We Really Can Make Glucose From Fatty Acids After All! O Textbook, How Thy Biochemistry Hast Deceived Me!

Visit Us
Follow Me

Biochemistry textbooks generally tell us that we can't turn fatty acids into glucose.  For example, on page 634 of the 2006 and 2008 editions of Biochemistry by Berg, Tymoczko, and Stryer, we find the following:

Animals Cannot Convert Fatty Acids to Glucose

It is important to note that animals are unable to effect the net synthesis of glucose from fatty acids.  Specficially, acetyl CoA cannot be converted into pyruvate or oxaloacetate in animals.

In fact this is so important that it should be written in italics and have its own bold heading!  But it's not quite right.  Making glucose from fatty acids is low-paying work.  It's not the type of alchemy that would allow us to build imperial palaces out of sugar cubes or offer hourly sweet sacrifices upon the altar of the glorious god of glucose (God forbid!).  But it can be done, and it'll help pay the bills when times are tight.

All Aboard the Acetyl CoA!

When we're running primarily on fatty acids, our livers break the bulk of these fatty acids down into two-carbon units called acetate.  When acetate hangs out all by its lonesome like it does in a bottle of vinegar, it's called acetic acid and it gives vinegar its characteristic smell.  Our livers aren't bottles of vinegar, however, and they do things a bit differently.  They have a little shuttle called coenzyme A, or “CoA” for short, that carries acetate wherever it needs to go.  When the acetate passenger is loaded onto the CoA shuttle, we refer to the whole shebang as acetyl CoA.

As acetyl CoA moves its caboose along the biochemical railway, it eventually reaches a crossroads where it has to decide whether to enter the Land of Ketogenesis or traverse the TCA cycle.  The Land of Ketogenesis is a quite magical place to which we'll return in a few moments, but navigating the TCA cycle can be a nightmare.  Traveling down this route is particularly dreadful for three reasons.  First, every time the Biochemical Traffic Committee holds session it has the the cycle renamed and has the signage repainted.  As a consequence, everyone is always calling it something different.  Some call it the tricarboxylic acid cycle.  Others call it the citric acid cycle and yet others call it the Krebs cycle.  Second, the TCA cycle is a treacherous roundabout.  It puts the typical town traffic circle to shame with its eight exits, all replete with incoming and outgoing traffic.  Third, even our cheerful little CoA That Could has to feel a tinge of guilt gliding along a railway that's been used to mercilessly torture generations of memorization-impaired biochemistry students in universities everywhere.

If acetyl CoA navigates the TCA cycle flawlessly, completing a full turn of the circle without either getting into a traffic accident or wandering off along one of its myriad exits, it arrives at that sacred space wherein our cells make new glucose.  Presuming a bit of poetic license, let's call this space the Candy Factory.

Balancing the Carbon Accounts in the TCA Cycle

But now we come to the problem that biochemistry textbooks grapple with so simplistically, and as we'll see, so wrongly: it is mathematically impossible for acetyl CoA to yield a net synthesis of glucose when it arrives at the Candy Factory by this route.  In other words, it is impossible by this means for acetate to contribute to the production of more glucose than is used up just to keep the TCA cycle running.  We can see this illustrated in the following flow chart, taken from a 1957 review that discussed this matter in detail (1):

This diagram is greatly simplified so that we can see just the essential points.  On the left, we see the point where glucose enters or exits the cycle.  When more glucose exits the cycle than enters it — that is, when more glucose is produced than consumed — we are in a state of gluconeogenesis and our Candy Factory is fully operational.  On the top, we see the point where acetyl CoA enters the cycle.  Acetate is a two-carbon molecule, so it naturally brings only two carbons to the table.  On the bottom, we see that both of these carbons leave the cycle as carbon dioxide before the CoA train even reaches the Candy Factory station.  Two minus two is zero, so there are no carbons left for making glucose.

The only way to make sure acetate carbons get stuffed into any of the delectable delights produced in the Candy Factory is for other molecules to enter the TCA cycle at any of the many entry points not shown in the above diagram and thereby provide those two carbons that need to leave the cycle as carbon dioxide during each turn.  Indeed, careful experiments using radioactive tracers had already definitively shown that carbons could flow from fatty acids to glucose in this manner by the time that review had been published in 1957.

But that's not a net synthesis of glucose.  For every two carbons that fatty acids could provide for glucose synthesis in that scenario, two would have to be taken either from glucose itself, or from some other molecule that could just as easily have served as a precursor to glucose, just to keep the TCA cycle going.  Once again, two minus two is zero, and fatty acids cannot contribute to the net synthesis of new glucose in this manner.

Magical Things Happen In The Land of Ketogenesis

By the time the 1980s rolled around, however, it had become clear that fatty acid metabolism is more complex than this and that there are indeed ways that fatty acids can contribute to the net synthesis of new glucose.

When large quantities of fatty acids flood the liver during fasting, caloric restriction, diabetes, or the consumption of a low-carbohydrate, high-fat, ketogenic diet, our livers produce so much acetate that the TCA cycle suffers heavy traffic.  Any acetyl CoA with the foresight to listen to the evening traffic report would quickly decide to head straight for the Land of Ketogenesis, where the railways are open and the paths are free.  This is where our livers turn acetate into ketones, sending the ketones out into the bloodstream so our other tissues can use them for energy.

One of the ketones we make is acetone.  Acetone makes an excellent paint thinner, and is responsible for the “ketone breath” that some people get on low-carbohydrate diets.  It also happens to be a great raw material for making glucose.

In 1979, a group of researchers from Philadelphia studied acetone metabolism in fasting humans (2). These authors estimated that during a three-day fast acetone may constitute over a third of the ketones we produce, and that 50-70 percent of it undergoes further metabolism.  They used radioactive tracers to show that acetone could be converted to glucose in these subjects, and estimated that if in fact acetone contributes to the net synthesis of new glucose, it could account for just over ten percent of such glucose newly made.

In the mid-1980s, researchers showed that rats can convert acetone to glucose through two different intermediates, methylglyoxal and 1,2-propanediol (3).  These pathways were summarized graphically in a later review (4):

The pathways shown above represent several alternative methods of converting acetone to pyruvate, which can then be converted to glucose.*  Since acetone is formed from acetyl CoA, this directly contradicts the claims of even the most recent biochemistry textbooks, which plainly state as a matter of fact that “acetyl CoA cannot be converted into pyruvate or oxaloacetate in animals.”  We would expect each of these pathways leading to pyruvate to result in the net synthesis of new glucose.

In 1986, researchers administered radioactively labeled acetone to rats and examined the radioactive carbon “fingerprint” on the glucose molecules formed from it to provide additional evidence that acetone followed pathways leading to pyruvate that would indeed lead to a net synthesis of new glucose (5).  That same year, the Philadelphia group used a similar approach to show that acetone was following similar pathways in humans with diabetic ketoacidosis (6).  They estimated that in such patients at least ten percent of newly synthesized glucose may come from acetone.

Blood levels of acetone also rise appreciably in adults on the Atkins diet (7) and in epileptic children following a ketogenic diet (8), suggesting that it may be a normal state of affairs for humans to convert fatty acids to glucose when consuming a diet low in carbohydrate and high in fat.

For comparison, I compiled the blood levels of acetone reached in humans under various conditions in the following table:$

In July of 2011, a German research group revisited the question of converting fatty acids to carbohydrate by publishing a computational analysis of the most up-to-date information about human biochemistry available (9).  These authors identified 22 pathways by which acetone could be converted to pyruvate that they considered likely to be important, and concluded that these pathways would be less cost-efficient than making glucose from amino acids or glycerol, but are nevertheless biochemically feasible and likely serve as supplementary modes of glucose production.

Lo and behold, we have three decades of evidence suggesting that the Land of Ketogenesis is graced with its own Candy Factory.  Sure, the desserts conjured therein may be sold at higher prices than those made in the Candy Factory just off exit eight of the TCA cycle, but when the traffic is heavy there, what else are we to do?  Such is the law of supply and demand.

