Discordant insulin sensitivity on a high protein diet

So, we have another “high protein” weight loss study (Smith et al., 2016).  Or really, a “low (0.8 g/kg) vs. moderate (1.2 g/kg) protein weight loss study.”  In brief, it took ’em about 6 months to lose 10% of their starting body weight, then were given 4 weeks of weight stability before “after” measurements were taken.

Important: this was not a contest to see who would lose more weight; they kept going and adjusting food intake until both groups lost 10%.  Not really ad lib, but otherwise a good study design imo.  The intervention was relatively weak (eg, protein 0.8 vs. 1.2 g/kg), but on the plus side, that’s realistic and very “do-able.”  If you’re interested in super-high protein diets (3-4 g/kg), check out research by Jose Antonio.



Big yet not unexpected finding: the low protein group lost about twice as much muscle than the normal protein group.




The isocaloric normal protein group lost more fat and less muscle than the low protein group.

But then everyone freaked out because the low protein group experienced a significant improvement in muscle/liver insulin sensitivity whereas the normal protein group didn’t:




-The headlines were hilarious, like, “high protein makes weight loss not work anymore.”

-Then some critics jumped the shark and blamed it on “liquid calories,” because whey protein shakes are totes non-Paleo, and #JERF.

-TBH, I found more interesting the changes in adipose insulin sensitivity

The normal protein group had the most insulin sensitive adipose of all groups… yet they lost more fat mass despite eating just as much or even slightly more than the other groups.




Does this mean they’re doomed to regain the weight?  I don’t think so, as high dietary protein is one of the strongest predictors of weight loss success long-term.

HERESY!  the low protein group had: 1) lower basal insulin than the normal protein group; 2) lower adipose insulin sensitivity; 3) ate less (NS); yet lost less fat mass.



In other words, the normal protein group had higher basal insulin, more insulin sensitive adipose tissue, and slightly higher food intake (NS).  According to the insulin model, they should’ve lost less fat mass than the low protein group, but they didn’t.

Is this another chink in the armor of the insulin model?

The truth seems to be: people lose weight on both LC and LF diets by giving up junk food.  On LC, this is accomplished by giving up carbs; on LF, this is accomplished by switching to better carbs.  Some people adhere better to one diet or the other.  Maybe insulin sensitivity has something to do with it.

Insulin from high protein: not bad?
Insulin from good carbs: not bad?
Junk food: no bueno.
So maybe just maybe it’s not just ze insulin…



Back to the protein…

This was not sorcery; it’s been seen before in a variety of different paradigms: dietary protein has a profound impact on nutrient partitioning.

Yes, even when it’s liquid calorie insulinogenic whey protein isolate bro-shakes.

Yes, even when it’s not crazy-high levels of protein…  seriously, 1.2 g/kg is not “high”


+ + +  +   +     +     +        +             +                     +

Past blog posts on [the non-sorcery of] dietary protein:

Holiday feasts, the freshman 15, and damage control

Dietary protein, ketosis, and appetite control.

Nutrient Partitioning: …a *very* high protein diet.

Protein “requirements,” carbs, and nutrient partitioning

Cyclical ketosis, glycogen depletion, and nutrient partitioning

Meal frequency, intermittent fasting, and dietary protein

Muscle growth sans carbs

+ + +  +   +     +     +        +             +                     +

For full access to all articles and much more (or if you just like what I do and want to support it), become a Patron! It’s three bucks a month and there are many other options. Sign up soon because there are only a limited number of spots left at the $3 level. It’s ad-free and you can cancel if it sucks ????

Also, I’m open to suggestions, so please don’t hesitate to leave a comment or contact me directly at drlagakos@gmail.com.

Affiliate discounts: if you’re 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 HERETrueDark is running a pretty big sale HEREIf 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.

calories proper


Become a Patron!












