That 5000 Calorie Jokesperiment

Note:  I started this article with the intent of it being a single stand-alone before Smash the Fat's Sam Feltham had done his final summary of his overeating stunt.  Since then, however, he has written said summary and I've written two posts discussing this disconnect between the you-can-eat-all-the-calories claims and the current uber-restrictive low carb plans many long time low carbers are actually following themselves.  See:  Fast or Feltham and Fast or Feltham II

---

Apologies in Advance:  I wrote this with updates/edits over several days and with "new information" coming in the interim.  It could probably use a good bit of editing and crunching down, but as this is just a blog, I'm going to publish it up w/o wasting time on that.  Please forgive the length, repetitiveness, etc.  Thanks!

---

Further Preface:  Sam Feltham is the founder of Smash the Fat.  This was a publicity stunt.  Not that there's anything really wrong with that!  He did this to drum up business.  Again, not that there's anything wrong with that!  What there is something wrong with are those who latched on to this as if it were some sort of "scientific experiment" ... worthy of high praise for Feltham's supreme sacrifice,  yada yada.   This was not done to "push science forward", however, and any claims to that effect I do have a problem with.  Sam is a personal trainer whose business gimick is that they aren't like those "commercial fitness industry" outfits.  

It’s my ethos to make SMASH THE FAT Fitness & Fat Loss Boot Camps different from the commercial fitness industry. We are 100% honest with you and all of our systems and techniques for fat loss and health are 100% backed up with the latest scientific research studies. This is why we are able to offer a “Drop a clothes size in 28 days” GUARANTEE on everything we do.   {bolding his, red highlight mine}

Hopefully most readers will not be impressed by the red-highlighted nonsense.  I haven't seen any scientific research studies to accompany the explanation as to why his stunt did not net him some 6-7 kg excess fat.   100% honest?  Everyone else lies, we're honest!   The bolded part?  LOL, is anyone impressed with that?  Back in my low carb insomniac days I used to hear that guarantee on the wee morning hour infomercials.  Such a claim/guarantee is meaningless and perhaps 20-something newbie to weight issues might fall for this crap, but c'mon, nobody else should!    Boot camps?  Exercise??  Sorry, they sell cleanses for weight loss. I link to be complete, please don't follow that link and waste your money on  the "powerful and effective ten day course of five cleansing herbal products which collectively help cleanse the whole body of cellular waste", liquid chlorophyll and 12 billion bacteria to cleanse your body of toxins and lose 7 lbs.  You're smarter than that, and right there, I'm sorry, but Sam Feltham = purveyor of quackery.  I would like to see the scientific studies that back any of the claims made in conjunction with this product.  Smash the Fat!  No Situps, No Shouting ... No Shit?

---

The part I started a few days ago ... 

If you pay attention at all to internet diet and nutrition buzz, you've no doubt caught wind of the guy who ate 5000 cal/day on what he describes as a LCHF diet, for 21 days. Here's Sam Feltham's 10.5 Day, Halfway Report at HuffPo.

"10 days ago I started a 21 day experiment where I eat 5,794 calories of a low carbohydrate high fat to diet to see if a calorie is really just a calorie. I have come across some militant scientists, who say my experiment is bogus, and some very supportive ones, who have applauded me for trying to push science forward. The only premises that I'm starting with are that a calorie is a calorie, that if you eat more than you burn you put on weight and that 1lb (0.45kg) of fat is 3,500 calories."  (red emphasis mine)

Jokesperiment:  This is a word I made up for stunts and such that people try to pass off as some sort of scientific experiment ... at least there was no n=1 hashtag. Sorry, but to call this an experiment and/or laud Feltham's efforts in the name of science is, indeed, a joke.  It's not quite at the level of Jimmy Moore smothering half of a sweet potato in half a stick of butter and declaring himself intolerant to "so called safe starches" ... but, sadly, it's not really all that far behind.   You see, if the sugar industry commissions a study that is well designed and doesn't conclude sugar will kill you, that is bad science and dead on arrival ... because bias.  But if a personal trainer tries to drum up business with a publicity stunt and proclaims it in the name of science?   Can we get him on the show?!  LOL.

I guess by Feltham's options for responses, I fall in that militant scientist camp because I see this for what it is:  A stunt.  A gimmick.  Free advertising.  Kudos for the marketing prowess, but anyone promoting this just looks silly.

One thing this stunt has done is provide a context in which to address some of the many and varied strawman  arguments used by the folks like the pool noodled one and the keeper of the black box.

