Two things I’ve changed my mind about, part I: Butter

I like to believe I can update my beliefs when presented with new evidence or a better interpretation of the same body of evidence, rather than clinging to my old beliefs.  In the next two posts, I’ll discuss two examples where that has happened.  Years ago, I argued that butter may make positive contributions to health.  My thinking has gradually evolved since then, and I now think less refined forms of dairy are preferable to butter.  Here’s why.

The whole food heuristic

Heuristics are simple rules that get us most of the way to the truth.  One of the most valuable heuristics in nutrition is also one of the simplest: eat minimally refined foods.  This single-handedly eliminates nearly all of the calorie-dense, nutrient-poor ingredients and foods that contribute to obesity and noncommunicable disease.  The whole food heuristic is not only supported by a large and growing body of evidence, it’s also consistent with evolutionary considerations and common sense.

Food refining is a continuum.  For example, when considering sugar, whole fresh fruit is a minimally refined source, while white sugar is a maximally refined source.  Fruit leather and fruit juice are somewhere in between.  Where is butter on that continuum?  It’s not as refined as liquid seed oils, which are typically solvent extracted, degummed, bleached, and deodorized.  It doesn’t require industrial technology to produce.  But it’s certainly much more refined than milk, the whole food it’s derived from.  To illustrate this, consider that typical milk only contains 3-4 percent fat by weight when it comes out of a cow.  Making butter is a process of concentrating a small fraction of milk by 30-fold.  And as a result, a wide variety of nutrients are separated out and calorie density increases greatly.

Dairy fat may still have nutritional value, but why not eat it in a less refined form that is more nutritious and less calorie-dense?

Blood lipids

Circulating lipoproteins influence cardiovascular disease risk, and LDL is a particularly important one.  Things that increase the number of LDL particles, and/or their total cholesterol cargo (called LDL cholesterol) tend to increase cardiovascular disease risk.  The evidence supporting this is now extremely strong (1, 2).

As it turns out, butter increases LDL cholesterol and particle number more than most other fats (3, 4, 5, 6).  While butter also increases “good” HDL cholesterol, recent drug trials have questioned the causal relevance of increasing HDL cholesterol, meaning that it may not actually protect you from an increase in LDL cholesterol (7).

Recent evidence suggests that the effect of dairy fat on LDL cholesterol depends heavily on the context in which the fat is delivered.  Less refined forms of dairy such as yogurt, cheese, and cream have little effect on LDL cholesterol, even when they deliver the same amount of dairy fat as butter (8, 9, 10, 11)!  Some research suggests that this is due to a protective substance in less refined dairy foods called the milk fat globule membrane, which is lost during the butter-making process (12).

Calorie intake

Face it: the main reason we eat butter is that it tastes amazing.  And this is no surprise.  Butter is one of the most calorie-dense foods in the world, and the human brain likes calories (for an explanation of why, see my book The Hungry Brain).  Most of us eat too many calories, and butter isn’t helping.

Fat isn’t automatically fattening, but research suggests that it can be fattening when it meets two criteria: high calorie density and high palatability.  Butter gets near-maximal scores on both.  Why not get the same nutrients and more from yogurt, which is less calorie-dense, less extreme in palatability, more sating per calorie, and therefore more likely to favor a healthy calorie intake?


The whole food heuristic is probably the simplest and most effective diet rule we have, and it suggests that whole sources of dairy should be healthier than butter.  Research in the areas of cardiovascular disease and appetite control are increasingly supporting this prediction.

That said, butter isn’t toxic, and if you want to use a little to saute onions or mix into steamed vegetables, I doubt it will have much negative impact on an otherwise healthy diet.  But extra-virgin olive oil is probably better for cardiovascular health, and if you choose to eat dairy, it may be preferable to focus on less refined forms of it, particularly yogurt.

78 Responses to Two things I’ve changed my mind about, part I: Butter

  1. Looking forward to the next post. I haven’t read you in years. I just happened to stop following nutrition blogs several years ago. But I randomly checked out your blog today. And it’s great to see you or really *anyone* update their opinion in public. That takes courage. Cool.

  2. Hey Stephan,

    I like the phrase whole food heuristic. Did you just coin that?

    Would you consider olive oil to be more of a whole food than butter?

