Friday, September 14, 2012

More Thoughts on Macronutrient Trends

I had a brief positive exchange with Gary Taubes about the NuSI post.  He reminded me that there's an artifact (measurement error) in the USDA data on fat consumption in the year 2000 when they changed assessment methods.  Here are the USDA data on macronutrient consumption since 1970, corrected for loss (28.8%) but not corrected for the artifact:
The data suggest an increase in fat consumption, although I didn't make that claim in the last post because I'm aware of the limitations of the dataset.  What I wrote is that the data don't support a decrease in fat consumption, which is accurate.

Looking at the graph above, you can see the abrupt increase in fat consumption in 2000.  I corrected the data for this artifact in the graph below: 
When you correct for the artifact, the data show very little change in fat consumption between 1970 and 2006 (small increase).  This is consistent with NHANES survey data, which also show essentially no change (slight decrease).  Each dataset has its own limitations, but I prefer the USDA data for most things because they're yearly, more detailed, and more accessible.

Here's a graph of macronutrient intake since 1909, adjusted for loss and the artifact in 2000:

The data from the first decade or so are somewhat sketchy because they were estimated retrospectively.  The quality of the data improve progressively from the early 1920s through 1940, as more commodities were tracked.

The graph suggests that fat consumption has gradually increased over the last century (+ 30%), carbohydrate consumption has decreased and then rebounded (currently - 6%), and protein consumption has increased modestly (+ 10%).

The modern "obesity epidemic" began in earnest between the 1976-1980 and 1988-1994 NHANES survey periods, and corresponded with an increase in calorie intake.  According to the adjusted data, this increase was 350 calories per day since 1970.  65 percent of the increase in calories can be attributed to carbohydrate, 24 percent can be attributed to fat, and 11 percent to protein.

The excess carbohydrate and fat came mostly from processed refined foods such as sweetened soda and processed snacks.  

A Note about Reproducing Materials Published on Whole Health Source

I've noticed instances of other bloggers using my original materials without attribution (e.g. graphs).  My copyright policy, as always, is on the sidebar to the right (bottom).  Here's what it says:
Please feel free to reproduce the contents of this blog, on the condition that you:  
1) Attribute the work to me
2) Provide a link to the page where you found it
3) Do not use it for commercial purposes
I think it's a pretty generous policy, so please respect it like most people have.  Since the graphs in this post are probably going to end up all over the Internet, I've added my name to each to facilitate this. 


Richard said...

CLEARLY an increase in calories consumed has caused the obesity "epidemic." That is no mystery. The concern is in the nature of how to deal with that issue on an individual and population-wide basis. The first step is to at least attempt to put forth an accurate and helpful understanding of the known causes. It seems reasonably clear that an increase in carbohydrate consumption is one part of the problem. But fats? What is a fat in this analysis? Corn oil? Peanut oil? Any and all manner of fats? That's not helpful or informative, really. Combining an increase in PUFA consumption with excess carbs and while at the same time reducing saturated fat consumption is most likely exactly the problem.

Alternatively, we could simply reduce calories over-all and increase exercise, and weight goes down and health goes up...Cuba. Again, it proves that excess calories are the problem. We know that.

There are also at least two concerns here: One is the obesity epidemic and the other is (individual) optimum health...Not quite the same things, although related.

This information strikes me as not much more profound than correlating death with age.

The problem is that we know that carbohydrates by themselves, in primitive societies, can be consumed in widely varying amounts and yet weight and health seem unaffected. The best coherent explanation of that seems to be the "acellular" explanation in terms of the carbohydrates of choice in a modern society. That may in part explain the "success" of the pure potato diet.

Flour and PUFA may be the answer. It would be helpful to see something that addresses that directly, or even indirectly.

Richard said...

I neglected to include the pernicious effects of sugar, both sucrose and fructose, as well, both just another form of carbohydrate, and easily and conveniently consumed.

Dan said...

Stephan, what is this artifact? It looks to me like the fat intake data changed even prior to 2000 after correcting for it, but that could be that there are no intermediate sampling points between 2000 and the prior sampling point.

ann. said...

As with the effects of chemicals, plastics etc on humans becoming more detrimental when combined (synergistic effect), wouldn't it be the same with the reasons why humans are getting fatter?

Constant availability of unhealthy food, less everyday moving (walking, bicycling, outdoors playing etc), depleted farm lands, pollution, supersizing and this list could go on.

Isn't there also an accumulative detrimental effect over the generations that doesn't really show until about three generations have been fed on "non food"?

If graphs were made of other possible obesity causing factors and compared with the macronutrient graphs, I guess they'd show that the human character/brain is no match for that kind of development.

Sorry if this post could have done with some more "think through" but the sun is shining, the outdoors are calling and so are my tiny tots!

Unknown said...

You change colors between the graphs and the pie chart, that was a bit confusing. It would be nice if the colors would be consistent. :-)

Other than that, yes, this. Thank you for posting it.

Grinch said...

"CLEARLY an increase in calories consumed has caused the obesity "epidemic.""

So why start off with this statement and then contradict yourself throughout the rest of your post? Either calories CAUSE obesity or they are simply necessary for obesity to occur. This kind of statement is exactly what I think is wrong with the mainstream obesity paradigm. Your first statement indicates that calorie consumption is a behavioral problem, as if the CHOICE to consume more calories is the root cause here. Instead we should get away from these types of statements because 100 million people don't suddenly decide they want to eat an extra 400 calories per day just for the hell of it.