The LC mantra has been some varying explanation of the basic theme that carbohydrates are inherently fattening and they induce a hormonal horizontal growth disorder.  Nevermind the billions eating more carbs than your average overweight SADder yet remaining slim.  The marketing gimmick goes something like this:  Ever since we've been told to eat less and move more, we've gotten fatter and fatter.  ELMM has not only failed but is even blamed for the opposite effect!  Ultimately, calories are irrelevant.  This is somewhat modified by the "you are overeating because you're getting fatter, not getting fatter because you are overeating" -- which is a very strange concept in the context of deliberate overfeeding that I'll address at the end of this post.  

The Energy Balance Equation

Please forgive any issues with Lyle McDonald's rhetoric at times, but this is hands-down the best discussion of what I would call the comprehensive energy balance equation.  

Simple:  Energy In = Energy Out + Change in Body Stores

Comprehensive:  

Energy In is adjusted for digestion = Energy Absorbed
Energy Out = BMR/RMR + TEF + TEA + SPA/NEAT

BMR/RMR also called REE = all refer to resting energy expenditureTEF = Thermic Effect of Food, aka "thermogenesis" TEA = Thermic Effect of Activity, aka calories burned during deliberate activitySPA/NEAT = Spontaneous Physical Activity or Non-Exercise Activity Thermogenesis

Alternate Hypotheses and The Energy Balance Equation

The so-called "alternate hypothesis", or plural as it is, has gone through so many convolutions, it is difficult to give a synopsis anymore.   As calories go, the arguments oscillate between them being meaningless to the acknowledgment that the first law of thermo must hold and therefore overeating and/or sedentary behavior are the result of fat cells gone wild and accumulating stores leading to ravenous hunger and/or crippling fatigue.  What controls all of this is still the carbohydrate, and only the carbohydrate via its tyrannical rule of adipocytes by taking control over your pancreas ... nevermind that protein stimulates insulin as well.  

At the end of the day most, grudgingly or otherwise, acknowledge that a caloric surplus is a prerequisite for weight gain, and a caloric deficit for weight loss.  The chicanery lies in how this deficit is magically produced simply by altering the macronutrient composition of the diet.  The "late great" Dr. Robert Atkins had a great gimmick!  Oh yeah, as I've reminded my audience here at this blog on numerous occasions, Atkins was a CICO man!  The Diet Revolution Calorie Theory (pg. 94, paperback version, Sept. 1973 printing, ©1972 Bantam):

Calories in equals calories used plus calories excreted unused

This is THE reason Atkins had dieters raiding the diabetes section for ketostix and peeing on them -- from induction (getting into ketosis) to adding back 5g/day carb at a time while monitoring for the persistence of ketosis -- peeing out mass quantities of calories was the mechanism by which Atkins himself claimed his diet worked.   Sure, there was also some stuff about a never-identified (or to be seen again in later versions) fat mobilizing substance, and a little bit about insulin, but Atkins' schtick was simple.  He also mistakenly told readers that ketones were the result of incomplete fat burning, furthering the gimmick so dieters were convinced these ketones equated to unused fat.  (This is also why I recently bumped this post with links to a BBC documentary on the diet.  Innumerable calculations on excreted ketones put these caloric "freebies" at < 50 cal/day, generally in the 20 cal/day ballpark.) 

Through the years we've had the "metabolic advantage" arguments.  Oddly enough, the TEF argument would be to the detriment of today's uber high fat recommendations because fat is significantly less (about 1/3rd) thermogenic vs. carb, and current trends are to limit even protein which ironically is the most thermogenic of the macros (as much as 10X fat and 3X carb).   And when Eades stooped to laughably arguing that a few grams of "insensible water weight" loss during sleep was the MA of low carb, even Jonathan Bailor's clogged sink couldn't hold that argument!   Which leaves us to the "internal starvation" model of the hormonally induced horizontal growth disorder.  You know, where insulin locks away all of your energy and your starving cells cause slothiness and insatiable appetite.  Oddly enough ... CICO based arguments that are not discounted by NEOtastic™ stunts, not to be confused with "paleo" anything.

So the thing is ... all of this still boils down to some supposed alteration in the Energy Balance Equation.  Therefore, a well designed experiment controls for all of the terms not being analyzed.

A Well-Designed Overfeeding Experiment

Before I discuss Feltham, I'd like to discuss what a "perfect world" overfeeding experiment to test CICO & weight change should entail.

1.  Establish baseline parameters:  

• Weight and Body Composition by DXA
• REE by CO2 test
• Activity level with FitBit type device   (exercise and NEAT)
• TEF and/or periodic EE while consuming baseline diet 
• Baseline diet -- weigh/measure/quantify
• TDEE by double-labeled water method
• Verify weight stability and compare measured TDEE to calories consumed.
• Stool analysis for energy content.  

Remember, "perfect world" where there are no financial limits, etc.  Some of the above may be redundant, but that's also part of the point here ... to take possible confounders out of the equation and account for the drawbacks/inaccuracies of various methods of assessing the various terms.

