Tuesday, May 3, 2011

When Fat Burns In the Flame of Lean Muscle Mass -- Better Put That Butter Either on Steak or Potatoes

by Chris Masterjohn

My last post was political and got 17 comments just in the first six hours.  I suppose that means I'm due for another post about religion, or one about sex.  Given Stephan Guyenet's recent post about the dangers of hyperpalatability, though, I'm inclined to obey the proverb "don't take too much honey" and delicately sprinkle those sweet and enticing posts on the nourishing bulk of the usual biochemistry you get at this blog.  

I try to make this blog more like wine than grape juice -- you know, the adult flavors people enjoy in France where obesity rates are lower.

And besides, this post will answer the cliff-hanger I dropped at the end of Monday's post, "Let Us Honor Ancel Keys, Our Patron, As We Cherry Pick Studies to Bash Fructose." 

Peter Dobromylskyj over at Hyperlipid recently posted "Diabetic nephropathy and the lost Swede," wherein he discussed the ability of a ~95% fat diet to produce weight loss in mice.  

In this post, I'd like to take a look at what happened to food intake, hormones, and body composition in that study, and explain why eating butter with no steak, bread, vegetables, or potatoes under it isn't a very good idea.

And no, this is not just due to the "problems of traction presented by the butter-butter interface."  Though I admit that's a problem too.

Here's a link to the study for anyone who wants to follow the reference:
Kennedy AR, Pissios P, Otu H, Roberson R, Xue B, Asakura K, Furukawa N, Marino FE, Liu FF, Kahn BB, Libermann TA, Maratos-Flier E. A high-fat, ketogenic diet induces a unique metabolic state in mice. Am J Physiol Endocrinol Metab. 2007;292(6):E1724-39.
The authors compared a standard chow diet to a "high-fat" diet containing 45% fat and 17.5% sucrose, and to a zero-carb, 95% fat "ketogenic" diet.

They also put one group of mice on calorie restriction, where the mice ate only two thirds of their normal calorie intake.  The authors don't describe this diet as clearly but it was apparently a chow diet formulated to give the mice the same amounts of vitamins and minerals they would have gotten on any of the other diets.

The authors reported food intake rather strangely, reporting only day one, day 25, and the total over the first half of the study, stating only that the pattern was consistent in the second half of the study.

Nevertheless, it appears from what they reported that there was no difference in food intake but that at the very beginning of the study mice on the "high-fat" diet ate twice as much food as the others.  This is a sign that the diet was more palatable than the others.  This supports the idea that sugar makes a high-fat diet extra-palatable, but also shows that a 95% fat diet is palatable enough for a mouse to eat normal amounts of it. 

Now let's look at obesity:

As always, you can click on the image to enlarge it.

Interestingly, the only diet that produced obesity was the extra-palatable diet, the one that produced a doubling of food intake at the very beginning of the study but no difference in food intake over the long-term.  Although I personally have some data of my own suggesting that palatable diets do not always produce obesity, this observation is nevertheless consistent with the idea that palatability may have some effect on obesity independent of caloric intake.  Perhaps Stephan Guyenet's new series on palatability will make some sense out of this.

Now let's take a look at insulin and leptin:

Which graph corresponds to fat mass?  Clearly, the leptin graph and not the insulin graph.  Only the mice who got obese had dramatically elevated insulin, but the ketogenic mice had 90 percent lower insulin compared to chow control mice and no difference in fat mass.

As I wrote late last November in "Is Insulin Resistance Really Making Us Fat?", leptin resistance is much more closely associated with obesity than insulin resistance is.  LIRKO mice have no functional insulin receptors in their liver and consequently have 7-fold higher fasting insulin and 23-fold higher insulin after a meal yet are at least as sensitive to leptin as healthy controls and if anything slightly leaner.  Mice and rats with defects in leptin signaling are all insulin resistant and fat.

But uh-oh, let's look at lean mass:

Hmm, it looks like eating a 95% fat diet is sort of like starving yourself.  All of the weight lost on the super-ketogenic diet was lost as lean mass.  And this isn't just because these young mice failed to grow normally; they actually lost weight.

The ketogenic diet wasn't entirely like the calorie-restricted diet.  The calorie-restricted mice had lower testosterone, whereas the ketogenic mice had normal testosterone.  The mechanism of weight loss seemed different.  The ketogenic mice turned more energy into heat than all the other mice.

The authors concluded that the loss of lean mass was not due to muscle wasting because there was no decrease in lean mass in an isolated hind limb like there was in the whole animal.  Regardless of whether the loss of lean mass came from their internal organs, bones, hind limb muscles, or some other set of muscles, they still lost lean mass without losing any fat.

There is evidence that the elevated levels of free fatty acids that occurs on an extreme ketogenic diet helps divert energy towards heat production, at least in rodents.  But why would this result in lower lean mass instead of fat mass?

In 1895, a biochemist by the name of Rosenfeld coined the expression "fat burns in the flame of carbohydrate."  This was based on observations that cells could break down fatty acids into ketone bodies but without sufficient glucose the cells could not break them down fully into carbon dioxide and hydrogen.  

The aphorism offered a simple explanation for why ketones are elevated to extreme levels in diabetes: diabetics do not use glucose efficiently, so glucose levels in the blood rise; since their cells are starved of glucose, the fat stores of these diabetics release their fatty acids and their livers break the fatty acids down into ketones, but these ketones cannot be used in the absence of glucose, so ketone levels in the blood rise and then the ketones are finally lost in the urine.  Thus the diabetic is effectively in a constant state of starvation.

We now know that fat burns in the flame of oxaloacetate, which can be derived from either glucose or amino acids.
When we break down fats or carbohydrates for energy, we turn them into acetic acid, or acetate, which is a two-carbon unit.  A little shuttle called coenzyme A, which is made of pantothenic acid, carries the acetate around and together we call the complex acetyl CoA.  Pantothenate is also called vitamin B5 and is found abundantly in many foods but liver and egg yolks are among the highest (along with certain mushrooms, seeds, and yeast).

