I don't have a stance on the DDW thing, but I think this is a pretty uncharitable framing of that take:
> You can tell me physics demands that flapping your wings hard enough lets you fly, but If I just watched 10,000 people splatter at the bottom of a cliff trying it out I’m pretty justified in doubting you.
An example of a way in which CICO might not be exactly true is if there's a fat thermostat or whatever SMTM calls it, and people are subconsciously changing their rate of calorie usage (fidgeting, or becoming restless and moving around, or even just being warmer/colder) when their body is lower/higher on fat. Then it's still literally true that it's calories in/calories out, but the calories out part is harder to modify than you'd expect since your body is fighting against you. Similar is if people's hunger pangs kick in at different points, it's still calories in/calories out, but an overweight person starts suffering if they eat 2k calories and I basically feel fine.
For what it's worth, I don't exactly disagree; I talk about the same thing being at least possible elsewhere in the article. I think the effect is usually a bit exaggerated, though; about the biggest differences I've ever seen were really extreme cases like the biggest loser study, and I think they found the protective effect was something like 15-25% (don't trust this, it's been a while since I've read it).
Note that in the case of the biggest loser study, the people still lost weight; your body can't cut everything without killing you. I'd be a bit interested to see sealed-metabolic-room studies showing the effect, but I'm not sure anybody has done them - they'd have to do them over the course of weeks and I doubt anyone has been able to make a study like that work.
I would suggest the book Burn by Herman Pontzer. It teases out some of the nuances to CICO. For example, if you burn 600 calories exercising, your body will almost always compensate for some portion of those 600 calories by decreasing the energy it burns during involuntary processes and/or by decreasing other physical activity throughout the day. So maybe burning 600 calories gives you some room for additional calorie intake, but it might be more like 200 or 250 calories, not 600.
On this basis, I would suggest that the CI part of CICO is a surer way to lose weight than the CO part.
This is a great point and becomes more and more pronounced the leaner you become. In the bodybuilding community it's well known that the deeper you get into a cut, the less NEAT occurs to the point of extreme sluggishness. Conversely, many find they have abundant energy and move around much more during a bulk. In both cases, the body is actively working against the intended deficit/surplus and trying to maintain homeostasis in a largely unseen manner.
The body is incredibly good at down-regulating activity to make up for intake deficiencies and it's very difficult to quantify the CO portion outside of dedicated bouts of exercise.
That's well known in weight loss. Everybody from Weight Watchers to The Hacker's Diet will tell you that you lose most of your weight in the kitchen, not at the gym. The gym is so that you maintain relatively more muscle during your loss phase.
Yeah, I think that's accurate. The issue is that the general public, even when they've been exposed to diets that adhere to this premise, seem to wander back to the thought that diet and exercise are equal partners in the losing-weight equation.
Scott Alexander did a review of "The Hungry Brain" a while back; I assume you've read it since I know you've read most of his stuff but it does reference what I think is one such study:
"In 1965, some scientists locked people in a room where they could only eat nutrient sludge dispensed from a machine. Even though the volunteers had no idea how many calories the nutrient sludge was, they ate exactly enough to maintain their normal weight, proving the existence of a “sixth sense” for food caloric content. Next, they locked morbidly obese people in the same room. They ended up eating only tiny amounts of the nutrient sludge, one or two hundred calories a day, without feeling any hunger. This proved that their bodies “wanted” to lose the excess weight and preferred to simply live off stored fat once removed from the overly-rewarding food environment. After six months on the sludge, a man who weighed 400 lbs at the start of the experiment was down to 200, without consciously trying to reduce his weight."
To me, this is a VERY strong argument against the environmental-contaminants theory, because six weeks isn't nearly long enough to make a significant dent in the kind of long-term contaminants SMTM is hypothesizing.
It might be the case that environmental contaminants are a contributing factor; i.e. they somehow increase people's susceptibility to the ridiculously good-tasting food that's readily available these days - that would certainly help to explain things like the altitude difference. But I have a hard time getting away from "food tastes good so people eat a lot of it."
So I agree on most of this, but I would point out that SMTM's theory mostly has to do with the post 1980's period of time - I think he's positing that whatever it is causing it, it's a post 1980's contaminant, or something like that. I'd love to see them re-run the same experiment now and see what happens, in any case.
That would be amazing. But I'm pretty sure that kind of study is illegal these days. Or at least could never get IRB approval and therefore never be published in a formal journal.
Reading the first little bit of SSC's Hungry Brain review, I realize that the term CICO is probably more nuanced than I thought. I think of CICO as meaning only that over a sufficiently long time frame, CI>CO=weight gain, while CI<CO=weight loss. This is what I mean by "CICO" in other comments here.