2.  Overfeed:  Generally these protocols could come in one of a few forms, all of which "show" different things.   Keep in mind that any change from a person's usual diet that is considered "baseline" must be assessed as in #1.

• Increase usual intake -- same foods, eating pattern, etc. -- to some percentage excess of baseline, say 150% baseline intake. Three eggs for two, three apples for two, 9oz steak for 6oz.  Same meal frequency, etc.
• Increase intake over baseline by adding different foods.  Here the baseline diet is maintained but supplemented often by adding just one type of food or macronutrient.  So add some SSB's, or extra butter to each meal, or a whey shake or three.  
• Switch completely to a "formulary diet" and increase intake.  Formulary diets are generally composed of few "real whole foods" and rather rely on shakes or muffins for most of the caloric intake.  The shake or muffin allows for disguising content and leveling the playing field as to absorption, etc.  This especially would require a second "baseline" assessment.

3.  Intermediate or just an "After" evaluation of everything assessed in #1 at baseline.

Sam Feltham's Overfeeding Experiment Stunt

We don't really have anywhere near the due diligence and controls outlined above in Feltham's experiment.  To be fair, even many studies conducted in metabolic wards don't -- it is just not practical and the real world carries with it certain limitations, not the least of which is paying for all of it!  Feltham's "premise" is basically a strawman of CICO -- one which presumes no adaptations to drastic changes in intake, assumes all mass gained or lost is lipid, and equates weight with fat mass.  He "proved" his point, which was that he could "overeat" for a relatively short period of time and not gain exactly the amount of weight his version of CICO predicts.  But do his results mean anything?  I'd say pretty much no, except that he did gain weight by deliberately overeating.  

To repeat, I don't expect anyone except for maybe Dave Asprey, to assess everything in a little n=1 self experimentation, but Feltham falls far, FAR, short of any meaningful experimental design here.   He's a relative "kid", but he's not just some poor schmoe doing his best with the resources he has.  Of my checklist in #1, he certainly has access to more than a bathroom scale and a tape measure.   He's in the personal training business and should have access to some of these tools, even if he had to pay for them.  Still, he seems savvy enough to have traded exposure for the purveyors of such services in exchange for them.  That he fails to do so is a huge strike against Feltham here and pretty much renders his experiment useless.  A DXA scan, REE-CO2 test and some sort of BodyBugg device  are more than reasonable expectations here.  At the very least, he could have gone the cheapo route of determining his TDEE, which would be to eat his normal weight-stable diet for at least two, preferably three or four weeks and weigh/measure everything.   This way, a meaningful average daily expenditure could be determined.  

Instead, he provides his starting weight and waist circumference.  That's it.  We are given NO information as to Feltham's habitual diet.  I tweeted Sam to inquire about it, which he responded:  "Usually about 30% carbs, 40% fat and 30% protein...but I'll probably turn the fat up a notch now :)"   .  Still no knowledge of his usual caloric intake or the types of foods he eats/ate, an omission that is pretty damning towards calling this any sort of scientific experiment.  Instead, Feltham turns to a version of the Harris-Benedict equation to calculate his baseline TDEE.  From this he came up with a BMR of 1973 cal/day and applying a "moderate exercise" factor arrived at a baseline TDEE of 3058 cal/day.  Such equations are ESTIMATES.  Many times in diet studies I've seen them used as a starting point with some run-in tweaking period to establish weight stability.  This should have been the least Feltham did as it would have cost him nothing but a small amount of time.  He also used a less reliable formula based on total body mass vs. lean mass, and the fixed activity multiplier is truly a "fuzzy factor" here.  Likely a lean person such as Sam would have a higher basal metabolism by a formula based on LBM, and seems to be the type to have fairly high NEAT.  I could give several examples of how small "errors" in this formula or applying other general relationships can translate easily to 200 on up to 500 cal/day difference in the TDEE he uses.  

Before moving on, however, it is worth noting that for his calculated 3058 cal/day, given Sam's macro percents, this would work out to about 229g carb, 229g protein, and 136g fat which would put us well into Sisson's "insidious weight gain" realm.  This also seems quite high protein to me in absolute amounts.   

The bottom line is we have a totally unsubstantiated 21 day gross caloric surplus of 57,456 calories to work with, and the fact that this translated into only 1.3 kg of "weight".   With no special equipment, he could have gotten much closer to what the actual caloric intake was, and there's really no excuse for not making a better effort to include body composition.

