Insulin, dietary fat, and calories: context matters!

Jane Plain recently wrote a great article about the relationship between insulin, dietary fat, and calories.  There are a lot of data on this topic, which collectively suggest: context matters! 

For example,

Insulin and ketone responses to ingestion of MCTs and LCTs in man. (Pi-Sunyer et al., 1969)

14 healthy subjects, overnight fasted; dose: 1g/kg.

In brief, MCTs are more insulinogenic than corn oil.  But it’s not a lot of insulin.  Really.  Enough to inhibit lipolysis, perhaps, but that’s not saying much… & certainly not enough to induce hypoglycemia.

Pi-Sunyer MCT Corn oil

Stimulation of insulin secretion by medium chain triglycerides in patients with cirrhosis (McCullough et al., 1971)

100 grams glucose vs. 28.5 grams (30 mL) MCTs.

In this experiment, they gave about half as much MCTs as Pi-Sunyer and saw no insulin response in healthy subjects but a doubling of insulin in cirrhotics; still, however, this wasn’t enough to budge blood glucose.  Note that after 100 grams of glucose, insulin increased 6-7x.  That’ll leave a mark.McCullough

N.B. these aren’t direct comparisons; I know this.  Context still matters.

The coingestion of fat on the glucose, insulin, and GIP responses to carbohydrate and protein. (Collier and O’Dea, 1983)

Simple enough: healthy young subjects fed a boiled potato or veal, with or without butter.  Carbs stimulate insulin the most; this is blunted by dietary fat.  Protein stimulates less insulin; this, too, is blunted by dietary fat.  The insulin AUCs were pretty much unaffected by dietary fat.


Also note that 50 grams of carbs increase insulin ~5x (significantly more than corn oil and MCTs).  For other patient populations or different foods, check HERE & HERE.

One of the studies cited by Jane: Usual dietary fat intake and insulin concentrations in healthy women twins. (Mayer et al., 1993)

Mayer I

Mayer II

Most of the correlation was attributable to body weight, but even after controlling for a variety of potential confounders, there was still a relationship between dietary fat and fasting insulin (but not fed insulin).

Fat and carbohydrate overfeeding in human: different effects on energy storage (Horton et al., 1995)

When in positive energy balance, excess dietary fat is insulinogenic in obese people, but carbohydrate overfeeding is still way more insulinogenic (in everyone).





A ketogenic diet favorably affects serum biomarkers for CVD in normal weight men (Sharman et al., 2002)

Sharman diet

Much higher fat intake (157 vs. 50 grams), but much lower insulin levels.  Context matters!  #EnergyBalance

Sharman data


Keto-warrior dieters, 1-meal-per-day, etc. (unless you’re losing weight): insulin.  Even if it’s 100% fat, there will be some insulin secreted (Pi-Sunyer).  Probably enough to inhibit lipolysis, but not much else.  If you’re lean or losing weight, and eating a few LC meals throughout day, insulin levels will remain low (Horton).

Fat blunts the insulin response to meals; even if the additional calories aren’t accounted for.  In the Collier study, 50 grams of carbs (200 kcal) induced more insulin than 50 grams carbs + 50 grams fat (650 kcal).

If energy surplus and overweight (or post-obese, as per Jane’s analysis [I think]), then it’s also likely that higher fat –> higher fasting insulin (Mayer & Horton).

If you’re in relative energy balance, and not ingesting 2000 Calories all at once, then the old rules apply: low carb = low insulin (Sharman).

LCHF works well in obesity primarily because of the spontaneous reduction in food intake; insulin levels and body fat decline.

Context matters!


If you like what I do and want to support it, check out my Patreon campaign!

UPDATED Affiliate links: still looking for a pair of hot blue blockers? Carbonshade  is offering 15% off with the coupon code LAGAKOS and Spectra479 is offering 15% off HERE.

If you have no idea what I’m talking about, read this then this.

20% off some delish stocks and broths from Kettle and Fire HERE.