Insulin Regulates Gluconeogenesis From Fatty Acids

If we really do make glucose from fatty acids when times are tight as all of this evidence so strongly suggests, there should be a way for our bodies to regulate this process so that it only kicks in when we are in need of glucose.  Indeed, such a mechanism exists.  Let's take a look at a figure from the recent computational analysis (9) and focus in on the part I outlined in blue:

This part of the picture represents a complicated network of reactions that provide a multitude of ways to achieve the critical event needed to convert fatty acids to glucose: the conversion of acetone to pyruvate (you know, that conversion that the biochemistry textbooks categorically state can never happen).  Pyruvate is half of a glucose molecule, so once acetone has made it that far, the rest is downhill.  Let's zero in on this part of the picture and pay special attention to the part I circled in red:


We can see that despite the many different paths down which acetone may travel to ultimately wind up at pyruvate, they all start with the conversion of acetone to acetol, a conversion facilitated by an enzyme called cytochrome P450 2E1, or CYP2E1 for short.  Insulin suppresses the production of this enzyme, while acetone prevents its degradation (10). Thus, when insulin levels fall and ketone levels rise, as occurs when our carbohydrate intake is low, our cells increase their supply of CYP2E1 and thereby activate the conversion of fatty acids to glucose.  We've found our way to the expensive Candy Factory in the magical Land of Ketogenesis.
The authors of the computational analysis (9) calculated that the most cost-efficient way of converting fatty acids to glucose is by converting acetol to methylglyoxal, facilitated again by CYP2E1, and then converting methylglyoxal directly to pyruvate, facilitated by an enzyme called aldehyde dehydrogenase.  I've outlined this pathway in red here:

As Peter Dobromylskyj over at Hyperlipid has pointed out before, methylglyoxal inhibits the breakdown of glucose.  In a future post, I will cover methylgyloxal's inhibition of glucose consumption in greater detail, but for now it is worth noting that when this pathway is activated, we not only convert fatty acids to glucose, but methylglyoxal concentrations rise and inhibit the breakdown of glucose.  Thus, when glucose runs low and we begin subsisting primarily on fatty acids for fuel, we have a coordinated effort to both spare glucose and to make more of it.  When the glucose recession hits, our cells do what any other budget-conscious cells would do and spend less.

O Textbook, Why Hast Thou Deceived Me?

In the 1980s, at least two reviews were published outlining the evidence for the conversion of fatty acids to glucose (4, 11).  One of them emphasized that biochemistry students were taught that such pathways do not exist (11):

Students are often asked to describe a pathway by which a long-chain fatty acid is converted into glucose in mammalian liver.  This type of question is normally a trick one which, for most biochemists, would have a simple reply: such a pathway does not exist.  Nevertheless, recent studies point towards a role for acetone in the conversion of fat to carbohydrate.

It further emphasized that most textbooks wrongly classify acetone as a useless chemical that can't be metabolized at all (11):

Most biochemistry textbooks state that acetone is a non-metabolizable byproduct of lipid metabolism, which accumulates when there are insufficient glycolytic intermediates to effect the complete oxidation of the acetyl CoA generated in the degradation of fatty acids.  However, studies with 14C-labeled acetone in lactating cows, rats, guinea pigs, and humans have shown that acetone is not just excreted, but that it can be metabolized further.

The other review (4) noted that the first evidence for the conversion of acetone to glucose had been generated as far back as 1951.

Another quarter century has gone by, and the textbooks haven't changed their tune one toot.  At the beginning of this post, I quoted Biochemistry by Berg, Tymoczko, and Stryer, which plainly states in a boldfaced section header that “Animals Cannot Convert Fatty Acids to Glucose.”  Of acetone, all this book tells us is that its odor can be detected in the breath.  Another biochemistry textbook I keep on hand, the 2005 edition of Lippincott's Illustrated Reviews: Biochemistry, tells us that acetone is a “non-metabolizable side product” of ketone production.

What is most striking is that these textbooks do not even alert us to any controversy about this topic, let alone to the strong evidence supporting the opposing view.  This emphasizes the need to use what we learn from textbooks and academic classes as a starting point for further exploration of the primary literature.  If we don't have time for that, as is usually the case, we need to seek out the best arguments from opposing viewpoints and consider them with an open mind.  If we do not have even the time for that, I think it would be wise not to cling too tightly to any of our cherished beliefs.  If this applies to something as mundane as whether acetone can be converted to glucose, it must hold true all the more for the myriad topics that have emotion, politics, money, or ideology lurking within them.

None of this is to say that we should blame the authors of these textbooks.  Rather, we should admire them for undertaking such a gargantuan task, a task that no human or small group of humans could execute flawlessly.  What is important is the recognition that any such work, no matter how authoritative, is a human work and thus necessarily subject to error.

The Good News

On the bright side, this finding is a testament to the great versatility of life.  Biochemistry is enormously complex, and while activating any particular set of pathways might not necessarily be optimal, the plethora of possibilities contributes to the resiliency we possess as living beings.

And three cheers for the Little CoA That Could!

Join the Next Live Q&A

Have a question for me? Ask it at the next Q&A! Learn more here.


Subscribe or upgrade your subscription here.

Masterpass members get access to premium content (preview the premium posts here), all my ebook guides for free (see the collection of ebook guides here), monthly live Q&A sessions (see when the next session is here), all my courses for free (see the collection here), and exclusive access to massive discounts (see the specific discounts available by clicking here). Upgrade your subscription to include Masterpass membership with this link.

Learn more about the Masterpass here.

Take a Look at the Store


* This graph as originally published included a direct conversion of acetone to acetate and formate, and conversion of this acetate to glucose.  Acetate would be converted to glucose through the TCA cycle and would thus not yield net synthesis of new glucose.  I erased this pathway because we currently do not have evidence for acetone being converted to acetate without first being converted to pyruvate (Christoph Kaleta, senior author of reference 9, personal communication).  In any case, the authors discussed evidence that this pathway is not active unless supraphysiological doses of acetone were given.  In the latter case, they may have been observing conversion of pyruvate to acetate when the capacity for gluconeogenesis from pyruvate was saturated.

$ It is important to realize that 1) these values are taken from different studies over the span of 27 years using different experimental techniques, 2) none of these subjects were randomly allocated to specific treatments, and 3) plasma acetone will be influenced as much by acetone metabolism and excretion as by acetone production.  This table is therefore provided for purposes of rough comparison only, and no conclusions should be drawn about relative acetone production under these different conditions.  Moreover, inter-individual variation is not represented in this table.  See references for standard deviations, standard errors, or ranges given therein.


1.  Weinman EO, Strisower EH, Chaikoff IL.  Conversion of fatty acids to carbohydrate; application of isotopes to this problem and role of the Krebs cycle as a synthetic pathway.  Physiol Rev. 1957;37(2):252-72.

2.  Reichard GA Jr, Haff AC, Skutches CL, Paul P, Holroyde CP, Owen OE.  Plasma acetone metabolism in the fasting human.  J Clin Invest. 1979;63(4):619-26.

3.  Casazza JP, Felver ME, Veech RL.  The metabolism of acetone in the rat.  J Biol Chem. 1984;259(1):231-6.

4.  Landau BR, Brunengraber H.  The role of acetone in the conversion of fat to carbohydrate.  Trends in Biochemical Sciences. 1987;12:113-4.

5.  Kosugi K, Chandramouli V, Kumaran K, Schumann WC, Landau BR.  Determinants in the pathways followed by the carbons of acetone in their conversion to glucose.  J Biol Chem. 1986;261(28):13179-81.

6.  Reichard GA, Jr, Skutches CL, Hoeldtke RD, Owen OE.  Acetone metabolism in humans during diabetic ketoacidosis.  Diabetes. 1986;35(6):668-74.

7.  Beisswenger BG, Delucia EM, Lapoint N, Sanford RJ, Beisswenger PJ.  Ketosis leads to increased methylglyoxal production on the Atkins diet.  Ann NY Acad Sci. 2005;1043:201-10.

8.  Musa-Veloso K, Likhodii SS, Rarama E, Benoit S, Liu YM, Chartrand D, Curtis R, Carmant L, Lortie A, Comeau FJ, Cunnane SC.  Breath acetone predicts plasma ketone bodies in children with epilepsy on a ketogenic diet.  Nutrition. 2006;22(1):1-8.

9.  Kaleta C, de Figueiredo LF, Werner S, Guthke R, Ristow M, Schuster S.  In silico evidence for gluconeogenesis from fatty acids in humans.  PLoS Comput Biol. 2011;7(7):e1002116.  Epub 2011 Jul 21.

10.  Gonzalez FJ.  The 2006 Bernard B. Brodie Award Lecture. Cyp2e1.  Drug Metab Dispos. 2007;35(1):1-8.

11.  Argiles JM.  Has acetone a role in the conversion of fat to carbohydrate in mammals?  Trends in Biochemical Sciences. 1986;11(2):61-63.

Visit Us
Follow Me

You may also like


  1. Outstanding! As a nutritional biochemist and a severe type 2 diabetic, on a modified Atkins diet and have been puzzled by the amount of glucose I make. In wear a continuous glucose monitor. I just assumed that they are from glycogenolysis. Never considered acetoacetate and acetone as a possibility. I am currently engaged in writing a “nutrition” book-not a next book but a popular book and would be sure to add the information.
    I would assume that a high fat diet under diabetic conditions would suppress acetyl Co-A carboxylase and push acetate towards acetoacetate and acetone (perhaps increased cholesterol too).
    Could you write a follow up article right from FA to Glucoses pathway?
    By the way, one of my paper would suggest a non-enzymatic conversion of acetoacetate to acetone by oxygen free radicals.

  2. Thanks for some other informative blog. The place else may I get that type of info written in such an ideal approach?
    I have a project that I am simply now working on, and
    I have been at the look out for such information.