These graphs only tell us what we already know, that the key predictor in weight is calories. Still tells us absolutely nothing about causation. Nobody can adequately explain why we over-consume calories, but at least I praise Taubes, Atkins, and others for looking at this from a different paradigm, even though I think putting all their eggs in the insulin basket was just wrong.

Grinch said...

Oh and talking about all the changes in our environment in the past 100 years isn't good enough. There are still millions of people who eat whatever they want and stay thin without any effort whatsoever.

We must find a way to explain what is different about these people to sufficiency understand and treat obesity effectively.

Galina L. said...

Probably it is more easy to increase carbohydrates in a diet than proteins or fats. Before seed oils and margarine got so popular, fat was less available and more expensive, and fat/carbs combinations were not so cheap.For my grandma generation (she is 93)the cheapest source of fat was lard. My mom (she is 75) remembers the arrival of sunflower oil and the beginning to use it in a cooking liberally.
I now witness the beginning of obesity epidemic in Russia(during my visits). There is nothing new about plump middle-aged people and old folks with bellies, and young women dieting from time to time, but fat children is the new phenomenon. Sugar, seed oils, margarine were widely used before, but fat children started to show on streets only now. They are the group who has the least access to a junk food and fed by mothers and grandparents.Teenagers do go to fast-food places and look thin so far.

Megan Brooks said...

Is there data from which energy expenditure over the same period can be estimated?

Robert said...

If the macronutrient data isn't reliable until 1940, why do you use 1909 as the baseline? I hope it's not just to show a drop in carb intake.

In any event, the question of what contributed the most calories is not as interesting as the question what contributed the most hunger.

I had a brief positive exchange with Gary Taubes about the NuSI post.

Consider the possibility that all of your exchanges with Gary Taubes have been positive.

We can't help correlating our feelings with reality, but it's worth remembering that correlation is not causation. :-)

Unknown said...

Dave, as far as this.."There are still millions of people who eat whatever they want and stay thin without any effort whatsoever". I promise you that if you actually look at their intake, it's much less than you think.

With the exception of adolescents and young adults, few can literally overeat continuously and not retain fat. What you typically find is that when you really determine how much most thin people eat weekly, they don't consume as much as you think.

Thirty years ago, before this was such a hot and demanding topic, I began to observe myself, friends, family and colleagues that began to slowly add the weight. Many became obese and remain that way. What I noticed is that those who reacted quickly, myself included, had little trouble adjusting and remaining relatively thin, those that didn't developed a problem. And the way I described the problem then was, it appeared that the more energy one consistently takes in, the more the body expects it, and when restricted, it revolts almost violently with hunger pangs as it's main defense. I have to date seen nothing that actually convinces me that this phenomena it's not what actually happens in the day to day lives of people, not all, but most. And the key for those friends of mine that didn't become very over weight or obese, was deciding early not to get that way. There are exceptions, but I propose if virtually anyone allows themselves to get into this situation, their body will revolt, especially when restricted quickly.

I have and do work with a few obese people and take an approach to slowly restrict calories and attempt to loose weight at a rate much less than people want, so far it works if followed and without too much lean mass loss or violent hunger issues. Getting someone to understand the length of time it took to accumulate and the length of time it should take to change the body's behavior is difficult but is the key in my opinion. Those two, accumulation time and loss, are not one in the same, but slow is the way to go.

Anonymous said...

I think when it comes to weight loss, calories consumed versus calories out is still the most important factor. However, I also believe that the type of food consumed can influence how much or little one eats, hence leading to weight loss.

For example, low carbohydrate, high fat diets show great amounts of weight loss. This I think is based on the fact that fat and protein are more filling than carbohydrates. It's difficult to overeat a fatty food without starting to feel sick, for example a whole block of cheese or a whole tub of sour cream, for most of us anyway. Thus eating fat and protein are in a way 'self-limiting', and in turn you probably eat less calories overall compared to eating a diet with foods that are easy to over consume, especially sugary and starchy foods such as chocolate, biscuits, bread, pasta ... etc. and because anything that is sweet tends to make one want it more.

Also, low carb diets basically restrict almost everything from an entire food group - the carbohydrate group. There goes quite a fair bit of food you're allowed to eat and subsequently reduced a lot of calories you choose to eat. This coupled with the fact that most of us feel fuller and fuller for longer on diets with higher fat and protein, could be why it's easy to lose weight on a low carb diet.

I do agree with Stephan that insulin is not the sole cause of weight gain/obesity and that low carb diets that claim insulin is the one and only enemy are missing many other points in the weight gain, weight loss equation.

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

Richard, it strikes me as odd that Ian Spreadbury’s seminal paper

(Comparison with ancestral diets suggests dense acellular carbohydrates promote an inflammatory microbiota, and may be the primary dietary cause of leptin resistance and obesity)

is largely ignored on the leading health forums.

Spreadbury’s hypothesis explains and predicts all the ‘paradoxical’ observations. One can only hope that NuSI will recognize the profoundness of Spreadbury’s idea. It should be tested immediately, but I doubt this will happen. Confirmation would make all the current competing paradigms superfluous. It’s just too neat.

Key said...

Have you even gone through and read all his sources in regards to the acellular hypothesis?

Foods that are completely digested in the stomach like white sugar contribute very little to endotoxin loads. The study he uses to show that cream increases LPS also shows that orange juice(which is an acellular carbohydrate?) has no effect on inflammation and LPS.

