is more difficult to test. As Karl Popper noted, a good hypothesis is a testable hypothesis. Eating natural foods will make an enormous difference for the better in your health if you are coming from the standard American diet, and you can justify this statement based on the CHO, the palatability hypothesis, or even a few others – e.g., a nutrient density hypothesis, which would be closer to Weston Price's views. Even if you eat only plant-based natural foods, which I cannot fully recommend based on data I’ve reviewed on this blog, you will be better off.


Glenn said...

Nicely done, Ned.

gallier2 said...

Thank you, Ned at least some sanity back in the paleo-blogosphere.

Kindke said...

The problem is that the formulations I have seen of the palatability hypothesis treat the palatability construct as static, when in fact it is dynamic – very dynamic.

Exactly, I've been shocked how most people are not able to grasp that concept. There is nothing intriniscally rewarding about food X, its how your body REACTS thats important.

Its the physiological state that your body is in that matters.

I think the problem with carbohydrate based foods goes beyond GI and GL, its not easy to pick it out but I definitely have a problem with starch based foods.

It could be something to do with the incretin response.

Zooko said...


A very nice contribution to the discussion. Thank you!

Not to be a total Gary Taubes fan boy (okay, I admit: I'm a total Gary Taubes fan boy), but he made exactly the same point about testability at the end of "Good Calories, Bad Calories" (second edition, if I recall correctly).

He said the carbohydrate hypothesis has been criticized for being "simplistic", and he cited Occam's Razor, which is the principle that among alternate hypotheses, all of which have passed the tests so far, you ought to favor the simplest one.

Some people think that this is because the simplest one is more likely to be true, but that's not a good reason to use Occam's Razor. There are two good reasons to favor the simplest hypothesis:

1. As Taubes and you have said, a simple hypothesis is easier to test. It is easy to underestimate how important that is, especially if you don't have experience trying to implement those tests. A lot of even moderately complicated hypotheses are just never going to get properly tested, or at least not within the next few decades.

2. When you disprove a simple hypothesis, we all learn a lot more than when you disprove a complex one.

And, as you wisely point out at the end of this blog post, "which hypothesis should scientists be trying to disprove?" and "which hypothesis should I use to guide my meal planning?" may have different answers.



David Csonka said...


Regarding the table of GI/GL of various carbohydrate dominant foods - in the case of potatoes and yams, is the qualified in respect to the degree or type of cooking performed?

Cooking will certainly have an effect on starch availability, so I find this to be incredibly pertinent in a comparison such as this.


M. said...

Hi Ned,

The problem is that when Taubes keeps making statements such as nuts and cheese are fattening because of their carb/insulinogenic properties or that starchy vegetables like potatoes (low GL) will cause obesity in America, then a mythology is developed and perpetuated. This mythology needs to be cut away, but an almost monolithic form of “low-carbism” has evolved that is amalgam of various ideas that are sometimes contradictory to one another.

A few weeks ago Taubes was complaining about starchy vegetables and “carbs” from fruits and vegetables making up half of the new My Plate. This seems contradictory to your Glycemic Load argument. I think if Glycemic Load is the argument, then we should speak of the Glycemic Load argument instead of the Carbohydrate Hypothesis.

I think a big confounder in any type of glycemic measurement of a particular food item is that it is not the particular food item that necessarily matters; it is the meal. The glycemic impact from bread will vary depending on if you put butter on it or drink water with it or drink soda pop with it or wrap it around a big fatty piece of meat.

Maybe the problem is just that we eat so much food outside of meals (soda pop, candy, snacks, etc…), but I’m not sure where GL would fit into the “why” here.

M. said...

Hi Ned, one more point. It seems like caloric density would be a strong confounder to Glycemic Load as well. It seems like many high GL foods are more calorie dense, but calorie dense and “fattening” foods like ice cream, nuts, and cheese are not necessarily high GL.

Anonymous said...

It is summer, and as of late I have been eating a lot of fruit - not refined or processed in any way - and as a result I was getting fatter and 'fatter': if I try do describe the feeling, it's kind of sweetness of fruit makes me thirsty - I drink a lot (really lot) - water (or fat?) stays in me (mostly in the mid-section: one or two days of such eating makes no substantial difference - returning to my normal low-carb eating style flushes accumulated water/fat out, but, as I have noticed, prolonged periods of 'fruit binge' forces body to get 'structurally bigger' (new body fat set-point?) when it becomes much harder to slim down.

