The First Law of Thermo still doesn't violate The Second Law

Well, I've been meaning to weigh in on the Bray overfeeding study (Effect of Dietary Protein Content on Weight Gain, Energy Expenditure, and Body Composition During Overeating) that made a bit of a splash a little while back. George Bray is especially hated by ardent low carb advocates because he wrote a particularly unfavorable review of GCBC published in Obesity Reviews.   Bray is what Adam Kosloff refers to as a "Calorie Wizard":

Regardless of diet, it is a positive energy balance over months to years that is the sine qua non for obesity. Obese people clearly eat more than do lean ones, and food-intake records are notoriously unreliable, as documented by use of doubly labelled water.  Underreporting of food intake is greater in obese than in normal-weight people and is worse for fat than for other macronutrient groups.  Accepting the concept that obesity results from a positive energy balance does not tell us why energy balance is positive.  This depends on a variety of environmental factors interacting with the genetic susceptibility of certain individuals.  Weight loss is related to adherence to the diet, not to its macronutrient composition.

Interestingly, the emphasized passage is often glossed over by offended low carbers who insist that Bray is just blaming them for their obesity and accuses them of lying and whatnot.  Now while I don't necessarily agree with Bray on all points, he responded to the 10 Critical Conclusions in GCBC.  Newer readers might want to read my own response to those:  CarbSane on The 11 Critical Conclusions of Taubes.  (Note depending on the book or pubisher's site, there are 10 or 11 conclusions with slightly different wording).   This critique apparently ruffled the feathers of one Dr. Mike Eades who offered up his blog for Taubes to respond.

In any case, Dr. Richard "there's no such thing as an essential carbohydrate" Feinman, seems especially upset that somehow Dr. Bray didn't listen to him when he told George Bray how to do it right.  Really, is there any humility ever to be found amongst these low carb gurus?  Feinman is probably right that it was never about the dietary protein but carb vs. fat.  Still ... who the frig is this Feinman guy to think he should be telling Bray how to conduct his studies and what it is he's looking at anyway?  Get your own funding and conduct your own studies.   What ... the hit campaign on Hope Warshaw didn't generate tons of funding for your Nutrition & Metabolism Society or something?

So Bray is one of those perennial "bad scientists" targeted by the low carbers who know better how to do "proper studies".  Interestingly enough, studies like the one I'm about to discuss would otherwise be described as of the highest quality by the likes of Dr. Andreas Eenfeldt and J. Stanton ... the "gold standard" randomized controlled trial.    And a metabolic ward study at that.  Bray is always doing something wrong because he's not controlling for what the low carbers think he should control for and is hopelessly biased by his "flawed" acceptance of calories.  Ummm ... would it be fair at this point to say that there is often bias in the design and reporting of studies by the likes of Volek and Westman?  But I digress ...

So Feinman writes:

The debate is also about calories. Should you cut calories or just cut out carbs? Is it really “excess calorie consumption” and not the effect of excess carbohydrates ? “A calorie is a calorie” or not. Many of the gurus have gone beyond “claiming” to demonstrating that when carbohydrates are low, weight loss is greater than when carbohydrate is high and that the weight loss on a low-carbohydrate diet is primarily in fat stores rather than lean mass. In head-to-head comparisons, for however long they are compared, low-carbohydrate diets generally out-perform low-fat diets on other parameters as well, glycemic control, the features of atherogenic dyslipidemia. This has been the major challenge to traditional nutrition and the general approach has been to simply ignore this data and dismiss the researchers with innuendo as above.

Unless of course they are conducted in metabolic ward settings and both calories and protein content are controlled for .....   Now what's really odd is that in this first blog post, Feinman acknowledges that Bray holds carbohydrate constant, and somehow this is the reason that Bray could conclude that it was calories and not something else that impacted body fat gain.  As one can see (scroll down, first set of scatter plots), calorie increase over baseline correlates with body fat increases while protein intake does not.  Feinman concludes:

In some sense, Bray, et al. answered a question that we weren’t asking, but protein is important if more complicated than carbohydrate and fat.

I wonder if there is anything stopping these high brow scientists from doing the study they want to see done.  That is, just overfeed, just underfeed, control for carb or protein or fat and see what happens.   I'll refer y'all to Anthony Colpo's work demolishing Eades ... this is so "been there done that" already!

So, let's discuss the study that Bray et.al. did do over the objections of know-it-all's who think they know better.  In this metabolic ward study, they determined weight maintaining caloric intake over a period of 13-25 days on a standard diet P/F/C% of 15/25/60%.  The average maintenance calories for all participants was 2414 cals.  Then they overfed ~40% of calories which amounted to an increase of ~950 cal/day for total caloric intake averaging 3375 cal/day.  The carbohydrate content of the overfeeding diets was fixed by percentage at ~40%.  The remaining ~60% of calories were split between protein and fat.  So the three overfeed diets, low to high protein P/F/C% were:  6/52/42, 15/44/41 and 26/33/41 respectively.  The overfeeding lasted 8 weeks.