In order to fully harvest energy from acetate, we need to send it through the citric acid cycle, also called the Krebs cycle or the tricarboxylic acid (TCA cycle).  This cycle will break the acetate down into carbon dioxide and hydrogen.  In doing so, it will also release high-energy electrons whose energy can then be harvested to synthesize ATP, a major usable energy currency of the cell.  Entry into this cycle is dependent on a compound called oxaloacetate.

In the presence of glucose, we convert glucose to oxaloacetate.  Thus, as oxaloacetate leaves the Krebs cycle cuz it's got things to do and people to see, we can just use glucose to replenish it.  In the absence of glucose, we do the opposite: we turn oxaloacetate into glucose.  Thus, oxaloacetate gets depleted in the absence of glucose unless we have some other source of it.  We can make oxaloacetate from a variety of amino acids, but not from fats.  Thus, in the absence of dietary protein or carbohydrate, the only place to get oxaloacetate is to dig into the lean proteins found in our muscles and internal organs.

The authors of the mouse study didn't investigate this explanation, but it seems the most reasonable to me.

Thus, eating an all-fat diet isn't very smart.  Of course, virtually no one in the real world would actually do that, at least not at the extreme level used in this mouse study (95% of calories from fat).  Hyper-ketogenic diets used to treat disorders of the central nervous system, especially seizures, might get close at around 90% of calories.  In any case, biochemistry varies from person to person, and someone attempting a very low-carbohyrate diet who feels fatigued or is losing lean muscle mass might want to consider whether they should up their protein or carbohydrate content.

One thing I really like about  The Perfect Health Diet is that although the authors advocate a low-carb diet, they devote a lot of attention to the body's need for glucose, rather than coming up with some silly aphorism like "there are essential amino acids and essential fatty acids, but there is no essential carbohydrate."  The body may be able to survive without dietary glucose, but only because it can make glucose from protein.  Give it only fat, and it will make that glucose — and oxaloacetate — from lean muscle tissue.

Better get a steak or potato to go with that butter!

Read more about the author, Chris Masterjohn, PhD, here.


  1. Re: "Thus, eating an all-fat diet isn't very smart. Of course, virtually no one in the real world would actually do that.."

    Are you suggesting that noone 'actually' waddles over to Starbucks to wash down those 2 'Double Quarter Pounders with Cheese' with a 'Peppermint White Hot Chocolate' (that slice of chocolate cheesecake notwithstanding)?

  2. Vilhjalmur Stafansson (in _My Life with the Eskimo_)


    said he'd been in the position on rare occasions (while travelling) both of having not enough fat and of having not enough lean, and he said you wouldn't want to be in either position, although he thought the first state slightly more unpleasant.

    Those are unusual kinds of experiences of course, because most of us move in environments where carbohydrates aren't difficult to kind.

  3. The aphorism *isn't* silly. The nutritionist's definition of "essential," when they aren't posing in the media to further the low-fat/low-cal cardboard rabbit food agenda, is "necessary to the body's survival but cannot be made by the body." Is glucose necessary to the body's survival? Yes--at an amount of about two teaspoons for the entire body. Can glucose be made by the body? Yes. Therefore, glucose is not essential.

    Jay, do you really believe the meal you've described is an ALL-fat meal? Really? And since when does Starbucks serve Quarter Pounders?


  4. I'm unable to find the exact studies, but I recall that severe carbohydrate restriction shifts the balance between rT3 and t3 in the dejodinase in humans - at least when combined with calorie restriction. In that way a very low calorie diet would increase rT3, but combined with a strict carbohydrate restriction this effect would increase even more. A higher level of rT3 and a lower level of T3 will significantly lower the metabolic rate.

    These two papers is what I'm able to find now.

  5. Dana,

    The brain can get at most ~75% of its energy from d-3-hydroxybutyrate, as some dendrites are too narrow to accomodate mitochondria. The rest has to come from glucose. Red blood cells get energy only from glucose. Ditto parts of the liver.

    The human body uses more than two teaspoonfuls (~9g) of glucose per day. I'd guess ~50g per day minimum for a sedentary person. The liver & kidneys can synthesise that from amino acids, glycerol, pyruvate & lactate but it isn't ideal, *especially* if the person isn't sedentary!


  6. Dana,

    I think Jay meant washing their MacDonald's down with starbucks. Still, you are right, the food he described, while very high in fat, is also very high in sugar and protein.

    Jay, the diet Chris is referring to would be one composed almost entirely of butter, coconut oil, almond/macadamia nut oil, fish oil, and animal fat with all noticeable protein trimmed away. No burger buns, nor even much meat (since it deliberatly avoids protein).

    I can't see any other way of being remotely healthy while eating 95% fat, since you'd want to avoid eating large amounts of industrially refined seed oils. So yeah, there are undoubtedly precious few people actually eating this way outside laboratories.

  7. eating VLC say aprox 20g of carbs a day interspersed with the inevitable 1 day a week of pigging out on home baked goods or that plate of pasta you have been avoiding all week is probally more similar to what a low carber does than any sort of %95 all fat diet that these studies are measuring.

    i dont fear loosing lean muscle because only the most hardcore disciplined low carber will approach %90 and above for fat levels, it just seems too extreme.

  8. Response to everyone.

    Jay, I'm thoroughly confused. Do any of these foods, or any others at Starbucks, contain 95% of their calories as fat?

    Mike, great anecdote, and further supporting the suboptimality of eating close to 100% fat.