The Dietary Intervention:  

So let's look at what this dietary experiment actually entailed.  His diet is detailed at the half-way mark, the totals come to:  5,794 Calories, Carbohydrate=85.2g (10%), Fat=461.42g (53%), Protein=322.2g (37%).   Now, I'm not even going to harp on his calling this a 5000 calorie experiment when his intake is much closer to 6000 calories, because that would be too easy and it is the least of the problems with all of this.  Let's break it down:

Carbohydrate:  This is a VLC diet by percentage, but over the usual 50g limit to be ketogenic.  There's something fishy on his carb counts that I only mention as regards the accuracy of his data.  Perhaps in the UK fiber is accounted for differently but it would seem odd if it were counted in the green beans but not in the nuts.  For each item, Feltham's values are given with values from nutritiondata.com in parentheses.

Green beans 180g:  60 calories (47), Carbohydrate=18g (11g total, 5g fiber), Fat= 0g (0g), Protein=2g (2g)     Walnuts, 150g: 1,058 Calories (981), Carbohydrate=4.6g (21g total, 10.5g fiber), Fat=102.75g (97.5g), Protein=25.95g (22.5g)      Pecans, 150g: 1,059 Calories (1065), Carbohydrate=8.25g (21g total, 18g fiber), Fat=105.15g (111g), Protein=16.35g (13.5g)     Almonds, 150g: 967 Calories (896), Carbohydrate=9.75g (28.5g total, 18g fiber), Fat=83.7g (79.5g), Protein=38.1g (33g)

Like I said, the carb content is likely not an issue here on balance, but the reported values are inconsistent in terms of whether or not net or total carbs are reported, etc.  For someone's Fitday personal log?  Anyone reading this with the free time to enter the data into whatever nutritional analysis package you use, I'd love it if someone could verify the intake.  I have the photo upload active so you can share a screenshot if you like in comments!  

Protein: I have admittedly not followed this experiment closely, nor read a whole lot about the web on this.  Has anyone else noted the absurdity of his protein intake??  While limits on protein ingestion may well be higher than generally accepted, over 300g/day is cause for further examination!  To put this into context, that is over one dozen cans* of tuna!   That's a LOT of protein folks. (*a can of tuna these days only 5 oz)    

Another way to look at this is in the context of general protein recommendations are.  Usually you'll see anywhere from 0.8 to 2.5 g/kg total body weight for recommendations depending on body composition and activity.  More accurate needs are established vs. lean body mass.  Now Mr. Feltham looks quite lean and weighed ~86 kg to start.  This puts his protein intake at a whopping 3.75 g/kg body weight and likely somewhere over 4 g/kg LBM.

Fat:  Don't tell anyone, but at "only" 53% fat, Mr. Feltham doesn't qualify for HighFat™ status, though this is certainly a high fat diet by all other dietary standards.  His diet includes a whopping 461 grams of fat.  

Here's the first problem with that.  As was pointed out in comments on the midway progress report -- of those 461 grams, 292 grams of fat came from nuts (450 g or about 1 pound!).  This may have been slightly overestimated (see Carb section), but I'm going to give the benefit of the doubt that on whole this is inconsequential.  What is very consequential, however, is that the digestibility of nuts, the fats in them in particular, is highly variable.  That fully 2624 calories came from nut fat injects a HUGE confounder into this "experiment".  This would be the equivalent of 29 tablespoons of regular mayo (almost half of a standard jar).   I would note that despite the overfeeding here, the only "added fats" involved were 15g or about 1T of coconut oil at each meal, this was only about 400 calories.

CICO, The Calorie Surplus & Weight

I am going to refer you back to Lyle's article on the energy balance equation and all of the terms as it would be wasteful to repeat everything in detail here.  However, the calorie surplus = CI - CO so I will touch on some problems as they  pertain to Feltham's stunt.

Calories In:  Cutting to the chase here, I think the so-called "missing calories" may well be explained by this single term.  I've seen people accuse folks like myself and various commenters (particularly on the HuffPo piece) as grasping at straws and changing our tunes as to "is a calorie a calorie" and how this was part of Feltham's point all along.  Look, this is the oldest strawman of all.  Atwater himself did NOT just burn foods in a calorimeter!  Don't believe me?  Read this:  THE ATWATER SYSTEM OF CALCULATING THE CALORIC VALUE OF DIETS.  The 9-4-4 system are now and have always been estimates.  It may well be fair to say that a shift from whole foods to more processed/refined foods is at least in part responsible for the obesity epidemic by increasing the NET energy intake for the same GROSS intake.  At least part of the advantage of LC vs. LF diets can also be explained by this as in diet trials, LC generally includes meats and veggies while the LF diets generally include cereals and bread.  Did you know that even eating ground beef vs. a steak can alter the absorbed energy?  

