This whole series was kicked off after I first read the NY Times article: After ‘The Biggest Loser,’ Their Bodies Fought to Regain Weight, and came away feeling pretty disheartened at the notion that biology is stacked so unfavorably against these Biggest Loser contestants.
That despite their massive weight loss, they would be condemned to an eternity of weight re-gain that would erode all of their substantive weight loss and make leave them more obese than when they started!
The article left me with a few questions which I tried to explore in a followup post, “The Biggest Loser Revisted,” in which I wasn’t quite satisfied with the answers that were provided, so I decided to look at the source material.
Luckily this is a free article and can be found here:
Let’s start with some vocab!
Adaptive thermogenesis and metabolic adaptation = the slowing of resting metabolic rate that comes with weight loss that is more than expected based on changes in body composition. This acts as a counter to weight loss and is believed to contribute to weight re-gain.
As we know, the weight loss regimen during the Biggest Loser competition involved 30 weeks of intense exercise and caloric restriction (I’ve actually never watched an episode… so I can only confirm this through everything I’ve read).
The researchers noted that the contestants lost mainly fat mass while preserving muscle (fat free mass).
We can see that on average, the 14 contestants evaluated in this study weighed 148.9 kg before the competition and weighed 90.6 kg at the end of the competition for a whopping loss of 58.3 kg!
Their initial fat mass was 73.4 kg which decreased to 26.2 kg for a loss of 47.2 kg.
Their initial fat free mass was 75.5 kg which decreased to a lesser to degree to 64.4 kg for a loss of 11.1 kg.
The average body composition started at 49.3% body fat which fell to 28.1% at the end of the composition.
I just want to pause here to recognize what enormous changes these are in the span of 30 weeks.
Ok, let’s move on.
Metabolic Adaptation / Adaptive Thermogenesis
Immediately after the competition, the researchers observed that along with this weight loss and change in body composition the resting metabolic rates also decreased from 2607 calories/day to 1996 calories per day. Now without any point of reference this doesn’t look that bad right?
If you’re obese and you lose weight, you lose both fat and muscle, as we see with these guys. With a decreased amount of muscle mass, you’d expect your RMR to also decrease, due to the loss of metabolically active tissue.
The researchers then calculated an expected RMR for each individual based on body composition, age, and sex, and found this average to be 2272 calories/day.
Comparing the MEASURED RMR of 1996 calories/day to the PREDICTED RMR of 2272 calories/day, we see that on average, there is a deficit of 275 calories/day. The contestants were burning 275 calories/day LESS than predicted.
This is metabolic adaptation / adaptive thermogenesis.
Ok, now I want take another pause to reflect on calculated RMRs.
Calculated Resting Metabolic Rate
The first point I want to make is, I’m not sure where they got their equation for the calculated RMR from:
Baseline data from all 16 subjects were used to generate a least squares best-fit linear regression equation for RMR as a function of FFM, FM, age, and sex (R2 5 0.84):
RMR (kcal/d) = 1001 + 21.2 x (FFM in kg) + 1.4 x (FM in kg) – 7.1 x (Age in yr) + 276 x (Sex, F=0, M=1)
Ok, lets just take a look at this equation. Everyone gets a baseline of 1001 calories. Then they take their fat free mass times 21.2 added to their fat mass times 1.4. They’re assuming that fat free mass is 15.1x more metabolically active than fat mass.
It’s interesting to note that they assume 1 kg of lean mass burns an extra 21.2 calories per day whereas the researcher in the second NY Times article said that 4.5 lbs = 24 calories per day, which comes out to 11.8 calories per kg of lean mass, almost half of whats assumed in this article. This begs the question, which assumption is true?
Ok, so moving on, they then subtract their age times 7.1, because as we get older our metabolism slows down… apparently by 7.1 kcal/day for every year that passes.
Then, if you’re a male, you get a 276 calorie bonus, and if you’re a female you don’t get any bonus.
On the surface this all makes sense, but I still can’t figure out where they got this equation from!
