Does dairy make you fat?

Does dairy make you fat?

The primary proteins in milk are unique to milk, not found in any other tissues. Broadly speaking, there are two categories of milk proteins, the main group is the casein proteins ( about 3 or 4 depending on the species ), with the remaining proteins in milk grouped together as the whey protein group. All of these proteins are highly insulinogenic, meaning they produce a pronounced insulin response, even in the absence of carbohydrate. Keep that last bit in mind, it will be pivotal later.

Several studies have been performed to determine the effects of milk proteins on insulin levels, and a really interesting one compared the effects of ingesting cream vs. caseins. Cream is basically all fat, the lactose and proteins in milk have been processed out via centrifuging. So, we’re essentially comparing the insulin effects of ingesting milk fats vs. ingesting milk proteins, no carbohydrates allowed.

Take a look at the below table ( cribbed from Peter @ Hyperlipid¬†) and see what happens to your insulin when you ingest casein. One and two hours after ingesting casein, insulin levels are threefold higher than baseline, and this is a statistically significant result. Three hours later and we’re still more than double the baseline insulin level. But remember, we have not been ingesting any carbohydrates, so, by rights, our hapless experimental subjects should all be in hypoglycemic comas. But, happily, they’re not, because glucose levels have not budged from baseline.

Insulinogenic Effects of Casein

The burning question then becomes, how do you keep glucose levels constant in the face of tripled insulin levels and no dietary intake of carbohydrate? The one word answer is … glucagon.

The chart below comes from a different study¬†that had subjects eat four different isocaloric meals that were high in protein, fat, carbohydrate, or alcohol. So, we’re eating the same amount of calories in each meal, but were changing the macronutrient ratios. Take a look at the insulin and glucagon response panels at the bottom of the diagram, the bit we’re interested in is the delta-AUC, or, the change in the area under the curve, which we can interpret as the overall effect of the meal. Well, as I’ve highlighted, every meal, with the notable exception of the high carbohydrate meal, causes both insulin and glucagon levels to increase in approximately the same magnitude. The high carbohydrate meal spikes insulin while at the same time suppressing glucagon, in fact lowering it somewhat over baseline.

Insulin / Glucagon Response to Different Macronutrient Isocaloric Diets

Glucagon has the opposite effect of insulin on lipolysis, that is, while insulin inhibits lipolysis, glucagon promotes it. So, in general, net net it is pretty much a wash in terms of lipolysis … one step forward, one step back. Only in the case of high carbohydrate meals do we have a net suppression of lipolysis, but not due to insulin secretion, but rather to the suppression of glucagon.

Now, we can finally answer the question: does dairy intake make one fat via the action of dairy on insulin and that hormone’s peripheral effects on lipolysis? No, because there is a compensatory secretion of glucagon by the pancreas in order to maintain normoglycemia and the peripheral effect of glucagon on lipolysis is to increase it. I tend to think that the anti-dairy paleo faction is probably overstating their case.

  • Pingback: Does dairy make me fat? | Mark's Daily Apple Health and Fitness Forum page()

  • Pingback: Weekly Round Up « Highbrow Paleo()

  • Pingback: Lactic acid calories? | Mark's Daily Apple Health and Fitness Forum page()

  • Pablo Tioseco

    This conclusion is what I have thought all along. Thanks for sharing your research and thoughts. Terrific content on this site btw

  • Pingback: Lacto ovo paleo experiment | Mark's Daily Apple Health and Fitness Forum page()

  • David Beers

    A lot of holes in your argument, at least from this post.

    • Pablo Klopper

      Hi David,

      While I appreciate you taking the time to read my post and comment, your comment doesn’t do much to help me to get better. If you have specific things to rebut in the argument, I would love to see them and to have a chance to address them.



      • David Beers

        Thanks for responding! Didn’t want to write out something long if you weren’t looking at comments any longer.

        My issue is the casein testing vs the protein testing involving glucagon (admittedly, I’ve just began your blog, so you may have said this elsewhere). Based on this post, how can I be sure that casein has the same rise in glucagon as does the protein meal? If it doesn’t, then that would make casein extremely fat inducing, correct?

