the insulin-obesity hypothesis is under attack

…but it isn’t dead, imo, because that would be really hard to do.  Like, seriously.

 

 

side note: please consider the modern views of Taubes, Lustig, Gardner, Attia, and others on Carbs™.  They’re less “Carbs-cause-obesity, keto-for-all, etc.,” and more thinking it might not be Carbs™ per se, but rather processed and refined foods.  And #context…  And I tend to agree at the moment (nuances and caveats are subject to change, as more evidence accumulates).

 

disclaimer: I haven’t seen the full text of Hall’s recent study, but that’s not really relevant to what I want to discuss.  In other words, I don’t think the full text will provide any additional details on this particular point.

 




 

Tl;dr: this study was not designed to prove or disprove metabolic advantage or the insulin-obesity hypothesis.

It’s in the study design:  four weeks of low fat followed by four weeks of low carb.  We KNOW that weight loss slows over time (especially if calories are controlled, as they were in this study).  It has to do with the order of treatments.

Weight loss-slowing over time in the Minnesota Experiment:

 

 

Minn-Starvation-weight

 

 

 

Now, what did Hall et al. do?  Low fat diet first: rapid weight loss.  ANY diet second: slower rate of weight loss…

however, by the last 2 weeks of low carb, fat loss looked JUST AS RAPID as it was during the low fat period*, even though it was handicapped by being the SECOND diet.  *this bit might be clarified in the full text

again, I haven’t seen manuscript yet, but for now I’m just talking about the study design which shouldn’t differ much between the poster/talk and manuscript

My question: what would’ve happened if they had a group on low carb FIRST?  They would’ve had the added benefit of simply being in the beginning phase of a weight loss diet.  This study showed a low carb diet was just as good as low fat even when it is handicapped by being the SECOND diet studied.  (Maybe.)

Alternate #2: they could’ve just kept a group on low fat for the last four weeks.  I bet you’d see the metabolic slow down common of typical diets; the very same metabolic slow down that low carb gimped.

In other words, this study didn’t prove or disprove the insulin-obesity hypothesis, not because of the actual results (which we still don’t really know), but because it was not designed to do so.

 

calories proper

 

Become a Patron!

 

 

Save

Be Sociable, Share!
Share
  • Wab Mester

    “not designed to do so”‘

    Sure it was. But it wasn’t implemented well.

    The lead-in diet was supposed to be in energy balance, as Hall mentioned in the video and as stated in the design here:
    https://clinicaltrials.gov/ct2/show/NCT01967563?term=kevin+hall&rank=2

    “we will measure changes in energy expenditure in response to 4 weeks of inpatient feeding of a eucaloric, very low carbohydrate, ketogenic diet (5% Carbohydrate, 15% Protein, 80% Fat) immediately following an inpatient period of at least 4 weeks of consuming an energy balanced standard American diet (50% Carbohydrate, 15 % Protein, 35% Fat).”

    • “It’s in the study design: four weeks of low fat followed by four weeks of low carb. We KNOW that weight slows over time (especially if calories are controlled, as they were in this study). It has to do with the order of treatments.”

      http://caloriesproper.com/prelude-to-a-crossover-ii/

      • Wab Mester

        You might want to edit that to read “weight loss slows” instead of “weight slows,” but my point of course was that the goal was no weight loss on the lead-in diet. They probably could have executed better by keeping the subjects in the metabolic chamber, but they would have screamed their heads off. 🙂

        Had the lead-in lead to no loss, it would have been slightly more interesting, but frankly few people I’ve talked to believe in the “energy sequestration” aspect of the CIH.

        I still want to hear what Hall says about the appetite effects of the low-carb diet.

      • Guido Vogel

        This is what the study protocal says: (https://clinicaltrials.gov/ct2/show/NCT01967563?term=kevin+hall&rank=2)

        “To determine 24hr energy expenditure, respiratory quotient, and sleeping energy expenditure following a 4-week eucaloric ketogenic diet as compared to 4 weeks of consuming an energy-balanced standard diet.”

        and

        “we will measure changes in energy expenditure in response to 4 weeks of inpatient feeding of a eucaloric, very low carbohydrate, ketogenic diet (5% Carbohydrate, 15% Protein, 80% Fat) *immediately following* an inpatient period of at least 4 weeks of consuming an energy balanced standard American diet (50% Carbohydrate, 15 % Protein, 35% Fat).”

        1) Bad design?
        2) Why did nobody at NuSi reject this design?

    • CynicalEng

      Shame the “eucaloric, very low carbohydrate, ketogenic diet” wasn’t eucaloric either. Shouldn’t the order of diets be randomised in these studies, which would address Bill’s issue about the rate of loss vs time of “restriction”.

      • Wab Mester

        Randomized order would have been ideal, but it wouldn’t have been a factor if the SAD had correctly maintained weight. The main points were to give the LC diet more time for “adaptation” vs his previous experiment and to test for increased EE on the LC diet. Those objectives were met, but they were tainted by the weight loss on the lead-in diet.