    • Hi Todd,

      Extra-virgin olive oil is on the same level as butter in terms of its degree of refining. It can also probably contribute to overeating calories and undereating essential nutrients. But it doesn’t have the same LDL-raising effect as butter and its overall effects on cardiovascular health are probably beneficial.

    • One could say virgin olive oil is closer to milk than it is to butter. That is, it is not just an extracted lipid, it comes bundled with phytonutrients.

      • But so does butter, if it’s made from grassfed milk. Nutrients such as vitamin A, vitamin K, CLA

  3. Kudos for openness to change and favoring heuristics over adherence to dogmatic theories. Both help minimize cognitive dissonance.

    Butter spiked my ldl more than anything. Another difference between butter and EV olive oil that I haven’t seen investigated is that butter is heat-pasteurized, whereas EVOO isn’t significantly heated. Might be irrelevant, but maybe not.

  4. At 75 years of age and low carb for most of the time, it is heartening to hear a respected voice not tied to nutritional dogma; low carb or otherwise. I’m glad to see that Mr. Guyenet has the savvy to recognize and state that what appears to be solid and accepted information, may not be at all. People of my age need the nutritional truth as our lives and health depend on it.

  5. I absolutely agree.
    I think that those who defend the consumption of butter do so with weak arguments, trying to hide the fact that they eat it because it’s one of the most palatable foods of all.

    I also agree that in the optimization of food consumption, the best choices are whole foods. And in the specific case if fats, whole vegetable foods (avocado, walnuts, almonds, seeds, and olives) and oily fish among the best sources of fat. And I prefer them over unrefined olive oil.

    Anyway, nice article: short and to the chase.

    Greetings from Argentina.
    Ramiro Ferrando.

    • Hi Ramiro,

      Thanks, and I agree with your comment of choosing whole foods over olive oil. I offered olive oil as an alternative because I know many people will want to use some kind of added fat, but I think whole foods probably tend to be better than any added fat.

      • Hi Stephan,
        Thanks for the answer.
        Just to clarify. My comment wasn’t destined to criticize your words.
        On the contrary, I know where you stand regarding the comparison of whole foods vs. processed foods (even olive oil).

        I’ve always thought that in the end, the main difference between any kind of highly refined oil and olive oil in terms, is the content of bioactive compounds (and perhaps the lack of some negative byproducts of the refinement process). Because, in terms of macronutrients, they are a concentrated source of fat. But since olive oil retains those compounds, its effects have some, let’s say, “pros”.
        Actually, my main point when people compares olive oil with a highly refined oil is to give the example of refined white sugar and some less processed cane sugar. In the end they are almost the same, but the less processed may preserve some micronutrients. But those micronutrients, at least for me, even when they make them winner vs. its very refined counterparts, are not enough to put them in some kind of nutrition pedestal.

        Anyway, thanks again for the answer.
        And by the way, I’m a big fan of “The Hungry Brain” and all of your work.
        Big hug.

  6. Having reviewed a few of your old posts on the fats and heart disease, and looked into both the epidemiology (French Paradox) and the mechanisms, I think it’s highly unlikely that butter contributes to CVD.

    Higher LDL per se isn’t problematic in the context of low oxidized LDL, and while olive oil, due to its higher oleic acid content is better at resolving oxLDL, butter wouldn’t contribute to it. The fatty acid profile of butter contributes to higher LDL (temporarily), but would be less likely to contribute to the pathological process.

    As someone who occasionally eats buter straight, there is no more satiating food, IME.

    As always, your posts are thought provoking, so thanks!

  7. I’m not super hip on the science of cholesterol, so take my comment with a grain of salt (or two). Hope to hear from Chris Masterjohn on this.

    People with high levels of circulating LDL tend to increase risk of CVD. This doesn’t mean that high levels of LDL increases your risk of CVD. It could be that those high levels of LDL are being oxidized within the framework of high levels of inflammation and PUFA being incorporated into the cell membranes of those LDL particles. Also, of course particle size needs to be taken into account.

    I’m sure you know this so I’m surprised to see your use of it is the argument against butter. You could easily apply the same argument to grains, e.g. epidemiological studies showing whole grain eaters have lower CVD risk.