  3. Scientists have virtually no understanding of how nutrients affect cells. We are all groping in the dark and taking the same risk as walking through a dangerous minefield. Even biochemical celluar operations in general, themselves, are very poorly understood. There is a monstrous level of uncertainty in the nutritional field.

  4. Hello,
    First of all nice Post.
    Im a undergraduate bioinformatics student, and at the moment im writing my Bachlorthesis on this subject.
    Im searching for evidence of Gluconeogenese from fatty acids via acetone, in the gene expression profiles of hibernating/fasting/starving mammals. GNG from fatty acids should play an important role here.
    The Pathways im checking are the ones S. Schuster & Kaletta discovered with their global metabolic network reconstructions and flux pattern analysis. And as you already marked, the Enzyme CYP2E1 is significally upregulated in starving rat liver and hibernating squirrels and bears. But i have a few questions: Cyp2e1 is a multifunctional enzyme, therefore its upregulation does not 100% mean that this is because, its used in the conversion of acetone to acetol, which is the first reaction following the aceton-pyruvate pathway.
    A combination of upregulating cyp2e1, and molecular evidence that GNG from proteins has a limit might be from greater importness.
    Ive searched the literature and the internet, but didnt find anything about limitation in GNG from proteins/AS, but as i was thinking there should be a mechanism.
    In long period of fasting, the organism needs it muscleprotein, therefore the GNG from it should eventually stop and a limit. This is the time when GNG from fatty acids via acetone could contribute glucose. Have u read about such a mechanism, and why do you think upregulating CYP2E1 is evidence enough?

    1. I don’t understand your question. I never claimed upregulating CYP2E1 is “evidence enough” of anything. There are papers cited within this post estimating 10% of glucose from acetone in human studies.

      1. Sorry if you dont understand my questions, Im not a native speaker.
        The really important one was about the limitation of gluconeogenesis from proteins. I have searched for papers but didnt findt any, and in your article you mentioned it once.

          1. Well Sorry, I got a bit confused. I didnt get that from your Post.
            But still, I talked with my professor (S. Schuster who calculated the best possible pathways for GNG from Fatty Acids with Kaleta) and we both agreed that there should be a mechanism that prevents the body from wasting all its muscle protein for GNG for example in starvation or hibernation.

  5. Hi,

    I was reading “Principles of Biochemistry” (Lehninger, Nelson, and Cox) to try and understand the connection between Kreb’s cycle and Glyoxalate cycle (considering the common intermediates in both).
    And the very first paragraph mentioned in that section is “Vertebrates cannot convert fatty acids, or the acetate derived from them, to carbohydrates. Conversion of phosphoenolpyruvate to pyruvate and of pyruvate to acetyl-CoA are so exergonic as to be essentially irreversible. If a cell cannot convert acetate into phosphoenolpyruvate, acetate cannot serve as the starting material for the gluconeogenic pathway, which leads from phosphoenolpyruvate to glucose. Without this capacity, then, a cell or organism is unable to convert fuels or metabolites that are degraded to acetate (fatty acids and certain amino acids) into carbohydrates.”

    When I read this, the very first question that came to my mind was, “Does it mean fatty acid-driven gluconeogenesis does not work at all in the human system?” Proteins (considering their versatility in physiological functioning) are way too precious to be utilized for glucose production. So, if proteins are meant to be the last resort for glucose production and (as per the text), fatty acids cannot be converted to carbohydrates, then how does the body derive glucose for itself – in case glycogen stores are all used up? I searched for an answer and I ended up at your blog. Needless to say, my query was perfectly answered.

    Thank you for enlightening me with your research and illustrative pathways. This also implies that textbooks are not the be-all and end-all of everything. Textbooks are basically meant to make you think so that you can find your own path further. Case in point – The above cited paragraph (from the text) makes NO mention of acetone in any manner (forget about labeling it as a “non-metabolizable” product or anything else). Therefore, if it weren’t for your blog, I would not have realized the significance of acetone in this regard.

    Thanks once again 🙂

    1. “Proteins (considering their versatility in physiological functioning) are way too precious to be utilized for glucose production.”

      Our bodies have very little ability to store excess protein. We can grow new lean by a very slow process and that’s nearly it. So even though protein is high value, excess protein grams get burned for fuel. And about half that energy comes through gluconeogenesis.

      Ketosis is primarily a successful predator diet response from low carbing. Much of the blood sugar while low carbing would come from this source.

      Ketosis is secondarily a starvation response. Lean is burned first to reduce long term caloric demand but blood sugar levels never do fall and that can’t really be explained by lean loss. The amounts don’t match.

  6. Hallelujah!

    Fantastic article!

    This would explain how virtually every prophet of every major religion “FASTED 40 days and 40 nights” without dying! (But they did drink lots of water, or in extreme situations possibly urine… but we won’t dwell on that except in EXTREME emergencies. Astronauts have learned how to “purify” urine!?)


  7. Hallelujah!

    Fantastic article!

    This would explain how virtually every prophet of every major religion “FASTED 40 days and 40 nights” without dying! (But they did drink lots of water, or in extreme situations possibly urine… but we won’t dwell on that except in EXTREME emergencies. Astronots have learned how to “purify” urine!?)


  8. Even though I am not schooled in chemistry, let alone biochemistry this was still fascinating to read. You have a real talent and I suspect that is because you work hard to maintain a wide perspective in all things.

    I thought your comments regarding text books were extremely diplomatic. They also hit to the heart of some very critical issues that we are all going to have to re-examine if we going to have a shot at sustaining large human populations. I know you get this, so I won’t go on. I would encourage you, if I may be so bold, to not change a thing about the way you present.

    Love your work – thanks for sharing

  9. Chris, to add, I’m shocked (anew) at the lock-step march of pro-glucose slant and flat-out misinfo in the biochem texts. It’s almost like the major scholastic tech pub house editors have to check the “Does this accord with gov’t dietary guidelines?” box.

    But what surprises me most (again, anew) is how much we don’t actually know. Even seemingly very basic things.

    The past few days, in an attempt to answer the question:

    “Can the body meet its glucose needs solely from gluconeogenesis from fats?”

    I tried to get some kind of handle on these two questions:
    1) What is our absolute minimum daily glucose need? Once we fully transition to keto?
    2) How much glucose can we generate from fats/TAGs alone?

    From the beginning, it was an uphill struggle. I found nothing to address even basic questions like: What are the daily glucose requirements for red blood cells, our (only?) obligate glucovore cells?

    I finally tossed in the quant data towel (for now) after I read Carl Zimmer’s post at Nat Geo from late 2013:


    where he makes clear we can’t even narrow down to *within a magnitude* how many cells make actually up our bods proper! (Separate from our microbiome. And also our RBCs?):

    “…estimates sprawled over a huge range, from 5 billion to 200 million trillion cells…If scientists can’t count all the cells in a human body, how can they estimate it? The mean weight of a cell is 1 nanogram. For an adult man weighing 70 kilograms, simple arithmetic would lead us to conclude that that man has 70 trillion cells. On the other hand, it’s also possible to do this calculation based on the volume of cells. The mean volume of a mammal cell is estimated to be 4 billionths of a cubic centimeter…Based on an adult man’s typical volume, you might conclude that the human body contains 15 trillion cells…”

    Chris, it’s made one thing clear to me. About that question:

    “Can the body meet its glucose needs solely from gluconeogenesis from fats/TAGs?”

    no-one's gotten far, and may not currently be able, to quantitatively answer this. But we can live VLC-VHF, get plenty of tests, and see how it goes long-term. –Bryan

  10. Chris, for years I’ve read your work here and over at WAPF. Yours and Petro Dobromylskyj's over at Hyperlipid. But this is my first comment to you.

    Your intriguing post here coaxed–compelled–me out of the woodwork. I've lived VLC-VHF for over eight years. It may be the main factor to help me essentially normalize my BGs, way down from a winter 2006/2007 average of 170 mg/dL (7.0% HbA1c). I also lost 62 pounds (230 to 168 lbs) but had brought my BGs way down months before that weight loss.

    I can’t see how you showing acetone → pyruvate can’t be anything but HUGE. I’m amazed I don’t see more on this topic since you posted this potential earth-shaker exactly four years ago.

    Chris, do I understand correctly that our bods can keep cranking out acetyl-CoA regardless of oxaloacetate’s status? That, even under low BG/insulin conditions when the bod’s wicking off OAA for GNG, it can keep making Ac-CoA? And so Ac-CoA → ketones → pyruvate keeps a keepin' on so long as we have FFAs to feed to the beta-oxidizers? Thus, from fat-derived acetyls, we see one path to ketones PLUS two paths to pyruvate? If all that is or could be true, then Chris that’s amazing. And a testament to our bods’ evolution to robustly maintain homeostasis despite hugely varying exogenous conditions.