But  harping on this point is to distract (I believe intentionally) ... To believe in CICO is not to believe that 100 calories of table sugar is equivalent to 100 calories of apple, or for that matter that 100 calories of whole peanuts is equal to 100 calories of peanut oil.   Atwater and others after him have always been quick to point out that context mattered and that these values generally averaged out over the course of time for your basic "mixed diet".  There have been a ton of experiments done demonstrating that a fully whole food vs. a refined food diet can "add up" to a few hundred calorie per day difference, and that even the same food invarious forms can equate to different calories in.  In animal studies, especially, it is possible to vary diets to even greater extremes (way past what would be ethical in humans, which should be food for thought).  But this is why well designed overfeeding studies do something Feltham also did not:  They maximize accuracy of caloric intake by either standardizing foods and just increasing amounts and/or using formulary type diets that utilize maximally absorbed forms of the nutrients.  They would also analyze energy content of feces, but I'd have given Sam a pass on that had he chosen either or preferably both of the aforementioned approaches.  While losing a bit in translation back to "real world" scenarios, stacking the deck in favor of highest absorption with lowest variability is the way to go if indeed calories are being tested.

I think I'll dub Feltham's experiment NuttyC.  Its a ridiculous diet pretty much any way you look at it ... I don't know of a single person who advocates eating half their calories in the form of nuts to lose weight, although after this maybe there will be a new NutterBusters diet.  Such a diet, rich in phytates and omega6 oils and such, is not likely to go over all that well ... maybe in the mainstream.   Feltham's "test diet" included not just some nut calories, but over half of his caloric intake, 53%, was in the form of nuts.  This is far too big a variable to discount, and essentially renders the entire experiment moot.   Even the best chewer and digester of nuts is unlikely to absorb all 3000+ calories ingested.    I'm pretty sure this was by design ... to get the desired outcome.  

These studies are with almonds, but there are likely similar factors at play with pecans and walnuts.  Feltham consumed 150g of each or roughly one pound of nuts per day.  

• Discrepancy between the Atwater factor predicted and empirically measured energy values of almonds in human diets (hat tip David Kennedy in comments from here) -- Using currently accepted 9-4-4, adjusting for fiber, etc., almonds contain 6.0-6.1 cal/gram, while in this study the net energy absorbed for the same almonds was only 4.6 cal/g.  Using the 6.0 value, this amounts to only 77% absorption.  This study involved a maximum 84 g/day "dose" of almonds (Feltham consumed over 5X this in nuts).  But using the 77% figure, his 3084 cal/day of nuts might be estimated to equate to only 2364 calories.  This is a 720 cal/day * 21 = 15,120 cals.
• Manipulation of lipid bioaccessibility of almond seeds influences postprandial lipemia in healthy human subjects.  In this study, the postprandial content of plasma was assessed for test meals of almonds, almond oil + defatted flour, and a control oil (sunflower blend).  The "fat exposure" for the whole almonds was 74% lower than that of the almond oil ... this for a 50g fat load.  Even presuming similar chewing intensity and whatnot for 150g nuts at a time, the 2624 cal/day of nut fat was equivalent to only 682 cal/day -- a 1924 cal/day * 21 =  40,782 calorie shortfall from reported.

There are many more studies, and as you can see from just these two, we're talking significant unreliability of looking at the effect of fat calories from Feltham's choice of nuts as vehicle. Given that his portions were larger and as he offers up no information indicating he even made an effort to chew the nuts well, it is not unreasonable to chalk up the entire outcome of his stunt to malabsorption of nuts calories.

(Aside: If eating more and absorbing less is what you're after, you might want to try more nuts and see if you are "lucky" not to absorb the full nutrition.)

Calories Out:

One of the most common strawmen erected by what I term calorie deniers, is that CO is not independent of CI.  Tis a strawman indeed, because no serious person believes that the calories expended in a day is a fixed value for an individual, even if we could control activity to the same level day in and day out.   First, we have the baseline TDEE to deal with ... who knows if that is even accurate.  Indeed nuts aside, a goodly portion of the "missing" calories might not have been missing at all if Feltham's TDEE is higher than estimated (which I suspect it may well be).  In addition, we have:

• TEF:  Since he limited carbs quite severely, that usual suspect is unlikely at play here.  Ditto fat that has a very low TEF of only about 2.5%.  But the protein intake here is considerable at  This is unlikely to occur with normally fluctuating excesses (or conversely shortfalls) from eating a lot here or there. Then there's the matter of protein, which at the aforementioned 322 g/day  exceeded his usual intake of 229 g/day by 93 g/day -- a 41% increase over baseline if we take his general macros and estimated TDEE at face value.  This is bound to have a significant TEF impact regardless of the rest of the overfeeding.  
• REE:   Though less predictable, the overfeeding in general may translate into an elevated REE.  In this overfeeding study, only 90-95% of overfed fat calories were accounted for (vs. only 75%-85% of overfed carbs, BTW).  While this is a bit of an overlap with TEF, this is more of a sustained increase in energy expenditure to "waste" some excesses.  It should be measurable in the REE, but might require periodic EE measurements over a 24 hour period to assess fully.
• TEA and NEAT:  I'll take Feltham at his word that his voluntary activity "exercise" didn't change, but especially in overfeeding contexts, there's reason to believe that NEAT could increase considerably.  Too bad this wasn't assessed.
• Ketones:  Missing from Lyle's equation, because normally this is so negligible, but at that high a fat intake and short duration (not enough time to adapt fully), it is quite possible that significant energy -- perhaps 100 cal/day? -- was lost in urine and breath.

Calories Stored:

Here is where the CICO deniers have their biggest field day and where they are most intellectually dishonest.  Long chain fatty acid triglycerides contain roughly 9 calories per gram.  This is an average based on chain length and such, but the chemical energy stored in the bonds that is released when these bonds are broken during the oxidation process is essentially fixed.  It is the same for me, or you, or Gary, or Jimmy as it was for Jesus Christ himself whether or not you believe he was the Son of God.  We can go around and around all day as to what portion of adipose tissue is liquid, non-lipid components of adipocytes, or even non-adipose cells (e.g. macrophages).  A pound of lipid stores in that tissue (or anywhere else in your body for that matter) contains roughly 3500 calories -- to translate this to a pound of "fat" is tricky, but there's not a whole lot of water (10-15% if memory serves) in conjunction with it.   One can be pretty well assured that the vast majority of excess energy in this experiment is lipid accumulation.  

Carbohydrate stores are far trickier, especially in the short term.  Depending on the start level and capacity of your glycogen stores, you could gain significant "weight" before gaining any fat if you begin depleted and carb overfeed.  If we use a round number of total stores of 500g (a little over 1 lb),  a common figure of 3g water per gram glycogen is bantied about.  Thus a 1 lb swing in carbohydrate, amounting to 2000 calories, will register as a 4 lb swing on the scale.  There's a phantom 6000 "calories" or a little less than 2 lbs of fat that have "gone missing" if one drains glycogen while overfeeding.  With Feltham's protein intake, I'm not sure the extent of glycogen depletion achieved, but it could factor in here.  Protein storage is even trickier, but just eating more protein doesn't translate to protein storage.

So ... there is ALWAYS the question of where the calories go in overfeeding, where they come from in under feeding, and the general partitioning.  As I've discussed in the Insulin Paradox, the high insulin state, as would be the case in such a high protein overfeeding as Feltham's,  favors lean mass building over fat sequestration, but lack of dietary carbohydrate may limit the former.

As the discussion of glycogen eluded to, not all weight is the result of energy storage.  Fluid balance plays a big role in weight, I've seen people gain or lose 5 or more pounds overnight and it is not likely humanly possible  to do so in terms of "real weight" or "solids".  Low carbohydrate diets, in addition to depleting glycogen and associated water, are also dehydrating in general.  There is no fact-based individual who will deny that most of the early losses on Atkins are water weight.    

In making no attempt to evaluate and document body composition, Feltham's experiment is again lacking.

During the course of my writing of this post, Feltham posted a recap:  Why I Didn’t Get Fat From Eating 5,000 Calories A Day.   The net result after 21 Days of actual Nutty-K or is that Nutty-C overfeeding?   He writes:

If a calorie is just a calorie when it comes to eating food, over the 21 days I should have put on 7.3kg ending up at 92.9kg from my starting weight of 85.6kg. However, after a 56,654 calorie surplus over the 21 days I ended up putting on 1.3kg ending up at 86.9kg, a relatively large discrepancy to say the least to the tune of 6kg. I also measured my waist which started off at 79.5cm and I ended up at 3cm less at 76.5cm. Not exactly congruent with the linear weight gain and waist increase the calorie formula shows in my results graph and photos below.

I was going to trick folks and reverse the order to see if I could get anyone to take the bait that he looked like he lost weight.  But rather I decided to highlight his neck.  It stuck out at me to see that regardless of  what the scale said -- only 1.3 kg = 2.86 lb gain -- the three week result wasn't pretty.

Perhaps he lost an inch or so in the waist, but that's a pretty interesting, rather disturbing, distribution of the fat he did gain going on there.