The only evidence he has against fructose is based on diets of pure fructose which is not completely absorbed so it feeds bacteria.

The acellular theory only makes sense if you don't understand the science, it's clear that he cherry picks studies to fit his hypothesis. I think endotoxin probably has a big influence on health in modern day cultures but his acellular hypothesis doesn't hold up imo. Allergens causing bowel dysfunction makes much more sense.

Grinch said...

@mark I know that thin people don't eat alot, that's kinda my point. Despite having unlimited access to high calorie, palatable foods, they have no desire to over-consume like others do. They can eat a modest sized meal and actually feel too full to eat more. We need to understand why some people get full after eating their meals and some people are chronically hungry.

Puddleg said...

There are NHANES data about polyunsaturate consumption. An increase in PUFA and decrease in SFA seems to be a good match for increased caloric intake in women.
Compared to no change in men.

Stephan Guyenet said...

Hi Dave,

Genetics explains more than half of variability in body fatness in modern populations. Genes determine how susceptible people are to a fattening food environment. We see the same thing with hypertension. The genes involved are mostly but not exclusively acting in the brain in the case of obesity, although most of the genetic factors remain to be identified. Basically, some people have a hypothalamus that defends against fat gain more effectively. Other brain regions are certainly involved as well.

We still have much to learn but we do have an outline of why people differ in susceptibility to fat gain, even if put in the same environment.

Unknown said...

Perhaps Stephen, we should find the gene that regulates vanity! Or maybe there one and the same?

Serious though, do you find the genetic predisposition is variable among siblings?

If half of the variability is genetic, then how difficult given access to food is it going to be to reverse a growing problem, one that didn't exist a short time ago?

Does this indicate that, in the context of easy access to food/excess calories, the ultimate answer will likely be pharmacologic?

Grinch said...

@Stephan since this is the case about genetics, do you think that the "eat less, move more" approach to treating obesity has any validity? To what extent do we really have any control over our long term weight status?

Anyways I'm really waiting for the day that we can determine precisely what are the obesogenic properties of food and which really aren't.

Gretchen said...

"Here are the USDA data on macronutrient consumption since 1970, corrected for loss (28.8%)"

Loss of what?

Stephan Guyenet said...

Hi Mark and Dave,

The important thing to remember about genetics is that genetic variability in many cases determines the response to a specific environment. If you change environments, you can get a different response from the same genome. That's why genetic variability can have such a major impact on body fatness, yet genetics nevertheless doesn't explain the modern obesity epidemic.

The precise 'defended' level of fat mass isn't encoded in the genome-- what is encoded is the degree of susceptibility to a fattening environment (with some occasional exceptions). The point is that yes, we can impact our weight status regardless of genetic susceptibility, but it's a lot easier to prevent fat gain than to correct it.

Hi Gretchen,

USDA data are based on estimates of domestic food sales, but not everything that is sold is consumed. To correct for this, the USDA uses a (admittedly crude) correction factor that approximates the loss of food to waste between sale and consumption. That's what I mean when I say "corrected for loss".

Stephan Guyenet said...

To add to my previous comment, I don't mean to imply that everyone has the same genetic capacity to be ripped. Some people have a genetic tendency to carry more fat than others, and becoming '6-pack' lean will be nearly impossible. However, I do think almost everyone has the genetic capacity to have a metabolically healthy and fairly lean body. To fully express this capacity requires a healthy diet and lifestyle from pre-conception throughout adult life (and perhaps even in the previous few generations).

Unknown said...

"The point is that yes, we can impact our weight status regardless of genetic susceptibility, but it's a lot easier to prevent fat gain than to correct it."

Stephan, as I stated earlier, I have found this to be true for many years. It's easy to say that those who chose early not to let themselves get too far down the excess body fat road have a predisposition for that decision. I am not so convinced, virtually all I know that took that path did so out of vanity, which of course may be one and the same.

Again, I'm not yet convinced that if you take the majority of people, and over a long period of time keep them in a hypercaloric state, that when you change that environment, the body wouldn't react as violently as I have witnessed with many that try to enter a hypocaloric state after becoming obese. I have witnessed over years those that become obese have hunger pains that are very strong and very overwhelming, and they recognize the difference over time. I also used to think they were entering a state of hypoglycemia but I am no longer convinced as one I now work with to lose weight has done so after being diagnosed with T2, and he keeps very close eye on his levels, and I used to think he was having that problem when witnessing his ravenous hunger, but I now know that's not the case.

Anyway it's a very interesting topic, an unfortunate situation for many, and thank goodness people like you are working on the problem.

Anonymous said...


The Institute of Shortening and Edible Oils (ISEO) has some detailed data on fat consumption from 1970 to present which permits the percentage of animal and vegetable fats to be estimated. The data shows only a slight increase in total fat consumption between 1970 and 2010 (+4%), but shows that during that period consumption of fat of animal origin decreased from 61% to 40% of total fat consumed.

Going further back, it is reasonable to estimate that a century ago animal fat constituted 80 to 90 % of total fat consumed.

Decreased consumption of animal fats has resulted in a large decrease in intake of vital nutrients such as vitamin K2 and vitamin B12. Recent studies suggest that vitamin K2 deficiency is associated with increased risk of diabetes.

The increased intake of linoleic acid (LA) from vegetable oils results in increased inflammation which contributes to heart disease and other problems.

These factors may contribute to the increase in obesity caused largely by increased intake of calories.

Anonymous said...