    As “low-paying” as it may be, can acetone → pyruvate conceivably serve as the ultimate homeostatic backstop to ensure steady and sufficient BG regardless of life conditions and macro balance in the diet? Could it help–even fully–explain why researchers find that, for subjects in ketosis, net GNG appears to stay the same regardless of protein intake? You can read about that here:


    Chris, please, I very much hope you’ll do a follow-up post on this vastly important, even pivotal, finding. To give us the latest info you’ve found on acetone → pyruvate, esp numbers. And even on whatever other lipid pdts → pyruvate/glucose pathways you’ve since come across.

    Thank you so much for this Chris. I’ve thought about little else this week. If I’m over-excited and wrong about part or all of this, I look forward to your tempering comments and corrections. –Bryan

  11. Hi Chris,

    I am a dietetics student at UGA and am hoping to get into nutritional biochemistry research. I was just wondering how pure fats are metabolized without ANY carbs or protein? I know some insulin will still come in but I was just wondering if such a protocol would mimic fasting even in a positive energy balance for patients who are trying to "starve" cancer?

    I have tried to google it and have searched all of the nutrition databases that I know and have not found an answer.

    Thank you!

    1. Back about 2001 I conducted an experiment on myself by drinking pure 50-50 canola-olive oil for breakfast. I estimated the calories of my standard breakfast, converted it to calories of oil and got about a jigger of canola-olive oil. I only did this for one meal of my day while otherwise eating a common 50 total gram per day low carb plan.

      The first day I drank the oil for breakfast. The second day I didn’t want to but could swallow it. The third day I had to hold my nose to push to swallow. The fourth day I choked and gasped rather than swallow.

      I conclude that at least for me, pure fat is not doable. It’s self limiting.

      Yet mixing enough protein (cream cheese) or carb (butter cookie) makes it easy to eat almost unlimited fat. it’s only very nearly pure fat that is self limiting.

      1. Get your fats from whole, real foods such as coconut meats, olives and seeds. Pressed oils are at the bottom of nutrient dense foods. Besides the oils, whole foods contains micro & macro nutrients a as well as soluble and insoluble fiber. Don’t use your body as a chemistry experiment!

  12. whoops forgot to add a 200 level nutrition course with yet another review of the kreb's cycle to the list of pertinent classes…again taught what you're saying is some new insight. I'm just KYNA confused…

  13. I have taken anatomy, physiology, myology (with a review of the krebs cycle), kinesiology and even general biol (dont ask why lol) with yet another review of the krebs cycle). I have been told each time, gluconeogenesis and in fact making all the macronutrients can happen from all/any the macronutrients. lipids to aminos, glucose to lipids, lipids to glucose, aminos to glucose. I first took physio in 1989. I last took General Biol in 2013. None of these went beyond 200 level classes. Hmmmmm, was I taking courses at a very progressive school lol I will be taking biochem when i get some more 200 level chem courses under my belt, but Im wondering about this making what we need from what we have stuff now…

  14. The idea of texts books being inaccurate to any extent I would think has much to do with the times. If a study is done half a century ago and documented thus – then that's what we knew and we wrote it down – and the rest of us read it – most times as fact. The other factor is that there is little homogeneity left among humans. Our diets and our varying environments afford a multitude of varying study results. And then there is evolution of a species. Most time "evolution' implies things that evolve slowly over great lengths of time (millennia?). But in this last 130 years the American diet especially has been all over the place allowing for potential dramatic changes perhaps in how the human body works, functions, reacts and processes, metabolizes X,Y, and Z. And with each passing decade, the human gene pool expands thus exponentially. This is in part, the problem with drugs, and the desperate need to create "smart drugs".
    Drugs aside, how can any book, text or otherwise, be truly accurate or factual? They should all be titled – writings on a perpetual work in progress…

  15. Hey Chris,

    It seems that adipose and visceral fat can be stored by means other than by storing glucose as glycogen in fatty tissue, but I cannot seem to find this pathway. Could you, or someone else, steer me in the right direction to find this information?

    PS: Awesome article!

  16. Please note the oxaloacetate can be replenished from any of the cycle intermediate members from their various counter glucogenic amino acid. So acetylCoA from fatty acid can react with oxaloacetate with net glucose production effect.

    I now consider a beta oxidation of fatty acid to succinyl CoA which is one of the 4 anaplerotic reactions till found by humans, which can contribute net glucose production from fatty acid. Isn't it beautiful ?

    I do not know this reaction can happen in human body. But the most important anap. rxn of us is pyruvate to ox.acetate pathway.

    Cheers !

    1. This is correct and in fact, if you do an energy balance considering the ATP derived from beta oxidation, a significant amount of glucose is derived from Fat. In fact, if all the carbons source for driving the TCA cycle came from lipids, which is not too far off in a fasted state, then by the 3 rd turn of the TCA cycle 5 of the 6 carbons of citrate come from fats. So stoichiometry is very misleading. Energy balance and 13C tracer studies are more informative.

  17. The interesting thing is, eventhough gluconeogenesis is happening in the liver cell, the TCA cycle can be run with above pathway as long as alanine and fatty acid are available within the body and/or liver mitochondria.

    I have never seen this my explanation at any biosource.


  18. Of course some of oxaloacetate from alanine go directly to glucose, but some of them (you said 10 % of the glucose portion generated from total genesis?) could meet (collide) the molecule of acetylCoA from fatty acid, create citrate and go into CTA cycle.

  19. Fatty acid can go to glucose with net formation.

    The amount of oxaloacetate formed from glucogenic amino acid such as alanine and etc. thru pyruvate for gluconeogenesis is bigger than the amount of acetylCoA formed from the same starting amino acid.

    This is easy understanding to biochemist, because the two pathway enzyme dynamics are different and the former is fast.

    So the acetylCoA formed from other fatty acid can meet new oxaloacetate from alanine and form citrate. The final destination of this gluconeogenesis pathway is glucose generation in various human cell including, of course liver cell.

    Is that clear ?


  20. Could have knocked me over with a feather. I've been wondering about this for about 40 years, and it was only my recent interest in and effective use of a 'Paleo Diet' that drew me to your article. I've been asking Biochemists for year about this evolutionary oddity, and I thought the lack of available carbs out in the Paleo world could have accounted for it, like primates not making Vitamin C. Fascinating new thoughts evoked by your narrative. Thanks!

  21. I agree that this plug-in is a great booster for blogs. for the point of view of the blog owner although is not a set and forget. If ppl love their blogs they should really take care of them on a daily basis in order to don’t stack up on spam and crap posts. on the point of the visitor, this is a great motivator to encourage ppl to make a decent post on any specific subject, contributing this way to enrich the blog content. so it is a win win scenario in my POV.
    Thank you for sharing The Daily Lipid.

  22. Fascinating post Chris! I came across your blog after watching your video presentation for AHS 2011.

    I recently wrote an article for my blog on keto-adaptation for military training, partly because I'm in the military, and partly to get my head back into the nightmare world of biochemistry prior to applying for medical school!

    One thing I wonder about though, is whilst it may be possible to reverse this supposedly irreversible step in cellular biochemistry, is it worth the effort so to speak? Given that consuming fat in our diets, or utilising Tg stored in our adipocytes, we have a good source of glycerol that can readily be used to generate glucose in hepatocytes. Are these other means of obtaining glucose from acetone significant? Once keto-adapted, and the demand placed on gluconeogenesis is reduced (and thus OAA is spared from this process), is the generation of glucose from glycerol sufficient to keep red blood cells etc… happy?

    As a side thought, is it wise to explain 'normal' biochemistry based upon evidence from ketoacidosis? For a long time extrapolations have been made about what is best for those in the centre of the bell curve based upon the extremes…

    Again, brilliant article, and great to read in conjunction with Lodish et al!

  23. Chris, thanks very much for this post. Some time ago, I read this from Professor John Yudkin, from an article, "Why Low-Carb Diets Must Be High-Fat, Not High-Protein", which Barry Groves has at his site, and wondered about the mechanism. Your post helps much with the explanation. 🙂


    "Dietary proteins are converted to glucose at about fifty-eight percent efficiency, so approximately 100g of protein can produce 58g of glucose via gluconeogenesis.[v] During prolonged fasting, glycerol released from the breakdown of triglycerides in body fat may account for nearly twenty percent of gluconeogenesis.[vi] Body fats are stored as triglycerides, molecules that contain three fatty acids combined with glycerol. The fatty acids are used directly as a fuel, with the glycerol stripped off. This is not wasted. As the glycerol is nearly ten percent of triglyceride by weight and two molecules of glycerol combine to form one molecule of glucose, this also supplies a source of glucose."

    Thanks again for another great post!

  24. Thought you might like to know about what looks like misuse of this research by the Life Extension Foundation in an effort to make Green Coffee Bean Extract seem like a worthwhile supplement…
    Danial Becker is the author

    Quote from

    The enzyme glucose-6-phosphatase (G6P) helps to produce dangerous
    after-meal blood sugar spikes in two ways. It releases glucose from
    its storage area in the liver, and it promotes formation of new
    glucose molecules from non-sugar sources.78 The latter process is
    called gluconeogenesis.