Be Sociable, Share!
  • Daz

    Looks like the ‘high protein’ group were getting two of these per day to get their extra protein;

    Ingredients: Whey protein isolate, fructose, natural flavors, cocoa powder (alkali processed), soy lecithin, xanthan gum, acesulfame K and sucralose. Contains: Milk, soy, tree nut (kola).

    Would any of these ‘other ingredients’ have any influence on the insulin findings. Impossible to say perhaps…?

    • Sky King

      Good question! It was probably donated to them or they used it because it’s what’s used in major hospitals and research facilities.

      Bros in the know doing RT would never use that crap. They’re more likely to use one that features JUST whey, such as this one, the one I personally use:


    • unlikely

    • TechnoTriticale

      re: Would any of these ‘other ingredients’ have any influence on the insulin findings.

      Insulinotrophic response to whey is reasonably common, and part of why dairy elimination trial is often suggested by diet programs for weight loss stalls.

      The fructose is an eyebrow-raiser. It won’t provoke BG, but is a known fat magnet, and of negative benefit in, for example, NAFLD.

      Indirectly, I see several more that could materially distort microbiome.

      Soy lecithin and xantham gum are emulsifiers. They can be mucous lining disruptors, cause intestinal permeability, and alter bowel flora resulting in increased blood insulin, blood sugar, and inflammation. And yes, these sort of emulsifiers/thickeners are pervasive in processed food-like substances generally.

      Sucralose is another suspect microbiome antagonist, and one of the stars of the long running saga “Why you can’t lose weight with ‘diet’ foods.” AceK I’m less certain about.

      For nutrition trials, Unjury looks to me to be packets of concentrated cohort confounder at the very least.

      • the concentration of those ingredients is so so low, and they’re present in many other foods in *both* diets…

      • Sky King

        Well said..!!

        • TechnoTriticale

          Well thanks, but Bill has a point too.

          I sometimes wonder how we’re able to tease any useful conclusions at all from the vast majority of contemporary and historical nutrition trials.

  • NY

    Does FFM=Muscle? Is it not better to have more insulin sensitive muscle than having insulin sensitive fat which was achieved by low protein group? Could it be because the high protein group lost more fat so they achieved more fat insulin sensitivity and vice versa?

    • FFM is exactly what it says, “fat-free mass,” but we can fairly safely assume most of the changes in FFM are due to changes in skeletal muscle

  • NY

    Looking at the tables at http://www.cell.com/cell-reports/fulltext/S2211-1247(16)31286-4

    Carb intake is 45%+. Is that right? Protein intake from 19-31%.
    High carb low/moderate protein and low fat-diets.

  • Sky King

    People (mostly obesity researchers) are still looking for that unified theory (the Holy Grail) of explaining the root causes of obesity. Any theory they come up with should be consistent in explaining ALL these seemingly contrary observations.

    As I’ve pointed out numerous times in the past here and elsewhere, me thinks Ian Spreadbury has the best theory thus far!


    • Tuck

      “As I’ve pointed out numerous times in the past here and elsewhere, me thinks Ian Spreadbury has the best theory thus far!”

      I don’t think so. How then to explain polynesians who traditionally ate poi, which is a processed carb, and basically liquid. Very acellular…

      Yet they were lean and healthy?

    • Kindke

      The unified theory is that fat tissue growth can be caused by many factors, but all fat tissue growth is essentially the same, i.e. its “obesity” AND fat tissue growth is irreversible, atleast in vivo.

      Until obesity researchers start to acknowledge this and the role of adipogenesis in obesity and its many initiators, theres gonna be no progress.

      lowering insulin through diet only works on adipocytes that had increased storage through increased insulin, but there are many different lineages of adipocytes and some store alot of fat regardless of insulin.

  • Food intake et al.

  • Tuck

    “The truth seems to be: people lose weight on both LC and LF diets by giving up junk food.”

    Yeah, bingo. I think that’s all there is to it. Now the question is, are there any ingredients in junk food that are novel to the human diet that might explain that observation?

    It ain’t wheat, although the novel goatgrass hybrid isn’t a bad candidate, as even in sensitive people it doesn’t produce the MetSym…

    • ref for goatgrass vs regular wheat?

