A brief explanation of Hall et al., ie, THE LOW CARB WAR

“Examination of acute shifts in energy balance by selectively reducing calorie intake from one macronutrient.”

Intro (1/2): please don’t read this study with the media headlines in your mind.  Don’t even pay any attention to the study’s title, abstract, intro, and discussion.  In no way did this study put low carb proper on the chopping block, regardless of what you’ve seen online or elsewhere.  Mmmkay?


Intro (2/2): if you want a lesson (or refresher) in Advanced Nutrition, check out the Supplemental Information: in formulating his mathematical models, Dr. Hall seemingly reviewed every single biochemical pathway and physiological variable ever invented.  Read it, for science.  Really.



Tl;dr: this study didn’t break low carb (that’s a ridiculous conclusion).  But it was pretty cool, and probably included the most accurate and comprehensive assessment of energy balance EVER.


Outline: baseline diet for a few days, than randomized to either Low Fat or Low Carb for 6 days.  Washout, then Crossover.  All in a Metabolic Ward.





Please stop complaining about the composition of the intervention diets!  they weren’t designed to mimic a proper low fat or ketogenic diet. They were designed with a very specific goal in mind: to selectively reduce the calories from ONE macronutrient without increasing any others, so that any effects could be attributed to the selective reduction of said calories.  Other studies compensate, for example, but swapping carbs & fat, but instead this study left us with a 30% caloric deficit.  Which is fine.



Step 1.  Constructing the low fat diet.

Reducing fat from 109 grams (981 kcals) to 17 grams (153 kcals) is pretty much the biggest reduction possible, and amounted to ~828 fewer kcals from fat.


Remember, it was a 30% energy deficit: 828 / 2740 = 30%.


Theoretically, they could’ve tested as little as zero carbs on the Low Carb diet, but they were limited by how much they could decrease the fat, and you can’t get much lower than 17 grams of fat/d.  THAT IS WHAT SET THE BOUNDARIES FOR STEP 2 –>


Step 2.  Constructing the low carb diet.

There are 2 limits that step 1 put in place (which is really just 2 different ways of saying the same thing): 1) 30% energy restriction from ONE macronutrient; and 2) ~828 kcals that need to be disposed of.

840 kcals of carbs = 210 grams.  So the low carb group went from 350 grams carb at baseline to 140, a reduction of 210 grams (840 kcals).  It’s not ketogenic (see Table 4), but it is “low carb” – come on guys, they slashed carbs from 350 to 140; that’s gonna leave a mark.






Why not zero carb?  Because 350 grams of carbs = 1400 kcals, and you can’t possibly reduce fat kcals by that much because there were only 981 fat kcals to begin with.  If they did go lower carb, protein would’ve had to decline in the Low Fat group to keep the diets isocaloric… and that would’ve been a MAJOR confounder.


comprehensive diet


This study rigorously controlled “calories in,” and rigorously measured “calories out.”

Energy expenditure declined in both groups (as expected), slightly more so in low carbers.  After 6 daysEbbeling showed that this is reversed by 4 weeks, or approximately enough time to adapt to the new fueling scenario.


The changes in body composition are not worth mentioning, but have a look if you’d like:


Body comp & energy metabolism


Popular criticisms:


-17g fat/d is impossibly low fat (8% of calories).
I agree, but whatevs! my stance is that regardless of how the authors & journalists spin it, this study wasn’t designed to test real life diets.  Ad libitum (ie, real life diets) is completely different from rigorously controlling every calorie that goes down your pie hole in a Metabolic Ward.
Full stop.


-This study was comparing super low fat to modestly low carb.
Again, I don’t care [in the context of this study design]!  This was the only way they could do it, mathematically… so if this is your criticism, I think you’re targeting either the authors or journalists, not the actual study  –> better title may have been: “Examination of acute shifts in energy balance by selectively reducing calorie intake from one macronutrient.”


-The biggest actual flaw, imo, is that they didn’t achieve the reductions in insulin they were shooting for.
Fasting insulin & c-peptide weren’t different between the groups, and only when 24h c-peptide was assessed did they reach statistical significance.  Maybe lower carb would’ve done it, but as mentioned above, they went as low as they could go given the restraints of the study design.  Maybe longer duration would’ve done it, but: 1) this was already the most expensive study EVER; and 2) you can’t really keep people locked up in metabolic chambers for much longer… so they would’ve had to sacrifice accuracy, and I don’t think Dr. Hall would’ve been very happy about that.