    Traditionally, scientists have assumed that amino acids from proteins
    were the only precursors of glucose in gluconeogenesis.79 Recent
    discoveries, however, suggest that fatty acids are also important
    precursors of this dangerous source of excess glucose.79,80

    Fatty acids contribute to gluconeogenesis by at least three
    mechanisms. First, excessive fatty acids stimulate gluconeogenesis and
    provide G6P the energy it needs to convert traditional substrates such
    as amino acids into glucose.81-84 Second, when triglycerides break
    down to fatty acids, glycerol is released, and then converted into
    glucose by gluconeogenesis.85 Finally, advanced computer models have
    revealed new pathways by which fatty acids are directly converted into
    glucose; the final steps in those pathways involve G6P.79

    These discoveries provide further insight into the causes of the
    deadly after-meal glucose spikes. They also provide further incentives
    to block the enzymes that participate in gluconeogenesis, such as G6P.
    Green coffee extracts are an excellent source of G6P-blocking

  25. Beautiful.
    Methylglyoxal is the active antimicrobial ingredient in New Zealand Manuka Honey, so its formation may also be related to innate immunity, and desirable increases in production may result from loss of appetite in response to infections.
    Critics of Atkins (and proponents of Calorie Restriction) claim methylglyoxal is an AGE precursor (it's characterised as such in Aubrey de Grey's book, Ending Ageing), but it is obviously much more than that.

  26. Stabby, Great idea. Thanks! I look forward to your next comment. 🙂

    Evelyn, You're welcome, and I'll email you reference 6. I agree with you on the carb-to-fat issue. Thank you for the link — I'll try to look at it when I can.

    I don't understand what you mean about fat X>carb by running in reverse, so I'm not sure if what I said here refutes that. The argument I'm familiar with, which I delineated here and agreed with, is that net synthesis of carbohydrate from fat cannot occur through the TCA cycle. I agree with that, and what I presented here doesn't refute it. It points out that alternative pathways are active that do allow net synthesis.

    I don't have much time to pay attention to sports right now, though I followed soccer a bit when I was living with an MVP soccer player, but I vicariously participate in enthusiasm for the Huskies through facebook statuses. 😛 Oh and I know the chant too. Have a great weekend!


  27. Hi Chris, Thanks for the welcome and thoughtful reply. If you get a chance I would like the full text (carbsane at gmail dot com), just for my curiosity, but your explanation of the numbers makes sense now.

    The post of mine Anonymous cited really wasn't about this topic per se, it was about the whole excess carbs turned to body fat making us fat thing, and Nora's even more exaggerated claim that all body fat comes from glucose. I am shocked that Nora has not only written what she does of metabolism, but remains steadfast that it came from a reputable text. I consider this paper to be one of the best reviews on the whole carbs to fat thing: https://www.ajcn.org/content/74/6/707.full Taubes has talked to Hellerstein …

    In any case, I actually edited the post the other day in response to this one because I did have a definitive statement in there about converting fat to carb. It's funny, I took two biochem classes as an undergrad. One in the bio department was more metabolism oriented, the one in the chem department was more O-chem reaction focused (that one was actually one of those 400 level type classes taken by seniors and grad students alike). My recollection was that carb -> fat via Acetyl CoA, but fat X> carb by running things in reverse. I don't see this acetone route as really debunking that. Perhaps all that's needed is some qualifiers? The difference in fed/fasted acetone metabolism is very interesting to me … more out of geeky curiosity than anything else 😉

    Enjoy the weekend. Go Huskies! (Dunno if you're a fan, or if you know this but I'm a UConn alum)

  28. Sweet, well then I guess I'll just believe that some omega-3s are good and leave the saturated fat thing up in the air, just like -some- people in the world of health should have been doing all along.

    Cheers. Looking forward to the next post, interview or whatever else you do.

  29. Hi Stabby,

    Looking over that study briefly, it seems to show a benefit of DHA supplementation, but I don't see how they could possibly blame stearic acid or saturated fat for the reduction of FMD that occurred at 240 minutes, given the heparin treatment and the fact that the drink used had chocolate powder, emulsifiers, and skim milk powder. Their bias in favor of blaming "saturated fat" can be seen in Table 3, where they reveal that plasma free fatty acids were 39 percent saturated at baseline, 46 percent saturated after the saturated fat treatment, and 44 percent saturated after the the saturated fat plus DHA treatment. At all time points, saturated fat is a minority of free fatty acids, yet other types of fatty acids aren't even included in the table! So, while, based on a cursory glance through this study, I agree it is supportive of low-dose DHA supplementation (~500 mg), I'm not so sure about the interaction with "saturated fat."

    Hi John,

    Thank you, I'll have to take a look at that study when I can!


    You're welcome. Interesting thought!


  30. Thanks for a very interesting article! Now I understand a mechanism behind run-away diabetic ketoacidosis! No insulin = accumulation of acetone from fatty acid nethabolism, then rapid conversion of that to glucose which then gets stuck due to a complete lack of insulin.

  31. Chris,

    The only head-to-head comparison I've seen was a rat study that compared "lab chow" to a 10p90f diet and a 10p78c12f diet, but they were looking at cerebral metabolism. The results are interesting anyway.


    "Cerebral levels of glucose, glycogen, lactate, and citrate were similar in all groups. 2-Deoxyglucose studies showed that the ketogenic diet did not significantly alter regional brain glucose utilization. However, rats maintained on the high-carbohydrate diet had a marked decrease in their brain glucose utilization and increased cerebral concentrations of glucose 6-phosphate. These findings indicate that long-term moderate ketonemia does not significantly alter brain glucose phosphorylation. However, even marginal protein dietary deficiency, when coupled with a carbohydrate-rich diet, depresses cerebral glucose utilization to a degree often seen in metabolic encephalopathies. Our results support the clinical contention that protein dietary deficiency coupled with increased carbohydrate intake can lead to CNS dysfunction."

  32. Thanks, Chris! Never ever ever take the abstract's word for it is the lesson, that's why I usually ask someone.

    Okay round two, and final round because I know you're busy, what about this one? https://www.ajcn.org/content/94/3/742.abstract Stearic acid, not the ever maligned palmitic acid, but a saturated fat in animal fat nonetheless. And judging by the abstract they claim a direct effect on postprandial reactivity, but that DHA prevents this. There is also heparin infusion to elevate FFAs which could confound things, but I'm not sure, so I'm asking for your opinion.

    Thanks in advance.

  33. [Continued]

    One of the points in these papers was that glucose gets labeled predominantly on carbons 1, 2, 5, and 6 from 14C-2-lableled acetone, which suggests a 3-carbon pathway rather than addition of acetate via the TCA cycle. It had been shown definitely in the 1950s that the acetate carbons make their way to glucose, so that wasn't the issue. The issue was net synthesis. A 3-carbon pathway argues for net synthesis. The argument for physiological meaningfulness is less air-tight, but it is supported by the high plasma acetone concentrations under ketogenic conditions, the role of insulin in regulating the process, the high rate of acetone metabolism even in healthy fasting subjects, and the substantial contribution to gluconeogenesis calculated by the diabetic ketoacidosis paper.

    Yes, it is interesting that you get much more acetone with dietary fat than with fasting. I wouldn't read too much into it because the comparison is not direct, but the difference is large and suggestive.

    I agree it is absurd to say that fatty acids must be converted to glucose before they can be converted to fat. I was referring to the more sensible statement about glucose being needed for the glycerol backbone. We seem to be in agreement on that topic.


  34. Hi Evelyn,

    Thanks for writing. Welcome to the comment section! 🙂

    I agree with your general approach, that we need to carefully distinguish between what is possible and what is physiologically important. However, I believe this data suggests that conversion of acetone to glucose is likely to be physiologically important under conditions of low carbohydrate and high fat intake.

    The two textbooks I have state categorically that conversion of acetyl CoA to glucose can not happen in a way that yields a net synthesis of glucose. The "in silico" paper cited one that I own (Berg et al) and two others that also say this. The thrust of these papers was assessing this claim and providing evidence otherwise. A less central goal of these papers was to assess the physiological relevance, which was a necessarily less definitive judgment.

    I agree that insulin plays a role in directing the fate of acetyl CoA, but my understanding is that the main regulatory point in directing ketogenesis is actually the presence or absence of oxaloacetate. Ketogenesis occurs when the supply of acetyl CoA to the TCA cycle surpasses the supply of oxaloacetate, which is necessary for its entry into the cycle. Carbohydrate provides a continual source of pyruvate, which acts as a source of acetyl CoA (through pyruvate dehydrogenase) and oxaloacetate (through pyruvate carboxylase). Thus, there is always a sufficient supply of oxaloacetate to prevent any meaningful diversion of acetyl CoA into ketogenesis. Fatty acids, however, supply acetyl CoA in bulk but do not lead to oxaloacetate (except by these pathways, via ketogenesis). The situation with acetone leading to pyruvate is, of course, quite different. Acetone is a quantitatively major product of ketogenesis, and the pathways favor its conversion to glucose when glucose is lacking.