If you want the benefits of  ‘shrooms but don’t like eating them, Real Mushrooms makes great extracts. 10% off with coupon code LAGAKOS. I recommend Lion’s Mane for the brain and Reishi for everything else.

Join with this link. Get free magical internet money!

calories proper



Be Sociable, Share!
  • charles grashow

    The longer and higher the insulin levels, the higher the insulin
    resistance. The higher the resistance, the higher the insulin. This is what sets into motion the time dependent effects of obesity. The fat get fatter. The longer you have obesity, the harder it is to eradicate. Insulin Causes Obesity.

    Everybody knows about these time dependent effects. However, most current thinking about obesity completely ignores these effects even though they are plainly obvious to anybody and everybody. Since type 2 Diabetes is all about insulin resistance – this also leads us to the inescapable conclusion that Insulin Causes Diabetes.

    Insulin causes both obesity and diabetes. This is the new science of Diabesity. With this new understanding, we are led to entirely new possibilities for the cure of diabesity. If high insulin levels are the cause, then the cure is to lower insulin levels. Cure type 2 diabetes? Can it be true? Yes, but we still have some work before we get there.

    To be continued…

    • Wenchypoo

      Then you have people like Dr. Neal Barnard (just today) shreiking to high heaven about how vegetarians and vegans have lower blood pressure, but he never mentions blood sugar! Sure, these people will be thinner, but at an expense of muscle mass. Sure, they’ll have lower blood pressure, but at an expense of electrolyte imbalance.

      I wonder if anyone’s ever done a study of vegetarians, vegans, blood sugar, and the incidence of diabetes and/or Alzheimer’s in later years…

      • Charles Grashow

        Actually there are studies that show a link between higher cholesterol in mid-life and Alzheimer in later life.

        From this study, the researchers saw that those individuals with total cholesterol levels between 221 to 248 mg/dL were more than 1.25 times as likely to develop Alzheimer’s disease later on in life than those with normal cholesterol levels.

        When total cholesterol levels ranged between
        249 to 500 mg/dL, these individuals were 1.5 times as likely to develop Alzheimer’s disease.

        • Wenchypoo

          As former Paleo, and now keto eaters, we eat tons of meat and sat fat, but our cholesterol #’s are fine–well below 200. So I’m not seeing the connection between cholesterol and Alzheimer’s, and where a vegetarian diet would serve to improve things. Those plaques are made of sugar, just like the plaques in your arteries.

          Have you seen the study where some Alzheimer’s patients were given a dose of nasal insulin, and they experienced cognitive improvement starting that day?


          We’re in the middle of an experiment to try to reduce LDL by changing the chain length of our fats taken in–more medium chain and less long chain.

      • +1

      • The vegetarian diet is horrible for mental health. Abstaining from animal food is basically asking for neurological degeneration.

    • Barbara Corkey gave a very compelling presentation about this at EB in 2011. Her focus was not on carbohydrates, but rather on environmental agents which cause insulin hypersecretion (which implicates carbohydrates as well, albeit indirectly).

      Banting lecture 2011: hyperinsulinemia: cause or consequence?

  • johnnyv

    I only eat two meals a day, high fat ad lib and I maintain or improve in composition for years now.
    Dropping breakfast and moving to a smaller eating window was a fantastic move for my wife and myself.
    I guess the fasting insulin picture maybe significantly different when you throw in RT and HIIT.
    RT and HIIT did little under high carb, moderate fat but there was visceral fat at the time which confounds things.

    • Thanks!

      High intensity exercise brings it to a whole new level; Jane Plain calls it a “glucose sink,” and I agree… also, the accompanying sympathetic nervous system activation should further suppress insulin.

      • Wenchypoo

        I have to concur, seeing as how exercise eliminates my hot flashes for about 3 days. They come back again unless I exercise again.

        At first, it seems hard to do, because you’re thinking “any sort of effort will set off another flash, and I’ll be trapped under weights AND sweat clothing AND a roomful of other people generating heat, but if you do it outside (as in vigorous and lengthy lawn-mowing marathons), it isn’t so bad. Snow shoveling also works (minus the coat where you can get away with it).