But a lot of people apparently tack on additions to this (including theories about how voluntary CI and CO are), such as Stephan Guyenet, who in Hungry Brain defines CICO as:
"the idea that our body weight is determined by voluntary decisions about how much we eat and move, and in order to control our body weight, all we need is a little advice about how many calories to eat and burn, and a little willpower. The primary defining feature of this model is that it assumes that food intake and body fatness are not regulated. This model seems to exist mostly to make lean people feel smug, since it attributes their leanness entirely to wise voluntary decisions and a strong character. I think at this point, few people in the research world believe the CICO model."
Yeah, I don't think Guyenet's definition is fair. IMHO, most CICO people understand that akrasia/willpower can be a real problem - the reason most people give up on their diets is that it's very *hard* behaviorally, but it's important to separate out:
(1) Diets are hard to do, especially over the long term, from:
Although I think the rest of his work is BS, I liked this analogy from (I believe) Gary Taubes. Imagine you were in a room at a museum, and it started filling up with people, and you asked your companion, "Why is it so crowded in here all of a sudden?" If she were to respond with, "More people are entering the room than leaving it," this would be true but not helpful. There's no question that people gain weight when they take in more calories than then expend, or lose weight when they expend more than they take in, but that's just the mechanism. It doesn't explain WHY that is happening.
I wish I'd recorded the citation, but I read a few years ago that someone tracked weights in a population over time. I think they found that those who became obese did it at a rate of around 100 excess calories (i.e., KCal) per day. Seemingly not a lot, but compounded over months, years...
Also, I think it is important to note that the mechanism for metabolizing alcohol and refined sugar is different than most other calories. They must be processed by the liver. There is a fair amount of evidence that Type 2 diabetes is the result of years and years of pushing the liver to metabolize more of these substances than it evolved to handle (e.g., leading to fatty liver, pushing sugars back into the blood stream, weight gain, and diabetes). "Researchers examined five nationally representative surveys about food intake in the U.S. from 1977 to 2010, and found that added sugar consumption by American adults has increased by about 30% in the last three decades."
Could the mystery be as simple as refined sugars and a more sedentary lifestyle?
I'm not sure what "standardized activity" could mean in this context. I can't see any practical way to force the rats to move around more or less in their downtime.
You can give them toys! And set the lighting and temperature so they think it's day/night/whatever. There is obviously huge individual and strain variation, but we assume the mouse strains behave roughly the same across decades (otherwise the comparison would be garbage).
"But it’s not clearly weird that eating a caloric intake that had reliably kept you at a certain weight would reliably return you to that weight."
After six weeks? I mean, sure, even in a naive CICO approach, maintenance calories go up the heavier you get, so you'd expect a certain caloric intake to have a particular equilibrium weight associated with it, but dropping 10 kilos in six weeks implies like a 1800/day deficit.
More broadly I think the theory undergirding SMTM's work is the idea that your body "wants" to be at a certain weight and will modulate your appetite or metabolism to get there. This explains why it's easier for people to lose the first ten or twenty pounds, or why subjects in these overeating experiments drop the accrued weight quickly.
(I am personally sold on CICO as a philosophy since counting calories is the only approach that has consistently worked for me to lose weight.)
Unrelated: my pet theory is that the obesity epidemic is at least partly driven by people eating out much more than they used to. The trends start at roughly the same time in the US and there are good reasons to believe that you're going to eat more calories at a restaurant than you would if you had cooked for yourself.
I note that the obesity increases over time correlate with the trajectory of Americans going on disability over the same time period. It's hard to compare apples to apples though, as most charts I could find on disability during my admittedly cursory search are in raw numbers rather than a percentile basis, so I'm not sure if the correlation is really there. And, of course, the number of obese people far outnumbers the number of people on disability, some of whom are not obese. And, causation (if any) could be going the other way, i.e. obesity-->disability.
That's an interesting point. There are two good effect chains at work there (no pun intended):
1: People get injured and can't work, and also can't work out.
2: More people claim unemployment and stop going into work, so stay home more and stop moving around as much, stop caring as much about how they look because they don't have coworkers' opinions to care about, etc. Lots of social pressure goes away, and possibly a lot of depression creeps in.
I get really annoyed when I hear "basically nobody loses weight permanently" when I know people who used to be fat, changed their lifestyle and habits, then stopped being fat for the rest of their lives. Apparently my social bubble is full of scientifically unexplainable phenomena.
I think the whole thing is polycausal, and while contaminants could be a part of the story, a larger part would be the changing nutritional profiles of food. We know grass-fed beef has a completely different omega 3 : omega 6 ratio. What about commercially available vegetables vs. the ones you get from your garden? Dairy? And let's not get into heavily processed products...
I recall a publication where they attempted to compare changes between 1900 and 2000s diets but they found the products themselves changed so much there wasn't really a way to make an apples-to-apples (heh) comparison.
At some point, and I have trouble finding it, someone tried to make a database of people who had successfully lost weight and kept it off and the diet-and-nutrition people really didn't like it much.