Yasin said...

Great post Ned. It's too bad that in 2005, the authors still believed SFA to be a contributor to the diseases of civilization. Otherwise it is great that the subject is getting some much deserved attention by the scientific community.

gwarm said...
This comment has been removed by the author.
gwarm said...
This comment has been removed by the author.
gwarm said...
This comment has been removed by the author.
gwarm said...

"Even if you eat only plant-based natural foods, which I cannot fully recommend based on data I’ve reviewed on this blog, you will be better off. "

Did you ever watch (11.5 hrs)

Posted in comments in earlier blogposts around 1 year ago?
Email I received today:

Paleo??? said...

Ned, I just watched the entire video series that gwarm posted. It was very compelling and totally undermines everything that paleo is about. I'm very scared and confused right now. What do I do?

As a start, I threw away all my coconut oil and gave all my strip steaks to the dog.

David Moss said...

I'm a bit confused Ned, the glycemic load can't be for a food per se, but only for an amount of food. It is, as you say, the amount of carbohydrate x the GI. So a Kitavan eating 1000 calories from high GI sweet potato, must surely be eating the same glycemic load as a westerner eating 1000 calories of an equally high GI food. So the only way there can be a distinction between "natural" and "refined" carbohydrates would be if i) the GI's were difference (which, ex hypothesi, isn't the case, since rice and potatos are high GI) or ii) if smaller total amounts of carbohydrate were consumed, thus reducing the GL. But we know that the latter isn't the case either, because the kitavans, asian populations etc eat most of their calories from potato/rice, so their GL must be very high.

In short, refined vs natural foods with the same GI, can't have different GL's, because the GL is dependent on portion size. So if you compare 100g of yam to a 100g of rice krispies (and we assume they have the same GI), then if the 100g of yam has less carbohydrate then the GL will be lower. But if you eat all your carb calories from the two equally GI foods, then they'll be equal GL.

Also, on this idea of yours and Kindke that reward or palatability must be a dynamic, context-dependent concept. This only seems interesting if it varies dependent on factors other than hunger/energy balance. If energy balance is the variable, then we can still define rewarding/palatable food based on reward/palatability for a given energy balance. Only if other variables make, for example, carbs more palatable at one time and fats more palatable at another, is palatability dynamic in an interesting way. Luckily I suspect it is dynamic in this way, because I certainly find carbs/fat very much more or less palatable depending on context (say, whether I've eaten more or less of one of them recently). Those who think that palatability is a useful and universal property would just have to show that these variations aren't so significant, which might be the case.

Anonymous said...

With respect to abnormal glucose and insulin responses after consumption of CHO--do you have information on what a "normal" insulin response is for an OGTT? I have seen "normal" glucose curves but have been unable to find data on what a "normal" insulin response curve would be when insulin values are taken along with glucose values. Any information would be appreciated.

Ned Kock said...

Hi Glenn, gallier2, Kindke, and Zooko. Thanks for your comments and additional insights.

Ned Kock said...

Hi David. Cooking can indeed have an impact on the GI and GL of certain foods. For example, potatoes that are boiled and chilled have a lot more resistant starch, which is essentially a form of insoluble fiber.

Ned Kock said...

Hi M. I agree that the important is the meal. However, I am not sure it is such a good idea to target the GL, and this is not the point of this post. The GL is one of the many attributes that can be used to show the discrepancy between natural and unnatural foods.

For targeting GL, one can play with macro proportions, do things like add butter to bread etc.

Still, I think that the most sensible way of dealing with the problem is staying away from industrialized foods as much as possible. At least until the food industry finds its way.

Ned Kock said...

Hi Anon. I am not sure you can flush accumulated fat as you said. Btw, have you seen this post?

Ned Kock said...

Hi Yasin. Indeed.

Ned Kock said...

Hi gwarm. Do these posts make a difference?

You may also want to click on the “China Study” link at the top of this blog. All I can do is point at evidence.

Ned Kock said...