    That being said, I agree with the basic premise. Especially in today’s conditions. That being said, for people who control for their calroies it isn’t a very convincing argument.

    Looking forward to part two.

    • Hi Carl,

      LDL oxidation seems to be important, and so does inflammation. Cardiovascular disease is driven by multiple mechanisms. But that doesn’t change the fact that all else being equal, higher LDL-c and/or particle count increases cardiovascular risk. The reason we know is that changing LDL levels by a variety of different means (either genetic or pharmacological) has fairly predictable effects on cardiovascular risk. That doesn’t mean changes in LDL always cause the expected change in CVD risk, regardless of how it is achieved, but it does suggest that we should be pretty suspicious of things that raise LDL.

      Inflammation in the arteries is at least partially a downstream consequence of high LDL and the vascular damage it causes. Also, oxidized LDL seems to increase with total LDL level.

      • So what if they have the arrow of causation is backwards? What if damaged arteries leak cytokines that cause a slight increase in LDL that would account for the weak correlation of LDL with heart disease?

        What if there are several medicines that lower LDL yet have no effect on mortality-heart deaths etc ( they exist ).

        I’ve not found a study that shows that LDL is related to CAD if oxLDL is held constant – the data exists to figure this out. One thing that causes oxLDL to increase is Linoleic-acid.. most of what is found in the PUFA the cardiologists were recommending until lately.

        What if – monocytes are recruited to damaged arteries – becoming macrophages – that end up as foam-cells because the artery is already damaged?

        What if the LDL mantra has drained off research funding that might have found the initial cause of CAD – (apparently an initial thickening of the inner intima – one narrative is due to insulin levels).

        A cut or other injury causes inflammation, increase in local LDL (apparently to fuel the repair process.) What if your cardiologist has the normal repair process confused with causation?

        What if the only drug that appears to slightly reduce CAD events ( NNT 81 for 5 years is not great ) only lowers LDL as a side effect?

        • Hi Xtronics,

          It’s easy to make speculative “what if” statements but that doesn’t mean they are well supported by evidence. If you look at the Mendelian randomization studies on LDL (one of which I cited in my post), the results are very difficult to explain unless you believe that LDL causes cardiovascular disease. The old arguments against the lipid hypothesis just don’t stand up to current evidence and need to be discarded.

  8. There is an important distinction to make: butter raises the big LDL particles (alpha type), not the small dense LDL (beta type). Only the latter correlate with atherosclerosis (total LDL only correlates if it’s mostly beta-LDL). Small dense LDL is driven mainly by sugar/fructose consumption (via the liver). This is why in western societies, it looks like “LDL” is contributing to heart disease. Nope.

    Already history shows you that butter can’t be detrimental at all, because it has been used by perfectly healthy traditional societies for millennia, to make milk storable. Strictly speaking, it’s not even a “refined” product.

    • Hi Cliff,

      I’m not a cardiologist but my understanding is that all types of LDL contribute to cardiovascular disease and that the best LDL-related predictor of risk is the particle count (small LDLs carry less cholesterol, implying that for the same LDL-cholesterol level, if your particles are small your particle number will be higher and thus your risk higher). For example, people with familial hypercholesterolemia tend to have big LDL particles and they develop severe cardiovascular disease.

  9. Hi Stephan,

    Thanks for the interesting article. A question. You write:

    “Things that increase the number of LDL particles, and/or their total cholesterol cargo (called LDL cholesterol) tend to increase cardiovascular disease risk. The evidence supporting this is now extremely strong (1, 2).”

    I’d be interested to hear your view on the increasingly common claims that it is not so much overall LDL particle numbers that matter, but rather the number of small, dense LDL particles (which allegedly increase risk) and “large, fluffy” LDL particles (which allegedly don’t). (For example, Rhonda Patrick and Ronald Krauss expound this view here :


  10. Hi Dr. Guyenet.

    I’d really like to konw where you stand now in regard to the diet-heart idea, from your 2009 articles ”Stuck at the starting gate?”. Given that you seem to accept the lipid hypothesis, and that in this post you argue that butter raise LDL which isn’t desirable, I’m wondering where do you stand regarding the issue of dietary SFAs and heart diseases risk (or even dietary SFAs rich-food in general in regard to health). Granted I’m well aware that different foods have different SFAs profile and hence different possible risk outcomes.