Several others have discussed some of the implications of this study, you might be interested in Prof. Andro's take and Anthony Colpo's.   The outcomes in this study were pretty interesting and I believe somewhat surprising ... to some at least.  Essentially the high protein group gained more weight, but the difference was in lean mass (the low protein group actually lost a bit of lean mass).   I'm not going to discuss that in this post although I have some thoughts on this for another day.

You see, a Bray study is bound to rile up the likes of Dr. Richard Feinman, and ... predictably it did.  And my friends, it has resulted in an acute bout with the Taubesian affliction of Egomonomaniacalis dunshütinfüt.  Because, you see, in his next blog post, Feinman concluded:  Bray, et al. Shows that a Calorie is Not a Calorie and that Dietary Carbohydrate Controls Fat Storage.  Really?  Umm... No.  Before I discuss Feinman's erroneous conclusions, let's review that masterpiece of thermodynamic literature that is "A calorie is a calorie" violates the second law of thermodynamics.  I contend that such an abomination concerning the laws of thermodynamics could only be published in a Nutrition journal where they were able to sufficiently baffle their peers with bull puckies.  Because so poor is the understanding, apparently, of the laws of thermodynamics that the authors (Feinman and Fine) contend that the First Law of Thermo (TFLOT) violates the Second Law (TSLOT).   The entire discussion of calories and thermogenesis presented as a TSLOT phenomenon is actually a perfectly valid TFLOT analysis.  I have discussed this many times before here.  Thermal energy, such as that generated in the metabolism of various macronutrient molecules, is a perfectly valid "calories out" term in a First Law energy balance equation.   Indeed F&F make a TFLOT argument in the paper which is the most perplexing thing about it.  

The crux of their thesis uses the average macronutrient thermogenic factors to make the case that "calories" of half fat/half protein do not equal an equivalent amount of "calories" of carb.  Basically, using Jequier's factors -- 2.5% for fats, 7% for carbs and 22.5% for proteins -- F&F do calculations based on a 2000 cal diet beginning with P/F/C of 15/30/55.  They then reduce carb replacing it with equal amounts of protein and fat to arrive at an effective caloric deficit of 100 cals when carbs are reduced to about 20%.  There are some errors in their calculations, and I've addressed this part of their paper before here:  Metabolic Advantage, Obesity and Eric Jequier.  It's worth mentioning that this whole argument falls apart and easily blows this part out of the water if one replaces carbohydrate with fat on a 1:1 basis -- because fat is less thermogenic than carb.   This is what I find so utterly and in flabbergastingly (is that a word?) disingenuous about this part of the argument for VLC/VHF diets, most especially in the recent environment of describing the diet as "high fat, moderate protein" and where protein restriction/control is all the rage for eating a "proper" low carb diet.

So, first let's do an F&F calorie analysis on the Bray overfeeding study.  In the spreadsheet below, I've used data from the study on intake and percentages to breakdown the macronutrient caloric intakes and absolute intakes in terms of grams.  The last column is for the "effective calories" that F&F argue for in their paper using Jequier's factors.  Note that values in ( ) are negative, L=low, N=normal, H=high, P=protein.

Now, if the argument put forth by Feinman and Fine in their Nutrition article holds, the high protein group should have gained less weight and fat than the other groups because using their thermogenic calculations, the effective caloric surplus was about 145 cal/day less than the normal and low protein groups.   Before moving on, I would like to clarify that from my understanding of how Atwater's caloric values were obtained, the thermogenic contribution is already built into the equation -- IOW the 4/4/9 were never "bomb calorimeter" values, they were for human metabolizable energy.  I could be wrong on this but when time and time again macros are varied by large percentages and CICO holds up, this tends to support that this is already accounted for.   In any case, folks can't have it both ways.  Either the calorie/metabolic advantage argument put forth in the F&F paper is valid or it is not.  It looks like the case made in the F&F paper is wholly contradicted by this Bray et.al. overfeeding study.  

When one takes the time to look at the actual compositions of the diets in the Bray study, we see that carbs are held constant in absolute amounts throughout the study.   That is, not only in the overfeeding arm, but even between the weight stabilizing and overfeeding periods.  It was fat and sometimes protein intake that increased during the overfeeding period with the percents P/F ranging from 6/52 to 26/33. 

So I'm not done with Dr. Feinman's brilliant analysis of Bray.   What he claimed in his blog post is one for the ages folks.  All of the diets -- weight maintaining and overfeeding -- consisted of about the same amount of carb.  In fact, if you look at the last rows of the second column from the right, you see that all groups even cut carbs a tad while overfeeding, and the group that gained the most weight cut the carbs the most!  Now here's what Feinman says, in his rejected Letter to the Editor of JAMA:  

In Bray’s experiment, carbohydrate was a fixed percentage of energy and therefore the level of this macronutrient is directly proportional to the number of calories. Thus, while the results are consistent with the primacy of calories, they can just as accurately be interpreted as control by carbohydrates.

Really?  Perhaps to a scientist like Feinman blinded by agenda, but not to any objective analyst.  Because it's true that for any given subject eating ~40% carb, one consuming 3400 calories would consume more carb than one consuming 2400 calories.  And it's indeed true that a subject in the low protein group is consuming fewer protein and more fat calories than a subject who is weight stable at the same caloric level who was assigned to the high protein group.  But when you look at the numbers, this study does not in any way, shape or form demonstrate that carbohydrates control fat accumulation.  The degree of caloric excess does not correlate with increase in carbohydrate consumption.  