The main equations I know of used to calculate RMR include the original Harris Benedict equation, revised Harris Benedict equation, and Mifflin St Jeor equation all of which only take into account weight, height, and age with differing coefficients, and no regard for body composition.
The Katch McArdle equation and Cunningham equation both only take into account lean body mass while ignoring weight, height, age, and sex!
When I did some of my own research I found this study that concluded that the Mifflin St. Jeor equation is the most accurate.
Here’s a study comparing measured RMR vs predicted RMRs from 11 different equation. They found that the equations are all over the place, and that some are more accurate for certain populations than others:
The results of our study suggest that the Cunningham and Park formulas are best suited for male athletic adolescent groups and the IMNA formula for male non-athletes. The Maffeis formula appears to most accurately represent the RMR among both athletic and non-athletic female adolescents groups. However, as observed in this study, predictive equations currently in use have several potential limitations for predicting RMR.
One more study I came across tried to figure out the best equation for 20 different subpopulations… and involves a lot of maths that turned my brain off.
All of this to say, I can’t figure out where the authors of the Biggest Loser paper got their equation from. Maybe I have to learn more about what a least squares best-fit linear regression equation is, but it’s too much maths for my brain.
The more I read, the more it’s apparent to me that calculated RMRs is an inexact science.
Doesn’t it seem strange that the same equation is used for someone who is morbidly obese and sedentary, then becomes super active and in active starvation for 30 weeks, and then stops again? I mean we know the hormonal milieu can change dramatically at each of these stage. In fact we SEE that the hormones change dramatically… and yet the same equation is applied. Strange.
In any case, the equation the authors used, wherever they got it from, and however inexact it is, lies at the heart of this study, because it’s what they’re basing their assessment of metabolic adaptation on.
Some Notes On Their Methods
Resting Metabolic Rate
They used a TrueOne metabolic cart to measure RMR. Just for kicks, I tried to look up how much one of these metabolic carts go for and I think it’s around $34k (Note to the BJJ Cavewife – It’s ok dear, you don’t have to worry about a mysterious $34k entry in our next credit card statement, even I’m not crazy enough to buy this). This was their protocol:
Participants rested supine in a quiet, darkened room for 30 min before making measurements of VO2 and VCO2 for 20 min with the last 15 min used to determine RMR.
I tested my RMR a few years ago too and remember being set up in a similar fashion.
Total Energy Expenditure
I found this part interesting. This is the total amount of energy expended per day. Resting metabolic rate is only the amount of energy expended at rest. I didn’t realize this could be measured outside of a metabolic ward, but they were able to do this after the subjects went home.
How did they do this? Well, they had the subjects drink some radioactive water, and then tested their urine at random times over the next 14 days.
Then they used some advanced maths to calculate the average total energy expenditure.
Physical Activity Energy Expenditure
This is the energy used in their normal activities of daily living. Walking around, going to work, playing with the dog, etc.
They took the Total Energy Expenditure and subtracted the resting metabolic rate from it and also subtracted the estimated thermic effect of food (the energy used to digest food).
The authors also divided this number by body weight because:
…most physical activities involve locomotion and therefore have an energy cost that is proportional to body weight for a given intensity and duration.
So this value is presented as calories per day per kg.
They didn’t actually measure what the subjects ate, this was just estimated:
… the estimated thermic effect of food which was assumed to be 10% of energy intake and was calculated as 0.1 X TEE at baseline and 6 years.
Body Composition and Weight
To measure body composition they used a Dexa scan. The machine they used turned out to be the same model as when I had my Dexa.
Of course, out of curiosity I had to look up how much something like this costs… it turns out you can get your very own Dexa machine for the low low price of $38,345.00 (Again, don’t worry dear, I’m not THAT CRAZY).
What Happened After 6 Years?