      • TechnoTriticale

        Here’s one:
        “Imazamox-resistant hybrids resulted from a cross between jointed goatgrass and an imazamox-resistant wheat…”

        Wheat has apparently been deliberately and accidentally crossed with goatgrass many times throughout history, and geneticists are toying with it yet, last I knew (although the focus may be on GMO wheats now, none of which have gone to market yet).

        • sorry, I meant regarding metabolic syndrome

          • TechnoTriticale

            re: sorry, I meant regarding metabolic syndrome

            Modern wheat and metsyn? If you want to go there … what was that movie line?
            “We’re gonna need a bigger blog.” ?

          • haha!

            metabolic syndrome: goat grass vs. regular wheat?

          • Tuck

            The toxic qualities in modern wheat come from a recent cross w/ aegilops tauschii, aka Tauschii’s goatgrass. There are older crosses, as Techno notes, but only the most recent cross created modern, hexaploid wheat with the problematic D genome.

            I looked for a long time for a link between gluten sensitivity and MetSym, and couldn’t find one. So no links for that. Causes lots of other things, but it typically results in weight loss, not gain, other things being equal.

            Here’s the link for goatgrass and modern wheat:

            “Diversification of the celiac disease ?-gliadin complex in wheat: a 33-mer peptide with six overlapping epitopes, evolved following polyploidization.”

  • John Lushefski

    I believe I have seen before that an increase in protein increases muscle insulin sensitivity and lowers adipose insulin sensitivity. It could’ve been a glutamine supplement–can’t remember exactly.

  • Dungaree

    thank god we have you BIll, to put all of us poor misunderstood keto advocates in our place! how dare we advocate moderate protein! Oh wait i mean LOW protein (so you have something to rail against). And thanks for the radical assertion that we lose weight by avoiding junk food! Your insistence on quashing all things low carb/keto for whatever deranged reason you feel is necessary is tiresome and disconcerting. Truth be told low carb and keto in any incarnation is miles ahead of any idiotic vegan, vegetarian or grain based SAD diet and yet you still insist on crapping where you eat instead of investing your energy into something useful. keep on keeping on being lost in the weeds.

    • Man

      Can you preach somewhere else ? Thanks.

      • Dungaree

        You’re cute

    • “investing your energy into something useful”

      I actually found the results of this study pretty interesting, regarding changes in insulin, adipose insulin sensitivity, and fat loss between the groups.

  • Man

    Big yet not unexpected finding: the low protein group lost about twice as much muscle than the normal protein group.

    Hey Bill. FFM is not just muscles but anything that is not fat. How do you know that the low protein group lost twice more muscles than the higher P ? I’d be interested to understand that point because a lot of the FFM lost during dieting is definitely not muscles only, especially if you practice some form of regular physical activity.

    • http://disq.us/p/1ctg9zs

      I’m pretty confident skeletal muscle comprised a good deal of the change in FFM because:

      1) they weren’t exercising

      2) the absolute change in FFM was fairly small in both groups (ie, could’ve easily been muscle given the time frame)

      3) turnover of skeletal muscle is faster than other tissues

      4) this interpretation is in agreement with previous findings and it makes sense…

  • tkent26

    “…on LF, this is accomplished by switching to better carbs.”

    Better carbs could still be operating through insulin. More fibrous carbs = slower digestion and absorption, so less insulin spike, and less sugar/fructose = less insulin resistance.

    Clearly there’s more going on than just insulin, likely glucagon is playing a role to counter-balance insulin in the higher protein group.

  • Jason Jodway

    I don’t even see what is interesting about this study. More protein preserved more FFM = expected. People who lost more fat ended with greater adipose insulin sensitivity = expected. This was an energy controlled study to enforce a specific % body weight loss so we cannot take it that this applied ad lib, meaning it has 0 relation to the primary argument of the CIH. Everyone knows if you force starve people regardless of composition they will in fact lose weight, for a time.