  • I very much enjoyed this blog entry, Bill.

    However, few point of contention!

    1)Re: MCTs and insulin, again context matters as you would point out hehe. Perhaps the insulin generated from MCT is insufficient for real hypoglycemia in a non wt reduced person without vulnerability to hypoglycemia, but what about someone very wt reduced with hypotrophic adipose who depends on constant unopposed lipolysis for satiation and energy? I can say in my case, MCTs do very much stimulate powerful appetite and symptomatic hypoglycemia. Although, I notice if I am deeper in ketosis (thus, perhaps, lower insulin and less vulnerability to lipolytic suppression) the MCTs no longer provoke hunger/fatigue/shakiness.

    I have noticed coconut products, in spite of glucose instability, seem to very much help with wt loss so I suspect the increase in thyroid hormone, metabolism, and reducing #s of gut microbes likely much more helpful than blip in energy metabolism.

    2) Re: adding fat lowers insulin AUC
    This might be definitely true for the immediate postprandal period but I highly suspect adding fat energy to pre-existing food actually raises 24h total insulin output or requirement (depending if IDDM or not). SUBSTITUTING carb for fat and leaving net energy yield to body the same, OTOH, likely would lower 24h insulin requirement, but if energy intake is increased above basal value particularly of fat (which tends to impair glucose oxidation and thus directly drive greater insulin need) we can expect over the longer term, post 3 hour mark, greater insulinemia. This is one of the reasons hard core keto dieters struggle with hyperglycemia.

    3) Regarding your summary:

    “When in positive energy balance, excess dietary fat is insulinogenic in obese people, but carbohydrate overfeeding is still way more insulinogenic (in everyone).”

    100% agreement 🙂 This is pretty much my thoughts, my blog entry mostly address the myth of “fat burning beasts” who can eat 10000 butter calories and never put on an ounce of fat. Not at all true, woo wishes it was! If I eat lots of butter, I need to do exercises 🙂

    I disagree the keto warrior diet of 1 meal a day is insulinogenic; while the meal itself is likely very insulinogenic, I im agine keto-warrior diet is far less insulinogenic than the same foods eaten over 3 or more meals/snacks. This is the advantage of fasting and why fasting dieters lose more wt – they make less insulin over a 24 hr period.

    “LCHF works well in obesity primarily because of the spontaneous reduction in food intake; insulin levels and body fat decline.”

    Yep! Cyclical relationship between lower insulin, lower food intake, leading to even lower insulin. Body fat is lost during this process until the “set point” of leptin signalling relative to CNS sensitivity is reached. Then plateau.

    My point in my article is not to argue that fat is bad, low fat is
    better, as I strongly disagree with this statement and usually it is
    much better to eat a high fat diet and SUBSTITUTE carb for fat. However,
    the dear keto/low carb myth that kerrygold is the body of christ and
    can do no harm is foolish and unscientific. All caloric energy increases
    insulin, dietary fat as well.

    “there was still a relationship
    between dietary fat and fasting insulin (but not fed insulin).” <–
    yep, fasting insulin is *most important* and specific for obesity.

    • Wenchypoo

      Kerrygold does ME harm–I’m allergic to dairy (among other things). I was using lard to cook with (coconut oil foams, so I can’t see what I’m frying), and now have switched to MCT oil. I use coconut oil in stuff that doesn’t need heat (like smoothies and puddings).

      Chocolate pudding with avocado in it is the only way I can get Hubby to eat avocados willingly, so he has it rather frequently. He also has trouble with straight palm oil, so I use MCT–a blend of palm and coconut oils. We also use macadamia nut oil for both hot and cold applications.

    • Thanks!

      1) agree. Weight-reduced, post-gastric bypass, etc., probably more sensitive to hypoglycemic effect of 2x insulin. Ketosis *should* reduce symptoms.

      3) yeah, I was referring to the bus load of calories; transient energy surplus = insulin.

