This is a great question.

Taubes framed his conclusion as an overall reaction to his immersion in the research. He also frequently notes his tendency to prefer the insights of clinicians, who actually have to treat obese patients, to those of researchers, who often lack direct personal experience with obesity or treating obese patients. So, in a sense, Taubes is speaking with regard to individuals who already have broken metabolic and hormonal function. And these individuals tend to tolerate carbohydrate poorly.

Guyenet - unlike Taubes - admits that he has always been slim and never really had to struggle with encroaching weight gains. So it is a little easier for someone like Guyenet to view the matter from a pure food reward perspective, which implies that removing one source of reward - be it fat or carbohydrate - will have similar effects.

We Paleohackers know - quite intimately - that a low carb diet rich in fat and protein can be very satisfying and enjoyable, once our fat burning pathways have been established. The literature suggests that low fat or calorie restricted diets, on the other hand, are very unsatisfying and unpleasant, and thus represent a less desirable strategy for treating obesity, even if in some strict sense Guyenet's food reward theory is correct.

One problem with both theories is that they are too simplistic and reductive to work as complete or overarching theories of what causes obesity or of what caused the obesity epidemic. That said, we have at least one reason to privilege Taubes here, because the obesity epidemic did not co-occur with a sustained increase in fat consumption, whereas it did co-occur with a sustained increase in carbohydrate consumption. Then again, in Guyenet's defense, there is co-occurrence with low quality forms of both macronutrients: processed carbs like HFCS and white flower, and industrial seed oils.

Taubes' notion that insulin is a fat storage hormone is pretty well established, but he frequently noted that researchers had that part figured out in the 1950s (if I recall correctly), and I don't know where the current literature stands on this. Taubes definitely placed an undue emphasis on insulin, however, as leptin is probably just as important. Since the same dietary protocol controls both, though, this shouldn't matter too much.

I think the basic idea is that - as mentioned by Cliff - chronically elevated insulin is the issue. And processed carbs certainly contribute to this state. When insulin levels are chronically high, we become less able to handle it, which is what insulin insensitivity is. I'd be surprised if Guyenet disagrees with this.

My conclusion from all of this is that for those who are like Guyenet and have healthy metabolisms, fat and carbohydrates can be thought of as equivalent sources of food reward. But for those of us like Taubes (and me) who have struggled with weight gain, and thus probably have broken metabolisms to some extent, we have to pay closer attention to our carbohydrate intake. We cannot deal with the insulin spike properly, which - skipping over a lot of biochemistry - is in part what leads to chronically elevated insulin levels in the first place, as well as the fat storage patterns that Taubes makes central to his explanation.

There is a difference between postprandial insulin spikes and chronically elevated insulin. The latter is what causes problems in people.

I think there are several good responses already, but I will throw in my 2 cents, maybe even a dime...

I do not view his recent comment regarding post prandial insulin as inconsistent with his past writings. If you look at his series on Chris Voigt and his all potato diet last year - http://wholehealthsource.blogspot.com/2010/12/potato-diet-interpretation.html - he actually started to give us hints regarding the Food Reward theory he has now developed fully.

Additionally, check out this article from 2008 (3 years ago!) - http://wholehealthsource.blogspot.com/2008/08/cardiovascular-risk-factors-on-kitava_17.html. Here is the money quote -

In my opinion, the most important finding in this paper is that a high-carbohydrate diet does not necessarily lead to elevated fasting insulin. This is why I think the statement "carbohydrate drives insulin drives fat" is an oversimplification. What drives fat accumulation is chronically high insulin (hyperinsulinemia), which the Kitavans do not have. With a properly-functioning pancreas and insulin-sensitive tissues (which many people in industrial societies do not have), a healthy person can eat a high-carbohydrate meal and keep blood glucose under control. Insulin definitely spikes, but it's temporary. The rest of the day, insulin is at basal levels. The Kitavans show that insulin spikes per se do not cause hyperinsulinemia.

Additionally, let's look at the Oracle of high fat in his potato post, Peter from Hyperlipid - http://high-fat-nutrition.blogspot.com/search/label/Potatoes%20and%20weight%20loss%20%281%29.