Hi David. As I said in my comment above, the GL is one of the many attributes that can be used to show the discrepancy between natural and unnatural foods. The most sensible way of dealing with the problem is to avoid foods for which we are maladapted.

Ned Kock said...

Hi Anon. Does this post on insulin responses help?

Ned Kock said...

A diet whose foundation was primarily palatability would lead to nutritional deficiencies, and eventually fail. The reason is that palatability is in part a function of the following ratio:

(A nutrient’s content in a food) / (The body’s need for the nutrient at a given point in time).

Where “nutrient” may refer to any macro- or micro-nutrient.

M. said...

Hi Ned,

I agree about avoiding industrial food. I am just very skeptical of an insulin-centric explanation.

I think the meme of Asian white rice eaters remains very problematic to CHO. It is especially problematic to Taubes’ various interview arguments talking about Glycemic Index and starchy vegetables being a problem, and I also don’t think Glycemic Load is the great vindicator that some people make it out to be.

There is something going on with Western diet and lifestyle, but I don’t know if getting caught up in CHO is that beneficial in figuring it out. I think someone compared Taubes to Copernicus in a comment on Nikoley’s blog. I have a feeling that Ptolemy is more appropriate. I cannot help but think that Glycemic Load is an epicycle in the insulin-centered universe.

(Sorry if I being too argumentative; I think I’m just not on the same page as everybody else in Paleo land.)

Ned Kock said...

Hi M. Actually you are on the same page, generally speaking, as some bloggers who are influential in the Paleo community, like Don Matesz and Stephan G.

My point on the post is that palatability will not likely replace CHO, but something else may. I myself think that the GL is one of many factors; e.g., natural foods require some serious chewing, and that has a number of health effects that have little to do with insulin.

Btw, my post is not aimed at Stephan G., as some readers have incorrectly guessed. He is actually the one doing the right thing regarding the notion of palatability – looking at it from different angles, and in a thorough and reasonably neutral way.

Note that Stephan G. has not proposed any diet based on palatability yet, and even pointed out the nutrient-related problem I mentioned above (which is not the only one) to a reader who apparently was about to embark on a diet of tasteless high-calorie liquids drank through a straw.

David Isaak said...

There is just so much that we don't understand...

For example: Some of the Israeli researchers who did the basic research on pomegranates and arterial health noted that even though pomegranates had high levels of sugar, diabetics didn't have the standard negative responses to the sugar dose--they didn't lose glycemic control, etc..

What does that mean? Similarly, a lot of people don't seem to respond negatively to the fructose in, say, berries--but high-fructose corn syrup is another matter.

I think there is something more important than simple nutrient concentration going on when things are refined--and there are probably better and worse ways of refining things.

Part of the problem is the paradigm researchers use. For example: 1)Fruits and vegetables seem to have some health benefits. 2) Why? Must be the antioxidants.

The net result are measures like ORAC and a lot of time spent barking up the antioxidant tree. Only gradually did a few people realize that it might be phenols, anthocyanins, and other complex chemicals that provide the benefits.

A couple of decades ago, any reputable scientist would have scoffed at the idea that highly colored fruits and vegetables might be good for you because of the color ittself. Now it is a standard piece of advice.

My conclusions?

a) We don't really know what the hell is going on.

b) Whole, traditional foods are safer bets.

David Isaak said...

A side question: A lot of low-carb dieters assert that their metabolisms are somehow "broken," and that the are forever hypersensitive to carbs per se.

Certainly obese people are usually in a bad metabolic state. But how seriously do you take the idea that people's metabolisms stay "broken" even after weight loss?

Kindke said...

A side question: A lot of low-carb dieters assert that their metabolisms are somehow "broken," and that the are forever hypersensitive to carbs per se.

Certainly obese people are usually in a bad metabolic state. But how seriously do you take the idea that people's metabolisms stay "broken" even after weight loss?

David Isaak,

Thats a big subject. As an avid low-carber myself that I consider has a "broken" metabolism, starch based foods are where all the problems come from.

I can tolerate fruit, Dairy ( lactose ), honey, even grapes which reportedly have a high GI dont screw me.

Starch (potato,rice,bread) based foods just mess me up though, I get a horrible flabby and bloated feeling around my waist, roller coaster hunger, and mood swings when consuming those foods.