    Any possible update on this issue?

    Thanks for your work!

    • Hi Frank,

      I can’t recall ever having rejected the idea that LDL contributes to cardiovascular disease, except perhaps for a few months after reading “Good Calories, Bad Calories”. But as you point out, I did reject the idea that saturated fat increases cardiovascular risk. I don’t feel as confident in that conclusion anymore and I’ve stopped writing about it for that reason, but I do still think the evidence leaves much uncertainty and saturated fat in general is probably not as harmful as it was originally claimed to be (and may not be harmful at all, although as I said I’m not confident about that).

      I’ve been gaining a better understanding of the sources of uncertainty. For example, as I explained in this post, only some forms of SFAs lead to a significantly harmful blood lipid profile, and this doesn’t just depend on the fatty acid composition but also critically on the food context. This means that a lot of the SFA we eat isn’t going to have the predicted impact on blood lipids. Also, people who are overweight or obese seem to have less of a blood lipid response to SFAs, meaning that the majority of the US population is not that sensitive to this mechanism. There are also remaining questions about the degree to which short-term changes in blood lipids observed in RCTs are reflective of long-term changes.

  11. “Most of us eat too many calories, and butter isn’t helping.”—YES! Love this.

    Reevaluating “refinement” reminds me of Angelo Coppola (, and his “Plant Paleo” approach.

    Looking forward to Part II. Thanks for your contributions, Stephan!

  12. Interesting. How does this jibe with previous animal research–that you wrote about back in 2009–about the possible benefits on health and metabolism butyric acid appeared to produce? [1]

    In your 2009 post, you noted that the butryic acid-fed mice were able remained lean and avoided metabolic problems unlike their high-fat diet (lard and soybean oil-fed)counterparts and that this particular study showed that “butyrate increased their energy expenditure…also massively increased the function of their mitochondria…lowered their blood cholesterol by approximately 25 percent…their triglycerides by nearly 50 percent…lowered their fasting insulin by nearly 50 percent, and increased their insulin sensitivity by nearly 300 percent.”

    Perhaps this is more a function of the way butter is processed than the actual effect butyric acid has on health markers?


    • Hi Michelle,

      It’s true that butter contains compounds that may have health benefits, including butyrate and vitamin K2. I do want to point out that the mouse study that was the basis for my post used unrealistically large doses of butyrate. But the point I’m making in the current post is that we can get those same compounds, plus more, from whole dairy.

      • “But the point I’m making in the current post is that we can get those same compounds, plus more, from whole dairy.”

        Even from yogurt?

      • Just to point out about butyrate for anyone reading comments. Butter does have butyrate, but we get more “bang for our buck” with whole foods such as beans, reheated potatoes, and green plantains. When we eat these foods, our bacteria make butyrate for us. For butyrate, best to focus on these foods over butter.

  13. Based upon your current knowledge, would you consider meat as part of a healthy diet?

  14. Dear Stephan, What new evidence were you presented with? The case against butter has been around for a very long time. Regards, Ron

  15. Stephan, what do you think about the controversial issue about caseins? There’s a great comment about dr.Masterjohn a while back about the bogus claims in The China Study. However, there are controversial perspective. Caseins trigger permeability in breast tissues because their accumulation gives the signal to stop milking.
    Someone argued that they may trigger leaky gut with the same mechanism but others think that pepsine and other enzymes may break down the problematic peptidic sequence.
    Furthermore, there’s a specific casein derived peptidic sequence that has shown to increase occludin expression reinforcing the tight junctions, a mechanism that suggests all the other way around from the aforementioned claim.
    On the other hand,There’s also the prolaminic nature of caseins that hints at a possible zonulin activation mechanism similar to gliadin. Is The cross reactivity with the latter just a sort of accident due to molecular mimicry of gluten or is there something more?
    Do caseins trigger leaky gut in calves to make them absorb macromolecules like necessary hormones or do they already have leaky gut at birth that is repaired by caseins?
    I’d love to have your opinion since so far we just have some epidemiological evidence about dairy and only studies focused on isolated compounds that steer away from the heuristic perspective you mentioned.
    Would you like to write an article about it?