The key paragraphs -

The ultimate determinant of weight loss is fasting insulin. This determines how much lipolysis occurs during the period before the next meal. No one expects to lose weight during the 4 hours immediately after any meal. The following 8 hours, especially overnight, is when weight loss occurs. Post absorptively, without dietary glucose input, there is no stimulus for anything other than basal insulin secretion. Fasting insulin will be low because muscles are insulin sensitive so relatively little insulin is needed for glucose uptake. As fasting insulin levels drop lipolysis will restart. Free fatty acids will feed back to the liver to cause some degree of hepatic insulin resistance, decrease first pass metabolism and stop too profound an hypoinsulinaemia occurring. But fat loss will happen. So you have to ask whether an almost all potato diet genuinely leads low fasting insulin and subsequent weight loss. For my perspective the answer is yes. The precedent for this has to the Kitavans with fasting insulin levels of 4.0microIU/ml.

Finally, my go-to guy, Dr Kurt Harris in his most amazing Paleo 2.0 post that many here have read (repeatedly I'm sure) - http://www.archevore.com/panu-weblog/2011/3/30/paleo-20-a-diet-manifesto.html

In my own intellectual evolution, I have expanded and modified the carbohydrate hypothesis of the nutritional transition, as proposed by Gary Taubes, to one that does not indict a whole class of macronutrients.

Good Calories Bad Calories was a seminal book. It showed 50+ years of bad science and the arguable deception that led to the development and promotion of the Lipid and Diet-Heart Hypotheses. The vilification of saturated fat to me was demolished by the book. This does not make everything in the book correct, nor does an incorrect hypothesis that perhaps now is being challenged, bring into question the entire book.

I could be way off, but I'm gonna go with one of my favorite words - "context"

In a metabolically sound person, perhaps the insulin spike isn't such a bad thing. In my basic knowlege, I think it blunts cortisol and stimulates growth. Where the catch is, so to speak, is that very few people seem to be metabollically sound, thus, it get's more of a bad rap than it should, do to context?

Here is what Chris Masterjohn told me:

Here's the problem with the "insulin theory" of obesity promoted by the low-carb people, however. There's a very simple but definitive disproof of the idea that even elevated fasting insulin promotes obesity by inhibiting the release of fat from adipose tissue. Obese people have higher free fatty acids than other people. This shows that insulin is not doing its job in adipose tissue because it is not sufficiently suppressing lipolysis. Of course, one could argue that the increased free fatty acids are not because of insulin resistance in adipose tissue but because there is just much more adipose tissue that can release free fatty acids -- even though we know that adipose tissue becomes insulin resistant in obesity -- but this is irrelevant for the following reason: the influx of free fatty acids into the liver exceeds what can be burned at any given moment, so the liver turns them back into triglycerides. This causes fatty liver and elevated blood triglycerides. The greater the relative deficiency of choline, the more this will lead to fatty liver instead of increased blood triglycerides.

In short, the bottleneck is not at the level of adipose lypolysis, but at the level of fatty acid oxidation in the liver.

Why does fatty acid oxidation shut down in the liver? Because the liver is suffering from energy overload. This is not necessarily excess calories, but could also be deficient nutrients or deficient oxygen or deficient exercise or deficient thyroid hormone or some other factor needed for energy metabolism. In any case, it's too many calories relative to the capacity to burn them.

Part of the increase in blood triglycerides, however, is due to the other part of insulin resistance in adipose tissue -- adipose tissue becomes ineffective at clearing the triglycerides from the blood.

Thus there is NOT ENOUGH insulin signaling to effectively promote adequate fat storage. That is one of the primary reasons for all of the metabolic abnormalities associated with obesity -- increased free fatty acids, fatty liver, and increased blood triglycerides, and, the big one, abnormal glucose metabolism.

This is a homeostatic response of adipose tissue because the energy it needs to store exceeds its capacity for storage. Deficient oxygen supply to adipose tissue probably plays a role in this. It's not just excess calories, but deficient capacity to handle the calories.

But the idea that the adipose tissue accumulates fat because there is so much insulin around is, in my opinion, just plain backwards. Adipose tissue stops responding to insulin because it can't handle any more fat storage.

@Eric S

I think I brought too many factors into such a brief explanation. Sticking to adipose lipolysis, if the adipose tissue were to release more free fatty acids, they would generate more triglycerides, because the liver's capacity to deal with all this energy is already overloaded. You can show this, for example, by giving obese people heparin.