    Dana, the physiological essentiality of glucose is much greater than two teaspoons. I think you are confusing how much exists in the blood at one time with how much is used by cells in one day, which are two radically different values. The brain uses upwards of 150 grams of glucose per day (about ten tablespoons!), and it doesn't shift over to predominant ketone use unless one is practicing extended fasting or a hyperketogenic diet of the type used to treat seizures, which requires substantial protein restriction in addition to carbohydrate restriction, and borders very close to the conditions that could diminish lean mass. In fact, some authors argue that oxaloacetate depletion in the brain on a ketogenic diet may be part of the therapeutic efficacy, by shifting amino acid balance away from glutamate and towards glutamine and GABA. This is a therapeutic diet, and not one within the spectrum of traditional diets. The Inuit on their traditional diet, for example, had minimal ketone excretion in their urine, suggesting they had enough protein to supply glucose needs and use ketones efficiently, which would also keep the brain running to a large degree on glucose.

    The aphorism is silly because "essentiality" as defined by the biochemical establishment is not any index of optimality. There is no "essential" saturated fat, and no "essential" monounsaturated fat, but there are "essential" polyunsaturated fats. And, moreover, neither arachidonic acid nor DHA are "essential," but only linoleic acid and alpha-linolenic acid. By this logic, we should emphasize soybean oil and flax oil.

    smgj, thanks, I'll check that out.

    Nigel, good points, though I think the physiological glucose requirement is more than that.

    Bryce, exactly.

  9. Jeff, I agree that periodically increasing protein/carb could alleviate any potential problems of eating close to the problematic borderline for the other days of the week. Everyone's a little different though, so it's good to keep this biochemistry in mind in case a problem arises that it can explain.

  10. Hi Chris, nice post.

    Another point I would make is that the use of body fat as a source of energy requires protein, more specifically albumin.

    Without protein, the albumin pool goes down, preventing the proper transport of FFAs to the tissues that will use them.

    Does this make sense to you?

  11. Chris, thanks very much for this post. Fascinating subject!

    If you don't mind a question, is it possible to say how much protein is necessary for the necessary glucose if one if eating 30 or 20 or 10g/CHO/d, for folks who do not do intensive exercises which deplete glycogen in the muscles?

    I really enjoyed this post. Thank you!

  12. Sorry about the typing error. That ought to read:

    ...if one is eating....

  13. Fascinating post, Chris. I agree that The Perfect Health Diet is really a very impressive piece of work.

    In response to your "response to everyone", I came across this review of ketogenic effects on the body, which includes the brain glucose/ketogenic switchover you discuss, as well as a bunch of interesting effects. For anyone interested in learning more about this:


  14. donny of naivenutrition (http://naivenutrition.blogspot.com/2010/04/httpwww.html), which has been inactive recently though, has lots of interesting ideas regarding Krebs cycle.

  15. Hi Ned, thanks -- that sounds plausible to me, though I'd have to look at some in vivo research on the matter to see just how much it contributes to muscle wasting and/or inability to deal with fatty acids. I would think FFAs could be utilized without albumin carrying them, but lower albumin would seem to have the potential to exert quite a lot of physiological effects, and of course maintaining albumin status would require digging into muscle.

    Hi H, you're welcome, and thanks! I think it would be really complicated to calculate the need in grams and would be widely variable between people and conditions, but I can say this: most proteins have about 50-60% of amino acids as potential gluose precursors, so as a rough estimate you should consider 2 grams of protein to be worth 1 gram of glucose.

    Tuck, thanks! Thanks for the link -- I've seen Mary Newport, who runs that site, speak about her experience with her husband and it is quite fascinating.

    john, thanks for the link.

  16. Chris, thanks very much. There is always much discussion amongst those who eat very low carb as to optimal protein intake. I appreciate your help.

    Thank you for your kind work in helping folks to enjoy improved health.

  17. Chris - I doubt you need to be burning predominantly ketones for the brain to see some benefit from ketosis. Zooko sent me links to so,e interesting studies on twitter - I think I will do a post on it, actually. They are familiar papers but the graphs are very interesting.

  18. Emily, I don't doubt it, but the brain selectively gets the glucose when it is initially limiting, with muscles getting the ketones. Under more extensively glucose deprivation this shifts towards muscles using fatty acids and brains using ketones, which spares lean mass by reducing the need for gluconeogenesis. This is my understanding, at least, though I do need to look deeper at the primary literature on this and the differences between ketogenic diets and prolonged fasting. But my point in response to Dana was that even on a very low-carb diet with a moderate amount of protein, the brain, while perhaps using ketones to some degree, is probably still using primarily glucose. I'm very open to changing this view as I consider it tentative. I look forward to reading your post.


  19. I should say that a strict anti-seizure benefit would have to be from strict ketosis to up regulate the GABA reliably and consistently. But other conditions seem to benefit from dips into ketosis, or ketosis that just turns a ketostix. Well, post coming one of these days...

  20. Hey Dr. Deans,

    In my experience, I can get ketostix purple just by exercising intensely and feasting on carbs. I'd be willing to bet that my brain was running almost entirely on glucose when my ketostix were as purple as they could get (granted I was eating plenty of fat including coconut fat during this time). I'm definitely not claiming dips into ketosis won't be of benefit -- I'm just saying (in my response to Dana, which I'm assuming that's what you're responding to) that in calculating the physiological glucose requirement of the brain you can't assume that a low-carb diet is going to substantially reduce it. Most low-carb diets probably keep the brain running mostly on glucose is my guess, and that's all I was saying.


  21. Hi Chris,

    Great post, and thanks for the shout-out!