  16. Hi Stephan
    I don’t understand how consumption of butter increases ldl. My understanding is that dietary lipid is transported to cells in chylomicrons. I don”t think the liver is involved. So given that ldl is derived from vldl in the liver how would butter increase it? I can see how excess carbs cause ldl rises because de novo lipogenesis occurs in the liver and these fatty acids then exported via vldl. Am I missing the link?

    • Hi Gretchen,

      I didn’t say cheese is less processed; I said it’s less refined. Refining is the process of isolating specific parts of a whole food. The only thing that’s removed in making cheese is the whey, while making butter removes almost everything except the fat.

      • With respect to dairy, isn’t most of the nutrients (vitamin A,D,E,K,CLA,choline) contained within the fat?

      • I haven’t researched this, but I would assume that butter contains the fat-soluble vitamins that are in milk. But both yogurt and cheese (except cottage types) are incubated for a long time with microorganisms, and who knows what nutrients they remove in addition to obvious removals like lactose in yogurt.

        But it’s partly a question of semantics as I gather the term “refine” can be defined in different ways. One definition is “removing unwanted material.” In butter, that’s most of the protein and carbohydrate. In yogurt, it’s most of the lactose.

        • Yogurt-making doesn’t remove anything from the milk, but the lactobacilli convert much of the lactose into lactic acid. This depresses the sugar content of the milk, reducing the glycemic index from ca 40 to ca 20, but does not significantly reduce calories. In comparison, butter-making removes both protein and lactose almost competely, and cheese-making removes lactose and some protein. Cheeses coagulated from whey, such as gjetost and real ricotta, have very different nutrient profiles from regular cheeses. IMO cheese-making represents a much higher degree of processing than butter-making, due to the sophisticated end-products.

  17. Thought-provoking post. I would like to comment on your statement: “Some research suggests that this is due to a protective substance in less refined dairy foods called the milk fat globule membrane, which is lost during the butter-making process (12).”

    Your ref 12 is to a paper by Rosqvist et al. Their research used butter oil (ie. ghee) of 98.7% fat, not butter. The heating used to make butter oil from butter does remove the fat globule membrane that is in the butter. I suggest that this part of your argument against butter is therefore not valid.

    • I would second this, as an amateur gourmet.

      The remnants of the MFGM are what give butter many of its culinary properties. Butter sauces that do not include cream or eggs depend on the MFGM.

      It is not correct to state that producing butter eliminates the MFGM, though it does reduce it. This reduction may or may not be significant enough to make a substantial difference in health outcomes, or even blood parameters.

  18. I am not having any luck in finding an older post of yours, the gist was IIRC, most of the so called “industrial” food oils are not all that harmful, trans fats excepted. I wanted to reference and provide a link on a diabetes site I follow. We are always having someone post a quasi-science post on the evils of the same.

  19. Thanks for this post. It goes in the same direction I’m going.
    I only use olive oil / coconut oil / ghee in small quantities in the pan so that my meat/fish doesn’t stick to it 😉
    Never thought pouring some over my food after that made sense. I’ve enough lipids through whole foods like avocados, meat, fish and nuts (and not too much processed dark chocolate :)).
    It also shows again the nonsense of bulletproof coffees and such.