      Different story for fasting insulin:
      Funny thing, the participants in this study were supposed to maintain their body weight. They couldn’t. They even reported being hungrier but STILL couldn’t maintain their body weight.

    • Jack Kruse

      MCTs lower the redox potential in the inner mitochondrial more than saturated fats so this blunts the ROS and RNS. Hence the reason why MCTs are very different than other fats.

      • Jack Kruse
        • Jack Kruse

          About the Sinclair paper…..the researchers have also found the antidote or Rx in mice for this “old mitochondria” problem…….That is a homogenized solution of C60 carbon buckyballs and olive oil. This increase their life spans by an average of 90%. Long term data is still pending on this affect past the original studies. I have the sense that the buckyballs alters the interface of the water hydration state around mitochondria to make it able to handle poor proton outflow from failing mitochondria via some of the things that Pollack has found about these structures interaction with water.

          • Jack Kruse

            C60 might also directly lower superoxide production by destroying the redox potential in mitochondria.

          • Jack Kruse

            remember that hydronium ion speed in water is exceptionally fast when the electrical field is intact (see Piccardi 1962) and if the water is not structured this drops and the protons have to go somewhere. The high mobility of protons is directly tied to the rapid transfer of them to the chains of hydrogen bonded water molecules in the reverse micelle around mitochondria to begin with. That “somewhere” become very destructive to the cell and surrounding proteins if the water is not there to handle the load of the mitochondrial production.

          • Jack Kruse

            proton migration is where Peter (Hyperlipid) is headed. Bill read the work or Aiello 1973. He describes jump conduction and tunneling of protons. This is where the shuttle Peter has spoken about G3Pdh works in relation to water and the mitochondrial membrane. It turns carbs into saturated fat. Proton tunneling is a solid state activity not one a biologist normally sees and requires a SQUID. That is what the mitochondrial membranes are essentially. The phenomenon is even more powerful when it is coupled to a membrane, which in this case it is because it links the mitochondria to a 3 D liquid crystalline water structure in the reverse micelle around the mitochondria

          • Jack Kruse

            60 undergoes six reversible, one-electron reductions to C6?
            60, whereas oxidation is irreversible. The first reduction requires is ~1.0 V (Fc/Fc+
            ), indicating that C60 is a moderately effective electron acceptor.C
            60 has a tendency of avoiding having double bonds within the pentagonal rings which makes electron delocalization poor, and results in the fact that C
            60 is not “superaromatic”. C60 behaves very much like an “electron deficient” alkene and readily reacts with electron rich species like ROS and RNS. C60 can be hydrogenated and homogenized suggesting that a modified buckminsterfullerene called organometallic buckyballs (OBBs) could become a vehicle for “high density, room temperature, ambient pressure storage of hydrogen”. These OBBs are created by binding atoms of a transition metal (TM) to C60 or C48B12 and then binding many hydrogen atoms to this TM atom, dispersing them evenly throughout the inside of the organometallic buckyball. In this way the buckyball becomes a proton sump…….and that is likely how it extends life or allows one to stay in senescence longer than would be expected.

          • Jack Kruse

            Woo’s belief’s are tied to her not understanding the solid state surroundings of the reverse water micelle. This is also why she struggles with CT. It is also why we disagree on leptin. FFA all liberate protons……..if the water is there KerryGold can be “the Holy eucharist” for some who understand how it works. Most do not because they do not understand solid state physics. See Albert Szent Gyorgyi 1941 statements on biologic semiconduction. He discovered Vitamin C and knew from the mechanism that semiconduction had to be involved. His 1968 book Bioelectronics is a masterpiece. Sadly no one in biology knows what this Nobel Prize winner found……short of Robert O Becker and me. Hyperlipid however is the most close biologist I have seen to getting it.

          • Melissa

            Damnnnn, Wooey got pwned by da Kruser!

          • Jack, could you please send me a link to Aiello 1973 and Pollack? I’ll definitely read them.