    Here are a few thoughts I had while reading:

    (1) I’m also looking forward to Stephan’s series. I think there will have to be an important distinction between palatability and “superstimuli” that disorient the metabolic control mechanisms.
    I believe our tastes evolved to help us be healthy and fit, and so the healthiest diet will be highly palatable, yet not fattening. Since our diet is rich in starch and fat, which makes a highly palatable combination, I’m a bit at odds with the lean-meat-and-vegetables for appetite suppression strategy. And I definitely agree with you – not all palatable diets produce obesity.
    I tend to attribute the fattening effect of fructose-sugar and polyunsaturated fat combinations to toxicity effects, but they may also be “superstimuli” that derange the brain. I expect to learn a lot from Stephan’s series!

    (2) The leptin graph is very interesting. Leptin is higher in part due to increased fat mass, but leptin levels went up 7-fold, far more than fat mass I believe. Is the leptin – fat mass relation nonlinear?
    Obviously leptin resistance plays a major role, and that is the key fact to explain. But what causes leptin resistance on normal diets?

    (3) On the issue of higher heat generation in the ketogenic mice: I had a similar experience when I ate a very low-carb ketogenic diet. Body temperature was very high, appetite was ravenous, but I kept losing lean mass. The cause turned out to be scurvy. Once I supplemented vitamin C, body temperature and appetite immediately returned to normal and lean mass was regained.
    Now these mice won’t have scurvy, but I think it’s possible that micronutrient deficiencies could factor in. What would be body logically do to remedy a chronic micronutrient deficiency: increase appetite to get you to eat more food, and turn up waste heat generation to dispose of the extra calories.
    In this case I think you’re right that it was a protein deficiency that was the issue, but it may be that a shortage of protein for collagen formation was most important and was behind the lean mass loss. That might explain why the hind limb didn’t lose mass – it wasn’t a whole-body loss of protein, but an inability to remodel tissue, and the hind limb didn’t need much remodeling because the mice were in a shoebox cage and not exercising much.

    (4) The “fat burns in the flame of oxaloacetate” discussion is an elegant summary of why 95% fat diets are unhealthy. Re Dana’s point, the body consumes a lot of glucose whether it is consumed or no. If you want to exclude carbs from the diet, you’d better eat sufficient protein! Even if you do, it’s far from clear that manufacture from protein is the optimal way to meet glucose needs. I don’t believe it is.

    Re ketogenic diets, I’ve migrated to the view that the optimal ketogenic diet even for clinical use should have minimal excess ketones, few excreted ketones, and a fair amount of dietary carbs and protein. Once ketones are being excreted, few tissues are taking up marginal ketones. I think the clinical ketogenic diets often overshoot the optimal amount of ketones, and undershoot the optimal amount of carbs and protein.

    Re Ned’s point, I’ve eaten a moderately ketogenic low-protein diet for a long time and my albumin levels have never budged. Of course I’ve never eaten a 95% fat diet … but I would imagine you would lose a lot of lean muscle mass before serum albumin levels would start to decrease.

    Anyway, great thought-provoking post!

    Best, Paul

  22. Great post. I found so many interesting stuff in your blog especially its discussion.There are several types of cholesterol in the body.can you give some information about type of cholesterol

  23. Hi Chris,

    You said: ": most proteins have about 50-60% of amino acids as potential gluose precursors, so as a rough estimate you should consider 2 grams of protein to be worth 1 gram of glucose."

    Do some kinds of protein more easily convert to glucose at a higher rate of conversion or is that 2:1 ratio pretty fixed across the board?

  24. Another great post Chris.

    The way I see it is that fat can burn in the flame of anything that keeps the TCA cycle going. Even though oxaloacetate is drawn off, you can replenish it via every intermediate in the TCA cycle, because it will eventually be turned into oxaloacetate anyway.

    Glucose and amino acids are just two of the TCA cycle replenishing options.

    Another option would be odd-chain fatty acids. They can burn (somewhat) in their own flame, because the last cycle of their β-oxidation will provide propionyl-CoA, which can be turned into succinyl-CoA, a TCA cycle intermediate (not shown in the TCA cycle diagram BTW).

    And what about the substance that gave the cycle its name: citric acid. Perhaps that's the reasoning behind the lemon diet: burn fat in the flame of citric acid. The citric acid (and some sugar) from the lemons should be able to keep the TCA cycle going, so that would spare protein.

    But when you know all this, why go for the sour taste of a lemon diet? Using a sugar drink should do the trick just as well.
    Hmmmm, burning fat with a sugar drink? That sounds a lot like the Shangri-La diet by Seth Roberts.


  25. Paul,

    You wrote: "But what causes leptin resistance on normal diets?"

    My guess is that elevated fasting insulin is an option for this, because insulin seems to act on the same neurons in the arcuate nucleus as leptin (NPY and α-MSH). And as an energy signal, insulin is postprandial/urgent, while leptin is long term/less urgent.
    So to me it sounds plausible that insulin signalling will override leptin signalling. No problem when it's only postprandial, but when it's still overriding it in the fasted state, it might be that the hypothalamus doesn't "see" all the fat. So it thinks there isn't enough fat, and it switches the energy balance to a starvation mode: "eat more, move less".

    One option to get elevated fasting insulin levels seems to be fatty liver, which Chris has written about extensively. A high calorie, low nutrient diet (e.g., lacking choline) seems to bring about fatty liver fairly consistently.

    Of course, if a diet is highly palatable and nutrient poor, it seems very easy to get fatty liver eventually: eat all the energy rich empty yummy stuff, and flood your liver with energy it cannot process because of lacking nutrients. And the leptin resitance resulting from the fatty liver will then cause even more eating, so a vicious cycle of metabolic derangement begins.
    On the other hand, if the starting point was a highly palatable nutrient dense diet, none of that would have happened, because the liver would have stayed healthy, and so no leptin resistance, no long term overeating etc.


  26. === I believe our tastes evolved to help us be healthy and fit

    wel,, I see evolution has produced many thousands of insect species who are adults only long enough to breed. So, my own guess is that the partial-pressure of evolution on hominids has been to help be fruitful and multiply IRRESPECTIVE of our health or fitness or longevity.