    If you'd like, I can email you a scanned copy of reference 6. They estimated minimal proportions of plasma glucose due to acetone were between 0.5% and 4%, with an average of 2%. Based on other research suggesting that gluconeogenesis accounts for a fifth of plasma glucose during diabetic ketoacidosis, they calculated that the average minimum of 2% of total plasma glucose would suggest an average minimum of 10% of gluconeogenesis. They put this in perspective by noting that it is about half of that accounted for by alanine. So, it is roughly on par with a glycogenic amino acid. The reason they say "minimal" is because radio labeled plasma glucose did not come even close to plateauing in this study, which would be necessary for the estimation. Moreover, if they took the 4% figure, it would have been double, and even in the subject with the 4% value, radio labeled plasma glucose had not reached a plateau. So 10% is a conservative estimate for the minimal contribution.

    Yes, ketoacidosis is pathological, but as you can see from the table I made, you get comparable accumulation of plasma acetone on ketogenic diets. So I do not think that you can assume they contribution is much lower under other conditions just because they are not pathological, if they indeed have comparable plasma acetone accumulation.


  35. Responses to Stabby and John.

    Hi Stabby,

    Thanks! I try. I don't have the time to read that study in detail at the moment, but the high-fat meal is McDonald's, and there is no low-fat control meal. So I don't think you can say from this study that "fat" impairs endothelial function. If the only antioxidant included in the fish oil is 3 IU vitamin E, and there aren't additional synthetic antioxidants, then it seems unlikely that would be a physiologically relevant confounder, though it would be nice to see the effort in all fish oil studies to normalize for all added ingredients, especially the antioxidants, which this study does not appear to have done. I would have to mull over whether normalizing FMD to shear stress (they did not get a positive result otherwise) is the best thing to do here. Assuming it is, though, it does seem like they saw a positive effect of the supplement, so I suppose this might make a good justification for giving out free fish oil capsules with these meals instead of a packet of "McStatin," as has been suggested previously in the literature. Thanks for sending it along!

    Hi John,

    I have never seen the two diets compared head-to-head that I remember, have you? Theoretically, with respect to my point in the other post, the main difference would be that in the carbohydrate-inclusive diet, all of the 5% protein can go towards the protein requirement, whereas in the ketogenic diet, a portion of it will be diverted into the TCA cycle as a source of oxaloacetate.


  36. Oops … sorry I finally saw the comment referencing my blog post. The issues with Taubes' and Gedgaudas' claims are separate.

    Nora's claims are, indeed, absurd. I have asked her for the specific page number for the all body fat comes from glucose — on her blog she attributes that to Textbook of Med. Physio so I asked which edition and got "I didn't make it up". She's not claiming glycerol has to be made from glucose, she's claiming all body fat is made from glucose. So you don't have to read the blog post, here is the quote of her version of metabolism from her blog:

    "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)."

    Taubes … sigh … I won't go there here.

  37. Hi Chris,

    I'm a long time reader, but I'm pretty sure this is the first time I've come out of lurkdome to comment.

    Great post, and very entertaining to boot!

    I had someone ask my take on my blog, and I've had questions on that in silico study before as well.

    It's been a long time (decades) since I took biochem, but my recollection from my texts and teachings was more similar to Gretchen's quotes — more that it's not physiologically important or relevant. Not that it can't happen, but for all intents and purposes, it doesn't.

    If one just looks at biochemical reaction pathways, why don't we convert glucose to ketones? The pathway's right there but large excesses of glucose-derived acetyl coA go to DNL instead. Yes, insulin directs this traffic, but likely some ketones are made too, it's just not the physiologically significant path in the whole scheme of things.

    I don't even see the full text of reference 6 online so I'm going by the abstract. In humans with diabetic ketoacidosis (a pathologic state with acetone levels more than 10X those of normal adults overnight fasted or obese 3 day fasted) the abstract states: " Estimates of the minimal percent plasma glucose and PPD derived from plasma acetone averaged 2.1 and 74%, respectively." Your article implied at least 10%. Even 10% under that pathologic condition does not seem like enough to me to say that fatty acids are converted to glucose. It's more like a carbon that starts out as part of a fatty acid can make it's way to being incorporated into a glucose molecule.

    The elevated acetone after eating vs. in the fasted state are interesting as well — a toxin disposal path rather than a generative path to preserve lean mass? Humans do not seem to be producing a lot of acetone from endogenous fatty acids …

    I sure hope this revelation doesn't have the carnivores supplementing with nail polish remover any time soon :p


  38. Chris,

    Back on the 5% protein ketogenic mice, doesn't the same thing happen to mice on a low protein diet that isn't low in carbs–high metabolic rate but slightly lower muscle mass/fat ratio? Are there notable differences between the two?

  39. Great post, I love how you give us some general knowledge when you can.

    Since you're into nutrient-nutrient interactions, have you heard about this one? https://www.nrcresearchpress.com/doi/abs/10.1139/H10-020?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%3dpubmed A "high fat meal" seems to impair endothelial function, but only if omega-3 isn't available to protect against it. I remember your old post interpreting the study that claimed that saturated fat was particularly bad compared with omega-6, and agree with your assessment, but maybe there's something to that notion, or just fat in general, but only conditionally?

    So are you aware of this do you have any particular thoughts? If what I think is true this provides one very big limitation of the epidemiology of saturated fat intake. Not that we need any of that to tell us what is definitively healthy regardless!

    Also it would make the ol' Weston Price cod liver/butter oil thing look really good.

  40. Hi Moose,

    Thanks! Yes, it's possible someone could have more acetone due to decreased conversion, but there could be other issues too. For example, maybe they are utilizing other ketones for fuel at a lower rate, thus have a buildup of acetoacetate to synthesize acteone from. Or, they have lower activity of the TCA cycle, perhaps due to lower thyroid output, or oxidative stress, or something else, and they thus produce less carbon dioxide, which then increases conversion of acetoacetate to acetone because carbon dioxide is a product of that reaction. pH balance could play a role, or these could all work together to produce that result. So insufficient gluocose production could be an issue in that scenario, but you could have a lot of other negative things going on too.

    Montmorency, I don't think this makes that much of a difference because conversion from fatty acids is likely to be limited. If it forms say a tenth or a quarter, or, perhaps even a third of the glucose there, this doesn't fundamentally change the fact that you can make new glucose yourself, it just means that you might make enough while having a little less protein than we would otherwise calculate you need. But none of these facts argue whether glucose production is or isn't regulated perfectly on a low-carbohydrate diet. Certainly some people improve their blood sugar on such a diet, but to argue that, in all people, such a diet will necessarily lead to superior glucose handling is to make an argument that I find incredibly unlikely to be true.

    Hope that helps,

  41. Chris,

    As a low-carber (often very low) of around 4 years now, I've never worried that my diet wouldn't be able to produce any glucose that I really needed, e.g. from protein (not ideal), or from, e.g., the glycerol released along with FFAs when triglycerides undergo lipolysis). Low-carbing experts whom I respect have said that sufficient glucose will be available, provided of course that one is consuming sufficient calories from fat and protein.

    (I'm not an athlete, but I'm fairly active and have never suffered symptoms of hypoglycaemia)

    Moreover, I think they also generally said that only sufficient would be synthesised to meet the body's immediate requirements, i.e. not in excess.

    However, I have always wondered about that. Just say that for those of us who went low-carb because they were overweight (which is supposed to be a sign of a "broken" (or "bent") metabolism, what would be to stop our bodies synthesising "too much" glucose (say from fat, especially if we consume a lot of fat, as low-carbers are often encouraged to)?

    This could lead temporarily to rises in blood-glucose which could lead to rises in insulin secretion which could (a la Taubes, anyway) lead to fat storage.

    And could possibly lead to high blood-glucose in the longer term, or high blood-insulin in the longer-term, or possibly both, none of which is ideal.

    I think what I am leading up to is that while low-carb diets are often said to help with blood-glucose control, or bodyweight control, in some individuals, if it is theoretically possible to synthesise "too much" glucose from fat, then low-carb diets may be no more effective for these people for those purposes than higher-carb diets.

    Any thoughts?

  42. Hi Chris,

    Wonderfully written post.

    As you mention, it seems that only some people on ketogenic diets have noticeable acetone on the breath. Any chance these same people would be genetically or transiently poor at converting acetone to glucose? If these same individuals also tend to suffer from the side-effects of ketogenic diets (any association?), perhaps it is due to insufficient endogenous glucose production?