Body Composition and Weight
On average this is what happened to the contestants:
Weight: 148.9 kg → 90.6 kg → 131.6 kg
BMI: 49.5 → 30.2 → 43.8
%Body Fat: 49.3 → 28.1 → 44.7
Fat Mass: 73.4 kg → 26.2 kg → 61.4 kg
Fat Free Mass: 75.5 kg → 64.4 kg → 70.2 kg
Pretty much all doom and gloom as we saw in the initial NY Times paper. They lost a bunch a weight and gained almost all of it back over the course of 6 years, and almost all of the weight they gained back was fat.
The before and after pictures and the followup 6 year pictures pretty much confirmed this:
*Images from original NY Times article
Changes in Metabolism
Total Energy Expenditure: 3804 cal/d → 3002 cal/d → 3429 cal/d
Measured RMR: 2607 cal/d → 1996 cal/d → 1903 cal/d
Calculated RMR (using equation): 2577 cal/d → 2272 cal/d → 2403 cal/d
Physical activity: 5.6 → 10.0 → 10.1
Lets look at the measured RMR first. It dropped dramatically at first, and then despite the weight re-gain, it remained low… in fact LOWER than the measured amount at the end of the competition. This is in line with the alarming NY Times article.
Ok, now let’s go to their “predicted RMR.” It’s surprising to me that despite a loss of over 58 kg from pre-Biggest Loser to immediately post-Biggest Loser, the equation only predicted a drop of 300 calories. I mean these people lost 40% of their body weight! Seeing their actual RMR drop 600 calories per day actually makes more sense.
This is the first point where the authors compare measured RMR to predicted RMR and conclude that there was a metabolic adaptation of -275 calories/day.
I’m definitely having my doubts about their equation.
6 years later, these people experience a 45% INCREASE in their body weight from the end of the competition and the equation only estimates in increased of 130 calories/day. Again, that just doesn’t pass the smell test to me, but we’re in uncharted territory here, because we know that despite the increased body weight, the measured RMR dropped by 90 calories per day, which also doesn’t make intuitive sense.
This is where the authors take the measured RMR of 1903 calories/day and subtract it from the predicted RMR of 2403 calories/day and come up with the metabolic adaptation of -499 calories/day…. hence the burning 500 calories per day less than expected finding the NY Times article refers to.
Total Energy Expenditure
The changes here make sense to me actually. They started at 3800 cal/d, probably because they were spending so much energy moving around their body weight.
Once they lost weight, this went down to 3000 cal/d because they had 40% less weight to move around.
This bounced back up to 3400 cal/d because they regained 70% of the weight they lost.
Physical Activity Expenditure
I’m not entirely sure how to interpret this information. Pre-competition this was calculated to be 5.6 cal/kg/d and this increased to 10.0 cal/kg/d at the end of the competition. This would suggest that the energy used from their physical activity had doubled from before the competition which makes sense because these guys were working out like 8 hours a day.
But after 6 years, this essentially STAYS THE SAME at 10.1 cal/kg/d. What the heck does that mean? From the NY Times article, it seems like these guys have all moved on with their lives and are back to their old jobs… or new jobs doing motivational speaking. There’s no way they’re working out 8 hours a day.
This is why it’s important to look over the Methods section very carefully.
Remember, this is calculated from (TEE – RMR – estimated thermic effect of food) / (body weight).
Also remember that the estimated thermic effect of food is simply calculated as 10% of the TEE. It’s estimated but not actually measured.
I’m pretty sure this is where the variation is coming from. I can imagine a scenario where someone is eating food that requires less than 10% of the TEE to metabolize (lots of refined carbs vs steak and fibrous veggies?) which can spuriously skew the equation and result in a higher than expected physical activity expenditure.
Here’s a quote from the NY Times article:
His slow metabolism is part of the problem, and so are his food cravings. He opens a bag of chips, thinking he will have just a few. “I’d eat five bites. Then I’d black out and eat the whole bag of chips and say, ‘What did I do?’”
Not exactly very paleo is it?
In any case, I don’t think much should be made of this number.
Changes in Hormones and Metabolites
This was a fun section for me to look at.