  27. Trying to understand the meanders of metabolic pathways is a very nice thing. Unfortunately, we as a species cannot live with the case of “I don’t know yet” or “no conclusion on this issue yet” and we tend to replaces the hole in our knowledge with partial answers, based on whatever studies, personal experience and so on. Most likely these answers are NOT the right ones…

    While we don’t exactly understand what we eat, how it is absorbed, in what quantity, where is it transported, etc, all our “scientific data” is at the level of pure speculation. This research can’t be done with the level of technology we have now if we really want accurate answers, in vivo. So why do it?
    I was reading the other day a paper that compared ileac absorption of egg proteins, raw and cooked, with a double labeled method. The results were surprising… cooked eggs have a better absorption. So why is that? Varies with season?? quality of the eggs?? cooking temperature?? How about whites only vs. yolks? Quantity matters? See? Not only it gives me a partial answer but opens more questions that will lead to more speculation. Base line… pointless

    One more point, before I digress… We as an entity live together along with billions of symbiotic organisms that have their own metabolic pathways. Ignoring their part of the game bites massively in our understanding of how things really work. Yes, the citric cycle is a cool thing but that’s way deep in OUR cells and the game is not just us.

  28. I'll respond to the other comments later today, but a quick response to Mike of Paleo Village -- Mike, I think you are overthinking this a bit. It makes perfect sense that cooking whole eggs would increase the digestion of their proteins because raw egg whites contain trypsin inhibitors, which decrease protein digestion. The firm reality is that sufficient protein and/or carbohydrate is necessary to efficiently burn fats for fuel. My conclusion, therefore, is that if someone is eating a high-fat, low-carbohydrate diet and they feel fatigued or are losing lean muscle mass, they should consider increasing their carb or protein intake to see if it helps. It's a practical conclusion, not an esoteric one.

  29. Totally agree with you Chris. About the over-thinking :) Thanks for your answer!

  30. Something doesnt add up here, so without oxaloacetate, the TCA breaks down and energy production stops?

    According to the proposed theory of ketogenesis, we can bypass the CTA completely and make ketones that are either exhaled, oxidized or excreted in the urine, Acetyl-COA is all we need to keep energy production going, which we can get plenty from in fat burning.

    So, during carb restriction, all glucose is pretty much used by the brain, meaning all the mitochondria in your body will be completely deprived of TCA substrates and thus unable to produce any energy?!

  31. Kindke,

    Yes, without oxaloacetate (or some precursor of oxaloacetate), the TCA cycle cannot go forward (once pre-existing oxaloacetate runs out as it gets used for other purposes), and full utilization of energy stops.

    What is the "proposed theory of ketogenesis"? Proposed by whom?

    Of course you can obtain some energy from breaking fatty acids down into ketones in the first place, and then you can excrete them, but this obtains only a small proportion of energy stored in the fatty acid and if this was the primary way of obtaining energy from fatty acids you would probably starve to death quite quickly during fasting or ketogenic dieting.

    During carb restriction, the mitochondria are not necessarily starved, because if you don't give them enough protein and carb to derive oxaloacetate, they can break down muscle mass to obtain it. Of course, during more moderate carb restriction this doesn't become an issue because you already have enough carbs and/or protein.


  32. Kindke, Chris,

    The way I understand it (but I'm still learning), the story for ketones is a bit as follows:

    Ketone formation in the liver will occur if:
    a) The liver cells have excess acetyl-CoA, and,
    b) The circumstances are unfavorable for fat synthesis (else the excess acetyl-CoA would be converted to fat).

    Carb restriction accomplishes both: it causes the liver to do gluconeogenesis, for which it draws oxaloacetate from the TCA cycle. This slows the cycle down, reducing its consumption of acetyl-CoA, and so acetyl-CoA accumulates.

    The acetyl-CoA excess will then be converted to ketones, as carb restriction is unfavorable for fat synthesis. And the liver lacks a key enzyme to convert ketones back to acetyl-CoA (β-ketoacyl-CoA transferase a.k.a. thiophorase). So the ketones will be exported into the blood.

    Muscle and brain cells will take up the ketones, and convert them back into acetyl-CoA. Their TCA cycle will also be oxaloacetate deprived during carb restriction, so the cycle needs to be replenished. As I wrote in my previous comment, this can be done via any of the TCA cycle intermediates or their precursors. Supplying these via the diet is preferred, because else they will be supplied by breaking down protein in the body.

    Long story short: ketones from the liver are basically water soluble pre-oxidized fatty acids. The liver did the oxidation already, so the cells using the ketones don't have to anymore. They just need a replenished TCA cycle to use the energy.
    The energy yield from ketones should be comparable to that from the fatty acids they were made of. And the cool thing is: ketones can cross the blood-brain barrier, which fatty acids cannot.

    And if you don't want to burden your liver to make ketones, you could supply them in the diet as well. Vinegar is an option (note: acetyl-CoA is also known as activated acetic acid).


  33. Responses to Paul, high cholesterol drugs, Frosty, John, Gino, and Alan

    Hey Paul,

    Thanks! And great comments.

    I think the leptin-fat relationship is not linear here because the obese mice are leptin resistant. I will do a post soon on whether insulin resistance might cause leptin resistance, but on the whole I think this is doubtful except in the possible case of high trilglycerides. The evidence is better that leptin resistance causes insulin resistance. Obesity may cause leptin resistance, but this is somewhat unsatisfying because leptin resistance should cause obesity. Inflammation may contribute. Messing up the reward system could contribute, as in the theory Stephan is laying out.

    Interesting theory about heat generation, but the ketogenic mice didn't eat more food. They also had their diet matched for nutrient content.