Now, this is not to say that there is no role for insulin. For example, I have not made any case that insulin has nothing to do with the bottleneck at the liver. But it's quite clear from the selectivity of insulin resistance in the liver that the cause of it is not too much insulin. It is a highly controlled process by the liver to change its own sensitivity to some insulin pathways and not others. Specifically, it stops suppressing glucose output in response to insulin but continues to suppress triglyceride output. Thus, all energy is flowing out, an obvious adaptive response to energy overload. This response almost certainly occurs prior to elevated systemic levels of insulin. The pancreas begins secreting more insulin precisely because the liver won't stop releasing glucose (and won't take up as much). Extra insulin is required to maintain a normal level of blood sugar. Since the liver keeps secreting triglycerides in response to insulin, blood triglycerides increase.

There's some evidence -- it is not definitive, but it's intriguing -- that these triglycerides block leptin passage across the blood brain barrier and thus contribute to the "starvation despite abundance" phenotype. This would provide a rationale for temporary carbohydrate, especially fructose, restriction in order to normalize leptin sensitivity.

Thus, carbs and insulin are involved, but the high insulin is not causing a bottleneck at the point of adipose lipolysis, and it is the result, not the cause, of insulin resistance at the level of the liver.

@aravind

I agree with you that they are two different processes, but I was trying to assess what the relevance is for obesity. Yes, insulin promotes fat storage at adipose tissue. It also promotes satiety in the brain, muscular hypertrophy in skeletal muscle, synthesis of lipids for the rest of the body in the liver, glucose uptake in many tissues, and plays a role in the leptin-thyroid axis communicating abundance and the consequent growth and sexual reproduction that should result. Promoting fat storage at adipose tissue is a very good thing under normal physiological conditions because it allows us to regulate our energy intake and expenditure and it is required for leptin signaling. I was simply addressing whether people get fat because insulin is suppressing the release of fatty acids from adipose tissue, and I don't think this is the case because the evidence indicates that this is not where the fat-burning bottleneck is taking place in obesity.

@arlokk

I'm not sure I understand what you're not convinced by, but eventually I'll write a referenced blog post on this, as I'm not really prepared to convince anyone of anything right now anyway. But I don't think your theory makes much physiological sense -- although it's true that insulin decreases blood sugar by stimulating GLUT4 migration to the cell surface in muscle (which is involved in promoting glucose uptake), tracer studies indicate that a very centrally important way it decreases blood sugar is by suppressing the release of free fatty acids from adipose tissue. Elevated free fatty acids will decrease glucose uptake.

I think Stephan is correct, most of us did not get fat by eating plain rice or potato, we got fat eating very palatable rewarding combinations of fat and carbohydrate. I also think that he is correct in saying that a low carb diet for a lean person is not as good as one that includes some starch.

As a person who lost their weight with a low carb diet this was a difficult realisation for me, and it took a while to change my mindset on carbs, it was accelerated when I noticed that my low level depression was ameliorated when I added some starch to my diet (hat tip to Robb Wolf for making me discover this by mentioning it on his podcast)

However, I know Stephan's plan to eat unrewarding food doesn't provide a better solution than a low carb or low fat. I tried it for a couple days, he says it takes some willpower, well it takes more than I have, it induces great feeling of deprivation in me.

I think any good plan will reduce food reward by distracting you from the fact that you are reducing food reward, in that instance, the 'story' of insulin being a bogeyman is probably helpful.

I personally buy into the food reward theory, as I've lost weight both while eating lots of carbs with extra sugar (vegetarian on Shangri-la Diet) and while eating few carbs with no sugar (VLC Paleo) but never saw changes in abdominal fat come as fast as when I recently began experimenting with low-reward starches.

I absolutely accept that insulin is a major factor for folks who are diabetic, or borderline diabetic, but I just can't reconcile my experience with thinking that it's the major factor for most people.

I think a big reason that the carbohydrate/insulin hypothesis (and the fructose panic) have a certain narrative power is because carbohydrate consumption has been a reasonable proxy for tracking consumption of processed foods in recent decades, and therefore a good proxy for higher food reward diets.

I do think that Taubes did a great service in writing GCBC and communicating the idea that the popular thinking about consumption and weight is deeply flawed. The really great part of his work is that he got the big idea right: that it's no more rational to blame fat people for storing fat than it is to blame teenagers for growing too damn fast. Growth is controlled by hormones.

I think the majority of the paleo community can agree that managing weight loss comes down to finding ways to manipulate our hormonal systems. The low food reward hypothesis is easy to test, and the proof will be in the bland, tasteless pudding. We'll all be able to see the results of a few weeks and months of low-reward experiments in those willing to try them soon enough.