  43. Hi thaneverbefore, thanks!

    thaneverbefore, David, and Anonymous, I will write a post directing people to the review. I wanted to wait till it was indexed for pubmed, which occurred this morning. I will write that post when I get a chance to write up a lay reader-friendly summary.

    David, not a bad idea.

    Anonymous, that analysis was done before I was in this lab so I don't actually have the brands on hand, but I think it would be unethical to disclose that in any case, at least in this context. Sorry.


  44. I agree with "thaneverbefore" that your review of green tea and NAFLD ought to be mentioned. (I'm not so sure that it's even off-topic.)

    Given that people have been speculating in this comment trail about possible dangers to the liver from VLC diets, I plan to wash down any future low-carb meals with green tea and several egg yolks.

  45. Hi John,

    It doesn't change it too much, except the technical explanation. It is false that fatty acids can't supply oxaloacetate. However, their supply may be limited, as it is constrained by NADPH, whose supply is limiting during conditions of gluconeogenesis, so it is extremely unlikely that fatty acids alone could supply sufficient oxaloacetate to keep themselves burning in the TCA cycle without additional protein or carbohydrate. So all this does most likely is reduce the amount of protein from what we would otherwise calculate by some margin. How large that margin is, we don't know from current data. I do plan to put an update in that post soon.


  46. Chris,

    Does this change any of your thoughts shared in your post* about the ketogenic mice with a 5% protein diet? I wonder what would have happened if the mice ate 15% protein instead…would their body composition be different?


  47. Responses to DM Mitchell, Amber, Dana, Robert Andrew Brown, Anonymous, and Anonymous.

    DM Mitchell, that's very cool, thanks for sharing!


    You're welcome. Yes, this would be a source of glucose in addition to that generated by amino acids. I have trouble believing anyone can quantify the rate of gluconeogenesis so straightforwardly. Exercise increases gluconeogenesis from methyglyoxal; I'm not sure about from amino acids. But I do think that a ketogenic diet could result in glucose deprivation if someone's gluconeogenic pathways are not operating optimally. I also think that here we are dealing just with possibilities, but they have little or no bearing on what is optimal. For example, is it better to get glucose from glucose, or from gluconeogenesis? That's a question that quantifying the maximal operation of the pathways can never answer.

    Dana, Yes, I think it is a partial explanation. However, I think this effect results from multiple inputs, all geared towards reducing reliance on glucose and increasing reliance on fatty acids.

    Robert Andrew Brown,

    You're welcome and thank you for your appreciation. That's very interesting about the dolphins. That would be interesting to calculate whether their protein utilization could explain that amount of glycogen or whether it suggests additional sources. Thanks for passing it along!


    Thank you very much. I only pretend to be humble in hopes it might rub off on me, but thank you for appreciating my effort. 🙂 I do think it helps prevent the rush to conclusion.


    Thanks for passing that along. I do think this idea is very wrong, but it is not absurd, nor is it unanimously refuted by medical texts, because up to recently the biochemistry textbooks were saying exactly that — that adipose tissue had to derive the glycerol backbone from glucose. That is false, and Gary Taubes has admitted it, but I cannot blame him for believing it when I have biochemistry textbooks lying on my desk right now that state this as if it is fact. Glyceroneogenesis was "discovered" recently, although the evidence for it dates back decades.


  48. Hi Chris,

    this reminds me of a post that CarbSane wrote on her blog here. https://carbsanity.blogspot.com/2011/09/dietary-source-of-body-fat.html

    She was picking up on the fact that several low carb zealots like Taubes and Nora Gedgaudas had suggested that fats can be converted to glucose, and to a degree where it matters. To quote Gedgaudas:

    "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)."

    CarbSane points out that this is absurd and is clearly refuted in medical textbooks.

    1. I believe that quote was from the original version of Nora's book and was corrected in the newer version (Primal Body, Primal Mind). She no longer says this…I heard her mention it in a lecture.

  49. Chris:

    Great post–always learn from you and appreciate that you challenge conventional wisdom–academic or otherwise. Also, your humility and commitment to science appears to keep you from making rushed conclusions unlike many other bloggers.

  50. Robert Andrew Brown said…

    ^ Dolphins are seriously paleo


    "Dolphins naturally ingest a high protein diet (estimated to be 73% protein, 24% fat, 3% carbohydrate)"

    and would appear to appreciate glucose as a fuel source


    "The high concentration of glycogen in the skeletal muscle of the dolphin indicates the ready availability of the substrate for glycolysis during muscular activity, which is needed for the energy-yielding reactions enabling this species to swim at speeds around 25 knots."

    "The average glycogen content of dolphin skeletal muscle was 0.98 per cent as compared with 0.16 to 0.20 per cent for rat skeletal muscle."

    The obvious question is where is the glucose coming from?

    Thanks again Chris for all your excellent work. (-:

  51. Thanks Chris

    Fascinating post.

    I have long wondered about glucose production in marine mammals and particularly those with big brains.

    Intuitively it seemed unlikely that nature would have restricted the ability to make glucose to breakdown of protein, given that marine mammals must have a relatively limited carbohydrate intake.

    Will the glucose metabolism of marine mammals may be a subject of future interest?



  52. Could this partially explain (I'm sure it's not the entire explanation) why some people experience slightly higher blood sugar when they are on a VLC diet? *That* would be interesting…

    1. The most likely explanation is that they're eating too much protein as the basis of their diet, which easily converts to glucose via gluconeogenesis.

      I follow a cyclic VLC diet called Carb Nite where the majority of my calories come from fat, so my glucose level stays at baseline.

  53. Hi, Chris. Thank you for working out these details. As a biochemical layperson, I am often frustrated by what appear to be contradictions in various sources about what is actually happening, especially under ketogenic conditions.

    Am I understanding correctly, that this would be a source of glucose in addition to hepatic and renal gluconeogenesis from amino acids?

    I have seen arguments that splanchnic gluconeogenesis is rate limited to about 400g per day. I'm not sure if that's true, but I've seen that assertion, in turn, used to argue that a ketogenic diet is an impairment to certain special needs.

    The argument goes, for example, that if someone used up their capacity by being particularly active, or by regenerating glycoproteins, then they would be impaired for the rest of the day. There are other reasons that I don't find this line of argument compelling, but if glucose production from acetate alone could account for another 50 grams, that could be significant.

  54. It seems to me that when I was in a very restrictive place, there was an old edition of Encyclopedia Britannica, possible 1983 or 1984. I read the whole section on carbohydates, metabolism, etc. From that experience I came away with the belief that fats could be converted into glucose. Actually, I came away believing that the body could convert carbs, protiens, and fats to glucose if needed.

  55. Gretchen,

    Thank you for quoting these other texts. That's very interesting. However, the references I cited suggest that the vast majority of acetone is available for metabolism. If the authors read those papers, I do not think we can construe their statements as simplifications of the truth to prevent students from being confused by quantitatively unimportant pathways. I think it would speak much better of them to assume that they had not read those references. And I don't see why we would assume they had, since the primary research an entire biochemistry textbook is based on must be more than one person can read in a career, assuming they have their own research responsibilities.

    Likewise, the references I cited suggested that gluconeogenesis from acetone is quantitatively important. The paper on diabetic ketoacidosis estimated 10% of gluconeogenesis at a *minimum.* This is comparable to a glucogenic amino acid.

    Acetone is a quantitatively important metabolite of beta-oxidation — this is normal metabolism, not obscure side pathways.

    Again, my point here is not to detract from the skill or credibility of the authors, just to point out that the statements in the textbooks are almost certainly wrong.


  56. Hi Vlprince,

    Thank you for the suggestion. That's very perceptive of you, and a good question. I was going to write a section on dietary implications and I erased it all because it was so speculative and wishy washy. In it, I had a paragraph on CYP2E1. I realize that increasing CYP2E1 could possibly increase oxidative stress and vulnerability to chemical toxins, but there are two issues: first, everything I said above to Deb about pathological thresholds and alternative compensations for methylglyoxal applies to CYP2E1; second, there is some conflicting information about CYP2E1 because not all studies suggest it is harmful; and third, it's harm is almost certainly context-dependent. For example, maybe oxidizing ethanol generates free radicals and oxidizing Tylenol gives a toxic quinone, but if you are not drinking and taking Tylenol, oxidizing acetol might not be harmful. So I agree it is possibly harmful, but that is a giant "if."

    Hope that helps,

  57. Chris, you might be interested that I was looking at a book called "Metabolism at a Glance," published in 1994, 2nd edition 1999, and in the chapter titled "Mammals cannot synthesize glucose from fatty acids," they have a section entitled, "Possible gluconeogenic pathways using fatty acid precursors in mammals." They mention branched-chain FAs (which some other books also do) and omega-oxidation to succinate. But they say, "However, gluconeogenesis from these fatty acids is unlikely to be quantitatively significant in physiological terms."

    That makes me suspect that authors are aware of possible exceptions but leave them out to avoid confusing students.