Not to mention I can consume GIGANTIC amounts of them before I eventually feel satiated/gratified.

Thats a fleeting feeling with those foods though, 2 hours later im back looking in the fridge for something else.

Also people's metabolism's are indeed still broken after weight loss and this stuff is well documented, e.g. low leptin due to fat cell hyperplasia, upregulation of orexigenic hormones in the brain etc.

J. Stanton said...

David Isaak:

Mitochondrial dysfunction generally, and a specific inability to oxidize fat for energy, is well-documented in the formerly obese, as well as the obese. I cover this topic (among others) at length here:

When Satiety Fails: Why Are We Hungry? Part 4


Michael Barker said...

David Isaak:

I have never been obese or overweight. Yet, I am a type 2 diabetic who has a very big problem with carbs.

I've found this to be the case with most people who have abrupt onset type 2 diabetes.

Given my perspective, I've never seen obesity as the issue in diabetes. Weight, for me, is a response to underlying issues of metabolism that probably have more to do with the CNS.

I wanted to write a response to you on my blog but this seems to be as good a time as any.

The question I wished to ask has to do with genetics.

There is a very strong correlation between peoples of certain background and diabetes. These people tend to be in the minority, however. My thinking was: if this minority had been in the majority would this correlation still hold true? Wouldn't, at some point, the majority population move away from problematic foods as they would from any behavior that showed dire social consequences?

I bring this up because in my research group one member felt that my diabetes, being considered relatively rare, should not be figured into the diagnostic mix when talking to physicians. This led me to think of when does a group show concern at a statistical level.

This might be a bit of a mish-mash but I bring this up because it seems for any given group that might have problems with diet there seems to always be another that doesn't have a problem with the same diet.

I have to wonder whether the groupings of people that we are making judgments about diet may simply be incorrect. If I and my kind were the standard of human metabolism would many of our diet insights be relevant or would they have to be couched in ways that statistically grouped certain attributes and genes?

Ned Kock said...

Hi David. There is some recent evidence that T2D can be reversed, and that seems to have something to do massive weight loss. The evidence is still a bit ambiguous in my view. Moreover, massive weight loss carries its own dangers, one of which is nutrient deficiencies that can bring a host of other problems.

Ned Kock said...

Hi Michael. I’ve long believed that some of what we call diseases are in fact traits that enabled increased reproductive success somewhere along our evolutionary paths, and those paths become increasingly different as we approach the current date. After all, a trait can evolve in a few hundred years if the right conditions are in place:

Ned Kock said...

A gem by Gonzalo Lira, via Peter (Hyperlipid), on how our bodies have been changing

Lerner said...

Greetings, to Ned and to all.

I'd first heard of Taubes several years ago when I was still reading Eades' blog. So I went to Amazon and saw a review which included the GCBC claim that it was the 'processed' grains that caused a big glycemic effect, while the unprocessed ones did not. I instantly knew that was WRONG WRONG WRONG. The reason that I knew Taubes was wrong was that my father had been very ill and was put on the corticosteroid prednisone, which created insulin resistance. At the same time, he was given syringes and insulin - to use when BG > 200 - and a glucometer. So it was not a matter of trying to sell diet books but instead of serious real world daily experience.

I plunged into learning about glycemic response, and very quickly encountered the counter-intuitive examples. Probably the biggest were that typical whole wheat bread is about the same as white bread... that the most processed pasta, white as can be, beats both. That raisins are bad but prunes are good. That brown versus white rice is not much of a difference, but high amylose beats high amylopectin rice (which usually means long grain versus short grain). And by the way, that stone ground whole wheat is good. These are not theories, they are simple measurements. Thoughts about *why* stone ground whole wheat is low glycemic would probably come under the heading of theory, but the numbers themselves don't.

When he ate the way I designed for him, he never needed an insulin shot. But when he went into the hospital and had their official menu, he needed insulin every time.

Notice that clinging to the idea that 'processed' equals high glycemic would have led to bad results. In fact, it's not unusual for people on prednisone to end up having diabetes. Whatever benefit may or may not be involved with whole foods is of course a matter mostly separate from glycemic effect. Hey, highly processed commercial yogurt with artifical sweetener is low GI and has probiotics.