Glucose and Insulin
Glucose: 95.7 → 70.2 → 104.9
Insulin 10.4 → 3.9 → 12.1
The drop to 70.2 at the end of the competition really reflects how much working out and how little they were eating, I mean if they were working out all day and subsisting on skinless chicken and lettuce with a large caloric deficit, they were most likely ketogenic.
The bounce back up to 104.9 after 6 years confirms to me that they weren’t eating a low carb diet. They’re now essentially prediabetic and more insulin resistant than before!
This is confirmed with the changes in their fasting insulin which dropped to 3.9 and bounced back higher than when they started.
The HOMA-IR which is just an estimation of insulin resistance calculated from glucose and insulin also confirms a worsened degree of insulin resistance than baseline.
I really get the sense that these contestants went back to eating whatever they were eating before the show.
Lets pause again for a moment here. From what the NY Times article wrote and what these authors concluded about the significant adaptive thermogenesis that these subjects underwent, what changes to thyroid levels would you expect?
A compelling narrative would have been that these contestants had normal thyroid levels, then did the 30 weeks of Biggest Loser and slammed their bodies so hard that they became hypothyroid, which then persisted for 6 years, causing them to regain most of their weight back… but alas, this was not the case.
What actually happened was that their thyroid levels remained relatively stable.
T3: 9.42 → 5.31 → 11.15
T4: 7.3 → 6.95 → 6.18
TSH: 1.52 → 1.42 → 1.93
These numbers are all within the normal range. I’m not going to bother going into what T3, T4, and TSH do, but if you’d like an explanation, you can check out one of my older posts here (in fact if you’d like to see everything I’ve ever written about my own thyroid you can go here).
The participants clearly weren’t hypothyroid is the only conclusion that we can draw.
However, it’d be interesting to see what their Reverse T3 levels were because high Reverse T3 can actually cause a slowing of the metabolism…
We all know how much I love talking about cholesterol, but I’m going to skip over this part because I don’t think it’s relevant to the metabolism of these contestants at this point.
One quick note is that the LDL bumped up from 105 to 126 at the end of the competition which probably reflects their weight loss and the increased stress/inflammation from their grueling regimen.
2.46 → 4.69 → 7.29
Prior to writing this, I didn’t know much about adiponectin, so I had to do a bit of research. It’s a hormone produced by fat cells and is found to be inversely correlated with body fat percentage.
Here’s a great article on adiponectin: New Insight into Adiponectin Role in Obesity and Obesity-Related Diseases
Adiponectin plays a pivotal role in energy metabolism; concentration of both total adiponectin and HMW decreases in obesity and increases after weight loss.
In addition, total and HMW adiponectin oligomers are inversely correlated to BMI, glucose, insulin and triglyceride levels, degree of IR, and, importantly, visceral fat accumulation
This is what it does to the following tissues:
- Muscle cells – Increases insulin sensitivity, burns fat
- Liver – Activates glucose transport, inhibits gluconeogenesis, increases fatty acid burning, increases insulin sensitivity, and decreases inflammation
- Pancreas – Stimulates insulin secretion
- Fat cells – Increases glucose uptake. In vitro studies show that it decreases fat burning, but gene knockout studies in mice show that it increases fat burning.
This seems like a great hormone right? But what’s going on in our obesity study? Adiponectin consistently INCREASED with the initial weight loss and then with the subsequent weight re-gain!
So much for the inverse correlation with body weight.
Recent studies have found that adiponectin can play a direct role on appetite by activating the hypothalamus.
Maybe what’s going on here is the rapid weight loss induced some sort of peripheral adiponectin resistance within muscle, liver, and fat cells. This would cause fat cells to secrete more and more adiponectin into the blood stream which would then have a greater and greater effect on their brain and appetite.
This is just the my own personal hair-brained hypothesis that I haven’t researched at all… but it’s an interesting thought isn’t it? (A quick search shows that this may in fact be a thing, here, here, and here).