Speculation:  OK, so if LF actually DID beat LC (in 6 days) (which it didn’t), here’s a [fringe] hypothesis about how it may have happened: at baseline, participants were weight stable: eating 109 g fat per day & burning 109 g fat per day.  Suddenly drop fat intake from 109 to 17, but body is still burning 109.  It takes a few days to downregulate fat oxidation to match intake & re-establish homeostasis.  Glucose oxidation doesn’t have this problem: drop carbs from the diet and glucose oxidation shuts down rather quickly.


Fasted blood metabolites and hormones


This is only a problem if your study is 6 days.  The authors almost addressed this by showing the RQ dropped when switching to low carb, which reflects decreased carb intake, but it hadn’t quite reached the new Food Quotient which should be just south of ~0.76 on a 50% fat diet with 30% energy deficit.


respiratory quotient


calories proper


Be Sociable, Share!
  • George

    Nice speculation Bill. Mine was slightly different and left in my comment on Zoe Harcombe’s blog.

    However, was there really any reason, beyond cussedness, why fat and carbohydrate weren’t evenly balanced at say 40%E in the baseline diet?
    That would have allowed even reductions of both macronutrients.

    • http://americalaidoff.com/ MichaelGWS
    • http://www.caloriesproper.com/ Bill Lagakos

      “That would have allowed even reductions of both macronutrients.”

      Run the numbers; I don’t think this is the case.

      Gotta play by the rules: can only change 1 macronutrient per diet. Isocaloric energy deficit of ~30% in both groups.

      • http://www.caloriesproper.com/ Bill Lagakos

        correction: if the baseline diet is tweaked, more “reasonable” LC & LF diets can be constructed,
        although I don’t think this would’ve made a big difference because there’s still the problem of ACUTE changes in “Fat Balance”
        (see “Speculation,” above).

      • Antonio

        The numbers ruled http://goo.gl/XD35CR

        • http://www.caloriesproper.com/ Bill Lagakos
          • Antonio

            not acute enough?

          • http://www.caloriesproper.com/ Bill Lagakos

            Decreased fat intake on the LF diet is gonna create a bigger ACUTE “Fat Imbalance” than ?Fat Oxidation on the LC diet.

  • Pingback: Why the New “Low Fat Beats Low Carb” Study Is a Clear Sign of the Apocalypse | Escape From Caloriegate()

  • http://americalaidoff.com/ MichaelGWS

    Bill –

    If the participants were insulin resistant or diabetic do you think it would have made a difference?

    • http://www.caloriesproper.com/ Bill Lagakos

      not very much (in this context):
      -6 days
      -metabolic chamber
      -controlled food intake

      doesn’t leave much room for differences in adherence or appetite

      • http://americalaidoff.com/ MichaelGWS

        Got it. After reading the negative press on low carb the last few days I’ve decided to adopt 30 BAD.

        Can’t deal with the negativity 😉

        • http://www.caloriesproper.com/ Bill Lagakos

          [that escalated quickly]

  • valerie

    Hi Bill,

    Thanks for posting some tables from the paper (I don’t have access). In table 3, there are two lines I don’t quite understand:
    (Change in) Fat mass
    Cumulative fat imbalance

    Do they explain what each means? Why are they so different? Where did the fat imbalance go, if not in fat mass?



    • CynicalEng

      Cumulative fat imbalance is calculated from respiratory data and food intake, change in fat mass from weighing and DEXA.

      • valerie

        Any idea why they don’t match? (They should match, right?)

        • CynicalEng

          in women they were close – about 10% out, in men a long way apart. Difference between a weight / DEXA measurement and a mathematical estimate from respiratory gases. Short time and small numbers don’t help, make errors more significant.

    • http://www.caloriesproper.com/ Bill Lagakos

      Fat Mass = DEXA results

      Fat Imbalance = Dietary fat intake – fat oxidation (calorimetry)

      “Where did the fat imbalance go, if not in fat mass?”
      Two ways to cause a greater fat imbalance are lower fat intake and higher fat oxidation.

      It might be easier to think about these numbers as ‘change from baseline,’ instead of immediately comparing RC to RF.

      • valerie

        Sorry, but that didn’t help…

        Let me try this again. Cumulative fat imbalance = fat in – fat out. That should be exactly equal to the change in fat mass, right?

        Why the big discrepancies? Why more discrepancy (lots more) in RC than RF?

        I am guessing that the measurement tools were inadequate (errors/uncertainties too big) for the kind of measurements they were trying to make. Any other explanation?