  20. I do not think anybody disputes nutrition is complex and multifactorial. We need food to survive. The need for food security so processing at some level is likely as old as we are as a species.
    Groups on pre-westernised diets were largely free of western non-communicable disease including heart disease. Some such groups included butter in their diets for example Sikhs, and were renowned for the health and physiques.
    In the UK at the turn of the 1900s butter and beef were part of the far from ideal British diet, but observed cardiovascular disease was rare.
    It is very clear that oxidative stress and connected inflammation is a major factor in non-communicable disease including cardiovascular disease, which brings into focus the balance between the oxidative and anti-oxidative capacity of foods. I have written on the arguable impact of excess oxidised linoleic acid on western disease in chapters 27-32 here – some libraries may have a copy.
    Like virgin v refined olive oil, all butter is not created equal; butter from grass v grain fed animals will not be equivalent in lipid profiles or fat soluble antioxidants. The nutritional impact of both butter and olive oil will depend on their origins and degree of processing.
    Butter contains cholesterol; in butter that is rich in antioxidants the risk that the cholesterol will be oxidised in processing is reduced. Oxidised cholesterol has a significant presence in placque and is a cardiovascular risk factor.
    We fail to accurately or consistently define what is meant by saturated fats. For example often lard is equated with saturated fat but due to grain feeding of animals may contain 10% and more of linoleic acid.
    Saturated fat is not a primary component of vascular placque, and is relatively inert, whereas linoleic which is easily oxidised acid is a major component of vascular placque. Linoleic acid is a minor component of butter, but forms a greater proportion of olive oil, however in virgin olive oils it is likely protected by antioxidant components.
    Sir Robert McCarrison observed that butter has nutritional benefits for rats on a diet of refined flour, and included butter in the diet of his lab rats designed to copy that of the Sikhs. These rats were extremely healthy.
    Weston Price used butter / butter oil to good effect in encouraging remineralisation of cavities in teeth.
    So in conclusion I would argue as ever things are complex including because of the lipid profile of butter will alter according to the lifestyle and fodder of the cattle, as will the butter content of other important fat soluble nutrients it can potentially contain. Human outcomes will depend on nutritional quality of the butter and wider diet, level of intake including of other lipids, degrees of processing, wider dietary factors and individual genetic profiles.

  21. Hi Stephan –

    I love reading your work, it is always interesting! I also want to thank you for bringing the Instant Pot to our attention.

    Where does homogenized milk fit in your continuum? It is less refined than butter, but more processed. Given the lose of the MFGM, I would place it near the butter end of the spectrum. We need to consider skim compared to whole as well.


  22. why is a single parameter so important for butter? High cholesterol, and to some extent fairly high LDL, is associated with a longer life. The purpose of eating well is to live longer, not to avoid a particular type of death. The 2015 mega- meta-analysis can be found here

    In particular look at Figs. 1-1 and 1-2. So long as your LDL is above a certain threshold your survival does not change much. Ghee will keep you right in that range.

  23. Interesting post Stephan,

    I wonder as to why Butter seems to be one of the only fats that mobilize my digestion and ensure a BM hours after consumption. Even in comparison to other high SFA oils like Coconut Oil. Being of British decent, I wonder if Butter is more in line with my genealogy and the diets of my ancestors.

    Just speculating here of course.

  24. A problem with the «whole food heuristic» is that we are eating foods we may not be so adapted to from an evolutionary point of view, including whole grains and dairy. Even low carb veggies isn´t really consumed much by hunter gatherers. And for this reason sometimes a refined food can be better than a whole, or a whole food like cow milk or low carb veggies can be improved by adding other (refined) foods to make them more compatible with the requirement of the human species.

    If you give human milk to a calf, it may well die due to the low level of protein, phosphorus and other nutrients, whereas if you give only cow milk to an infant it may also not grow well due to the high level of phosphorus and protein.

    As an example, if to 700 ml cow milk is added 2 egg yolks, 4 servings of fruits or berries, 4 tbsp honey/sucrose, 1 tbsp coconut oil, 1 tbsp butter, 2 tbsp olive oil, 1 ounce walnuts and 1/2 tsp fish oil, it becomes quite similar and provides just about all the nutrients found in 2000 kcal human milk. In human milk there´s a lot of saturated fats (including approx 25 grams of palmitic acid/2000 kcal – same as found in 100 grams of ghee) but also some highly unsaturated fats, so it´s possible that problems with butter in regards to LDL cholesterol, insulin resistance etc, can be mitigated by fats found in walnuts and fish, and also omega-6 arachidonic acid (found in organ meats, egg yolks, poultry, and abundantly in human milk). It´s all about balance, too hard (only butter) or too fluid type of fat (only fish and nuts) is not ideal in excess. Olive oil, in between, may be a good all round fat with a similar fluidity as found in human milk, assuming the person has obtained the minimum requirements for polyunsaturated fats.

    Likewise the lecithin/choline in egg yolks may make a fat rich diet much less harmful. Interestingly the human milk example above is very similar to traditional or Haagen Dazs ice cream which is typically composed of skimmed milk, cream, sugar/honey, egg yolks, fruits/berries and nuts. Our modern types of cheap ice cream with questionable ingredients may however be quite harmful.