It's good to see that you're referring to Brand-Miller and not Taubes here. Taubes gets such a fundamental thing all wrong. Even Neal Barnard (famous MD/researcher advising for low-fat diet in diabetics) is finally now saying that white and whole wheat bread are effectively the same. It's a new dawn.

As far as the epidemic of weight gain? When people have available lots of "tasty" foods, they eat more and get fat. Like in Japan, Okinawa, and on and on. Seems quite simple for the majority of cases - but wouldn't sell many diet books. It would seem easy to test the opposite: confine people to non-tasty food and they won't eat so much or weigh so much. (I bet you could even confine people to their one favorite tasty food - and only that - and they'd not gain weight because they'd get sick of it.)

Btw, my main aim has to do with minimizing atherosclerosis, and my current idea is that low-glycemic food is a big component of eating that will result in low non-HDL particle numbers, which is the way for me to go. Low caloric density matters, I'd think. And calories do matter, probably most of all. Hopefully such things are or will also be discussed here.

gwarm said...

It is clear you did not watch the videos (why?). It encompasses more than heart disease. And nowhere in the videos is China Study mentioned, and it is mentioned in video meta-analysis on protein as good for bone health (completely reversing/debunking old school's of thought). Watch for 11.5 hrs (2007-2011)

And in your protein link you link to "Chris Masterjohn does not agree with this idea" and link to a broken Weston A Price link. Do you not think Weston A Price and followers are very culty (similar to T.Campbell except with "nourishing traditions"... Tribal Gang wars between these two). I notice Guyenet respects this person/foundation as well. I remember first hearing about this dentist in Omnivore's Dilemma and how silly it all sounded.

I wonder how Guyenet explains role in fat-loss of all of these other catabolic and anabolic hormones Taubes cites in GCBC book:

It has also been shown that low-carbohydrate is bad for anti-inflammatory butyrate: (links Melissa McEwen blogpost, which links to PDF dissertation: "Most microorganisms in the colon prefer to ferment
carbohydrates and switch to protein fermentation when fermentable carbohydrates
are depleted.19 While carbohydrate fermentation generally leads to healthpromoting
SCFA production, protein fermentation yields branched-chain fatty acids
and potentially toxic metabolites (e.g. ammonia, amines, N-nitroso compounds,
phenols, indoles and thiols).20")

Ned Kock said...

Hi Lerner. Thanks for sharing your experience helping your father.

Processing tends to produce foods that most of us are maladapted for; some may tolerate them better than others. Glucose metabolism problems complicate the picture further, requiring a lot more experimentation.

Now, more to the point of the post, a diet whose foundation was primarily palatability might work in terms of short term weight loss, but would lead to nutritional deficiencies. If a diet leads to nutritional deficiencies, it will likely fail in the long term.

And the long term is what matters most.

Ned Kock said...

Hi gwarm. The work of the WPF has gone well beyond the original work by Price, which is truly remarkable (have you read his book?), and Chris’s research and writings are a good example of that progress.

I don’t have to agree with everything that comes from the WPF to recognize their invaluable contributions. It’s great to have them, as it is to have people like Taubes and Guyenet contributing to the discussion as well.

Roberto said...


"Regarding the CHO, it seems very clear that refined carbohydrate-rich foods in general, particularly the highly modified ones, disrupt normal biological mechanisms that regulate hunger."

I think this is still mostly attributable to the presence of refined carbohydrate in processed, highly palatable food.

I've never met a person who binged uncontrollably on bowls of dry, white flour on its own.

If you are going to suggest that refined carbohydrates disrupt our appetite - for reasons other than reward - it would be much more convincing if you gave examples of people becoming obese eating bland, repetitive, nutritionally complete diets that include refined carbs.

Otherwise, the 'carbohydrate hypothesis' cannot be applied to them, and the 'reward hypothesis' is still the best explanation for why refined carbohydrates might be problematic, in a certain context.

A case in point: In the 'tube-feeding'study Stephan cited, the subjects lost weight dramatically eating ample amounts of refined carbs and sugar, but in a bland, low-reward context.

Roberto said...
This comment has been removed by the author.
Roberto said...

It was a small study, but the fact that morbidly obese people returned to a much leaner state eating ample amounts of sugar and refined carbs makes it very hard to believe that refined carbs "disrupt normal biological mechanisms that regulate hunger" - unless you're referring to their reward value overwhelming our natural appetite.