  • RIchard Feinman

    In science, we usually try to address the current problems. Volek has done a classic experiment clearly bringing out the value of low-carb in people with metabolic syndrome. If you think there is a limitation that could be improved with a metabolic ward, that’s the one to do. These guys define low-carb by what they do. Their disdain for low-carb researchers is not matched by any good experiments. The data in Hall’s study is reported as means and SEM, if he Showed individual data, you would probably see that nothing was accomplished here. To get a rough idea, convert SEM to SD and ask yourself about the spread of data. And he doesn’t cite Volek because he doesn’t have to. IMHO waste of expensive methodology but maybe I am wrong. I will ask him to publish the individual data points.

    • Stipetic

      Good point, Dr. Feinman. The individual data is key. A single outlier could be providing the entire advantage seen in the low fat side. Regardless, it still won’t address the issues surrounding the short duration of the study, and its implication over the conclusions, but would provide useful interpretive information. And Volek’s study is a classic.

  • CP

    Just noticed the TG:HDL ratio was quite good for the test subjects. So, despite being obese, the test subjects were likely relatively sensitive to insulin. As you’ve pointed out before, IS people tend to do better on LF diets. However, most obese people are IR…

    • Thomas Hemming Larsen

      Exactly what I was thinking. IS could be a confounder.

      • CynicalEng

        I spent too much time in the supplementary materials, but at least I didn’t read any of the press. The women were insulin resistant, and the men insulin sensitive. The women’s fat loss (DEXA) was not significant. http://lowcarbuk.blogspot.co.uk/2015/08/short-term-weight-loss-in-insulin.html

        • http://www.caloriesproper.com/ Bill Lagakos

          I don’t think insulin sensitivity was involved here…

          -crossover study

          -6 days
          -metabolic chamber
          -controlled food intake

          doesn’t leave much room for differences in adherence or appetite

          • CynicalEng

            I bow to your greater knowledge. My observation was that DEXA didn’t see a fat mass change in women (HOMA-IR 3.3 vs 2.2 in men) and I’m reminded of the study where IS and IR subjects did better when matched to the LC or LF option that matched

          • http://americalaidoff.com/ MichaelGWS

            If they’re all insulin sensitive how does the cross over fix that?

          • http://www.caloriesproper.com/ Bill Lagakos

            in that case, it doesn’t matter.

            In theory, the crossover either fixes it (if mixed population) or doesn’t matter (if homogeneous population)

  • Thomas Hemming Larsen

    Good write up Bill. Somewhat off topic, do you have any references on how good Hall’s model is at estimating caloric expenditure, body weight changes etc. over the long term?

    • http://www.caloriesproper.com/ Bill Lagakos

      Check out the Supplemental Information.

      • Thomas Hemming Larsen

        Ah sorry, I could have figured that out myself. It looks like its pretty accurate over a range of time periods.

        • http://www.caloriesproper.com/ Bill Lagakos

          actually, I highly recommend reading the Supplemental! It’s pretty epic :-)

          • Thomas Hemming Larsen

            I have looked at it before and went over it briefly again. Its a pretty big model. The problem is that I understand the math but not the biology :)

            Again, I was surprised to see how accurate it has been in predicting weight, metabolic, oxidation etc. changes.

          • http://www.caloriesproper.com/ Bill Lagakos

            I liked all the stuff about specific rates of lipolysis, ketogenesis, protein turnover, etc, in actual “grams per day.” Probably not 100% accurate in every context, but pretty cool to have ballpark figures.

          • Thomas Hemming Larsen

            Yeah, its pretty cool that you can see it in grams. It reminds me of the economic models I used to study at university.
            I think I have to print for a rainy day :)

  • CynicalEng

    My issue with the study is that it does not achieve a steady state – calories are being drawn from carbohydrate reserves, suppressing fat loss. The Reduced Carb diet massively increased fat oxidation but the glycogen depletion reduced the demands placed on fat stores.

  • CynicalEng

    What about the carbohydrate depletion ? 400 kcal on day 1 and still 140 kcal on day 6 of RC. That’s about 8.5 days to equilibrium or some 200 grams of fat equivalent during the six days.

  • This Old Housewife

    This article/news flash was just a piece intended to keep low-fat diet recommendations front and center in people’s minds–probably instigated by the low-fat Heart Hypothesis crowd (what remains of it). Everybody who’s anybody knows it takes longer than 6 days to become fat adapted, so I’m looking at a lot of the “issues” as cheating in the same ways Ancel Keys did way back when to “prove” his point about sat fats, cholesterol, and heart disease.