One factor that effects food intake is the hormone cck produced in the gut in response to fat that reaches the illium., (The illial brake.) This mechanism reduces speed of passage of food through the gut assures that absorption of fat is maximized.

As people age sensitivity to cck increases and as a result the length of time required for digestion can increase by 50% in those over 70 and result in excess unintended weight loss.

Anonymous said...

re preceding post: spelling correction.

It is "ileal brake" and "ileum"

Jonathan said...

Before I comment, I'd like to disclose my bias towards QUALITY of calories versus QUANTITY. I used to be entirely in the calories in - calories out camp, but after a significant amount of research and self-experimentation I've changed my mind. I'm still open to flipping back to CI-CO or settling in between if someone can provide compelling evidence.

My issue with the assumptions that "The modern 'obesity epidemic' began ... and corresponded with an increase in calorie intake" and "CLEARLY an increase in calories consumed has caused the obesity 'epidemic'" is that there's clearly confusion with regards to correlation and causation. One could just as easily say "CLEARLY a [significant] increase in carbohydrates [or a modest increase in fat/protein] consumption has caused the obesity 'epidemic'."

In fact, if you do the math, beginning with the year 1977 as a caloric baseline (reasonable considering the implementation of the new USDA guidelines and the chart found in and ending in the late 2000s, the simple CI-CO model seems to fall apart.

(I don't have the raw data, so this is pretty simplified) A linear yearly increase in daily excess calories, reaching 350 extra/day in 2009 (about an extra 11/day/year) leads to a cumulative excess of 2,107,875 calories over the course of 30 years. At 3,500 cal/lb of fat, that's an additional 602 pounds of fat per person. So the average 30 year old weighing 160lbs in 1977 now weights almost 800lbs?

Can anyone reconcile that issue?

Anonymous said...

Perhaps we shouldn't be looking at the percentages or amounts of macronutrients people are consuming, because they really haven't canged that much.

Perhaps we should focus more on the FORM that those macros come in. ie processed vs whole food. And how that has changed over time.

Someone should graph that.

Stephan Guyenet said...

Hi Jonathan,

It is not an assumption that increases in calorie intake cause fat gain; it has been demonstrated experimentally many times. Mathematically, the increase in calorie intake can single-handedly explain the US obesity epidemic, according to the best available models.

The issue you raised is resolved by a more complete understanding of the relationship between calorie intake and body fatness. It is not correct that a person will continually gain 1 lb of fat per week if he overeats 3500 kcal per week. What happens is he will reach a new higher plateau, because energy expenditure increases as body weight increases, and eventually the two equal one another and weight gain ceases.

In 2012, we are at a new weight plateau where our energy intake and expenditure are both about 350 kcal/d higher than in the 70s. I should write a post about this because it's a frequent source of confusion.

Dan said...

Stephan, with respect to Jonathan's post -- you recently had a blog post where you reviewed a paper that showed that different macronutrient ratios resulted in different metabolic expenditures. It is true there are caveats to that study, namely it was for metabolically deranged subjects, but nonetheless the fact that metabolic rate can change by 300 calories is good evidence that calories in/calories out is not strictly true. Sure, if you gorge and consistently eat 500+ excess calories per day, no matter their "quality", you will still gain weight.

p.s. what is the adjustment for year 2000?

Puddleg said...

Of course not all fats sold are consumed, and this is especially true of fried foods. Restaurants have grease traps, some people put used deep-frier oil in their cars.

In China so much oil is wasted that recycling it is an underground industry.

Possibly the sale of oils and fats is a very poor match indeed for their dietary consumption.

Gretchen said...

I wonder if anyone has ever tried to see if different types of people waste more or less food. For example, when you're rich, you can waste a lot, stop eating when you're full.

When you're poor, you probably don't want to waste. So poor people might be eating a higher percentage of various nutrients than the average.

@Dan "metabolically deranged subjects." I'm thrilled to discover I'm deranged. My friends always suggested it, but now I have proof.

Grinch said...

@Gretchen, look everybody thinks they are different, they are the outlier, but it simply can't be true otherwise more of these outliers would be observed in studies. What we do know is that most people are incapable of objectively observing their own dietary behavior. That includes me. I know if I were to count calories, the number I would come up with would be way different than if Stephan counted my calories.

There are about two dozen metabolic ward studies proving that calories determine weight status. Period.

With that being said, we seriously need to stop saying that calories are the CAUSE because that implies they are the root cause and that calorie intake is NOT regulated by the body.

Gretchen said...

@Dave I don't think I'm different. I was referring to that rather large "loss" that Stephan cited. It could skew results if not uniform.

Studies select participants, and those who have some metabolic difference may be rejected for that reason. I think any blanket generalizations about weight gain/loss are apt to be wrong. There are always exceptions. We really know so little.

Metabolic ward studies show that different people lose/gain weight at different rates.

Stephan Guyenet said...

Hi Dan,

The Ebbeling et al. study was the first credible suggestion that diet composition might be able to influence energy expenditure in a meaningful way under specific circumstances (i.e. weight maintenance after fat loss). I look forward to seeing that independently confirmed and then explored further, and as I said in my discussion of it, I suspect we'll eventually find multiple ways by which diet influences energy expenditure in humans (as we have in animal models).

I don't even know what people mean when they say "calories in, calories out" anymore. It's often used as a derogatory label without defining it. If it means the first law of thermodynamics appplies to humans, then yes it's correct. If it means energy balance is not physiologically regulated and is purely under the control of voluntary behaviors, then it's not correct.

To adjust for the artifact in 2000, I simply calculated the difference between 1999 and 2000, and subtracted that value from the 2000-2006 data points. The true change from 1999-2000 is lost due to this adjustment, meaning it introduces some error into 2000-2006, but it's less error than there was without the correction.

Jonathan said...

re: First Law of Thermodynamics


I agree with you that evidence suggests that calorie regulation is not entirely under voluntary control; however, I would argue that the use of the First Law of Thermodynamics in the context of calories in-calories out is at best misleading and at worst totally incorrect.

Peter Attia does an excellent job of explaining the short-comings of the 1st Law as applied to the human metabolism in his post I don't subscribe to all of his views on fitness and nutrition (particularly regarding the ketogenic diet), but he does offer a pretty compelling counterpoint to CI-CO. Both Whole Health Source and The Eating Academy are on my list of daily required reads... it's good to get a variety of viewpoints. Attia holds an M.D. and degrees in mechanical engineering and applied math. He ~probably~ has a firm grasp on the intended application of the 1st Law.

Stephan Guyenet said...

Hi Jonathan,

You said "I would argue that the use of the First Law of Thermodynamics in the context of calories in-calories out is at best misleading and at worst totally incorrect"

I don't know what you mean by "calories in, calories out". That phrase is used in so many ways and almost never defined.

The fact that changes in the energy content of the human body (reflected mostly in fat stores) depend strictly on the amount entering minus the amount leaving is not in dispute. That's what the first law says. I'm not sure what position you're arguing against-- perhaps the way in which some people have interpreted the first law?

Grinch said...

Personally I think when LC'ers bash the first law, its more of a red herring than anything. They use it as a means to justify irrational beliefs about weight loss that are not supported by science.

For example, Taubes says the laws apply, but that Cals IN and OUT are dependent on each other. This is true to an extent, but he abuses the law by saying the composition of Cals IN (ie. fat vs. carb) can greatly determine Cals OUT, and metabolic ward studies dispute this completely.

This is coming from someone who eats LC by the way.

Jonathan said...

We might be arguing two sides of the same coin, I'm not even sure at this point...

I use "calories in-calories out" and "a calorie is a calorie" synonymously to refer to the common belief "...that it is impossible for two isocaloric diets to lead to different weight loss."(1)

If that were strictly true, then a 3000 kcal diet of only protein would lead to equivalent fat gain given a 3000 kcal diet of only fat, thus proving that "a calorie is a calorie". However, a more plausible outcome might be that the 30% thermic effect of protein vs. 3% thermic effect of fat would lead to lower fat deposition in the protein diet (ignoring hormonal, satiety, palatability effects, expenditure, etc.) because fewer calories are available for storage. Taking it one step further, the protein diet, when increased to 3100 kcal might still lead to lesser fat gain then the 3000 kcal fat diet, despite containing more calories. (You could even argue for the possibly of fat loss from a hypercaloric protein diet). If this is true, it seems to partially invalidate what you called the experimentally demonstrated observation that "increases in calorie intake cause fat gain."

Am I screwing up the logic somewhere?

(1)R Feinman,

Stephan Guyenet said...

Hi Jonathan,

I see what you're saying. I think it's putting it strongly to say that "a calorie is a calorie" means "it is impossible for two isocaloric diets to lead to different weight loss". Scientists generally avoid the word "impossible". I think most people would put it more like this if you asked them to be precise: "calorie intake is by far the main dietary determinant of body fatness, and if there are other dietary influences they have yet to be demonstrated scientifically".

As you mentioned, it's well known that there are certain aspects of diet composition that have a modest influence on energy expenditure through the thermal effects of food. The thermal effect is greatest for protein, intermediate for carbohydrate and lowest for fat, but in practice only protein has enough of an effect to suspect that it might be significant. But studies where calories are controlled and body fat is directly measured suggest that the effect of protein is so small that it has no detectable impact on fat mass independently of calorie intake. It's possible that longer studies with more subjects would tease out a small effect, but it's clear that it's a distant second to calorie intake.

Protein has been shown to benefit body composition though because it helps build/preserve muscle mass, but that is independent of its effect on fat mass. The other thing high-protein diets do is decrease overall calorie intake in free-living people, so they are helpful for fat loss in practice. But if you control for calorie intake the effect disappears.

Stephan Guyenet said...

I guess the simplest way to sum it up is that research has not yet turned up differences between different calorie sources when it comes to fat loss. The differences due to the thermic effect of food appear to be too small to matter in practice. However, diet composition does influence how much you eat, and that determines how much fat you carry, so diet composition is still important.

Ty Fyter said...

Hey Stephan,
i'm a first year bio student and have a question. on a cellular level, would there always be an 'excess of nutrients outside a cell'? it's something i heard, i think it means that there is always nutrients outside of a cell that it could take up. what are your thoughts on this?


Jonathan said...

OK, I'm satisfied with that answer, glad that's cleared up. I look forward to reading your next post.

Gretchen said...

In thinking about diets, there are two major questions: (1) How much of what you eat converts to fat (you can study this in metabolic wards, and it's of academic interest) and (2) What determines how much people eat when free living?

The latter is what really matters to dieters, and it's very difficult to study.

whitefox said...


those are the million dollar questions. Yet, having looked at a bunch of the research/mechanisms for the second question, I feel like it has already been answered. Don't we know that volume of food, weight of food, fat vs. carb vs. protein composition all influence how much we eat based on a combination of paracrine/endocrine signaling? CCK, PYY, ghrelin, orexin, leptin, and insulin are all diet and hunger affecting hormones, with maybe some dopamine effects on the brain front (I know some of those other hormones can pass the blood-brain barrier as well). So as we know that fat releases a certain combination of hormones, and carbs release a certain combination, and we can do some energy density (cellular carbs vs acellular carbs vs total calories) calculations and we should be able to find some reasonably mathematical answer....

But we haven't. And while I cannot (yet) fathom a computation/metabolic ward study that complex, I know that it can and will be done. Humans are too smart to fail at such an important topic, however complex. The good news is that, barring genetics and a bit of epigenetics, we know what makes people healthy. Paleo-style eating with its focus on cellular carbs (whether this matters more for microbiota-induced weight gain or for LPS-induced allergies/sickness, I don't care beyond an academic interest, the result is the same). If Kitavans/Okinawans can eat a bunch of carbs and live longer than some low-carbers, it just goes to show that genetic goodies + natural carbs = long life. There are more benefits to a high-fat diet, however, like increased protein synthesis (if high-fat or high-carb the right way lets me live to 100, I want to be ripped until the day I die). Maybe I'll make a pros-cons list of the effects... though I think IF + carb refeeds after workouts and LC everywhere else is probably the healthiest so that neurotransmitters and muscle glycogen and leptin and body fat are all perfectly regulated. Thoughts?

Puddleg said...

What would happen if you consumed 3,000Kcal a day from alcohol alone?
My guess is you'd lose weight. Your face and liver would be puffy but the rest of you would be stick thin.
There is a SECOND law of thermodynamics people; energy, like time, cannot be created or destroyed, but it can be wasted.

Puddleg said...

Interesting paper comparing nitric oxide and prostaglandins on diets of
a) dairy fat
b) high linoleic sunflower oil
c) high oleic sunflower oil
in humans.

Puddleg said...
Appetite dysfunction in obese males: evidence for role of hyperinsulinaemia in passive overconsumption with a high fat diet.

Testing the null hypothesis seems to have worked out for these researchers.

Grinch said...

@George there is simply no evidence to support the idea that diet composition affects energy expenditure or waste in a meaningful way.

People who cite the laws of thermo to defend a metabolic advantage simply don't know what they are talking about. Its not that its physically impossible, its just that it hasn't been observed to happen in humans. If it hasn't been observed, there is no reason to claim it happens like LCers so often do.

Stephan Guyenet said...

Hi Jonathan etc.,

Here's a recent article by Marion Nestle that explains the "calorie is a calorie" position with more nuance than you usually see. My position is not as firm as hers on the role of calories in body fatness, although I do think it's the dominant factor.

Hi George,

The study you cited was not able to determine cause-effect relationships, it simply noted an association. People who have elevated insulin are typically different from people who don't in multiple ways besides insulin.

Am I understanding correctly that you think the null hypothesis is that elevated insulin causes fat gain? If you think insulin plays a role in body fatness, I'd say a better null hypothesis is the opposite: that insulin resistance (and consequently elevated insulin) protects against fat gain. Most studies have shown that when you look prospectively, elevated insulin at baseline predicts less fat gain over time.

Unknown said...

"insulin resistance (and consequently elevated insulin) protects against fat gain."

Puddleg said...

This was the null hypothesis:
"To test the hypothesis that caloric and fat intake in a pre-load meal have no subsequent effects upon blood glucose and insulin concentrations, perceived hunger, subsequent food intake and appetite control in lean and obese men."

This is the way to phrase a hypothesis if you want to avoid the appearance of bias; you are testing the hypothesis that there is no effect.
So in fact, the null hypothesis is just semantics, but it does avoid predicting your result before you find it.
The interesting part of the study was, that an obese person eats more if they also ate more earlier, and a non-obese person does not.
One parallel that comes to mind is alcohol; an alcoholic will drink more if they have already drunk earlier in the day, a non-alcoholic will not.
In the case of these obese subjects insulin was not a satiety hormone.
Apart from that, the research begs for more study variants to provide more data. It could be used to support almost any point of view as it is, but nonetheless it's interesting.

Are you going to post about the Roundup & GMO corn cancer mice?
It's good to be reminded that the quality of food and the types of toxins in it have been changing faster than macronutrient ratios.

Puddleg said...

The insulin level was taken before the test meal but after the pre-load, so was post prandial.
We do not know if they gained weight when the two larger meals were consumed, we only know they ate more.
If a calorie is a calorie, they would have gained weight whether insulin resistant or not.

Puddleg said...

Now we're talking!
I found the PUFA data I wanted - on this very blog:

Now we're cookin' with gas.

Stephan, you've forgotten more than I'll ever know.

Grinch said...

"In the case of these obese subjects insulin was not a satiety hormone."

So is insulin the only hormone that affects satiety? Is it possible that insulin IS a satiety hormone but that these obese subjects ate more because of defects in another part of the complex feedback loop of appetite and satiety?

Puddleg said...

It is not only possible but probable; indeed, IR alone could explain it.

Can IR protect against weight gain in overfeeding if a calorie is a calorie?
What a mindless redundancy that is.
Actions and choices have consequences.
If I eat 1,000 Kcal meal that leaves me satiated and energetic, it's less fattening than a 1,000 Kcal meal that leaves me sluggish and hungry.
We are talking about the combined effect of a series of events (meals), the actual effect of each (fat gain or loss) being too small or indistinct (water weight etc.) to measure with any degree of certainty.

Eat less and/or move more? Aren't there food and exercise choices that don't involve calorie counting yet still tend to produce that result in futurity more than others?

Anonymous said...

I find this blog immensely interesting, thanks for the great work. I must admit, though, that knowing what an ideal diet for a human being should look like has also caused me quite a lot of stress. It is very clear that it would be absolutely impossible to feed the world a high meat (or high-potato, by the way) paleo diet, also I think the way meat is rised nowadays is inhumane and ecologically terrlible. Also, as a person who loves food and cooking more than anything and as an italian who has basically survived on white bread and pasta and delicious food, knowing that grains and wheat in particular may be harmful and that reward (i.e. deliciousness) itself is key to obesity and other serious helath conditions is rather disturbig, especially because I used to consider my almost vegetarian and almost sugarless diet much healthier than average. I feel very tempted to try out paleo, but also worried. I know it wouldn't suit my taste and if I found out that it did improve my well being so much, it would deprive me of my favourite thing, good traditional food. I guess this kind of extreme diet changes are much more difficult in a country with such a strong food tradition like Italy and impossible/unethical on any large scale. I don't know what to do or think.

ann. said...

Grazie rigalimoni!

I think this is exactly how many of us feel about all of the conflicting nutritional information around. I guess it's always been like this except before the all-embracing internets, the stress of eating "right" wasn't so in everyone's face?

But who was there can ever forget the fat phobia of the 80's? My mom fed us margarine and other weird chemical butter ersatz concoctions and I had reoccurring bad eczema to the point where I'd go outside and push my arms in snow to find relief from the itching. Only after I started eating real butter did the eczema disappear (for good it seems so far).

I can't help but compare that fat phobia with the current and escalating anti carb hysteria. Will wheat also be in this same maligned category with fat and carbs in a couple of years?

rigalimoni, I hope you saw my links to some articles on food written by Desmond Morris, in Stephan's latest post on calorie restriction? They're a bit of relief to a "right nutrition" frazzled mind.

On CR, I'd recommend the BBC Horizon documentary on CR/Intermittent fasting "Eat, Fast and Live Longer" ( ).

Honestly, intermittent fasting in any variation that you can stomach it seems to be the best solution.

Puddleg said...

@ Rigalimoni,
as long as YOU are healthy your diet is healthy, might be a good default to take.
In the case of wheat, I think the evidence is conclusive that a) it is very harmful to some people, and b) it is too often impossible to tell which people those will be until it is too late.
Someone who is sick on wheat is very sick indeed and might die. Coeliac had a 30-40% childhood mortality.
The case against saturated fat was always conjectural (the lipid hypothesis), that against wheat begins with an undisputed clinical entity.
No-one needs to feed the world, just one person or possibly a family. Vegetarians do not eat cereals alone, and many vegetarian choices seem very wasteful of land and water, not to mention fossil fuels, which are essential for bulk cereal farming, but much less necessary for pastured livestock (how economic are lettuces and cucumbers?).
The demand for cereals and vegetables also destroys the livelihood of pastoral peoples in the 3rd world, and the supply of cheap calories increases over-population; the agricultural revolution is the very reason there is a food crisis to begin with.
Vegetables are also the main market for weedkillers and pesticides.
I think we are free to make our own choices, and should not not feel obliged to solve the world's problems on our own.

Anonymous said...

This graph is very clear to understand. It is clearly showing that increased in calories consumed is the caused of obesity. One more intersting blog of heath is

Jane said...

'.. I think the evidence is conclusive that [wheat] is very harmful to some people ..'

Do you know of any evidence that wheat is harmful to people who have eaten whole wheat and no refined wheat all their lives? If you are to conclude 'wheat is murder' as you did on Hyperlipid recently (yes I have spies everywhere) you must provide this evidence. I expect you know that there are people with coeliac genetics who never get coeliac.

You also know (because I told you) that there are intestinal proteases activated by metals removed from white flour. See 'The activation of intestinal peptidases by manganese'. Some of these peptidases even break tricky proline bonds. Obviously, if gluten is degraded it isn't a problem.

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

It is really important to take care of our Health, Always eating of Process Foods specially to those with preservatives is not good and advisable for our Health <a

Puddleg said...

@ Jane, coeliac genetics (or most other disease genetics) is a very complex business; as per the title of a recent book (which I don't mind recommending as a primer) "It Takes A Genome". "Marker" genes are like marker lipids; not always reliable; a statistical correspondence, not an absolute one.
Even if gluten intolerance is due to some unknown flaw in processing,
or mistakes in weaning, or micronutrient deficiencies, why bother eating a food with that inbuilt risk, and no special virtues, that can so easily be replaced by something safer?

Anyway, this is some evidence linking obesity and linoleic acid plus carbohydrate:

or is that excess calories from linoleic acid and carbohydrate?
In the experiment, excess linoleic acid drove the consumption of excess carbohydrate (calories) which lead to obesity...

Puddleg said...

@ Jane,

assuming that eating whole wheat is safe (which I don't), you can't make bread or pasta without refining wheat in some way.
I have never seen whole wheat in a dish.
Pumpernickel bread is as close as it gets to wholegrains, and that's rye.

People started refining grains for a reason, it wasn't just a perverse desire to decrease micronutrient intake, but to make them more diggestible, an attempt to lessen their intrinsic harm.
Keep trying, peoples.

Unknown said...

Being healthy is no t being obes, we need to take care of our health dont always eat process food and foods with presevatives.

