High protein magic

Remission of pre-diabetes to normal glucose tolerance in obese adults with high protein versus high carbohydrate diet: randomized control trial (Stentz et al., 2016)


Duration = 6 months

Diet: all food provided.  Mucho gusto!


Critique #1: if my calculations are correct, we’re comparing low protein (0.675 g/kg) to adequate (1.35 g/kg) (not “high”).

The diets were decent:






Author’s conclusion was that this was due to high protein alone, but I’d say it was at least partially due to weight loss.  BOTH groups lost weight and improved insulin sensitivity.  Statistically significant in both groups.



Glucose (A) and insulin (B) in response to a 75g OGTT (red is high protein, blue high carb):



I still say weight loss was the primary driver, but must concede, however, that protein did have a little magical effect: high carb group actually lost slightly more weight, but insulin sensitivity improved more in the high protein group.  The high protein magic: reduced insulin secretion yet still greater reduction in glycemia.



Well, maybe not magic…



Despite having more insulin, high carbers lost slightly more fat mass but way more muscle.  THAT’s high protein magic lol


However, the meal tolerance tests show a slightly different trend:



We expect glucose and insulin excursions to be greater in HC (blue), because they had a high carb meal whereas the HP group had a high protein meal.  From this perspective, if we graphed the results as “change from time zero,” I think the reduction in glycemia from baseline to 6 months would be similar in both groups suggesting weight loss as bigger factor.  We’d still give some props to high protein because it lowered glucose just as much despite having less insulin.  High protein magic.


Note to self: gotta stop saying this was “high protein.”  1.35 g/kg is not “high,” seriously.  But still, High protein magic haha


Oh and one other thing, high protein usually induces greater weight loss:

Randomized trial on protein vs carbohydrate in ad libitum fat reduced diet for the treatment of obesity (Skov et al., 1999)

High protein vs high carbohydrate hypoenergetic diet for the treatment of obese hyperinsulinemic subjects (Baba et al., 1999) (not ad lib)

Comparison of high-fat and high-proein diets with a high-carbohydrate diet in insulin-resistant obese women (McAuley et al., 2005)

The effect of a low-fat, high-protein or high-carbohydrate ad libitum diiet on weight loss maintenance and metabolic risk factors (Claessens et al., 2009)


calories proper


Become a Patron!











Be Sociable, Share!
  • Me

    I thought protein requirement is usually expressed as g/kg “ideal” b/wt not actual? So maybe 135g is more than adequate on that basis. It certainly seems quite a lot of protein (at least by non bro stds) to me.

    • Malcolm Campbell

      @me Good point.
      I made some guesses and came to an average of around 135kg bodyweight so about 1g/kg which doesn’t sound bad for HP. Then reading the text it mentions 1800 calories for 100kg bodyweight which suggests 1.33g/kg, although I wasn’t sure if I was interpreting it correctly. Either way, perhaps the HP group were getting more than 0.6g/kg.

  • When i saw the BMIs both dropped similarly, but the HC group had such shitty glucose control improvements, I knew much of the HC fat loss was muscle.

    More reasons the HP group fared better:
    1) protein needs are even higher in negative energy balance, but HC protein intake is deficient for even eucaloric diet (thus more muscle breakdown)…
    2) the energy deficit is likely greater in an isocaloric high protein diet, as protein does not convert to usable energy as readily as equal caloric intake of carbohydrate ( so both groups are eating “1800” calories, HP group is getting less energy yeild which may have helped their glucose tolerance, as CR is known to do)

    And of course, the most obvious reason glucose parameters improved, the HP group is eating substantially less carbohydrate. Duh.

  • Wendy G

    Math error? You are describing 135 g protein as 0.6 g/kg of body weight, in Critique #1, but applying that math works out to 225 kg body weight, or 495 lbs. The body weight gets even higher if this is 0.6 g/kg lean mass (vs body weight, or even ideal weight).
    These subjects were avg 100 kg each (and body weight was capped at 350 lbs for participants). I work this out to be 1.35 g/kg in the HP diet, and 0.675 g/kg for the HC arm. The HC arm does seem low to me, and IIRC Volek & Phinney recommend 0.8-1.0 g/kg.
    Am I misreading your characterization, or otherwise failing at maths today?
    Cheers. Love your work, thanks for keeping at it.

    • thanks! I miscalculated. So it’s low vs. average protein intake. Still not “high.”

      V&P advocate low(ish) protein to help keep ketone levels up.

      For “high,” see recent studies by Jose Antonio, >4 g/kg!

  • valerie

    How do you reconcile the table with the bar chart?
    The table says that both groups los around 10% of their starting weight, but the bar chart shows something completely different.
    Am I missing something?