    I agree manufacturing glucose from protein is unlikely to be the optimal way to get it. For whatever reason, some people at least seem to do well on zero-carb diets. Obviously it's not the historical norm.

    I think "ketogenic" is somewhat misleading in that on these diets part of the rise in ketones in plasma and urine is due to ketone resistance. Still, some people believe part of the benefit is from oxaloacetate depletion in the brain. I'd be interested to see your thoughts developed in a post.

    high cholesterol drugs,

    Maybe try here: http://www.cholesterol-and-health.com/LDL-HDL-Good-Cholesterol-Bad-Cholesterol.html


    This is an average. Some dietary proteins vary from others. But ultimately this just represents the potential. You can't count on all the protein being converted to glucose because that depends both on the body's need and its capability. Maybe I'll try to do a more in-depth post on this in the future.


    Good points. I singled amino acids and glucose out for their dietary abundance. Glycerol is also a gluconeogenic substrate, but I didn't mention fat because the glycerol-to-fatty acid ratio is pretty much fixed. Odd-chain fatty acids tend to come along in small amounts for the ride. Citrate is interesting -- many people claim, anecdotally, that citrate abolishes the effect of low-carbing on fat-burning. This makes sense to me as cytosolic citrate turns on lipogenesis and turns fat-burning off. Not sure if it would do that before making it down to the mitochondria! If you know of any in vivo studies, please post!


    Most estimates I've seen are higher than that.


    Good points, though we are not insects. Humans are in part distinguished by their longevity, so it seems like something we are, to some degree, "supposed to have."


    Most of that sounds right to me. Thanks for posting it.


  34. Hi Chris,

    Thanks for the feedback. My citrate idea was me getting a bit carried away with information. My biochemistry knowledge is still in its infancy, with bits and pieces here and there, and gaping holes elsewhere. But I'm working on improving this.

    One gaping hole in my knowledge was that citrate seems to activate acetyl-CoA carboxylase. So it indeed makes more sense that lipogenesis would occur. In that regard the lemon diet doesn't make a lot of sense. Maybe someone whose biochemistry knowledge was as limited as mine came up with the diet ;-)

    I'm looking forward to your post about leptin resistance, as I'm someone who currently thinks insulin might be capable of interfering with leptin signalling, and I don't mind cognitive dissonance :-)

    Part of my insulin idea is based on Robert Lustig's work [1]. The other part is just me thinking it makes sense: both leptin and insulin appear to stimulate and inhibit the same neurosecretory cells in the arcuate nucleus. I would think that insulin would get priority over leptin, because it requires immediate action. Insulin's message might be: "Stop eating, because your blood sugar is getting too high!". After insulin drops off to a normal fasting level, leptin gets back in control (of energy signalling). In this speculative model, it seems plausible that if insulin does not drop off completely, it might cause the hypothalamus to not "see" all the leptin.
    Again, just me doing more speculation. I definitely prefer well researched science as you always present it over speculation (even my own).


    [1] Lustig R.H. Childhood obesity: behavioral aberration or biochemical drive? Reinterpreting the First Law of Thermodynamics. Nat Clin Pract Endocrinol Metab. 2006 Aug;2(8):447-58. http://pmid.us/16932334

  35. Hi John,

    Citrate is basically a communicator of the energy state of the cell. As ATP and NADH increase, which indicates plenteous cellular energy, isocitrate dehydrogenase is inhibited. This causes citrate to accumulate. As citrate accumulates, it is converted outside of the mitochondria. Once outside of the mitochondria, it activates fatty acid synthesis.

    Incidentally, ROS inhibit aconitase, which converts citrate to isocitrate, so ROS also cause citrate accumulation by acting as a mitochondrial overload signal.

    Insulin communicates that the body is in the fed state and thus not in need of ketones. It also activates fatty acid synthesis.

    Thus, the energy needs of the body (insulin), the energy needs of the cell (cytosolic citrate), and the capacity of mitochondria to deal with more energy (ROS) all converge to dictate whether acetate is broken down to produce more ATP or is instead rebuilt into fatty acids for storage.

    Dr. Lustig sent me five papers providing the basis for his idea that insulin causes leptin resistance so I will include that information in my review of the hypothesis.


  36. "Better get a steak or potato to go with that butter!"

    Sounds tasty! ;)

  37. Thanks Chris,

    Funny thing is: I knew that high ATP and NADH indicates energy sufficiency, and them inhibiting isocitrate dehydrogenase. But I didn't connect the dots. After I read your reply it makes perfect sense to me that if citrate starts flowing out of the mitochondria, this means that ATP and NADH are plentiful.
    So thank you for the aha moment.

    I didn't know about ROS inhibiting aconitase though.

    Which reminds me: in your reply to Paul you wrote that some people believe the benefit of ketogenic diets to be oxaloacetate depletion in the brain. You probably know that there is also the idea that high fat ketogenic diets increase UCP2 expression, which would cause a reduction of ROS in the brain.

    This UCP2 idea was mentioned in a review paper about uncoupling proteins and dietary fat [1]. It discusses how high NEFA lead to increased UCP2, and how this may do several things: lower the insulin response to glucose in pancreatic β-cells, prevent ROS damage in the brain, and reduce atherosclerosis risk by decreased ROS in endothelial cells. The review certainly is a nice resource for several interesting references on the matter of uncoupling proteins.


    [1] Fisler J.S., Warden C.H. Uncoupling proteins, dietary fat and the metabolic syndrome. Nutr Metab (Lond). 2006 Sep 12;3:38. http://pmid.us/16968550

  38. Hi John,

    You're welcome for the aha. Thanks for the reference, and your comments. I know about the UCP hypothesis but I haven't seen that paper. Others suggest that they help increase glucose uptake into the brain too. Lots of competing hypotheses.