I think that there are quite a few realms of science where the evidence is really bad in a way you'd expect to consistently give you bad data where science sees that and says "Well, it's the best data we can easily get so let's treat it like it's real." I include basically all survey data in this bucket.
In my experience, most people who lose weight long term do so as a result of a lifestyle change. They become borderline obsessive bikers or give up meat and sugar for ethical reasons or something. It seems to be much harder to lose weight long term if that's your only goal.
I've read SMTM's whole series and, honestly, I'd like to think it's environmental contaminants, because historically we've been pretty good at cleaning those up (do you know anyone who has black lung disease?). But I actually think it's simpler than that. We've progressed technologically and economically and can now create food that is almost arbitrarily delicious. As animals, we evolved with mechanisms to keep our weight in a steady and appropriate range (as long as enough food is available, of course), but we didn't evolve surrounded by Taco Bell and Popeye's and McDonald's. Our bodies respond to that food as a hyperstimulus ("omgosh a concentrated source of fat AND carbs AND salt?! this must be super rare, eat as much as you possibly can!") and we overeat. Because people vary, some of us are more prone to this than others.
I'm obese myself. I dieted some years ago for a few years and dropped about 50 lbs (leaving me still obese), but it was difficult and required a constant focus. I basically had to make diet and exercise my life - not in terms of the time commitment but in terms of mental space. When I let go of it, the weight came back. I'm not willing to enter that obsessive mental space again, but I do stay a bit thinner when I eat normal, not-that-exciting food at home.
I do try to stay fit through exercise, and to eat as well as I can make myself, but for the rest, well. I just remind myself that having one of my major problems in life be "I can constantly access incredibly delicious food in any quantity I want, and as a result I am less healthy and attractive than I would like to be" is something most of my ancestors would envy, and barely qualifies as a problem in the overall scheme of things.
Yeah, this is a good point. The fact that we tend toward obesity (storing energy) is the same thing that keeps us from starving quickly without food. If we were hummingbirds, we'd die of starvation the first time we skipped lunch.
a) Thank you for footnote 2. Brings to mind Bryan Caplan's "Social Desirability Bias." (I'm not sure if Caplan came up with SDB, but he talks about it a lot.) [Sheesh, putting in all these caveats all the time is exhausting.]
2) You may have treated this already in the above piece but if so I missed it. The argument that CICO doesn't work because someone ate X calories and stayed the same weight (or continued to gain weight) is easily countered by the fact that if the person continued counting calories in the same way (i.e. keeping the measurements wrong in the same way as before) and ate RELATIVELY LESS than before, they would probably start losing weight. If there is still no weight loss, then try eating relatively less than you did on the second try. At some point, weight loss will occur. The implicit argument that "I counted calories wrong so physics is wrong" is the same as the argument that "I wasn't speeding because my odometer is broken." If you eat zero calories, you die. If you eat the calories you're eating now, you gain weight. It follows that somewhere between zero and what you're doing now is a level where you can stay alive but lose weight. There are no fat people in concentration camps. This is true even if chemicals are messing with our metabolic set-points.
d) I made a lot of comments to this post. I know you like to respond to everyone's comments, but feel free to not respond to all or any of mine. I don't want to put extra burden on you.
I loved reading the original arguments, and really appreciate such a thorough critique/rebuttal.
My two cents is that in my experience people who are not overweight consistently "diet." They are always aware of the effects of eating too much of the wrong kinds of food, and of the need to stay active. Some people I know that have gained weight, recognize it and then permanently change their lifestyle and eating habits they can get back to a much thinner and healthier physique.
So, the statement that "diets don’t work", is only correct in the sense that short term attempts to restrict calories does not work. What does work and always has is for people to change their lifestyle, permanently. It works for me. It works for my spouse. For those in my family with the determination to make it work.
My hypothesis on the obesity epidemic is simple. The American diet and lifestyle lead to eating the types and quantities of delicious foods that make us gain a pound or two per year for 50 years. The rest is math. If you add a propensity to avoid routine activity, you get average Americans looking more like Costello than Abbot.
To me this is partly a story about the limitations of our science.
1. We have tightly controlled lab studies on diet which run briefly, and so are limited.
2. And we have longer intention-to-treat studies where nobody knows who did what, and so are limited. And
3. We have ex-post population studies where we look at loads of people over long periods and there are too many moving parts. Again: limited. (Especially ones looking at genetically uniform populations of hunter gatherers!)
I'm not saying the answers *can't* be in these studies. But that the right answers are not seen clearly enough to stand head and shoulders above the wrong answers and wave their arms. Leaves enough room for him to write 15,000 woids and you to pen 5000 in response and a lot of people to end up scratching their chins.
imo the biggest distraction is the short-term lab studies. I'm trained in economics, we focus on stocks vs flows. A slow flow over a long period of time can turn into a huge stock. Following certain practices, you could add weight well below levels of statistical significance in a sixteen week trial and still end up wearing an XXXL hospital gown in the cardiac ward age 55. Short-run studies aren't telling us much, and yet look most like science.