    • the bar chart is “percent change in lean mass” and “percent change in fat mass.” The participants had much more fat than muscle, so a small percent change in fat mass has a big impact on body weight. The small increase in lean mass has a relatively small overall impact on body weight.

      I agree that it would’ve been clearer for the authors to give us absolute lean & fat mass, but for some reason they went with BMI & percent changes in lean & fat…

      • valerie

        “The participants had much more fat than muscle, so a small percent change in fat mass has a big impact on body weight.”

        How can a 3% decrease of some part lead to a 10% decrease of the whole?!?

        Can you come up with a numerical example that approximates the table AND the bar chart? I sure can’t.

        • daz

          i know what you’re saying valerie (i think) >

          & all i can think of, is that in this study,

          Lean Mass + Fat Mass does not equal Total Weight.

          (maybe water weight has been excluded?)

  • tydale

    Hi Bill,
    In the context of DHA and turmeric of a previous post.
    Do you think it’s agood option to eat chia seeds with turmeric and fish oil supplemnt together to increase DHA in the brain?

  • Matthew Klein

    Hi Bill, I want to address the recent polarization between ZCers and other low carbers I obviously am speaking for myself not the ZC community. I think we all need to take a step back and realize the primitive emotional nature of the discussion as of late. Everyone thinks they’re right and have the sources to prove it. It is dangerous to believe we have more answers than questions and also dangerous to outright dismiss opposing ideas without thorough analysis. I am ZC for personal reasons but have come to realize there is not sufficient data to make blanket ZC recommendations to the public (although many have and can benefit greatly), but also that to say that ZC is dumb or unbalanced is equally ludicrous because there is simply no data to suggest so. RDAs (all constructed from people eating SAD) are likely irrelevant to ZC, we can test this hypothesis. My point is if we don’t try to understand another group’s opinions we stand to lose valuable connections as we start to view those with opposing views as terrible people and even less human. I am all for the hunger free diets, I am against diet dichotomy it doesn’t have to be an us vs them if we’re able to just see that the other side has come to their conclusions by simply being human. This is what inspired me to write this: https://www.youtube.com/watch?v=hOnefFVBEb0 its a heart warming and encouraging story, I hope you listen to it!


    • TechnoTriticale

      What polarization? On one of the very low net carb (keto optional) sites I follow, I’m not seeing any hint of that.

      Personally, I see ZC as just a limiting case of KD, and as with KD, people might need to be more mindful of key micronutrients, prebiotic fiber (RS, soluble fiber) in particular. The human RDA for carbs might be zero, but the beneficial gut bugs need their substrate, or the host won’t get their benefits.

      • Shameer Mulji

        “The human RDA for carbs might be zero, but the beneficial gut bugs need their substrate, or the host won’t get their benefits.”

        Contradictory statement don’t you think. If gut bugs need their substrate, then the human RDA for carbs isn’t zero since carbs are the only way to get prebiotic fiber.

        • TechnoTriticale

          There are basically 3 kinds of carbs:
          A. “net carbs” – digestible, either simple sugars, or cleaved by human enzymes or other digestive juices and become simple sugars
          B. prebiotic/resistant/soluble “fiber” carbs, which are indigestible by the host, but can be fermented by microbiota
          C. insoluble/roughage “fiber” carbs, that even the gut bugs ignore, and just pass thru (we need some, but not a lot of this)

          People doing KD and ZC understandably seek to avoid A, but without specific attention to it, are apt to be deficient in B and perhaps C as well.

          Further, choices in fiber type B are limited when the net carb target is below 20 grams/day (as for KD), and even more limited for ZC.

          The ideal amount of microbiome substrate is not yet established. It’s unlikely to turn out to be zero. More likely it will be somewhere in the 20-150 grams/day range, depending on microbiome spectrum (and ideal spectrum is also conjectural at this point).

          RDA for A might be zero. RDA for B & C are not. KD, and perhaps ZC, might be valuable tools, but those considering them need to investigate the gut flora issue for themselves. If I were to do either, I’d still attend to microbiome support.

      • daz

        since the term RDA was mentioned (for carbs)…
        The US does actually have DRI’s (Dietary Reference Intakes) for macronutrients.
        It gives a range for carbohydrates; 45-65% of energy.

      • Matthew Klein

        This was happening on twitter jain plain woo and others are up in arms about things posted by Shawn Baker MD, AFAIK.

  • Phil Thompson

    Taking 60 grams a day of carbohydrate out of the HP diet may also have been a factor, but as always in these diet things I don’t know how you disentangle that.