Ned Kock said...

Hi Roberto. Sure, but we need to keep the big picture in mind. I have no doubt that a diet that promotes a calorie deficit will lead to weight loss, with more or less lean body weight loss depending on the circumstances.

The question is: Will that diet work in the long term?

If anything, it appears that all diets that are approached as “diets”, as opposed to permanent lifestyle changes, tend to fail. It usually takes a couple of years, and the results are worse than if no diet was attempted – e.g., higher body fat percentage than before.

Moreover, I think that the right type of exercise makes an enormous difference. So what we put in our mouths is just part of the equation, although a very important part.

Roberto said...


"I have no doubt that a diet that promotes a calorie deficit will lead to weight loss..."

"If anything, it appears that all diets that are approached as “diets”, as opposed to permanent lifestyle changes, tend to fail."

The term 'diet' is a slippery concept. In essence, what does it refer to? Simply what you choose
to eat on a regular basis. If you cut out fast food and sugar you are embracing a new diet - and
I'm sure we can agree that that diet won't have any drawbacks. If by 'diet' you mean deliberate
calorie restriction, then i agree that is doomed for failure - the same way trying to hold your breath
is doomed for failure.

I would also argue that any extreme macronutrient restriction will lead to failure in many people. But given the choice, I'd go whole-foods low-fat before I'd go low-carb. The thyroid impairing effects of the low-carb lifestyle are appearing more and more real.

A low-reward approach to eating does not involve deliberate calorie restriction, nor extreme
macronutrient restriction. If a calorie deficit and weight loss occurs on such a diet - eaten ad
libitum - then it is not against the body's will and is likely to be maintained for as long as a person can maintain that way of eating. That is why the low-reward concept has gained so much traction. Too many people have been burned by low-fat, low-carb and calorie restricted diets.

There may be negative psychological imacts to low-reward dieting. I am quite concerned about the possibility of appetite fatigue if taken to the extreme - i.e. Stephan's 'level 5' approach.
Appetite fatigue appears to be a very real phenomenon when people are restricted to the same few foods day in and day out. It is very dangerous and is not a sign of improvement.

Johnnyv said...

Interesting, I have a problem with this food reward theory in that foods I find the most delicious do not encourage overeating.
For example I love rare steak with lots of butter plus salt and pepper, the first bite is beyond amazing but it decreases in taste the more I eat until it becomes unappealing. It is then strange that I can eat the same portion of steak plus a large serving of fried potatoes and not feel anymore satisfied than If I just ate the steak alone(except for a more bloated stomach feeling).
Macadamia nuts are my favorite nut but I can't eat many of them before they lose their palatability, cashews I like less but can eat vastly more in a sitting without feeling satisfied.

I feel that eating a bland liquid diet would indeed encourage automatic calorie restriction and would work for weight loss if the participant was iron willed or locked in a ward with no method of cheating. Relevance to most of the population that needs to lose weight is very low just like the normal low fat calorie restriction diets.

Anyway in terms of the GL of foods I have an observation relating to pigs. My parents owned a market garden in New Zealand, we also kept pigs for consumption that would be feed mostly excess vegetables from the garden.
Certain years we would have large excesses of certain vegetables which would comprise the majority of the pigs diet. When The pigs ate mostly Kumara they had a very large amount of subcutaneous fat compared to when they ate mostly carrots which gave fairly lean meat. Kumara is a sweet potato but has a higher GL than normal sweet potatoes apparantly(GI 78, GL 19.5).

Personally I would rather eat raw carrots than raw kumara :)

hemul said...

What amounts are used in the GL table? Servings? 100 g?

As David already said, I don't understand how any given food can have a higher GL than another in itself since GL is the product of GI and the amount eaten. I may eat 150 g of potatoes, but I hardly take 150 g of sugar in my tea (I don't take any at all but to make a point...).

Can someone please explain?

Ned Kock said...

Hi Johnnyv. Thanks, interesting.

Ned Kock said...

Hi hemul. The GLs refer to 100 g servings of the foods.

hemul said...

Thanks Ned. But I still don't understand - we don't eat the same amounts of different foods, so what's the use of comparing them like this?