    It is worth noting that many of the people that try to lose weight or change their diet, are middle aged or older, of which probably the majority would have diabetes or pre-diabetes and likely not optimal kidney function (even if the blood test says they´re okay). Walter Kempner (1903-1997) used a very refined diet of mostly white rice and sucrose (, apparently with great success on thousands of people, deling with conditions like hypertension, kidney problems, diabetes and obesity, and the idea was that if the kidneys could be relieved of the burden of too much protein and electrolytes, they could function normally, and a host of health problems would improve. Today it´s common for people with kidney disease to reduce phosphorus down to 800-1000 mg and potassium to 2500-3000 mg, also sodium and protein. However, the type of diet many people adopt later in life with large quantities of whole grains, low fat dairy products, low carb veggies, and chicken breast, with very little sugar or fat, not even any potatoes or fruits, may well cause some serious problems down the road, even if the diet has some immediate positive effect on weight or other metrics, especially if the person has been «acclimatized» to a relatively refined diet from an early age.

  25. I have been fascinated by the copper-deficiency theory of heart disease ever since reading your post from 2010, I believe.

    Do you feel that copper consumption has any weight here? Maybe butter consumption is not that bad if a person eats copper rich food regularly, like liver or dark chocolate?

    I’ve read about it and found that it’s also responsible for thyroid health, although this is not well known. The people that follow extreme low carb diets and have symptoms of hypothyroidism and high cholesterol may just have copper deficiency.

  26. Always good seeing a post of yours, Guiseppe. Butter, of course, does not create any insulin resistance, to the contrary it helps normalize it, and LDL is not a problem, since above a certain threshold (not below) it is only weakly correlated with mortality.

    It seems to me that there are plenty polyunsaturated fats, in the meat and eggs we eat, and also in avocado, lard and tallow, olive oil, and nuts of course. Butter itself has a fair amount. Why even worry? On a variety of fats things will average out.

    • Hi glib,

      I also don´t see much problems with butter and also don´t worry about getting enough pufa since I eat some fish, chicken, eggs, occasional nuts etc. I avoid vegetable oils except for olive oil. A study published last year suggested that it wasn´t so much difference between a 73% fat diet and a 30% diet over 12 weeks when they supplied the same amount of protein and other nutrients:

      Saturated fats can be turned into monounsaturated of course, but this requires a healthy organism and a good diet.

      I do feel better with olive oil than butter or coconut oil, and also my skin and hair looks better. But this may also be related to the fact that I usually consume olive oil along with vinegar in the form of salad dressing, which has some additional benefits. It may be that a key problem in regards to too much fat is the common lack of fermentable carbohydrates/soluble fiber on such diets (in human milk there´s a lot of special oligosaccharides which may act a bit similar as soluble fiber). They will create the organic acids via bacterial fermentation. So fat along with vegetables or potatoes (rich in soluble fiber compared to whole grains), is less of a problem. In traditional diets, high fat dishes is typically accompanied by some sour foods and also often egg yolks (mayonnaise for example is made of oil, vinegar and egg yolk), or they are just added to low carb veggies, or potatoes.

      Also thinking about this old potato experiment where the two participants obtained about half of the calories from butter/lard and the rest from potatoes (120-150 gm fat and 1-1.7 kg potatoes), yet over the course of 6 months didn´t gain any weight and actually felt very well.

  27. I find it strange that you say, “Things that increase the number of LDL particles, and/or their total cholesterol cargo (called LDL cholesterol) tend to increase cardiovascular disease risk. The evidence supporting this is now extremely strong (1, 2).” The ref 1 study declares in the “Background” paragraph “LDL-C is causally related to the risk of CHD. So that study is not even questioning whether LDL-C is bad or good, it declaring it bad right off the bat.

    On 23May17 Glib above has linked a long PDF that appears to exonerate LDL of any wrong doing.

    I would site Ravnskov U,Diamond DM, Hama R, et al.Lack of an association or an inverse association between low-density-lipoprotein cholesterol and mortality in the elderly: a systematic review. BMJ Open 2016;6: e010401. doi:10.1136/ bmjopen-2015-010401. This appears to suggest that LDL-C is more likely good for us than neutral and little chance of it being harmful. I don’t understand how you arrive at your anti-cholesterol stance.