    This is a clear sign of desperation from an entire legion of people: the medical establishment, the processed food makers, Big Pharma, Wall Street, and our economy as a whole. This runs along the lines of the recent Coke article saying that people were obese because of insufficient exercise, not sugar consumption.

    I, for one, will be glad when this last vestige of the Matrix we’ve been living in for the last 70 years or so comes crumbling down to expose the little man behind the curtain called CORRUPTION and FRAUD.

    • James IV

      Pull out your tin foil hats!

      • http://americalaidoff.com/ MichaelGWS

        I love Tin Hats.

        • http://www.caloriesproper.com/ Bill Lagakos

          and blue blockers

  • Pingback: 7 Gründe warum echtes LowCarb doch besser ist als LowFat | Leichter Gesund Leben()

  • Guido Vogel

    Though I agree with many of the comments here, the study demonstrated that you CAN lose fat without cutting carbs. Right?

    In a 30% calorie restriction with extreme low-fat (7.7%) and baseline carbs (29%) fat is burnt, weight is lost and insulin goes down.

    If so, what is the metabolic pathway? Either via

    1) Reduced insulin secretion (can be done without carb reduction)
    2) Another metabolic pathway
    3) Combination of 1) & 2)

    • CynicalEng

      Good questions, and yes the men on reduced fat lost about the same as they did on reduced carb. Their fasting insulin went down on RF but not on RC. RC so a drop in 24h insulin secretion.

      • http://www.caloriesproper.com/ Bill Lagakos

        fasting insulin declined in both groups (Table 4)

        • CynicalEng

          I stand by what I said. *Men* on reduced fat and *Women* on reduced carb had fasting insulin declines. The opposite didn’t. Table S2.

    • http://www.caloriesproper.com/ Bill Lagakos

      “the study demonstrated that you CAN lose fat without cutting carbs. Right?”

      yes, just need an energy deficit… can do it on many different diets.

      • http://novuelvoaengordar.com Vicente

        The study demonstrated the authors’ bias. They want us to BELIEVE short-term fat loss is “fat loss”. No, it isn’t. Long-term success is the relevant outcome here. A 6-day experiment is a joke. Only zealots can’t see that.

        If low-calorie diets worked for long-term weight loss, I guess we would have already seen a successful weight-loss study. We haven’t. The most reasonable explanation is that it is a wrong idea.

  • Guido Vogel

    Table 4 shows insulin data:

    Insulin (pg/ml)

    Baseline 12.6 ± 2

    Reduced Carbs: ?2.76 ± 0.77

    Reduced Fat: ?2.04 ± 0.8

    So the low-fat group reduced insulin almost as much as the reduced-carbs group, with the same amount of carbs.

    It is possible to burn fat and reduce insulin without cutting carbs. That seems to be an interesting finding even though the low-fat diet is not a real-life diet.


    1) How does low-fat diet reduce insulin if carbs are constant in relation to baseline diet?
    2) Is reducing insulin still the only mechanism or are any other mechanisms at work? If yes which?

    If the researchers wanted to validate whether carb reduction was required to burn fat, why introduce a low-carb group? Why not all 19 on low-fat? That would have generated more robust data.

    • CynicalEng

      insulin wasn’t high to start with, the men were lower than the women (10 vs 15.5) initially. The women dropped insulin on RC and the men dropped on RF. An 800 calorie restriction is a lot less food which will be part of the story behind reduced BG and insulin.
      It was a crossover study, so most of the subjects did both diets.

      • http://americalaidoff.com/ MichaelGWS

        Fats can raise your insulin because free fatty acids have to be disposed from your blood stream and insulin in the regulator of such stuff. I know from personal experience (have T2D) but might have stated this incorrectly. This is why very low fat diets work for T2D also.

        I read Dr. Hall’s blog post on Weighty Matters and was surprised that in fact, this whole thing was just to debunk and make Taubes look like an idiot.

        This study effectively proves that Gary Taubes is the most hated man on the Internet.

        • http://www.caloriesproper.com/ Bill Lagakos

          The authors actually cited Taubes & GCBC in the paper! (I found that surprising for a study published in Cell Metabolism)

          some more info on dietary fat & insulin: http://caloriesproper.com/insulin-dietary-fat-and-calories-context-matters/

          • http://americalaidoff.com/ MichaelGWS

            Our tax dollars at work !