Your objections to his claims on diets are all sound but I don't think they address the big issue either: something meta has to have changed in the environment.
Again, note my training, but I'm wondering if it's the price of food. The price of (manufactured) food is a subtle nudge in the direction of eating more. Food is so abundant now. Hunger used to be literal, now it's a metaphor. It's small, pervasive, hard to notice in small trials, but matters a lot in the real world. Could tip everyone in the direction of eating 5% more, creating calorie surpluses. I'm not sold on this, but it's a factor to think about.
"He acts like it’s very weird that people who were previously weight-stable at a particular caloric intake return to that same weight-stable point once they return to that particular intake. But unless I’m missing something weird, could it be any other way?"
I think this deserves a bit more unpacking. If you are at equilibrium, and don't change your activity level, you should neither gain or lose weight. If you broke from equilibrium by eating more, you should then gain weight. But at that point, if you return to your old eating habits, why would you lose weight? Your old habits kept you stable, you should be stable again, albeit at a higher level.
This is not true, though, if being at a higher weight generally causes you to expend more energy just getting around. So the exact same level of food would indeed cause you to lose weight.
But going back again, this is still more complicated, since wouldn't we naively assume that if you now have a higher baseline energy expenditure, wouldn't you *be hungrier?* And therefore eat more, even if you weren't realizing it?
So being at a higher weight not only makes you spend more energy getting around, but it also just makes you *spend more energy*. You have to pump more blood, you have to maintain that fat, you have to keep it warm, etc.
In the case of "being hungrier", I don't think it's particularly relevant for the two studies he details, primarily because in one case the food intake of the prisoners was presumably set and because in the other the person doesn't describe what the weight loss process was like (did he eat less? did he exercise?). But it's also sort of a big part of the question SMTM is asking.
SMTM believes really heavily in set-point philosophy, very essentially that your body decides it likes certain weights and tries to stay there by any of a number of means (reducing your energy level and making you sluggish or trying to mind-control you into eating more hormonally, etc.). What you are describing is sort of a rudimentary version of that, basically assuming that your body will amp up hunger in an situation where it might be losing weight or that hunger is exactly keyed into maintaining weight.
I'm not sure any of that's false, but it's sort of the question; I'll get to it later but SMTM is basically saying "whatever set point mechanics might exist, lithium or whatever messes with them and makes that higher".
One thing that I think I definitely need to look into in the next article, or at least mention, is that the weight loss the people experienced was sort of *faster* than you'd expect. Dude said he lost 22 pounds over 6 weeks, which is about 1800 k/cal a day of deficit in terms of forcing the weight loss with pure thermodynamic-logic. That's a ton, and it's more than we'd expect dude to save from moving around at a lighter weight, or from even eating a little less if he grossed himself out on food or something.
It is more than one would expect from a pure energy-storage model, but from my experience I would expect that the weight loss would hit a brick wall from that point on until the guy resets his metabolism by eating at maintenance level for a week or two; 6-8 weeks is (IIRC) about as long as a cut can usefully go before the body down-regulates some autonomic functions to maintain status in what it assumes is a low-calorie environment. I experienced this myself last year, losing about 10 kilos in two months, then remaining at that same level (despite lowering my calorie intake to be at a ~500 kcal deficit for my new weight) for another month or so.
Another possibility might be that a significant part of that weight loss is water weight, which people in the fitness community will also mention when people ask why the first few weeks of weight loss are so dramatic, but then it seems to taper off.
I like to say "all diets work if you do them; the question is which one can you do, and for how long," and I think you're right that's what SMTM's argument shows. My wife and I watch Survivor, and you never see a person stay fat once they go on the 600 calorie diets that the show imposes. (I guess SMTM could argue that they lose weight because they're away from chemicals?)
If he wants to argue that chemicals are predisposing us towards diets that lead to weight gain and away from diets that lead to weight loss, well fine, I guess, but my null hypothesis is that the cause of weight gain is nothing but the main:
- Food is cheaper, easier to get, and more delicious than ever before, for a certain insidious view of delicious.
- Modern lifestyles encourage the consumption of said delicious, cheap and weight intensive food. Whether it's desk jobs, the increasing amount of leisure enjoyed by the working class, or welfare, there's something about our lives that make driving through McDonalds or grabbing a bag of Cheese-Its easier than it used to be relative to preparing food. (My kid knows how to cook and he just won't. If I ask him to cook dinner for the family, he will, but left to feed himself, he puts frozen food in the oven. I've never seen him fry an egg or boil pasta of his own volition.)
Good write up, glad to see you digging through this. It's funny how a few different posts around weight loss popped up in the past week or so... maybe people breaking out the summer clothes and thinking "Eh... I don't know if this a good idea."