Feltham calculates a 5794 - 3058 = 2736 cal/day surplus, thus multiply by 21 = 57,456 caloric surplus.  Dividing by 3500 cal/lb = 16.4 lbs or dividing by 7716 cal/kg = 7.446... or 7.4 kg.    He gained 1.3 kg, thus there are 6.1 kg of weight gone missing.  (6.0 by his math).   How to explain this?  Well, I've discussed several, but let's list the likely major contributors:

1. Calories In is overestimated due to the high reliance on nuts for fat and total caloric intake.
2. Initial Calories Out may be overstated.
3. Conversion to mass is 90-95% with fat overfeeding and 75-85% with carb overfeeding.  While there is no way to predict what fat + protein overfeeding might accomplish in a carbohydrate deficit, it is likely to be a lower percent than for fat.  Using 75-95% we're talking  0.4 to 1.8 kg right there IF all of the nut calories are absorbed.
4. Water weight:  Even on a lean guy like Sam, it's impossible to judge the role of non-"solid" mass that we're talking about here.  But whatever the caloric surplus that hasn't been accounted for by 1,2,3 or any combination of the above, can easily be accounted for by water weight.  First, although glycogen stores would be difficult to assess, we're talking a possible 1.5 kg water weight in addition to the 0.5g glycogen depleted from the liver and muscles.  Then there's the possibility for a few kilograms more weight in water due to the dehydrating nature of the diet (likely exacerbated by the high protein intake).

This stunt was a scientific "wash" at best.  

Suggestions for Improving on this Experiment:

There's really little that could be done about the current experiment to rescue it from all the flaws, but after several weeks of returning to normal eating, Feltham could go about establishing his usual TDEE.  He could then repeat this exact experiment only substituting nut oils and defatted nut flours equivalent to the 450 grams of almonds, pecans and walnuts he consumed -- similar to the second study from the nut bioavailability discussion.  This would put the speculation on likely the biggest confounder to rest (and as a bonus would eliminate the need for potentially health-compromising experiments in the future thought that might limit its usefulness for it's original publicity generating purpose).  Since the oil and defatted flours might be either hard to come by, difficult to choke down or both, I would suggest a nut-free repeat.  So ...

Better:   Return to normal eating and in September, repeat this experiment.  First , select simple easy to digest foods.  Lean proteins (canned tuna or chicken breast are reliable options) and eggs, oils, mayo, cream and better.  Throw in a consistent veggie like the green beans.   For however long it takes, eat C/P/F of 10/35/55 and establish baseline weight stable TDEE.  Then increase intake by 50% for 21 days.

Alternate Better:  Consume 5794 calories of Zoe Harcombe's diet prior to her conversion to eating meat.  Not the Harcombe diet, but what she ate:  Plain porridge for breakfast, cheese salad for lunch and butternut squash curry over brown rice for dinner.  

Better Yet:  Do a different experiment entirely.  Remember, Sam is a lean guy so there really is no reason to presume his current carbohydrate content is causing fat accumulation.

1. Between logging usual intake ratios for a few weeks and getting REE measured establish a baseline TDEE with at least two weeks weight stable intake of your usual ratios.
2. Compose your diets of lean proteins, isolated fats and easily digested starch for the carbs.
3. Have body composition determined by DXA
4. Now increase just the fats, keeping absolute protein and carb intake constant.  Suggested overfeed is 50% of baseline intake.
5. Have REE and body composition determined again at the end.

On second thought, this wouldn't work well as a publicity stunt because it's been done before.

BEST:  Don't link to Gary Taubes, Jimmy Moore and the Swedish Diet Doctor or rely on them for an accurate and complete understanding of biochemistry and human metabolism.  Save time and money and don't bother with any of the above.  Find a different venue for entrepreneurial energies or truly follow a passion for fitness without scamming.  Don't build a business on a house of cards that's barely standing ...

Final Thoughts

After completing this, and because Feltham seemed like a perfectly nice guy on Twitter yesterday, I almost felt I owed it to him to read his take in the "Why I Didn't Get Fat" summary piece before putting the lid on this.  Mistake?  Well, it is clear that Sam takes his cues from some purveyors of misinformation.  His limescale on the sink drain sounds Bailoresque (the not-so-Smarter Science of Slim guy, copyright infringement?), and he states some "facts" that just are not true, such as:

You may be asking yourself, why did you choose a high fat diet and why didn’t you get fat? Well the simple answer is dietary fat doesn’t really make you fat and in fact is the least likely macro-nutrient to make you fat from overconsumption. If you ask any body builder what macronutrient they manipulate in order to gain or lose weight, they’ll tell you carbohydrates. During a “bulking” phase a body builder will have a higher percentage of carbohydrate in their diet and during a “cutting” phase they will have a smaller percentage of carbohydrate in their diet. 