          • Guido Vogel

            In Dec. 2010 Gary Taubes hypothesized in this article (http://garytaubes.com/2010/12/calories-fat-or-carbohydrates/)

            “The science of fat metabolism strongly implies that the best way to do that [lose fat, GV], if not the only meaningful way, is by reducing the amount of carbohydrates consumed and/or improving the quality of those carbs we do consume.”

            One can discuss what “meaningful” means, but reading Taubes’ article, I think Hall tried to address many of the problems with nutrition research Taubes points out in his article. So I think it is fair to say Hall tested Taubes’ hypothesis (which is too vague to really test scientifically) about the necessity of carb reduction for fat loss.

            Both apparently overlooked existing research (or it is after Dec. 2010 so Taubes couldn’t know by then) that shows that meaningful fat loss is possible without reducing carbs.

            So Bill, please enlighten us and point us to the research :-)

          • http://www.caloriesproper.com/ Bill Lagakos
          • Guido Vogel
    • http://www.caloriesproper.com/ Bill Lagakos

      carbs aren’t the only insulinogenic nutrients. And carb reduction obviously isn’t *required* for fat loss; just need an energy deficit.

      • Guido Vogel

        If you say “carb reduction obviously isn’t *required* for fat loss; just need an energy deficit.”, does that mean the first research question didn’t make sense to start with (as the outcome has this has been demonstrated before)?

        “While the first law of thermodynamics requires that all calories are accounted, could it be true that reducing dietary fat without also reducing carbohydrates would have no effect on body fat?”

        • http://www.caloriesproper.com/ Bill Lagakos

          the authors put forth somewhat of a strawman, saying they intended to disprove the theory that carb reduction was REQUIRED for fat loss.

  • Pingback: Tedenski pregled #15()

  • http://twitter.com/mashdot ‘Mash (Tom Herbert)

    Morning Bill,

    “It takes a few days to downregulate fat oxidation to match intake & re-establish homeostasis.”

    This was an interesting comment, where can I look to understand where this comment comes from?


    • http://www.caloriesproper.com/ Bill Lagakos
      • http://twitter.com/mashdot ‘Mash (Tom Herbert)

        Were you being facetious, or mispasted? :)

        The context was “Suddenly drop fat intake from 109 to 17, but body is still burning 109.” which sparked my interest as I hadn’t read it like that before.

        But it makes sense in light of the VLC with CHO-loading before competition papers.

  • Christopher Kelly

    Hi Bill, thought you might like to see my latest podcast.

    Why report averages? It’s not that much data.

    To quote Dr. Feinman from the transcript:

    Well, it’s worth my attention because I know Kevin Hall and he’s a pretty smart guy and this is a pretty distressing paper. One of the reasons is that all of the data is reported as group averages. And the problem is that nobody loses an average amount of weight. The assumption of group averages is that there’s a normal distribution or at least essentially limited distribution and the underlying idea behind that is that people are roughly the same and that minor variations account for the spread of the data. Those aren’t good assumptions.

    But they’re okay if the data do not have big variation. Now, that’s not true in this study.

  • billslo

    First, metabolic “chamber” doesn’t equal metabolic “ward.” They were only in the chamber for half the 6 days of either diet. “…this study wasn’t designed to test real life diets” yet they disingenuously extrapolate out to 6 months of perfect adherence to either diet to get their estimated results.
    Also, real world is not represented well by such young healthy study participants.

  • Pingback: Latest in Paleo 141: Moore & More | – Humans Are Not Broken()

  • RIchard Feinman

    It is important to stress that fat loss was not measured. Fat oxidation was calculated from gas exchange data, that is, from indirect calorimetry. This is based on several assumptions laid out in the paper and the data can be accepted insofar as those are good assumptions but they are not tested and if anything else happens to fat than the processes mentioned, the data is not accurate. This would not matter much if the output was large but the output data changes are extremely small.

    The real problem is that, in most sciences, disagreement involves two parties talking to each other and defining what the problems are and what would be acceptable experiments. In nutrition there are two worlds, separate and largely hostile. This shows up, in this case, in Kevin Hall setting up a straw man, “weight loss REQUIRES carbohydrate restriction.” He suggests that inside the straw man is Gary Taubes but Gary has the freedom to talk loosely at least in most venues and no researcher thinks that he speaks for the field. It also allows him not to cite Volek, Westman, Dashti, and all the other studies that are currently of importance, a clear breach of scientific protocol.

    Just thought I would mention those points.

  • Pingback: Nuevo metaanálisis sobre dietas bajas en carbohidratos, adelgazamiento y riesgo cardiovascular | Metodo Quema Grasa()