One thing I have never seen anyone mention in regards to the obesity epidemic framing and the increases in obesity is the increase of life expectancy, particularly around cardiovascular problems. Weight is correlated with age, as are heart attacks. Whether weight and heart attacks are related is apparently pretty questionable, but people are definitely living longer, the proportion of old to young people is going up, and those older people are going to be heavier because people gain weight as they age. So assuming obesity doesn't affect longevity through heart attacks etc., just the fact people live a lot longer should make us expect more obesity.
If obesity does impact health significantly, either through heart attacks or something else, we should expect an even bigger increase in obesity as medical interventions improve. Whereas before people who were obese (for whatever reason) would get sick and die younger, thus taking themselves off the count, now they have a heart attack or whatever and live another 30 years. Population selection due to untreatable health issues used to select against obesity more before the 1980's than it does today.
By way of example, my grandfather died at about 60 right around 1980 of a heart attack, and some twenty years later my oldest uncle had a heart attack at 60 and lived to have some more. Both heavier guys, both not in great shape, but one died before he retired and one is still alive with his great grandkids. No matter why people are overweight, be it diet, exercise, genetics, what have you, the simple fact doctors can keep them alive when things go wrong has got to contribute a good bit to the increasing numbers of them.
And thanks for the plug, RC! The admiration is mutual! :D
Overall I take a similar tack to what you've done here — not criticizing the contamination hypothesis itself, but criticizing where SMTM doesn't provide the best evidence in favor of the standard explanations. Looking forward to your future posts on the subject!
Resetting the game board on SMTM vs self-helpers: For SMTM to be correct, it has to satisfy a few criteria:
1. That a certain chemical (e.g. Lithium) has to tamper with metabolism
2. That such chemicals also may likely cause mental health changes
3. That standard exercise and caloric control regime cannot improve metabolism when compared to "western lifestyle" as a default
4. That religiosity and "willpower" cannot defeat neurochemical and pathological changes without progressive detox (e.g. mineral intake, fluid excretion, treatment)
P.S. Forever chemicals, antibiotics, growth hormones and PUFA/MUFA need their own research on other weird or weak effects such as heart disease, immune disorders, and decreased virility.
What I'm tackling right now is the *deductive* side of SMTM's argument. So far, all I've touched is the point where they are saying "Hey, no matter what I propose as a cause next, it's *more likely to be true* because all these other things *can't be true*.
So where I get into the actual Lithium arguments they make, the inductive side, I might say "Hey, this is all super plausible and likely, I think it's this even though I said all that other stuff before". Nothing I've said so far keeps that from happening, it just points out that I don't think most of the deductive-side arguments are great.
A good deduction to do this in statistics is effect size of Lithium vs lifestyle change vs interactions of the previous two factors. Then we can deduce that one effect is stronger than the other. Problem also is that mind-altering drugs can warp experimental effectiveness (inability to finish trial is an issue)
I don't have a stance on the DDW thing, but I think this is a pretty uncharitable framing of that take:
> You can tell me physics demands that flapping your wings hard enough lets you fly, but If I just watched 10,000 people splatter at the bottom of a cliff trying it out I’m pretty justified in doubting you.
An example of a way in which CICO might not be exactly true is if there's a fat thermostat or whatever SMTM calls it, and people are subconsciously changing their rate of calorie usage (fidgeting, or becoming restless and moving around, or even just being warmer/colder) when their body is lower/higher on fat. Then it's still literally true that it's calories in/calories out, but the calories out part is harder to modify than you'd expect since your body is fighting against you. Similar is if people's hunger pangs kick in at different points, it's still calories in/calories out, but an overweight person starts suffering if they eat 2k calories and I basically feel fine.
CICO is still true, you just need to give fat people amphetamines so they start fidgeting too.
Coincidentally, that is a spectacularly effective weight loss treatment.
For what it's worth, I don't exactly disagree; I talk about the same thing being at least possible elsewhere in the article. I think the effect is usually a bit exaggerated, though; about the biggest differences I've ever seen were really extreme cases like the biggest loser study, and I think they found the protective effect was something like 15-25% (don't trust this, it's been a while since I've read it).
Note that in the case of the biggest loser study, the people still lost weight; your body can't cut everything without killing you. I'd be a bit interested to see sealed-metabolic-room studies showing the effect, but I'm not sure anybody has done them - they'd have to do them over the course of weeks and I doubt anyone has been able to make a study like that work.
I would suggest the book Burn by Herman Pontzer. It teases out some of the nuances to CICO. For example, if you burn 600 calories exercising, your body will almost always compensate for some portion of those 600 calories by decreasing the energy it burns during involuntary processes and/or by decreasing other physical activity throughout the day. So maybe burning 600 calories gives you some room for additional calorie intake, but it might be more like 200 or 250 calories, not 600.