I'm sure I'm missing something or maybe I'm just dense. The latter probably. ;-)

Ned Kock said...

Hi hemul. My guess is that the point you are trying to make is that we can combine foods to reduce the GL of a food combination, right? While the GL is not defined for food combinations, or for portions of a food, you can indeed do that.

Let us say you add any food that leads to no glucose response, cream for example, to another food with a nonzero GL (e.g., apples) to make up a dessert dish. The GL of the mixed dish will be lower than the GL of the apple portion of the dish, so to speak.

Adding protein also tends to reduce the GL of a mix, even more than fat in normoglycemic individuals, because protein leads to an insulin response without a corresponding glucose response.

M. said...

Hi Ned,

I think an issue is that for some of us GL does not seem to be very useful and almost seems to be a hand-waiving argument. Maybe you could do a post on it sometime and elaborate on it more.

GL just does not seem to endorse a carb or insulin-based context when you start looking at full meals or full-day consumption. I think it is especially problematic in the meme you mentioned of comparing the differences between a Western Diet with refined flour and sugar versus a more ancestral diet with white rice or potatoes.

The fact that rice has more water density and less calorie/carb density does not seem to have any bearing on insulin-based arguments. Furthermore, it seems logical that people would normalize their hydration, if not at a given meal then at least through-out the day. It would seem that one culture would not remain perpetually dehydrated compared to another. (You could speculate though that the culture eating the “drier” carbs like flour and sugar would be more likely to hydrate with calorie-filled beverages instead of water.)

It just seems like GL was a hand-waiving attempt to make the carb theory work with data rather than a real attempt to explain the data.

In the paper you referenced by Cordain and Eaton where they brought up Glycemic Load, they admitted that some dairy and even ice cream have low GL, but then mentioned that these foods had high Insulin Index. Talk about hand-waiving arguments trying to blame it all on insulin. If they want to talk about Insulin Index, then why bother with Glycemic Load? Because pasta has a very low Insulin Index (less than meat) and potatoes have a very high Insulin Index (up in candy bar territory). So why did they bring up Insulin Index?

Glycemic Load was the insulin measure that “fitted” the data best of all the insulin-centric options, but to me it does not really explain the data.

hemul said...

Ned, I'm not trying to make a point, I'm just trying to understand. GL in this context just doesn't make sense to me, so I don't "get" the message. I didn't refer to that we can combine foods, rather to that we simply don't eat the same amounts of different foods, say, sugar and potatoes. OK, I hear some people actually do, but generally speaking. ;-)

I guess one could one could say that the foods with the highest GL per 100 g are bound to be refined carbs and we know those are bad for us. But that seems like taking the long way to get to an obvious conclusion, so I still think I miss the point. Sorry for being so slow.Ned, I'm not trying to make a point, I'm just trying to understand. GL in this context just doesn't make sense to me, so I don't "get" the message. I didn't refer to that we can combine foods, rather to that we simply don't eat the same amounts of different foods, say, sugar and potatoes. OK, I hear some people actually do, but generally speaking. ;-)

I guess one could one could say that the foods with the highest GL per 100 g are bound to be refined carbs and we know those are bad for us. But that seems like taking the long way to get to an obvious conclusion, so I still think I miss the point. Sorry for being so slow.

hemul said...

Oh that went well! Sorry for the mess.

Ned Kock said...

Indeed ice cream is one of the most insulinogenic foods. The post below lists insulin and glucose responses, in terms of AUC, for various foods:

Ice cream leads to a non-negligible glucose response as well, but I guess the addition of protein and fat together lead to a reduction in GL.

Interestingly, if you take a careful look at those tables, you’ll notice that the insulin response to cheese is lower than that to beef.

M. said...

Hi Ned, sorry to keep harping on this, but..

If insulin is ultimately the problem, then why use insulin proxies like “carbs”, GI, and GL instead of Insulin Index?

The Insulin Index of refined wheat (pasta) versus white rice seems to pretty much make insulin-centric explanations very problematic if not non-viable, but is there really a good reason not to use Insulin Index?

Ned Kock said...

Hi M. The point you are making, as I understand it, is certainly a valid one. In fact, a focus on blood glucose often leads to folks not clearly identifying the period of hyperinsulinemia that normally precedes T2D. Hyperinsulinemia is not always associated with a high fasting blood glucose.