We know that adiponectin increased in these subjects after 6 years. We also know that insulin sensitivity worsened in these subjects. For some reason adiponectin isn’t doing what it’s supposed to do.
The authors didn’t address adiponectin at all except to say:
After 6 years, plasma leptin, thyroxin (T4), and TG remained lower than baseline while high-density lipoprotein and adiponectin were increased.
It’s interesting to note that the amount of adiponectin released doesn’t really correlate with fat mass, seeing as how it’s produced by fat cells.
41.14 → 2.56 → 27.68
Like adiponectin, this was another hormone I had to do more research on. Here are some good papers I found:
and for the more biochemistry inclined…
Leptin is also a hormone that is produced by fat cells and is directly correlated with fat mass. More fat = more leptin secreted.
In a normal system, leptin goes to the brain and tells the brain, “Hey, we have enough energy, time to chill out on the food consumption. Oh yeah, because we have so much energy available, you can go ahead and speed up our metabolism.”
When we lose weight and lose fat mass the amount of leptin decreases. With a lower amount of leptin floating around, the brain starts thinking, “Hold on, we are in an energy depleted state! GIVE ME MORE FOOD! Oh yeah, I also need to slow our metabolism down to conserve energy.”
This is actually what happened to the contestants who lost weight; their leptin levels plummeted from a sky high 41.14 ng/mL to a rock bottom 2.56 ng/mL (normal range is 1.2-9.5 in males, and 4.1-25.0 in females)
Fasting for 36 hr – 72 hr has also been shown to decrease leptin despite a non-significant change in fat mass.
So what happens in a faulty system, such as in an obese individual?
Well, there is a lot of fat mass which is pumping out gobs and gobs of leptin. The leptin is making it to the brain, but for some reason the brain isn’t quite registering it. This is called leptin resistance.
The fat cells are screaming “We have enough energy,” but because the brain is deaf to that signal, the brain still thinks it’s in an energy scarce state and signals the body to keep eating.
The NY Times article actually did a good job addressing this as one of the factors contributing to the ravenous appetite the contestants experienced after the Biggest Loser.
This is from figure 4 of the paper, showing the changes in metabolic adaptation in each of the individual subjects. Each dot represents an individual. The dots on the left show where they were at the end of the competition (30 weeks). These dots are connected by a line to dots on the right show which show where they were after 6 years.
The red arrows, courtesy of yours truly, point to two individuals who’s metabolism actually INCREASED more than expected and one individual who basically stayed the same.
Comparing the Different RMR Equations
Since the equation they used for the predicted RMR bothered me so much I decided to play around a little on a spreadsheet and compare their equation with some of the other equations out there. Since a few of the equations required knowledge of the subject’s height, which this paper didn’t provide, I did some literal back of the envelope calculations based on the BMI.
The average pre-Biggest Loser BMI listed was 49.5 kg/m². If the average weight was 148.9 kg, we can use some simple algebra:
Height = √(148.9/49.5) = 1.73 meters.
This is literally the first time I’ve square rooted ANYTHING since college… and I don’t plan on making a habit of it.
I double checked with the post-Biggest Loser and Followup at 6 Year values, and it consistently checked out at 1.73 m.
So now that we know the average height we can plug it in to the equations:
Here is the actual excel file if you want to play around with it yourself. The revised Harris Benedict and Mifflin St. Jeor both had male and female equations, so I just selected the male equation.
It’s interesting isn’t it? The numbers are all over the place.
All equations except for the one used in this paper seem to agree that at the end of the Biggest Loser competition there was a positive metabolic adaptation, meaning they were burning more calories than predicted.
At the 6 year followup, the supposedly most accurate, Mifflin-St. Jeor equation predicted a negative metabolic adaptation of -292.75 cal/day.
I’m not exactly sure what conclusions if any can be drawn from this except that this is truly an inexact science.
What Exactly Did They Find?
When the researchers ran the statistics they found that:
Weight regain was not significantly correlated with metabolic adaptation at the competition’s end, but those subjects maintaining greater weight loss at 6 years also experienced greater concurrent metabolic slowing.