    Another text notes that many books say you can't get glucose from fat, which is untrue because the glycerol backbone is a good gluconeogenic precursor. But again, although authors most likely know this very well, they're using a shortcut to help students learn.

    I agree with you that we have to accept that texts and professors are human and can make mistakes. I remember how shocked I was when I heard a professor state something I knew was not true. I thought they knew everything and that everything in print was accurate.

    I don't think I've ever quite recovered from the shock.

    1. A favourite quote that I offer to my son; who is currently at University…

      "It is important that students bring a certain ragamuffin, barefoot irreverence to their studies; they are not here to worship what is known, but to question it." — Jacob Bronowski

  58. Great post- and as one of the many 'memorization-impaired biochemistry students' you write of (I remember sitting with a stack of scrap paper, and six different highlighters trying to remember what went spinning off where) I was very amused by your perspective!

    I'm curious if you have any thoughts about the potential implications of increased CYP2E1 levels caused by fasting and/or high fat diets. In general, and perhaps potential production of glucose from lipids as you mention above. At least in the case of alcohol, the induction and increased reactivity of this enzyme (and the subsequent increase in ROS) is viewed as one of the major contributors to liver damage.

  59. I added this:

    "None of this is to say that we should blame the authors of these textbooks. Rather, we should admire them for undertaking such a gargantuan task, a task that no human or small group of humans could execute flawlessly. What is important is the recognition that any such work, no matter how authoritative, is a human work and thus necessarily subject to error."

  60. You might be interested that the 2011 edition of Lippincott's, which I have, has changed "a non-metabolizable side product" of ketone production to "a nonmetabolized side product," a subtle difference but suggesting it *can* be metabolized but *isn't.*

    I think textbooks are supposed to summarize "accepted knowledge," and it takes a long time for new findings to be considered generally accepted. I'm always surprised at how few texts use the term "epigenetics" today, even though it seems to be well established.

    The production process for textbooks is also slow. I used to copy edit medical books, some more than 1000 pages, and it takes a long time for that much material to be edited, then back to author, then proofread, then back to the author, then typeset. (No one does all that today, but it's still slow.) Any textbook is out of date by the time it hits the bookstores. So you don't go to texts for cutting-edge. You go for the basics, and for background that will allow you to understand the current research.

    You know that much of what you read in a text will be proved wrong in the future. But you need a general skeleton of the field on which you can hang new findings as they come out. If the texts tried to include all the exceptions to rules, they'd be so gargantuan that no one could get through them.

    So I agree with Dr. Feinman. I think you were a bit harsh on the textbook authors. I admire them for even attempting to cover such a complex field, including areas the authors are not experts in.

  61. Dr. Feinman,

    I agree with you that there are no "bad motives," and I did not mean to imply that. However, the evidence does not justify the categorical statement that "animals cannot convert fatty acids to glucose." This could easily be stated as "animals cannot convert acetyl CoA through the TCA cycle," with the text going on to explain this phenomenon, and then separately stating, "However, there is evidence that acetone is metabolized to glucose through several ways, and some biochemists believe it may make a quantitatively important contribution to net gluconeogenesis." I would not expect a textbook to be flawless, and I didn't mean to blame these authors. I just meant to say that we should not take statements in textbooks as definitive fact and assume that they necessarily represent the best science. I don't like questioning people's motives because I can't see inside their heart, but I think it is important to drive home the point that textbooks are sometimes simply inaccurate, not to bash the textbook writers, but to provide that realization to non-experts that may be tempted to take statements in textbooks for granted. Thank you so much for stopping by! You make good points, and it is good to have your contribution in the comments.


  62. Responses to Hannah, Anonymous, Tess, Deb, Tuck, Bill, Anonymous, Cavegirl, Anonymous, and Anonymous.

    Hi Hannah,

    Thank you for your appreciation. That is a good question about acetic acid. The explanation I have usually heard is that it slows absorption of carbohydrate in the intestine, but there is evidence that short-chain fatty acids like acetate and butyrate can have metabolic effects. I'm not sure what the most likely candidate(s) would be to explain benefits but it's definitely worth looking into.


    Your comments are very generous, more than I deserve, but very much appreciated. Thank you!


    I think we need to take textbooks as a starting point. No one can assemble such an enormous mass of information flawlessly, but we don't want to reject something just because it's imperfect, or else we will reject everything. So I think what we have to do, instead, is to regard everything as tentative and realize we are always learning. The textbook should be an entryway but not a destination.

    Deb and Tuck,

    I had started to write a section on dietary implications, but I erased it because I realized everything I was writing was such a stretch. I do not think this has any dietary implications.

    Ok, now individually. 🙂

    Deb, I am inclined to think likewise, that something that raises methylglyoxal levels is not good. However, there are problems interpreting this. For example, what happens to other dicarbonyls? Is methylglyoxal raised, but 3-deoxyglucosone lowered? To say this would increase AGEs would require a comprehensive screening of the known quantitatively predominant AGEs in at least plasma but preferably other tissues, and we don't have that data. Second, should we assume no threshold to methylglyoxal toxicity, or is there a threshold after which it becomes pathological? This is not known. If there is a threshold, is it exceeded on a low-carbohydrate diet? Again, this is not known. If the low-carbohydrate diet continues, does the increase in methylglyoxal continue, or does methylglyoxal normalize as other compensatory pathways kick in? Once again, this is unknown.

    Tuck, I do not think this changes the arguments to eat or not eat carbohydrate. Even if these pathways did not exist, if we assume gluconeogenesis is working optimally in a person, they can make glucose from protein alone. If fatty acids can account for some 10 percent of gluconeogenesis, all this means is that we have to eat somewhat less protein on a carbohydrate-free diet than we would otherwise calculate. This says nothing about the optimal carbohydrate intake, if such exists. It is easier to make fat from carbohydrate than carbohydrate from fat, so perhaps it is best to just eat carbohydrate and let the body make only as much fat as it needs. I don't believe that, but one could make either argument with equal force based on the biochemistry.

    Bill and Anonymous, I agree with Anonymous. I am suggesting that the net synthesis of glucose from fatty acids is common on low carbohydrate intake and is quantitatively important, perhaps accounting for a similar proportion of glucoeneogenesis as any particular glycogenic amino acid might.

    Cavegirl, De Vany's experience is peculiar to diabetes, so I think we should consider it very valuable, though we should be careful about generalizing to broadly. Thank you very much for contributing your comments.

    Anonymous, thanks! I'm glad the levity helped. I try my best to make the really boring stuff somewhat fun. 🙂

    Anonymous, thanks!


  63. This is very good post but I think your attitude on textbooks is slightly off the mark. Textbooks are written by guys like you and me who don't know everything and while textbooks do tend to be very conservative, there is an obligation to provide well established evidence and to avoid the superficial jumping to conclusions that we now have to deal with with the fructophobes. The texts I am writing will certainly incorporate this but it is important to remember that this tree is a side reaction on the main forest whose major points are hard to see. In particular, the point that you make that taking OAA out of the TCA cycle (that's what Krebs called it) means you have to put something back in is not obvious to students. (it is hard to understand that the TCA cycle intermediates are not really intermediates but rather catalysts). So, I don't think you want to imply bad motives on the part of textbook writers. On the other hand, if you want some unpleasant behavior, you can look at reference 7 which did not bother to research the chemistry you describe but latched onto methyl-glyoxal as a way of keeping patients from finding out the potential benefits of a low carbohydrate diet.

  64. very cool post! written with enough levity for a layman like me to follow, and dare i say… it was entertaining!

  65. Art De Vany has been saying we should let the liver produce just the amount of glucose we need for years after his experience with a type 1 diabetic son and first wife and his own n=2 experimentation with their diet and insulin requirements.

    1. Yeah, that works great as long as you don't have any kind of athletic goals.

      Fortunately it's not that simple, and Type-1 diabetics can make use of the GLUT-4 pathway for glucose uptake in the muscle after resistance training, during which time insulin is totally not required for processing of glucose from food.

  66. Bill, I think he's suggesting that it's not only possible, but is important, quantitatively and otherwise, under conditions of low-carbohydrate intake!

  67. Hi Chris, another intriguing post, very interesting stuff. To [over]simplify your post, would you say that the net synthesis of glucose from fatty acids is possible, but rare and quantitatively unimportant?
    Thanks, Bill

  68. Great post. The argument for consuming glucose takes a big hit if this is correct. Much better to let the body produce what it needs…

  69. wow…. so much for science-based textbooks; if they're wrong about this (and the evidence has been there for over half a century), what's gonna come next? thank you for ferreting out such interesting stuff, Chris!

  70. Chris, you always floor me with your intelligence, your humor, your incisiveness and your graciousness.

    What a gift you are!

  71. Fascinating post. So does that mean that the consumption of acetic acid (as vinegar or Kombucha) helps the body maintain a balanced blood sugar? Many diabetic have reported the need for less insulin when they include Kombucha as part of their regular diet.

Leave a Reply

Your email address will not be published.