So, NO, Dr. Feinman, you can't just replace the x-axis label of calorie excess with carbohydrate excess, as you did in your blog and letter to the editor.    I've hot-linked Dr. Feinman's contribution to fun with graphics manipulation below.  At least he didn't cherry pick data points from, say, half of the study participants.   But the table above clearly shows you cannot replace 

with

Because all of the excess calories for ALL participants came from fat and protein; entirely from fat for the lowest protein group, mostly from fat in the normal protein group (about 85%), and roughly 50:50 from fat and protein in the high protein group.  Too bad Feinman didn't actually do the math on the carbohydrate content of the diets, eh?  One does wonder how he got the numbers for the x-axis though ...

Bray's study ultimately demonstrates that gain in FAT mass is due to caloric excess, PERIOD.  Try as Feinman may, one cannot pin any gain on the lowly carbohydrate because, and I repeat, carbs are essentially held constant throughout this study.  I'm not sure if this was by design for the weight stable vs. overfeed phases, but nonetheless, this study looked at the effect of fat and/or protein overfeeding on body mass and composition.  Although they gained the least on the bathroom scale, the low protein group had slightly higher fat mass gains (not stat. sig) and actually lost lean mass despite being in considerable caloric surplus.  The high protein group by contrast gained slightly less fat mass, but considerably more weight as they tacked on lean mass.  I'll leave team TWICHOO to ponder over the calculations and worry over damaged metabolisms, somesuch and whatnot, and I do plan to speculate on a few things at some later date.

Oh ... and paging Mark Sisson, all participants were weight stable on roughly 350g carb per day -- that's off the primal carb chart right?  Lowering carb by percentage while increasing caloric intake resulted in all participants gaining weight consuming roughly the same 350g super fattening, insulin spiking, blood glucose surging carb per day (slightly fewer).  I'm repeating myself a bit here -- OK a lot -- but what Bray did -- intentionally or not -- was test some of TWICHOO!  Bray demonstrated:

• That one cannot be a protein and/or fat glutton and not gain weight.  Sure, TWICHOOB's will blame it on the "base" of 350g carb, but there's no getting around that these folks gained weight when protein and fat were added to their weight stable diets.  
• That higher postprandial insulin did not increase fat accumulation.  Rather the most insulinogenic diet (high protein group) did gain more weight, but this is because energy was partitioned to considerable lean mass gains (3.18 kg) in addition to 3.44 kg fat mass.  Meanwhile the lowest postprandial insulin diet (low protein group, 52% fat, 42% carb) resulted in slight loss of lean mass (-0.7 kg) while the group accumulated 3.66 kg fat.  
• That when the percentage of calories from carbohydrate in the diet was reduced by 20 percentage points (from 60% to 40%), the subjects gained fat mass because calories were increased by 40%.
• That dietary fat does influence body weight (fat mass) because it contributes calories!
• That it is the increase in calories, not carbohydrate that drives fat accumulation.

Of course, ol' Fat Head had to chime in with his own smear of Bray, and probably there are others to be found.  But since Fat Head has yet to publish (let's hope we never see that day!!) an article in a peer reviewed journal, we can let him off that hook.  However both he and Taubes are not off the hook for calling scientists liars and idiots when they continue to publish erroneous information ignoring the real implications of the science that has been done.

Whether Bray's study was designed to the likings of these curmudgeons or not, ultimately it turns out to provide us with some interesting and meaningful information that some of us might want to act on.  For starters, adding one more overfeeding study to the heap of metabolic ward studies demonstrating that among wide ranges of macronutrient composition, calories ultimately determine gain/loss in fat mass.  Not insulin levels, fasting or otherwise.  Not carbohydrate intake, absolute or percentage.  This study adds to the existing studies concerning weight loss, a study demonstrating what happens when you overfeed just fat and protein on top of a weight maintaining diet.  Fascinating results, eh?

What's also interesting here is that if you force folks to do so, they can gain fat on a "low fat" diet.  Which only goes to show that the high protein overfeeding diet, at over 3000 calories still contained a whopping 124g fat. So in no way was this a low fat diet in terms of absolute intake.


So in light of the absolute outrage hurled at the epidemiologists who conducted that recent red meat (for the blogosphere) study, including the insistence by Gary Taubes that the scientific community must police itself, I have two suggestions. 

1. Mr. Taubes:  It is time for you to come clean and correct your own hypothesis in light of all of the evidence now in hand.  
2. Dr. Feinman:  Please withdraw your article from Nutrition based on the information in the current study that clearly demonstrates no metabolic advantage for fat and protein over carbs in the context of overfeeding.

Men, police thine own selves.  If any of you in my audience today are members of NMS, I encourage you to write to Feinman and "police your own".  In the mean time, thank you Richard Feinman for bringing this study to my attention with your analysis.  It has restored my faith in JAMA that they rejected your letter attempting to misguide folks into believing fallacious claims.