One main reason for the focus on glucose that we see everywhere is that, currently, blood glucose is (much) cheaper and easier to measure.

But we can only make indirect assumptions about insulin levels based on glucose levels. For example, hyperinsulinemia is usually suggested by low blood glucose levels several hours (e.g., 3-5 h) after a high carb meal.

Lerner said...

a word on the GI of rice: that's probably the food that varies most, according to the variety - which varies according to the part of the world it comes from. There are two types of starches: amylose, which is lower GI, and amylopectin, which is higher GI. Rice varies a lot according to the percentages of the two type, while wheat does not. Short grain rice is usually composed more of amylopectin, and long grain rice is usually more amylose.

The reason is apparently due to structure. Amylopectin is more branched, while amylose is straight.

Imagine that you have a large, neatly stacked group of logs (starches) being attacked by a horde of beavers (enzymes). The beavers have to chew through the outer ones before they can get at the inner ones. Alternatively, if the logs were thrown in a random pile, like PickupStix, all the logs can be attacked and broken into pieces at once.

So, a study on rice grown in Australia (Jenny Brand-Miller's home country, btw) might not tell you about rice grown in South America. But if you got Uncle Ben's Converted rice, that is standard and the GI is known (it's fairly low).

David Isaak said...


"I'd first heard of Taubes several years ago when I was still reading Eades' blog. So I went to Amazon and saw a review which included the GCBC claim that it was the 'processed' grains that caused a big glycemic effect, while the unprocessed ones did not. I instantly knew that was WRONG WRONG WRONG."

I've read a lot of Taubes and I don't recall him ever claiming that unrefined carbs are fine and refined carbs are bad. In fact, your comment is kind of bizarre, as Taubes takes an immense amount of flak because he he is considered by many to be simplistically opposed to carbs per se.

In GCBC, Taubes spends several pages explaining why the glycemic index is misleading.

Perhaps you should read Taubes rather than drawing conclusions about what he says based on a review on Amazon.

David Isaak said...

@Michael Barker: "Given my perspective, I've never seen obesity as the issue in diabetes."

Perhaps not for some percentage of the T2 diabetic population. But the fact that weight-loss surgery reverses T2D in many people suggests that the link can run both ways in many people.

M. said...

Well, according to Peter’s recent post at Hyperlipid, it’s not Glycemic Load that differentiates between Western diet and pre-industrial diet; it is multi-generational exposure to high sugar that damages the metabolism first and then later makes carbs less bearable.

It is interesting that Stephan’s critique was of Taubes’ Carbohydrate Hypothesis (you know – some people have to beware the carbs in green leafy vegetables and all that), but many of the defenders of the “Carbohydrate Hypothesis” have much different interpretations in mind when they defend it.

I think this amalgam of low-carbism makes the scientific discussion very problematic because everybody is talking about different things.

Ned Kock said...

This epigenetic view of abnormally high glucose/insulin levels being a major cause of problems makes a lot of sense to me. From fertilized egg to full formed fetus, sensitivity to extraneous stimuli is very high, which is why there are so many food/drug label warnings aimed at pregnant mothers.

Maybe there should be similar mandated warnings on soda cans and the like.

gwarm said...

Mathieu Lalonde talked about glycemic load thrown off re: fructose intake affecting liver soaking up glucose:
Listen from 37min to 41min.

More Lalonde: he really challenged the paleo community recently on the preparations and types of lectins, saponins and more: gets good 9min in but his diet seems extreme and backwards to me

B.V. said...

Your insightful reconciliation of apparently conflicting theories is not only deeply satisfying but also, by providing a wider range of food choices to practiioners of carb restricted diets, make it easier for them to follow their diets.

Online pharmacy reviews said...

It was what I was looking for, actually it would be perfect if I could share it with my friends because this is terrific.

gwarm said...

"Diets high in protein6
and sulfate (derived primarily from food additives)
4 have been shown to contribute greatly to
the production of these potentially toxic products.
The production and absorption of toxic metabolites
is referred to as bowel toxemia.7" from

I found this link after listening to Mat Lalonde and his view on alcohol and dysbiosis: