The Dietary Source of Body Fat

Over on Stephan Guyenet's blog, in the comment section of his post on Humans on a Cafeteria Diet, a little discussion was started by one disgruntled reader (or I suppose ex-reader since this post apparently pushed him over the edge to unsubscribe) regarding where the fatty acids in our body fat came from.   Stephan wrote:  
When a diet of mixed macronutrient composition is eaten to excess, the carbohydrate is preferentially burned off, while the fat is mostly shunted into fat tissue. This makes sense, because why would the body go through the inefficient process of converting carbohydrate to fat for storage when it can just shunt dietary fat directly into fat tissue?
Said reader commented:  "This post has good info, but suggesting the fat is stored as fat is absolutely wrong and is bad science."    A discussion, contributed to by yours truly, ensued.  I think this is illustrative of just how damaging towards ultimate progress in the realm of understanding, preventing and treating obesity Gary Taubes has been.   On page 387 (according to Google books, my ebook page numbers are off) of GCBC, Taubes writes:

Now, Taubes' referencing is rather sketchy.  He cites that 1965 Handbook of Physiology for the quote on lipogenesis, but there's no indication of where he got the "perhaps 30 percent of the carbohydrates in any one meal" from.  I'm pretty sure he got that factoid from www dot pull-it-from-your-netherregions dot com.  Interestingly this is one of the many myths attributable to Taubes that simply was left out of WWGF.   Even before doing that, Taubes vaguely backed off on the whole carbs-to-fat angle in his recent lectures because Carson Chow and Kevin Hall -- those two young biophysicists at NIH -- had set him straight in that regard.  No doubt Taubes had help spreading this myth from a pair of Primal authors as well.  Mark Sisson is a big parrot of the glucose converted to fat meme, though I can't find where he cited both that 1965 text and GCBC as two sources for the same assertion on his blog.  Hands down, however, Nora Gedgaudas has my vote for the biggest fool anyone listens to on nutrition.  On her website she writes:  
Your body tries to take sugar from a meal out of the bloodstream as quickly as possible. The first order of business is to send glucose to your cells for immediate energy. If those cells are insulin resistant, then the sugar has to go somewhere (and energy cannot get into the cell). Your body sends some glucose to storage in the liver and muscle as glycogen. The rest of the glucose (i.e., most of it) goes to the liver to get converted into triglycerides so it can get sent to storage as body fat. Unless you have a very high rate of metabolism (not necessarily a good thing) you are likely to gain unwanted weight. This conversion to fat from sugar is a labor intensive process metabolically and takes a LOT of energy to accomplish. –It takes even more if a lot of fat was eaten at the same meal as the carbohydrates. Since burning the carbs off is priority #1 (and because it is impossible to burn fat AND sugar at the same time), whatever dietary fat is there also must be first converted to sugar before it can be re-converted to triglycerides and finally stored as body fat. (“All body fat is made from glucose”—Basic medical Biochemistry). This is a very energy INefficient process and takes an enormous amount of energy to do. ... (See pages 78,88, 99, 105 and 167 in PB-PM).
I don't have the book, but given the page citations, and having read some reviews of the book, it probably contains something similar.  Elsewhere on her blog, she's attributed that quote to the Textbook of Medical Physiology, which I assume is Guyton (earlier additions) or Guyton & Hall (later editions).  
Long chain saturates such as 18-carbon stearic acid (the most saturated fat in the body) is THE preferred fuel for the human heart. Stearates are commonly synthesized from glucose in the diet and are the primary storage form of saturated fat everyone wishes they had less of. If you don’t like it, eat less carbohydrate (the Textbook of Medical Physiology states “All body fat is made from glucose.”)
This is absurd!  Even more absurd is that a publisher picked up her book for re-release this year.   Which just goes to show you that publishers are rarely concerned with accuracy of their publications.   I emailed Nora when I first read that second blog post and asked her for the citation.  I have two versions of that text (it is hugely popular), hardcopy 6th ed from my college days, and the 11th ed  in PDF.  I searched the PDF for that quote -- and she puts it in quotations as a direct citation -- and can find nothing of the sort.  So I asked her about it and she stood by it, she was adamant that she didn't make it up.  Gedgaudas' telling of metabolism is even worse than that of Gary Taubes', and that's a high bar to clear!  Before we move on: 
  • Humans can't convert fats to glucose.  Period.  If we could, why would we bother wasting energy on gluconeogenesis from, amongst other things, amino acids?  Why would we even form ketones?

EDIT:  1/12/12  In light of Chris Masterjohn's post, I'm leaving the above original bullet point as is, but will qualify that here. 
  1. The pathway(s) are far from being a direct conversion of fats to glucose.  Whereas glucose → acetylCoA → fatty acid (via DNL), this new information does not state that fatty acid → acetylCoA → glucose via the TCA cycle.  
  2. To whatever degree acetone, derived from fatty acid metabolism, can be converted to glucose, it is not even significant in adults with diabetic ketoacidosis (his reference states minimum of 2.1%). 
  3. The possibility of pathway(s) in Chris' post, therefore, do not substantially alter the statement.  In future posts, I'll qualify the statement, and leave out the "period", but that conversion of fats to glucose doesn't occur in animals remains a functionally factual statement.
  • We are always burning fats and carbs at the same time, just in different proportions.  We are never burning only one or the other.
  • If your cells are IR, the glucose isn't getting into the liver and muscle cells to be converted to glycogen ... but why quibble ;-)
  • Human body fat is about 50% oleic acid, the primary product of hepatic DNL is palmitic acid which makes up roughly 20-25%.  Where does all the oleic acid come from Nora?
  • Even Taubes acknowledges the TAG/FA cycle and the uptake of dietary fats into adipose tissue is established fact.   
  • I have no idea what "the most saturated fat" means, all saturated fats are equally so.  Saturated means no multiple covalent bonds between carbons.  It's also not the predominant saturated fat in the human body making up roughly 5% of fat tissue.  
  • Why, as Stephan asked in his post, would we go through the -- as even Nora describes it -- wasteful/energy intensive process of converting fats to glucose only to convert them back again to fat to be stored?  
All in all, if internet discourse since the whole Taubes/Guyenet episode at the Ancestral Health Symposium is any indication, these myths sadly persevere -- carried forward by TWICHOOB's who have not received the "oh nevermind" memo from Taubes or who get their information to this day primally, no make that primarily, from the primal carbobbsey twins.  This is why I get so annoyed at these pseudoscientific "experts".  Please Ancestral Health peeps, can you not invite these people to speak next year?  The "buy in" should not come at the price of credibility and integrity.  Now that I got that off my chest .....

OK ... So let's talk where body fat comes from.

The vast majority of it does indeed come from dietary fat, and very little of it from carbohydrates.  I've listed some relevant prior blog posts discussing this below:
Nutrient Fates after Absorption
Excess carbs converted to fat?
Fat Futile Cycling from Carb Excess (or more lay-friendly)
Where do triglycerides come from? Part I, Part II, and Part III
Marc Hellerstein's group at Berkley has been looking into this for quite some time.  I wonder if he attended Taubes' lecture in 2007 :-)  I would like to briefly discuss three papers from this group here.

Measurement of De Novo Hepatic Lipogenesis in Humans Using Stable Isotopes  (1991)
The fraction of VLDL-palmitate derived from de novo lipogenesis was only 0.91±0.27% (fasted) and 1.64-1.97% (fed). For stearate, this was 037±0.08% and 0.47-0.64%. Precursor enrichments predicted from isotopomer ratios were close to measured SMX-acetate enrichments, indicating that SMX-acetate samples the true lipogenic acetyl-CoA pool. Stearate synthesis was less than palmitate and the two did not move in parallel. Estimated total VLDL-FA synthesis is < 500 mg/day. Thus, de novo hepatic lipogenesis is a quantitatively minor pathway, consistent with gas exchange estimates; fatty acid futile cycling (oxidation/resynthesis) is not thermogenically significant; and synthesis rates of different nonessential fatty acids by human liver are not identical in non-overfed normal men.

DNL is not the pathway of first resort for added dietary CHO in humans, at least on Western (high-fat) diets.  DNL can occur, but it generally does not. A `functional block' therefore exists between CHO and fat in humans, analogous to the absolute biochemical block in the direction from fat to carbohydrate in all animals. Therefore, the two major macronutrient energy sources are not interconvertible currencies in the mammalian organism; they must be considered separately and are probably regulated independently, by separate signals and toward separate ends. The major insight concerning DNL is therefore a negative one. Many questions related to this central observation still remain unanswered: what is the functional significance of DNL in adult life (Table 9)?   What are the ultimate limits of DNL in humans?   Is DNL only used as a final `safety-valve' for CHO in the organism?   What constrains DNL in human lipogenic tissues? Are there regulatory roles played by DNL that we have not yet identified?  Regardless of the answers to these questions, the metabolic and clinical consequences of the apparent functional block between CHO and fat are profound and have only begun to be understood.  

In this study they used radiolabeled water to determine the turn-over rate of fatty acids in fat tissue and the contribution of DNL-derived fatty acids.  This is the strongest citation in favor of any significant contribution of dietary carbohydrate to fat tissue accumulation I've ever come across.  
The turnover of adipose tissue components (lipids and cells) and the pathways of adipose lipid deposition have been difficult to measure in humans. We apply here a2H2O long-term labeling technique for concurrent measurement of adipose-triglyceride (TG) turnover, cell (DNA) proliferation, and de novo lipogenesis (DNL). ... Net lipolysis (TG turnover) was 50-60 g/day. DNL contribution to adipose-TG was ... ∼20% of newly deposited TG. ... In summary, long-term 2H2O administration to human subjects allows measurement of the dynamics of adipose tissue components. Turnover of all elements is slow, and DNL contributes ∼20% of new TG.
That sounds like a lot for an insignificant path ... however, consider that 20% of even 60g is 12g/day.  Also, by my math that means that by far, the largest contributor to new triglycerides is fats at 80%!!  There are some additional caveats to be made here.  This was conducted in weight stable non-obese people (BMI>28 excluded, subject characteristics can be viewed here).  So we're talking 50-60 g/day turnover is the TAG/FA cycle rate with a significant portion of dietary fat going to the cells for energy, 40-48 g going into the fat tissue.  If someone is consuming in excess of caloric needs, there does not appear to be much evidence that this alters the turnover rate.  The percentage of new triglycerides from fat will necessarily increase.  (I note discrepencies between this and the other studies in terms of amounts ... DNL in the fat tissue itself?).

All of this well predates GCBC, though it was done in the US after WWII.  There is absolutely no excuse for this part of TWICHOO.  Almost a third of carbs in a normal meal converted to fat? No excuse for that. None.   


psychic24 said…
The last citation is interesting. It doesn't mention carb intake for these volunteers and them goes on to make a conclusion about how much DNL contributes to fat borrow your expression, "there's no excuse for that".
MM said…
I own Nora's book, and I remember being kind of shocked that she wrote fat could be converted into sugar. At the time I gave her the benefit of the doubt and assumed she was referring maybe just to the glycerol part of a triglyceride. But I can see from what you quote above, that is clearly not what she meant at all. Not good!
Quarrel said…
Interesting post. Thanks.

I'm amazed with your patience with some people over on Stephan's blog. I keep on wanting to think they're trolling you, but unfortunately I don't think they are :(

That said, given the above, do you have any figures on the hepatic rate of fructose->triglyceride conversion? Lustig et al quote this as a major source of stored fats (assuming a diet rich enough in carbs to keep the liver happily full of starch).

If I drink my 2 litre coke/OJ/etc per day the fructose is presumably a large source of fats, even if the glucose gets dealt with as in your post?

Tonus said…
"(or I suppose ex-reader since this post apparently pushed him over the edge to unsubscribe)"

Silly Stephan, telling people things they don't want to hear. He'll never author a podcast at this rate!
CarbSane said…
Which just goes to show psychic that there's no pleasing some people. Don't we want to know how this works in the "real world"? The subjects ate their usual diet which presumably is rather high in carbs.

@Q: Might be interesting. This was a weight gain study. If in caloric balance at a normal weight, 2L Coke has 120g fructose. I think your body can probably handle that. The other thing we need to remember about VLDL is that only some of it is taken up (cleared) by the adipose tissue, the rest is acted upon by LPL to be transported into cells and burnt for energy. The triglycerides series dealt with increased VLDL being due to reduced clearance. Perhaps palmitic acid (the product of DNL) is preferentially taken up by visceral fat?

@MM: She self-published the original book from what I hear, so I guess anything goes there. But how such demonstrably wrong stuff can make it to press just amazes me. Judith Mazel on pemmican?
Anonymous said…
"whatever dietary fat is there also must be first converted to sugar before it can be re-converted to triglycerides and finally stored as body fat. (“All body fat is made from glucose”—Basic medical Biochemistry)."

I can't believe it actually says that. Amazing.
Anonymous said…
"Humans can't convert fats to glucose."

Aside frome glycerol, there might be a pathway to convert fatty acids to glucose:

Its in silico evidence but interesting none the less.
CarbSane said…
Hi Kurt. I was floored to read that and have Nora not only stick to her version but insist that it came from Guyton. Her blog posts stand now well over a year since that brief exchange. A friend is sending me a copy of her book so I'm going to see what it says in there.

@Lucas: Fair warning that the term "in silico" turns me off. I didn't know what it meant, and upon looking it up it's computer simulation basically. So, why not just say that? I've looked at that paper a few times and each time I still come away shaking my head wondering over the waste of brain power. Why? Because what they essentially did was establish that it was possible for this to happen, but we know it doesn't (or if it does it is so infinitessimally small). This much holds up well in studies and measurements of metabolism so even if someone were to detect it occurring, again, it wouldn't change a thing ... not even warrant a footnote it would seem.
Anonymous said…

Dismissing evidence only because is in silico is unfair. Most research uses bioinformatics for a great deal of different topics. In fact, molecular modelling is starting to be the start point in designing, for example, vaccines and drugs. Bioinformatics is also strongly related with biophysics, which I think is a very important area in molecular biology and has a wide range of applications. Papers like the above help us understand better metabolism and physiological regulation, providing theoretical findings and probable hypotheses. Remember that there is more than we don't know that what we know.

In particular, the study might not be relevant for the average joe trying to lose weight or discussing if calories matter. But it might be relevant for understanding other processes. Indeed the potential pathways identified are thermodinamically feasible.
CarbSane said…
Hi Lucas,

I'm not down on "in silico" per se, but two things about it that I'm not a fan of:
1. The terminology, sounds like some sort of physical method -- e.g. in situ. It almost could mean a tissue matrix of sorts ... without looking it up.
2. That something is theoretically possible may or may not be helpful to elucidating what does or does not occur or work.

That said, I see the most potential in the fields you pointed out e.g. drug/vaccine development. When I was in the industry, I worked at the discovery level. That meant an organic chemist started with a general structure and synthesized a bunch of different variations (a chlorine here, a bromine there, a double bond here, etc.) These were sent through various screens for mutagenicity and bioactivity. Those that passed went through basic tox screening in rodents, those that passed went for basic efficacy screening in same. Those last two might have been roughly simultaneous. Then they came to me ... see what happens in the rodent -- how it was metabolized, how to quantify levels, accumulation, excretion route, etc.etc.

That's a lot of man hours before it even gets identified as a candidate for development. If computer models could better predict which structures might be more effective, etc. that is a great application.

However, just identifying something as thermodynamically feasible in a living organism? I'm not sold on the usefulness of this. Yes, there's much we don't know, but the major metabolic pathways of carb, fat and protein utilization? Those are pretty well established. Thus, if it is theoretically possible it is of little interest. It is definitely possible for glycerol to be phosphorylated to make G3P for triglycerides, but adipocytes lack the enzyme for doing so -- they get their G3P elsewhere.

So I stand by my thoughts that even if it were to someday be shown that some miniscule amount of fat does get converted to glucose, it won't change anything.
Craig said…
Just before reading this post, I happened to finish reading a paper which discussed the association of visceral (as opposed to subcutaneous) fat with metabolic syndrome. After reading the two back to back, I'm left wondering if triglycerides created in the liver by DNL from carbohydrates are more likely to end up in the liver (or in the vicinity of the liver) as visceral fat, and thus are more metabolic damaging than triglycerides from dietary fats, even if they aren't the majority of the fats being stored in subcutaneous fat. I wonder what analysis of fats from liver biopsies would show, particularly for diabetic or prediabetic people?