        • Pablo Klopper

          The real question is how do you triple circulating insulin levels, yet maintain effectively constant blood glucose levels?

          One way to do that is to infuse glucose in order to compensate for what is being removed from the plasma due to the action of insulin. In the medical and research literature, this is known as a euglycemic hyperinsulinemic clamp, and it literally involves hooking subjects up to an intravenous glucose drip 80 minutes after injecting them with insulin. The whole procedure is intended to measure insulin sensitivity, insofar as if you are insulin sensitive, then we would need to infuse a whole bunch of glucose, as you were effectively removing it from circulation in response to elevated insulin. Conversely when you are insulin insensitive, we do not need to administer a whole lot of insulin, since your metabolism is resistant to insulin signalling, and therefore blood glucose hangs around.

          If you look at the data for the subjects who had ingested casein, you could be easily forgiven for assuming that they had been hooked up to a glucose drip, because blood glucose levels do not budge, despite a statistically significant tripling of insulin levels over a period of three hours. In fact, the blood glucose response to casein is virtually indistinguishable from the response observed when these same subjects ingested cream which did not result in a statistically significant increase in plasma insulin relative to baseline.

          How do we explain this? I can come up with at least two possibilities. One, there is something about casein that induces insulin resistance, or two, there had to be a concomitant release of glucagon in order to maintain normoglycemia. Given these possibilities, it doesn’t actually matter what happened, because in either case, the casein would not promote adipose tissue gain, as either the tissue was not responding to insulin signalling, or, it was, but it was also responding to glucagon signalling which serves to increase lipolysis.


          • David Beers

            I think I’m misunderstanding what we would expect to see from the glucose pattern if adipose tissue was gaining. What would that look like?

          • Pablo Klopper

            Let me put up a post on glucose disposal to help explain. The Disqus format is somewhat limiting, at least in so far as I can’t post pictures ( or at least I cannot seem to figure it out ), and the way that replies are indented makes the overall screen real estate limited.

          • David Beers

            Thanks! I’m thoroughly enjoying catching up on past posts now. One a day keeps the doctor away (haha :/)

          • Emilia Contressa Gomez

            have u put up the glucose disposal post yet? love to read it :)

          • Pablo Klopper

            Working on it … had to pull up some papers to substantiate the discussion!

  • Emilia Contressa Gomez

    ok, so if we just eat whey protein, with no carb at intake time, we’re doing good. but dairy “real” milk is not ok as it contains lactose. cream is ok as it not only doesn’t contain any kind of protein but no carb.
    do i have that right? milk is bad as far as insulintrophic properties but cream and whey protein is good?
    kinda sucky as whey protein is not really a food IMO, but a very processed product. cream is easily skimmed off the top of container left overnight.

    • Pablo Klopper

      Actually, the takeaway here is that dairy is no more and no less fattening than any other foods of roughly equivalent caloric content, whether that dairy is whey, cheese, cream, or what have you.

      People tend to oversimplify things when it comes to glucose disposal focusing entirely on insulin to the exclusion of other fundamental hormones like glucagon.

      Just as control of the speed of your car requires both an accelerator and a brake pedal, control of blood glucose is orchestrated via insulin _and_ glucagon acting in concert. The peripheral effect of glucagon is to increase fatty acid release from adipose tissue which counteracts the effects of insulin.

    • Emilia Contressa Gomez

      so its just coincidental that i eat cream and i feel very satiated for hours but if i drink milk i get hungry in 2 hrs? or if i eat cheese, which is not near as fat laden as cream, and i get hungry in 2 hrs?
      while glucagon is also involved in the response to food intake, i think the insulin connection to obesity is quite striking.
      i know personally, the only way i can budge weight off my body is by lowcarbing, and preferably, high fat intake. again, maybe anecdoctal but then again, there seems to be alot of peer-reviewed evidence that it’s not.

      • Pablo Klopper

        Remember that I said that dairy is no less and no more fattening than other foods of equivalent caloric value.