Puddleg said...

Around the whole excess linoleic acid
(current intake of 8% calories vs. 1900 1%) > elevated anandamide and cannabinoid tone > excess consumption of carbohydrate (calories)> obesity

Endocannabinoid hedonic hotspot for sensory pleasure: anandamide in nucleus accumbens shell enhances 'liking' of a sweet reward

Now, acetaminophen also elevates endocannabinoids. Has anyone cracked the numbers for Tylenol/Panadol plus omega-6 = obesity? I can't find anything online.
This is a plausible explanation for modern over-eating that may do more to address individual variation than most models. "healthy eating" parents who don't dose their kids or themselves with acetaminophen, use herbs or homeopathics instead, may not see so much obesity, despite adherence to the despised food pyramid.

Anonymous said...

Granted, the USDA dietary guidelines will not make you fat, but that does not mean that following the guidelines is healthy. I followed the USDA dietary guidelines for two decades beginning around 1980. I substituted margarine for butter, virtually eliminated eggs, limited cheese to one serving per week, drank only skim and low fat milk and ate mainly skinless chicken breast, fish and meat with all visible fat removed. I ate a lot of good vegetables and pasta. Breakfast was cold cereal with skim milk. I never drank soft drinks and ate very little sugar.

I did not get fat but I subsequently had many health problems as a consequence of the high carbohydrate diet and lack of nutrients that come mainly from animal fats. After ten years on the low fat diet my teeth began to break off which I now believe was due to lack of vitamin K2 needed for activation of osteocalcin.

I later developed calcification of the aortic valve and coronary arteries which resulted in valve replacement and a triple coronary bypass. The calcification was the probably mainly result of severe deficiency of vitamin K2 due to avoidance of animal fats, the main source of vitamin K2. The high carbohydrate diet, which increases endothelial inflammation that results in lower HDL, also contributed to calcification and atherosclerosis.

About ten years ago, after finding the website of the Weston Price Foundation, I completely abandoned the USDA diet and converted to a diet high in animal fats, including whole fat dairy, and saturated vegetable fats but no polyunsaturated vegetable fats. The increased fat intake and the decreased intake of carbohydrates and LA from vegetable oils greatly reduced endothelial inflammation as indicated by higher HDL cholesterol (from 55 to 85) and lower TG (from 120 to 60). My blood pressure dropped and my ejection fraction has increased to 75% which is about as good as it gets. I did not gain weight even though I increased my fat intake from less than 30% to over 50% of calories. (Wt. 155, ht. 6’1”)

The USDA dietary guidelines may not make you fat, but it is very unhealthy because it results in excess intake of LA and many, many nutrient deficiencies.

Puddleg said...

@Jack C,
that's a story I can relate to.
Maybe I didn't follow the guidelines as exactly as you did, but so many of the foods they approve were directly harmful to me - grains, soy, oils, some fruits and veges - and some they disapproved were beneficial - animal fats, other SFAs, meats, salt.

Here is the acetaminophen consumption graph. Does it match obesity?

Massive increase in consumption between 1978 and 1990.

Anonymous said...

@ Jack C

Interesting, that's similar to my situation. I'm a young female in my early 20s with high blood cholesterol, triglycerides, slightly elevated liver enzymes, but all other markers are clear. I had that blood test a couple of years back when I was eating a vegetarian diet that was quite high in carbohyrate, and moderate fat consumption. I was still active, brisk walked about 1 hour each and every day, yet I was confused and rather shocked about my blood test.

I have switched to a lower carbohydrate diet, higher fat and moderate protein diet with plenty of vegetables and a little bit of fruit. I've basically eliminated all grains and it has improved my IBS by a lot. i'm yet to take a blood test soon sometime, but since I have just been pregnant I'm letting my body recover or return to 'baseline'. I'm otherwise at healthy weight, 119lb at 5'5, actually quite skinny. But I don't have genetics on my side since there's a family history of cardiovascular disease, which is already starting to affect my father. I'm making sure I do all I can to avoid it as much as possible.. can't alter genetics, but can influence the environment.

Puddleg said...

Apologies, the Tylenol chart is actually autism incidence tied to tylenol landmarks. I'll keep looking.
But the critical event in Tylenol consumption is the Reyes syndrome aspirin scare of 1980. After that, acetaminophen was the only NSAID given to children.

Unknown said...

It is hard to tell exactly the cause of obesity because everyone is different, but being obese can truly affect the way you think and how you feel. In my case, its probably genetics and my lifestyle.

Anonymous said...

Jack C. (me) said: "Granted, the USDA dietary guidelines will not make you fat".

Apparently I was WRONG. According to a recent study (in rats) noted on Mark's Daily apple, the USDA recommendation to substitute of margarine for butter and vegetable oils for animal fats results in an increase in intake of LA which stimulates an increase in food intake which makes rats fat. I suppose it makes humans fat too. So excess LA not only causes inflammation that leads the CHD, shortens your lifespan (shortens telemer length) and causes ulcerative cholitis, it also makes you fat.

Anonymous said...

The PDF for the aforementioned study on the fattening effect of LA is:

Puddleg said...

And it's a doozy.
Emily Deans has written about it here:
and here:
I can see how this might reconcile all sorts of currently fragmented opinions; food reward, CIH, CICO, excess fructose and linoleic acid as food toxins, omega 3 deficiency - this anandamide model could be put to work to create a Unifying Theory of Obesity.
Which could serve as a prototype for future inclusive models.