  39. Just trying to get my head around the biochemistry here.

    So, once TCA is in motion, we only need Acetyl-CoA to keep it running, and oxaloacetate is 100% recycled. However, during oral carb restriction, oxaloacetate is instead sucked out of the TCA and used for gluconeogenesis in order to keep our blood sugar in normal ranges.

    Since oxaloacetate is delcared as the rate-limiting step for TCA, we could assume the other intermediate substrates are not supplied in any meaningful quantity from our diet to fuel TCA to any meaningful degree.

    Looking at the biochemistry pictures for ketogenesis, it seems all we need for ketones is lots of Acetyl-CoA, which we can get plenty from in fats.

    TCA can then be forgotten about at this point as we dont need it to produce energy?

    Also, doesnt it therefore make sense that if our goal is fat burning ( loosing weight ) we should be not consuming any potential TCA substrates so as to maximise our need for Acetyl-CoA?

  40. Hi Kindke,

    Oxaloacetate is recycled, but not 100% because there are other uses for it. During carb restriction, the demand for oxaloacetate increases in order to produce glucose from it, so its recycling should decrease further. But oxaloacetate always has some rate of leaving the cycle and thus always needs at least some replenishment.

    I single out oxaloacetate because it lies at the point of contact with acetyl CoA, which is where the energy from fats are coming in. You could indirectly replenish oxaloacetate with any of the other intermediates in the cycle, which can be derived from amino acids.

    So again, the point is that to fully break down fat for energy you need either some carbohydrate or some protein to go with it.

    As I said before, you cannot harvest most of the energy in a fat without the TCA cycle. If you do not use the TCA cycle, you harness only a little bit and the rest comes out in your breath and urine.

    The ordinary way of using ketones is that fatty acids are released from adipose tissue, then go to the liver where the liver breaks them down to ketones. Then they go back to other tissues such as muscle where they are burnt for energy in the TCA cycle.

    Theoretically, you could maximize weight loss by making sure you pee out your ketones instead of burning them for energy by restricting any supply of carbohydrate or TCA cycle intermediates, but this would likely lead to muscle wasting as your body isn't going to let you starve to death so easily, so it will get some oxaloacetate out of your muscle in order to help you use that fat for energy.

    Hope that helps,

  41. Lyle,

    Perhaps you could reformulate your comment in a more polite and intelligent manner so that the rest of us can benefit from it, if you'd be so kind.


  42. After exercise respiratory rate returns to normal after equalization oxygen debt-90 min. (ventilation is stimulated by increased concentration of hydrogen ions, lactate is converted into glycogen 80%, 20% CO2 and H2O-re-synthesis of ATP and phosphocreatine)

    Citric acid cycle and is controlled mainly by the concentration of carbon dioxide in the blood (increases) and the concentration of protons (inhibits).

    Laxity of the control of the Krebs cycle causes an abnormal postprandial production of ketone bodies, and acidification in diabetic metabolic syndrome diabetes

  43. Thanks Gino. That's basically what I meant by "ketone resistance" in previous comments.


  44. Kindke,

    In addition to what Chris said:

    In liver cells the oxaloacetate is seriously sucked out of their TCA cycle for gluconeogenesis. This causes acetyl-CoA to build up, eventually leading to ketogenesis.

    In muscle and brain cells, the ketones are taken up and converted back to acetyl-CoA, ready to provide energy via their TCA cycle. Muscle and brain cells will not be doing much gluconeogenesis, so their TCA cycle will likely have a lower oxaloacetate drain than liver cells.

    However, some TCA cycle intermediates are always drawn off to make vital substances. And muscle and brain cells are no exception to that. So there still is a deficit in their TCA cycles. And this deficit must be filled, because their TCA cycles have to accommodate their energy need. It is almost like the cell's energy need "sucks" TCA intermediates into the cycle, so it can run at the required rate.
    These "sucked in" intermediates have to come from somewhere: if the diet doesn't provide them, the body will.

    Not consuming potential TCA intermediates is unlikely to lower the energy need of muscle and brain cells. So their TCA cycles will still run at their usual rate. And because of the lack of dietary TCA intermediates the cycles will be replenished by breaking down bodily protein.


  45. Hi Chris,

    Interesting article. I stumbled on to your blog quite recently and I commend you for your great work.
    What I want to know is that if a person is following a VLC diet, say 30g of carbs a day, but is eating about 200 calories over maintenance where protein intake is just over 1g/lb of bodyweight and the rest of the calories are from fat, does he risk losing muscle mass? Assume that the guy weight trains in conjunction with this diet.
    I would think a lean mass loss would occur only if protein was insufficient, which generally isn't the case in people who weight train AND low carb.

  46. Hi Jarri,

    Welcome. I'm glad you like the blog, and thanks for the props.

    Risk? Probably, but that's more of a statistical concept. I can't say whether he *will* lose lean mass. As I indicated above, lean mass loss on a ketogenic diet isn't just about sufficient protein for muscle maintenance, it's about sufficient oxaloacetate for deriving energy from ketones. So, if you aren't eating much carb, you need extra protein to make up for it. 30g carb is probably not sufficient to supply physiological needs for glucose, let alone those needs *plus* TCA intermediates to keep energy utilization going. So the question is whether there is enough protein. And then the other question is what is the hormonal profile stimulated by the weight lifting. That could spare lean muscle mass. So I think you need to judge this on a case by case basis. If the person is losing muscle mass, I'd add another 30 g carb and see if it helps. If the person is in the prime of health, then the diet would appear to be working fine.


  47. That makes sense to me. Personally, I've been doing fine on a VLC diet in that I don't feel fatigued and keep progressing in my lifts. It's just that lean mass GAINS are hard to come by, but that could have more to do with genetics than the diet. Thanks for the reply.