What they’re basically saying is that the amount of weight that was regained had no relationship with the amount of metabolic slow down at the end of the competition.
This seems counterintuitive to the whole point of this paper doesn’t it?
The only relationship they found was that the people who kept the most weight off also had the most amount of metabolic slow down.
Um… ok Captain Obvious.
I thought after all the hullaballoo from the NY Times article that the authors would have concluded something like, “weight regain WAS significantly correlated with metabolic adaptation at competitions end, and the subjects that regained the most weight at 6 years also experienced greater metabolic slowing.”
So what can we conclude from all of this.
We know that on overage contestants lost a significant amount of weight and they regained almost all of it back.
Despite all the fancy calculations and numbers, there really was ZERO relationship between the amount of weight regained and the calculated relative metabolic slowing at the end of the Biggest Loser.
The only thing they found was that those that kept the most weight off also had the greatest degree of metabolic slow down, and that on average, there was an overall persistent ‘relative’ slowing of the metabolism.
I specify relative because the slow down is ‘relative’ to the predicted RMR which depends entirely on which RMR equation they choose to use (which in itself is an inexact science).
This also doesn’t even take into account individual variation, which as we’ve seen in this exact study, two subjects experienced a faster metabolism than expected and one remained essentially unchanged.
We can also conclude that their thyroid didn’t play a significant role in their weight gain, however it would be interesting to see if their Reverse T3 were out of whack.
Also, the whole physical activity expenditure doesn’t make any sense and this was never addressed. Why is their physical activity identical after 30 weeks of the Biggest Loser and at the 6 year followup?
Ultimately I’m not as sure about how big a role metabolic adaptation / adaptive thermogenesis played in the regaining of the weight, and this paper doesn’t help much. It’s obvious something is going on here. They’re almost back to square one.
The difficulty comes in explaining why.
To really understand what the heck happened I think it’s important to know more about what they were eating, the amount they were eating, how active they were, and what sort of exercise, if any, they did.
We know that whatever they were doing over the course of 6 years lead to worsening of their insulin resistance so that their fasting glucose and fasting insulin were higher than baseline!
I definitely think that leptin resistance and adiponectin resistance / dysfunction played a huge role in their weight gain, and it would’ve been nice for the researchers to address this issue more.
If they went back to a standard American diet consisting of highly processed foods filled to the brim with high fructose corn syrup, then this could explain a lot of it. That whole thing about blacking out and eating a whole bag of chips without realizing it makes me think this is probably what happened.
In fact, if I were to guess what happened it’s that after completing the Biggest Loser, these guys were so exhausted and so burned out that they couldn’t wait for the show to end.
When they end finally came and they were no longer forced to measure their food to the nearest nanogram and exercise for 8+ hrs a day, they could at last do whatever the hell they wanted without any immediate consequences.
No more drill sergeant to yell at them. No more threat of failure and humiliation on national television.
Their bottomed out leptin and out of whack hormones were screaming for them to eat and they obliged. Instead of eating the soulless ‘health’ food they were eating on the show, they went back to eating whatever it was that made them obese in the first place, with things like chips. And this binge didn’t stop for 6 years.
Maybe there were times when they tried to cut calories and increase their activity again, but their hormones never had a chance to fully recover, and so they eventually had to succumb to the signals of their dysfunctional endocrine system.
After reading this paper, I no longer think that if we lose weight we are condemned to regain all the weight we lost and more because our hunger would spin out of control and our metabolism would grind to a halt.
I can’t help but wonder what would’ve happened if the Biggest Loser contestants adopted a paleo/primal diet with nutrient dense whole foods, engaged in regular walking and resistance training, received adequate sunlight, and slept an uninterrupted and unmedicated 8 hours.
Perhaps they could’ve gotten their leptin and adiponectin and therefore their appetite under control. Then, maybe they wouldn’t have gained so much weight and found themselves more insulin resistant than when they started.