On this basis, I would suggest that the CI part of CICO is a surer way to lose weight than the CO part.
This is a great point and becomes more and more pronounced the leaner you become. In the bodybuilding community it's well known that the deeper you get into a cut, the less NEAT occurs to the point of extreme sluggishness. Conversely, many find they have abundant energy and move around much more during a bulk. In both cases, the body is actively working against the intended deficit/surplus and trying to maintain homeostasis in a largely unseen manner.
The body is incredibly good at down-regulating activity to make up for intake deficiencies and it's very difficult to quantify the CO portion outside of dedicated bouts of exercise.
That's well known in weight loss. Everybody from Weight Watchers to The Hacker's Diet will tell you that you lose most of your weight in the kitchen, not at the gym. The gym is so that you maintain relatively more muscle during your loss phase.
Yeah, I think that's accurate. The issue is that the general public, even when they've been exposed to diets that adhere to this premise, seem to wander back to the thought that diet and exercise are equal partners in the losing-weight equation.
Scott Alexander did a review of "The Hungry Brain" a while back; I assume you've read it since I know you've read most of his stuff but it does reference what I think is one such study:
"In 1965, some scientists locked people in a room where they could only eat nutrient sludge dispensed from a machine. Even though the volunteers had no idea how many calories the nutrient sludge was, they ate exactly enough to maintain their normal weight, proving the existence of a “sixth sense” for food caloric content. Next, they locked morbidly obese people in the same room. They ended up eating only tiny amounts of the nutrient sludge, one or two hundred calories a day, without feeling any hunger. This proved that their bodies “wanted” to lose the excess weight and preferred to simply live off stored fat once removed from the overly-rewarding food environment. After six months on the sludge, a man who weighed 400 lbs at the start of the experiment was down to 200, without consciously trying to reduce his weight."
To me, this is a VERY strong argument against the environmental-contaminants theory, because six weeks isn't nearly long enough to make a significant dent in the kind of long-term contaminants SMTM is hypothesizing.
It might be the case that environmental contaminants are a contributing factor; i.e. they somehow increase people's susceptibility to the ridiculously good-tasting food that's readily available these days - that would certainly help to explain things like the altitude difference. But I have a hard time getting away from "food tastes good so people eat a lot of it."
So I agree on most of this, but I would point out that SMTM's theory mostly has to do with the post 1980's period of time - I think he's positing that whatever it is causing it, it's a post 1980's contaminant, or something like that. I'd love to see them re-run the same experiment now and see what happens, in any case.
That would be amazing. But I'm pretty sure that kind of study is illegal these days. Or at least could never get IRB approval and therefore never be published in a formal journal.
Reading the first little bit of SSC's Hungry Brain review, I realize that the term CICO is probably more nuanced than I thought. I think of CICO as meaning only that over a sufficiently long time frame, CI>CO=weight gain, while CI<CO=weight loss. This is what I mean by "CICO" in other comments here.
But a lot of people apparently tack on additions to this (including theories about how voluntary CI and CO are), such as Stephan Guyenet, who in Hungry Brain defines CICO as:
"the idea that our body weight is determined by voluntary decisions about how much we eat and move, and in order to control our body weight, all we need is a little advice about how many calories to eat and burn, and a little willpower. The primary defining feature of this model is that it assumes that food intake and body fatness are not regulated. This model seems to exist mostly to make lean people feel smug, since it attributes their leanness entirely to wise voluntary decisions and a strong character. I think at this point, few people in the research world believe the CICO model."
Yeah, I don't think Guyenet's definition is fair. IMHO, most CICO people understand that akrasia/willpower can be a real problem - the reason most people give up on their diets is that it's very *hard* behaviorally, but it's important to separate out:
(1) Diets are hard to do, especially over the long term, from:
(2) Diets don't work, even if you do them.
100% agree.
Although I think the rest of his work is BS, I liked this analogy from (I believe) Gary Taubes. Imagine you were in a room at a museum, and it started filling up with people, and you asked your companion, "Why is it so crowded in here all of a sudden?" If she were to respond with, "More people are entering the room than leaving it," this would be true but not helpful. There's no question that people gain weight when they take in more calories than then expend, or lose weight when they expend more than they take in, but that's just the mechanism. It doesn't explain WHY that is happening.
I wish I'd recorded the citation, but I read a few years ago that someone tracked weights in a population over time. I think they found that those who became obese did it at a rate of around 100 excess calories (i.e., KCal) per day. Seemingly not a lot, but compounded over months, years...
Also, I think it is important to note that the mechanism for metabolizing alcohol and refined sugar is different than most other calories. They must be processed by the liver. There is a fair amount of evidence that Type 2 diabetes is the result of years and years of pushing the liver to metabolize more of these substances than it evolved to handle (e.g., leading to fatty liver, pushing sugars back into the blood stream, weight gain, and diabetes). "Researchers examined five nationally representative surveys about food intake in the U.S. from 1977 to 2010, and found that added sugar consumption by American adults has increased by about 30% in the last three decades."