  48. Jarri,

    It could have to do with genetics, or it could be the exact same principle operating. In other words, the muscle mass you might be "losing" is the muscle mass you aren't gaining. The only way to know is to experiment with your diet. If you're curious but satisfied, maybe it's not worth it. If you really want more lean mass gains, the best way to change your diet would be one thing at a time, incrementally, so that a) if you get negative results it's not that hard to reverse the changes and b) if you get positive results they're easy to interpret.


  49. re-reading this post (after spending some time at the WAPF site), it occurred to me to wonder -- has anyone specifically studied glutamate/aspartate damage of the hypothalamus, as a source of leptin resistance?


  50. @lylemcd:
    "You're not a mouse, retard."

    How do you know? Because of the issues discussed, many LIRKO mice are avid readers of this blog.

    And may I ask why you are so hostile toward mice? Are you by any chance a rat?

    M.M. Domesticus

  51. The first thing I thought of when I saw "95% fat diet", was Dr Atkins "fat fast".

    Of course, it was not meant to last very long, and only meant to be used in extreme cases to supposedly shock the system out of weight-loss stalls.

  52. G'day all,
    One factor is missing from the equation and that is water. Water acts on the ATP by acting like a hydroelectric pump on bound minerals and by hydrolysis increases the caloric factor of the food and opens up the messaging pathways of the nerves.

    Since the brain and muscles are mostly water it makes sense to keep hydrated as well.

    One good book on the subject is Obesity, Cancer and Depression, their common cause and cure by Dr Batmangelidj.


  53. "many people claim, anecdotally, that citrate abolishes the effect of low-carbing on fat-burning."

    hmmm... does this mean that if one were low-carbing and supplementing with Magnesium Citrate (recommended by some low-carbers as a very bioavailable source of Mg), it could be counter-productive to fat loss?

    In the context of low-carbers getting adequate glucose: If they were including fresh liver in the diet, would this contribute a significant amount of glucose (i.e. stored glycogen), if any?

    (I saw a video of a lecture by Loren Cordain recently, in which he said that, contrary to what some people think (and I thought), there is no glycogen left in muscle meat by the time we eat it).

  54. Hi Montmorency,

    Yes this is possible, but I have never seen any good clinical evidence that citrate in these doses will stop fat loss. I have just heard anecdotes.

    Liver provides a little bit of glycogen but it would be quite tiny and I don't think it would be very meaningful overall.


  55. Thanks Chris. Yes, doing some digging it's only about 3-4 g per 100g, so not very significant.

    Also found evidence that Cordain was right about muscle meat. The fact that rigor mortis takes place is a sign that the glycogen has been converted to lactic acid, which actually seems to help in the keeping and butchering properties of the meat.

    In passing, I was interested to note that (lamb's) liver contains about the same quantity of potassium per 100 g as 100g of raw banana. - Goodness, and actually more Vitamin C, and this is cooked liver. (Source USDA Nutrient DB).

  56. Hey Chris,
    I have heard that high-fat diet(limiting carbs to about 80g) will allow for ketosis and allow ketones to be the primary fuel for the body. However, in your article "We really Can Make Glucose From Fatty Acids After All!" you say that the ketones are converted into glucose. So are ketones really a source os fuel then, and if so are they more efficient? Or is, as Conventional Wisdom says, glucose the primarily and essential source of fuel for the human body?

    1. Hi Pranay,

      Their quantitative contribution to gluconeogenesis isn't very well characterized, but it is probably limited to roughly that of an individual glucogenic amino acid. So it is doubtful that the findings in my other blog post invalidate what I wrote here, but rather that they leave the principle the same and attenuate the magnitude of its importance. In other words, maybe the fact that fatty acids can be converted to glucose reduces the amount of protein and/or carb you need by 10% or so, but doesn't eliminate the need for one or the other. That's a rough estimation at this point.


  57. Thanks, but I meant to ask was whether ketone bodies are directly used as fuel or do they first have to converted into glucose in order to be used as fuel.

    1. Hi Pranay,

      Sorry for the confusion. Ketones can be used directly.


  58. Hello Chris. Interesting, but I do not fully agree with "Better get a steak or potato to go with that butter", unless you mean it as a controlled dose!
    The ketogenic group was of course kept on too low protein which resulted in the muscle waisting. The mechanism to shift from (preferred) autophagy recycling to burning useful proteins is probably blood sugar level; simply because the whole purpose of the exercise is just "blood sugar production". The study that was done probably didn't look at blood sugar, or blood ketones, but I am sure you can find other (newer?) that do. Our bodies probably use a rather low set point for this desperation level, as it is very much like burning the furniture to keep warm.
    Once autophagy is put into the picture with ketogenic diets it makes the whole ketogenic scenery more interesting again, referring to a previous posting yuu made about this that I cannot find right now.
    You also say that these mice were young which mean they had less -maybe hardly any - old waste to be recycled before the muscle proteins had to be consumed. Using older scruffy mice as well would have been interesting!

  59. Hi,
    A bit late, but came across this while googling for oxaloacetate. The discussion here seems relevant but the biochemistry is too detailed for me - is there some useful information here if we want to minimise ketone production? Is it as simple as eat more glucose/starch, or are there other things we can do as well/instead? Ideally I'd like to burn fat some of the time but produce fewer ketones.

  60. People who plan on increasing their muscle mass should be acquainted on the proper methods to prevent any health problem It is best to take supplements which can help in supporting your body and aid in building more muscle mass. JamiGoode

  61. I am highly impressed on this blog. To set yourself up for success, think about planning a healthy diet as a number of small, manageable steps rather than one big drastic change. If you approach the changes gradually and with commitment, you will have a healthy diet sooner than you think. Thank you.


To create a better user experience for everyone, comments are now moderated. Please allow up to one business day for your comment to post. In order to avoid the appearance of spam, please avoid posting links, especially to commercial destinations, and using all-caps.