        FWIW, my diet is relatively low carb. I figured there was perhaps a metabolic cost to ketone generation and to GNG, but I never gave that any consideration in terms of efficacy.

        Is that cost 100kcal? 300kcal? Are there other mechanisms that would increase EE on LC or decrease on HC? This study didn’t move the needle.

        • CynicalEng

          If the objective is to understand the energy expenditure in weight stability on LC vs SAD then there needs to be a run-in period on the LC diet that isn’t part of the dataset to move from one regime to the other. Otherwise as with his previous 6 day thing he’s analysing a period of change rather than comparing two periods of stable intake/expenditure.
          There seems to be a blindness to the fact that the subjects are effectively run-in (for months) on the SAD on account of it being closer to their habitual intake.

  • rs711

    Good analysis Bill

  • Nagar J

    Also, this was explicitly only a small pilot study, not at all intended to settle the scientific questions but rather to learn how to do the real, far more expensive, study. For that reason it’s fine that the study was completely uncontrolled. I’m certain the full scale study, should it take place, will be. You don’t need controls to study methodology. You do need them to test hypotheses.

  • I have noticed much of what Vox Day has written about social justice warriors seems to apply to these other fields. Taubes and friends go out of their way to be objective by funding Hall and his ilk, but Hall doubles down despite apparently having nothing to back himself up with. Even what little we know about the study design makes it unclear that it sheds any new light on anything- except that people like Hall deserve ridicule, not funding.

    • FromPA

      i think NUSI (I guess “NuSI” is the way they spell this) was trying to fund people who might not agree with what they were setting out to do. Not sure that makes sense, now. Perhaps Phinney and Volek would’ve been better?

      And I’m not sure this study is worth anything. Hasn’t this been done before? What was NUSI trying to prove? In particular, how does this study meet their goal of “CONCLUSIVE EVIDENCE IN THE NEXT DECADE”? To me, this study seems like a waste of money.

  • Bill –

    I know you’re a busy man and all but I had two questions for you –

    1) Why doesn’t the ICM support insulin as a satiety hormone? The brain is insulin resistant in a lot of obese people – and this sort of data would actually support ICM. But proponents are hostile towards the “brain” regulating fat/satiety. What’s wrong with that model?

    2) Why bet on metabolic advantage? Do we need to bet on it?

    I feel like the real problem here is that LC diets are not broadly accepted by medical docs and dietitians and are rarely in my experience recommended for people with Type II.

    Shouldn’t that be our focus? Instead of metabolic advantage???

    • 1) the science supporting “insulin as a satiety hormone” & IR brain in human obesity isn’t strong

      2) no, actual outcomes should be our focus

      • Thanks Brofessor!

        • interesting edit. LOL

          • Seriously Bill – you seem like such a regular guy and are so accessible I lose sight of the fact that you’re a legit scientist 😛 So I had to use the “Brofessor” honorific to show some respect.

  • valerie

    If there was a rapid weight loss on the first diet, it would have been during the first two weeks, which are excluded from the comparison.

    More importantly, they lost half a kilogram over two weeks (both in weeks 3–4 of the study, and in weeks 7–8 of the study). Do you really believe that losing one poor pound is going to slow their metabolism?

    • “If there was a rapid weight loss on the first diet, it would have been during the first two weeks”

      this usually refers to water weight on LC… but I’d still expect weeks 3-4 to show better fat loss than weeks 7-8

      “Do you really believe that losing one poor pound is going to slow their metabolism?”

      not “one poor pound,” but rather 3-4 weeks of caloric restriction vs. 7-8 weeks

  • Wab Mester

    Just for fun, I tried to estimate the increase in EE due to GNG and ketogenesis.

    Estimate for GNG efficiency is about 67%.

    Estimate for ketogenesis efficiency is about 80% (this comes via Kevin Hall in a communication to Peter Attia, BTW).

    Assuming the brain needs 500 kcal/day and ketones can provide 2/3 of that, we need a basal ketone supply of 335 kcal. That implies we’ll generate about 80kcal/day of waste energy just to supply brain ketones.

    Hall saw protein losses, but he didn’t give details. Phinney gives details. Max loss is around 25g/d in the first day, and that goes to zero after about 3 weeks. So GNG and protein catabolism are at least partly responsible for that peak in delta-EE in the first week.

    25g protein input into GNG would only waste about 33kcal in theory.

    So I can account for 80-113 kcal/day increase in EE for the keto period.

  • Colin Champ

    well put Bill, as usual

  • the study was published and my initial take remains relatively unchanged… except that it may have added the importance of dietary protein on a ketogenic diet. Keto is not muscle-sparing and may even be the opposite. Urinary nitrogen, urea, and ammonia were increased on keto relative to controls; this indicates muscle-wasting.

    Muscle growth sans carbs, anyone?
    http://caloriesproper.com/muscle-growth-sans-carbs/