          • Jack Kruse

            Pollack is the 4th phase of water. New book. Excellent. Aiello et al “Behavior of interacting protons: The average mass approach to its study and its poosible biologic relevance. In eds H Hakem & M Wagner, Cooperative Phenomena, Springer- Verlag. NY, pp 395-403

      • Jack Kruse

        “The specific feature attributable to the generation of ROS by mitochondria is related to the fact that the higher is the membrane potential (the larger is the difference in the concentration of protons inside and outside the mitochondria), the higher is the level of the superoxide anion production. As it was shown [29], there is steep dependence of mitochondrial superoxide-anion-radical generation on transmembrane potential (?). Even a small (10–15%) decline of ? resulted in tenfold lowering of ROS production rate.

        Therefore, the so-called mild uncouplers of oxidative phosphorylation are the substances which can move some of the protons inside the mitochondria and can possess an excellent oxygen-protective effect, although they are not antioxidants in terms of chemistry.” BOOM from here:

    • julia31

      “Get All The Support And Guidance You Need To Be A Success At Managing Your Weight Loss Goals!”

  • timothy

    context is meaningless, in the end calories are all that matters. maybe carb rich foods are more easily overeaten but if you look at jimmy moore et al they got b obese eating meat drenched in butter. you can’t argue with thermodynamics in the end it’s all about expending more than calories consumed

    • johnnyv

      Poor trolling attempt, will give 3/10 for Jimmy Moore reference.
      A constant low level insulin drip will change the laws of thermodynamics in your simple little world.
      How much calories does insulin have then?

      • “How much calories does insulin have then?”
        Assuming it’s fully digested, 4 kcal/g. JK!

        “A constant low level insulin drip will change the laws of thermodynamics in your simple little world.”
        reminds me of this:

        • johnnyv

          Yes, as you well know substrate utilization is greatly influenced by the hormonal milieu. And the hormonal milieu is influenced by the substrate and various other factors.

          How much negative calories does 2,4-dinitrophenol have per gram?
          I know, it increases mitochondrial willpower!

    • donny

      Jimmy Moore got in trouble eating various sugar alcohols, “low carb” breads, and pastas. His low carb diet wasn’t as low carb as he thought it was.
      Wait, did you say context is meaningless? Did you really say that?

  • I think that if you add a bit of high-intensity training to all of the context, insulin levels will be lowered. check Peter Attia’s experiment in the metabolic chamber when his insulin levels were higher during exercise and immediately went lower after the exercise as his body’s ketone levels increased.

    • sure. Anything that stimulates the sympNS should dampen insulin secretion.

      • Jack Kruse

        funny you should say that Bill…….cold works that exact way. eNOS directly inactivates the function of hypothalamic NPY! eNOS in cold is stimulated by beta adrenergic signaling. NPY is stimulated by carbohydrates in the hypothalamus and it drives carbohydrates cravings and food seeking behaviors. Why would mammals and humans have this hardwired into their DNA and in their brains? Evolution says it is biologically impossible to find carbohydrates in chronic freezing cold conditions because these foods require serious amounts of photons and electrons from sunlight.

        • Jack Kruse


        • Jack Kruse

          Bill you might enjoy this new book, entitled, COLD. there are a couple of eye opening comments in it.

          Look at this excerpt from a speech given by Prince Charles on Fiennes’ expedition in 1979. Praising the ambition and courage of the participants, he says:

          … Above all the human risks are still the same. They’re all there today, the frostbite, the loss of body fat because of the cold and the protracted bouts of shivering, especially at night …The cold findings are even in a public speech from the 1970s. So much for the paleo beliefs of “hormetic effects” attributed to cold. On page 100 he mentions the lowest survival body temperature ever recorded. It was “Swedish skier Anna Bågenholm who, in 1999, was trapped in water, and under the ice, for eighty minutes. Her body temperature dropped to 13.7 ?”. So many people believe the cold will kill you because they just do not understand physics. Yet in ATLS and PALS they always teach us that anyone who is in cold submersion should have extra long BLS ACLS recoveries because there is voluminous data on people “coming back” on long resuscitations.