        I am perfectly willing to accept that different kinds of dairy have different effects on satiety. The basic motivation behind this post was the frustrating debate that is ongoing in various nutrition communities regarding the uniquely fattening properties of dairy. This mindset seems to be particularly prevalent in the paleo / primal communities, but then, the prevalent analytical framework there seems to be WWGE, or What Would Grok Eat? My preferred analytical framework is the metabolic response to nutrients, and when we apply this to the fattening properties of milk, we find, unsurprisingly, that there are no such properties.

        To the extent that a given food causes a transient insulin spike, this will result in hunger after insulin levels normalize. At this point, if one does not eat again, then we would not observe any tendency towards increased adiposity. Of course, in our world, with instant on demand access to food, this does not happen, and people will “snack” their way into metabolic syndrome.

        Also note that I mentioned a transient insulin spike, in order to specifically differentiate that from sustained elevated insulin as we see with casein. In the short term, insulin is a satiety hormone in the hypothalamus, so I would expect that with a casein meal, one would experience a prolonged period of satiety. You could easily test this for yourself with various commercially available casein protein powders.

        As far as your personal experience goes, I cannot possibly argue against you. You are the ultimate arbiter of how you respond to foods. So my ultimate advice to you would be to enjoy those dairy products guilt free if they result in long term satiety for you.

  • Pingback: Glucose Disposal | Cogito Ergo Edo()

  • Brad

    Since I have been doing the pseudo-1-meal deal, I find myself craving things like cheese and salami after a workout. Often I’ll eat 200 grams of cheese and an entire liter of milk in addition to a ton of fruit. Basically my craving seems to be to eat as many calories (calorie dense foods) as soon as possible :-) LOL. That said, I am gaining muscle and staying lean (knock wood) so far. My n=1 seems to support the idea that dairy is nothing to be afraid of given an otherwise good exercise and diet program.

  • Larry Rotenberg

    for a diabetic what can one conclude about taking casein?

    • Pablo Klopper

      In the case of diabetes, we’re mostly concerned with the maintenance of normal blood sugar levels, normoglycemia. Insulin is merely a means to this end.

      Protein rich meals such as those containing Casein will result in both an insulin response and a glucagon response in normal individuals, but in diabetics, we have two distinct scenarios to deal with, type 1 insulin dependent (T1DM), and type 2 insulin resistant (T2DM).

      Let’s take the insulin resistant case first, because it is conceptually easier. In this case, the body is awash in insulin, so eating a substantial amount of casein would probably produce an exaggerated insulin response, which should effectively shutdown glucagon production via the paracrine effects of insulin on the alpha cells.

      The T2DM case is made easier by the fact that casein ingestion does not affect blood glucose due to the presumed suppression of glucagon release. In the T1DM case, we have no such suppression, because the beta cells do not produce insulin.

      For T1DM individuals, I think a different strategy would be needed. Specifically, try to limit protein intake in general, relying mostly on fats for caloric needs, allowing protein to be reserved for structural needs only. By eating a fat centric, ketogenic style diet, we would also presumably affect the kinetics of protein absorption, i.e. making absorption take longer, which would blunt the glucagon response.

      You could also try to control the glucagon response by administering insulin in a spaced out protocol rather than one large injection at meal time. The problem with this approach is that you are administering insulin peripherally, but hoping to simulate / stimulate a paracrine effect. In other words, the insulin first circulates around the body, and eventually makes its way to the pancreas. This is exactly the opposite of what happens in a healthy individual. In a very real sense, first and foremost, the beta cells of the pancreas talk to the alpha cells via insulin signalling and then they speak to the liver, and then only somewhat incidentally is this insulin conversation overheard by the rest of the body.

      I’m sorry about giving you a somewhat unsatisfying answer, because, well, it depends. I think the easier case to address by far is early T2DM, where we still have significant beta cell function, and in that case, the best advice probably is to stop eating for a while ( i.e. fast ).

  • Marty Kendall

    Based on the latest food insulin index test data it appears that insulinogenic effects of dairy products, along with other foods, are largely explained by their carb, protein and fibre content.

    See analysis at

  • Pablo Klopper

    Predicting the insulinogenic effects of a food based upon a linear regression model is interesting. However, it still misses the mark insofar as it doesn’t address the mandatory other half of the equation, which is of course, what is happening to glucagon. Without this information, we cannot really do anything meaningful.