Could the mystery be as simple as refined sugars and a more sedentary lifestyle?
The lab rats have standarized diets and activity, yet they're getting fatter too. This is the main piece of evidence in favor of SMTM's idea.
I'm not sure what "standardized activity" could mean in this context. I can't see any practical way to force the rats to move around more or less in their downtime.
I can think of ways to make them move around more, but then I grew up with barn cats :)
You can give them toys! And set the lighting and temperature so they think it's day/night/whatever. There is obviously huge individual and strain variation, but we assume the mouse strains behave roughly the same across decades (otherwise the comparison would be garbage).
"But it’s not clearly weird that eating a caloric intake that had reliably kept you at a certain weight would reliably return you to that weight."
After six weeks? I mean, sure, even in a naive CICO approach, maintenance calories go up the heavier you get, so you'd expect a certain caloric intake to have a particular equilibrium weight associated with it, but dropping 10 kilos in six weeks implies like a 1800/day deficit.
More broadly I think the theory undergirding SMTM's work is the idea that your body "wants" to be at a certain weight and will modulate your appetite or metabolism to get there. This explains why it's easier for people to lose the first ten or twenty pounds, or why subjects in these overeating experiments drop the accrued weight quickly.
(I am personally sold on CICO as a philosophy since counting calories is the only approach that has consistently worked for me to lose weight.)
Unrelated: my pet theory is that the obesity epidemic is at least partly driven by people eating out much more than they used to. The trends start at roughly the same time in the US and there are good reasons to believe that you're going to eat more calories at a restaurant than you would if you had cooked for yourself.
I note that the obesity increases over time correlate with the trajectory of Americans going on disability over the same time period. It's hard to compare apples to apples though, as most charts I could find on disability during my admittedly cursory search are in raw numbers rather than a percentile basis, so I'm not sure if the correlation is really there. And, of course, the number of obese people far outnumbers the number of people on disability, some of whom are not obese. And, causation (if any) could be going the other way, i.e. obesity-->disability.
That's an interesting point. There are two good effect chains at work there (no pun intended):
1: People get injured and can't work, and also can't work out.
2: More people claim unemployment and stop going into work, so stay home more and stop moving around as much, stop caring as much about how they look because they don't have coworkers' opinions to care about, etc. Lots of social pressure goes away, and possibly a lot of depression creeps in.
Yeah, the pain-depression-pain-depression loop can be brutal.
That's not a bad point, really. I'll think about it and maybe add a note once I'm more awake.
I get really annoyed when I hear "basically nobody loses weight permanently" when I know people who used to be fat, changed their lifestyle and habits, then stopped being fat for the rest of their lives. Apparently my social bubble is full of scientifically unexplainable phenomena.
I think the whole thing is polycausal, and while contaminants could be a part of the story, a larger part would be the changing nutritional profiles of food. We know grass-fed beef has a completely different omega 3 : omega 6 ratio. What about commercially available vegetables vs. the ones you get from your garden? Dairy? And let's not get into heavily processed products...
I recall a publication where they attempted to compare changes between 1900 and 2000s diets but they found the products themselves changed so much there wasn't really a way to make an apples-to-apples (heh) comparison.
At some point, and I have trouble finding it, someone tried to make a database of people who had successfully lost weight and kept it off and the diet-and-nutrition people really didn't like it much.
I think that there are quite a few realms of science where the evidence is really bad in a way you'd expect to consistently give you bad data where science sees that and says "Well, it's the best data we can easily get so let's treat it like it's real." I include basically all survey data in this bucket.
I think what you're talking about is the National Weight Control Registry.
It is, thank you! It's comforting to see they have never updated their geocities style page design.
In my experience, most people who lose weight long term do so as a result of a lifestyle change. They become borderline obsessive bikers or give up meat and sugar for ethical reasons or something. It seems to be much harder to lose weight long term if that's your only goal.
That checks out, yeah. Perhaps vanity is a weak motivation after all.
I've read SMTM's whole series and, honestly, I'd like to think it's environmental contaminants, because historically we've been pretty good at cleaning those up (do you know anyone who has black lung disease?). But I actually think it's simpler than that. We've progressed technologically and economically and can now create food that is almost arbitrarily delicious. As animals, we evolved with mechanisms to keep our weight in a steady and appropriate range (as long as enough food is available, of course), but we didn't evolve surrounded by Taco Bell and Popeye's and McDonald's. Our bodies respond to that food as a hyperstimulus ("omgosh a concentrated source of fat AND carbs AND salt?! this must be super rare, eat as much as you possibly can!") and we overeat. Because people vary, some of us are more prone to this than others.