          • Thanks, Jack.

            I just finished a chapter about cold therapy in “The Hibernation Response.” Not very technical, but pretty interesting. Devra Davis’ book is next, then I’ll look into COLD.

  • persistentone

    To some degree you are cherry picking studies. Two can play that game. 🙂

    Let’s ignore the MCT Oil studies, because many of those involved patients with horrible liver defects, and MCT is just a very complex thing.

    The “simple” case is ingesting saturated fats together with carbohydrates. Your selected study is “The coingestion of fat on the glucose, insulin, and GIP responses to carbohydrate and protein” (Collier and O’Dea, 1983)

    Okay, but now allow me to cherry pick a study that followed a very similar methodology and came to a far different conclusion. Mine is “Effects of meals rich in either monounsaturated or saturated fat on lipid concentrations and on insulin secretion and action in subjects with high fasting triglyceride concentrations”

    In my study the key graph is:

    White dots are carbs alone. Grey dots are monounsaturated plus carbs. Black dots are saturated fats (butter) plus carbs.

    See graph D. Insulin explodes with the saturated fat meal, more than doubling relative to the carb meal alone. That’s classic insulin resistance.

    Here’s the problem with your post: your are taking isolated studies and trying to then generalize their claims to larger populations. The problem with that is that you need to consider the metabolic defects of the populations being studied. Your study looks at healthy 21 year olds with no lipid or glucose metabolism issues. You could drop a hand grenade down those throats, and they would probably burb once and then go back to normal. 🙂 My study looks at people whose only known defect is high triglycerides, but probably you would suspect something in their glucose metabolism might also be impaired/involved in that defect.

    Some of us are prediabetic. Some of us are diabetic. Some of us have lipid disorders. People are different and metabolize fats in very different ways. It’s dangerous to take a study that uses 21 year olds in perfect health who can metabolically compensate for almost anything and generalize that to populations that include people with various kinds of metabolic disorders.

    In a perfect world, we would be able to measure our insulin and glucose together after specific meals and then make adjustments based on how we as individuals metabolize specific types of food. Since we have no way to measure insulin, we are left to guess. And generalizing from these types of studies is very dangerous and tricky.

    • Thanks for the link. And I agree; context matters!

      For example, while Horton’s study design was different from Muriana’s & O’Dea’s, she found “fat overfeeding” to be very insulinogenic in obese but not lean people.

      • persistentone

        I’m thin not obese, but definitely saturated fats cause me to get higher glucose numbers. MCT Oil definitely brings down my glucose rapidly. I’m prediabetic, so it is just another case of broken metabolism. But these things are non trivial, and I feel a lot of frustration that we have no way to measure insulin resistance directly.

  • Pingback: Lipogenesis and Ketosis – No Conversion of Carbohydrates to Fats in Ketogenic Diets | The Personal Blog of Cristi Vlad()

  • Thomas Hemming Larsen

    Hi Bill, sorry to bother you with another question. Is the reason why some people say that you can eat as much fat as you want, while keeping carbs very low, because insulin will be very low so fat storage is impossible?

    • I think the reason they say that is because they don’t understand Taubes et al. at a very basic level.

      Fat will be stored, even with very low insulin.

      • Thomas Hemming Larsen

        Thanks Bill. That was the post I needed to read 🙂 FFA turnover is basically the same even if insulin is increased tenfold. Even if someone on a very high fat diet could keep insulin low, it would be difficult to keep it low enough to prevent fat storage. And I still can’t get my head around how eating excessive amounts of fat “can’t” be stored…

        • it all comes down to satiety; CICO still ‘works’ 🙂

          • Thomas Hemming Larsen

            Yep. That’s also why I’m more and more starting to believe that your non-protein calories can be distributed in a range of ways. What’s optimal will depend on the context and individual.
            As I’ve mentioned previously, macros of the most healthy populations vary greatly.