Firstly the bodybuilder explanation has little to do with the claim.  The simple truth is that dietary fat can make you fat if you are absorbing it and in caloric excess.  Feltham gained fat in this experiment, you can see it, and if he kept it up, he would have continued to do so.  It is interesting that he seems to have done so in the neck/chin/pec region.  If he had chosen his fat in a more digestible form, he would have gained  more weight.  It is a fact that dietary fat is the MOST likely macronutrient to make you fat (that is, be stored as body fat) from overconsumption.  This is because your body need do nothing to the fat other than not need it for energy and take up the chylomicron (dietary fat lipoprotein) particles and store the fat.  Further, as has been shown previously, carbohydrate and protein stimulate metabolic rate to at least try to compensate for overconsumption.  Fat simply does not.   I'm not talking about ridiculously extreme diets here because these are not sustainable nor even advised for 99.99 percent of the population.  Feltham also wrote:

The reason for this is that when we store carbohydrates we can store approximately 2,000 calories or 500g worth on the average person before it spills over into your fat stores. This biochemical trick of being able to store carbohydrates in the muscle and liver can be useful when we need quick intense energy, along with being able to create fat from carbohydrate for the multitude of biochemical processes that we need fat for. This might be news to some low fat advocates but a lot of the high carbohydrate diet you’re eating actually becomes fat in your body, just like how a cow creates fatty milk from grass. 

Actually the substrate for DNL in the cow is SCFA but that's a minor quibble.  The conversion of carb to fat is an energy intensive process and uses about 25% of the energy contained in the carbohydrate.  We've discussed the minor role of DNL in Western diets quite a bit here (some would say too much).  I'd simply refer to this summary graphic (click to view large, from Discovering Nutrition by Insel, Turner &Ross).  Here is the bottom line folks.  Even if you want to believe that DNL is a major path in humans, carbohydrates would still be less fattening than fats because your body will always take the easy way out to store fat -- that is it will store the stuff that is already fat before it's going through the trouble to convert carbohydrate to fat.  On some level the carbs to fat crowd understands that protein to fat is an even less likely route and more energy intensive process than carbs to fat, yet they are blinded by bias to the simple fact that neither is really necessary when you have fat!

Feltham goes on to explain how his body "dumped" the excess fat.  Yep!  It probably did and it's a good thing he didn't do so in the forest or some modern day hunter gatherers might have mistakenly thought they had stumbled upon that paleo protein bar bush ;-)

Near the end we learn that in September Feltham plans to repeat his experiment with simple carbs, and in January with real whole carbs.  I strongly urge him to first repeat the experiment with dietary fat sources that have a "track record" for assimilation in prior experiments -- e.g. cream, butter and long chain oils such as olive oil.  If he is interested in science, perhaps he can raise funds to do it better the next time.  Then, and only then should he engage in a carbohydrate overfeed.  If he doesn't do this, he will surely gain more weight eating highly available calories in refined form which will have nothing to do with his biochemistry mumbo jumbo.  If he's not willing to level the playing field then he should skip the crap overfeed entirely and give an honest shake to the whole food carbs.   This would mean:

• Protein matched at 322g = 37%
• Fat intake at 10% -- that's right, you want this to be fair.
• Carbs = 53% largely from (whole) corn, green peas, other legumes and fiberous fruits like apples and melon.   Only about 400 calories = 100g should come from readily digested starch or sugar such as rice, bread or juice and yes, even potatoes.

This would be a serious attempt at some sort of meaningful comparison.  I realize it would be difficult for him to eat 5794 calories of this diet, but fair is fair.  Oh ... and having declared his biases, ALL future "experiments" should be verified to the highest degree possible.  All meals videotaped at least.

My last thought just before publishing here is that the 5794 calorie version of the  diet Zoe Harcombe stayed slim on for 15 years would really do a greater public service.   Forget what her diet is now or what the Harcombe diet is as written.  For 21 days I want Mr. Feltham to eat: plain porridge for breakfast (low fat, no sugar), a giant cheese salad for lunch (go ahead pile on that brie and drench it in the non-bodyfat promoting olive oil), and butternut squash curry over brown rice (make that rice steamed and plain, low fat again).  When he gains more weight than this current experiment he won't know which macronutrient to blame.... but we'll have dispensed with at least one more LC scam.

---

Hopefully some day I get around to clearing the air on some misconceptions about calories, the utility of "calorie counting" and the very real and clear drawbacks and issues with calories in general.  But the reality that all calories in their packages are not created equally shouldn't be used as smokescreens to push the proven nonsense that fat calories are somehow not fattening while carbohydrate calories are not.  Continuing to perpetuate this myth is fraudulent at this point and comparing nuts to sugar is no more a valid comparison than comparing oil to apples.   Folks can't really be that gullible to fall for this stuff over and over ... really ... say it ain't so!

Meanwhile, if you're looking for a snack that might satisfy you physically without adding too many "unwanted" calories -- whole almonds are a good option!