  28. It is worse than that, Tim. Low LDL is associated with increased mortality. The evidence supporting this is extremely strong. (Grass fed) ghee saves lives. It seems our friend Stephan has gone over to the dark side. Alas, this will help keep Social Security solvent.

    • I wouldn’t be so quick to assume the association between low LDL cholesterol and mortality is causal. It’s risky business making that kind of assumption about observational findings. Certain diseases like cancer tend to reduce LDL cholesterol levels, so it could be that disease lowers LDL cholesterol rather than low cholesterol causing disease.

      I’m not aware of any compelling evidence that low LDL actually causes an increased mortality risk.

      • Not causal by itself. But it certainly shows that LDL is not a problem. Other papers note an immune boost with increased cholesterol, so there is an explanation readily available that implies causality. I note that

        a) those survival curves have the general form associated with a necessary nutrient, they fall steeply going from low to medium LDL (medium by my standards, not by Big Pharma standards), and then flatten or slightly rise as LDL increases

        b) the large intake of full fat dairy by the high longevity Sardinians described in the Blue Zone book.

        c) in the paper linked above studies were from all over the developed world

        David Feldman has a blog that shows how to change both LDL and total cholesterol with a 3-day dietary change, which is useful if you are shopping for life insurance (and they require low cholesterol to lower your rate!). I tested that in one direction, finding a 60 points increase in both total and LDL just by testing after a 44-hours fast. Next time I will test after the 3-day gorging on fat that he prescribes, I expect the same swing in the other direction, but really, LDL is doubly meaningless, being a dynamic quantity and not being associated with mortality.

  29. Thanks for an excellent post.

    Wondering what this tell us about drinking whole milk vs. nonfat milk, conditioned on choosing to drink milk in the first place. The “whole food heuristic” would suggest drinking whole milk, but of course if you’re trying to watch calories nonfat milk has less calories by volume. (Although it may not be more satiating per calorie of course.) Similar question for yogurt.

  30. “.. butter isn’t toxic, and if you want to use a little to saute onions or mix into steamed vegetables, I doubt it will have much negative impact on an otherwise healthy diet.” is probably true for the minority of healthy people in our country. I can’t help but think that this is a slippery slope for a nation that’s in a health crisis of obesity and cardiovascular problems. What’s the line between “a little” and “too much?” For how many people will a heart attack be the 1st indication that they’ve crossed it?

  31. Question: how is butter more refined than cheese or yogurt? You can take fresh cream from a cow, cool it down and shake it in a mason jar for 10 minutes and end up with butter.

    • Hi Jeremy,

      By definition, refining is the process of taking a whole substance and extracting/concentrating a portion of it. Butter is a concentrated portion of milk so it fits that definition precisely. Yogurt has not been refined in any way because nothing has been removed or concentrated.

  32. ß-oxidation is much cleaner than glykolytic-oxidation!! Less free radicals more antioxidants!!

  33. Hi Stephan,

    Thanks for the blog post. Since you mentioned the “milk fat globule membrane” – I wonder how you feel about homogenization of dairy and it’s impact on health? Also, do you think low and no fat dairy is a healthy choice when trying to limit calories?

  34. Hi Stephan,

    I actually came across this post by searching for butter and vitamin K2 and have just read a couple of your earlier blog post entries.

    In particular in this earlier blog post you presented some very compelling evidence and studies that indicate that vitamin K2 may be synergistic with vitamins A and D (including a study in Netherlands comparing butter and margarine). I’m wondering if you can comment a bit on this aspect of butter consumption in light of the whole food heuristic.

  35. Why do you say that yogurt and cheese are less “refined” than butter. I think butter is the least refined of the three, or, at least, no more so than yogurt. Clearly, it’s less refined than cheese.

  36. The “refining” of butter is minor. A much more important distinction is whether the milk is from organic grass-fed cows or confined, grain-fed cows. The fatty acid profile is significantly different, as are other nutrients such as fat-soluble vitamins, including vitamin K2.

    “Averaged over 12 months, organic milk contained 25% less ω-6 fatty acids and 62% more ω-3 fatty acids than conventional milk, yielding a 2.5-fold higher ω-6/ω-3 ratio in conventional compared to organic milk (5.77 vs. 2.28).”

    But not all milk labeled organic is from grass-fed cows. The source is critical. See for their dairy scorecard.