I'm obese myself. I dieted some years ago for a few years and dropped about 50 lbs (leaving me still obese), but it was difficult and required a constant focus. I basically had to make diet and exercise my life - not in terms of the time commitment but in terms of mental space. When I let go of it, the weight came back. I'm not willing to enter that obsessive mental space again, but I do stay a bit thinner when I eat normal, not-that-exciting food at home.
I do try to stay fit through exercise, and to eat as well as I can make myself, but for the rest, well. I just remind myself that having one of my major problems in life be "I can constantly access incredibly delicious food in any quantity I want, and as a result I am less healthy and attractive than I would like to be" is something most of my ancestors would envy, and barely qualifies as a problem in the overall scheme of things.
Yeah, this is a good point. The fact that we tend toward obesity (storing energy) is the same thing that keeps us from starving quickly without food. If we were hummingbirds, we'd die of starvation the first time we skipped lunch.
RC, three things:
a) Thank you for footnote 2. Brings to mind Bryan Caplan's "Social Desirability Bias." (I'm not sure if Caplan came up with SDB, but he talks about it a lot.) [Sheesh, putting in all these caveats all the time is exhausting.]
2) You may have treated this already in the above piece but if so I missed it. The argument that CICO doesn't work because someone ate X calories and stayed the same weight (or continued to gain weight) is easily countered by the fact that if the person continued counting calories in the same way (i.e. keeping the measurements wrong in the same way as before) and ate RELATIVELY LESS than before, they would probably start losing weight. If there is still no weight loss, then try eating relatively less than you did on the second try. At some point, weight loss will occur. The implicit argument that "I counted calories wrong so physics is wrong" is the same as the argument that "I wasn't speeding because my odometer is broken." If you eat zero calories, you die. If you eat the calories you're eating now, you gain weight. It follows that somewhere between zero and what you're doing now is a level where you can stay alive but lose weight. There are no fat people in concentration camps. This is true even if chemicals are messing with our metabolic set-points.
d) I made a lot of comments to this post. I know you like to respond to everyone's comments, but feel free to not respond to all or any of mine. I don't want to put extra burden on you.
I loved reading the original arguments, and really appreciate such a thorough critique/rebuttal.
My two cents is that in my experience people who are not overweight consistently "diet." They are always aware of the effects of eating too much of the wrong kinds of food, and of the need to stay active. Some people I know that have gained weight, recognize it and then permanently change their lifestyle and eating habits they can get back to a much thinner and healthier physique.
So, the statement that "diets don’t work", is only correct in the sense that short term attempts to restrict calories does not work. What does work and always has is for people to change their lifestyle, permanently. It works for me. It works for my spouse. For those in my family with the determination to make it work.
My hypothesis on the obesity epidemic is simple. The American diet and lifestyle lead to eating the types and quantities of delicious foods that make us gain a pound or two per year for 50 years. The rest is math. If you add a propensity to avoid routine activity, you get average Americans looking more like Costello than Abbot.
To me this is partly a story about the limitations of our science.
1. We have tightly controlled lab studies on diet which run briefly, and so are limited.
2. And we have longer intention-to-treat studies where nobody knows who did what, and so are limited. And
3. We have ex-post population studies where we look at loads of people over long periods and there are too many moving parts. Again: limited. (Especially ones looking at genetically uniform populations of hunter gatherers!)
I'm not saying the answers *can't* be in these studies. But that the right answers are not seen clearly enough to stand head and shoulders above the wrong answers and wave their arms. Leaves enough room for him to write 15,000 woids and you to pen 5000 in response and a lot of people to end up scratching their chins.
imo the biggest distraction is the short-term lab studies. I'm trained in economics, we focus on stocks vs flows. A slow flow over a long period of time can turn into a huge stock. Following certain practices, you could add weight well below levels of statistical significance in a sixteen week trial and still end up wearing an XXXL hospital gown in the cardiac ward age 55. Short-run studies aren't telling us much, and yet look most like science.
Your objections to his claims on diets are all sound but I don't think they address the big issue either: something meta has to have changed in the environment.
Again, note my training, but I'm wondering if it's the price of food. The price of (manufactured) food is a subtle nudge in the direction of eating more. Food is so abundant now. Hunger used to be literal, now it's a metaphor. It's small, pervasive, hard to notice in small trials, but matters a lot in the real world. Could tip everyone in the direction of eating 5% more, creating calorie surpluses. I'm not sold on this, but it's a factor to think about.
"He acts like it’s very weird that people who were previously weight-stable at a particular caloric intake return to that same weight-stable point once they return to that particular intake. But unless I’m missing something weird, could it be any other way?"
I think this deserves a bit more unpacking. If you are at equilibrium, and don't change your activity level, you should neither gain or lose weight. If you broke from equilibrium by eating more, you should then gain weight. But at that point, if you return to your old eating habits, why would you lose weight? Your old habits kept you stable, you should be stable again, albeit at a higher level.
This is not true, though, if being at a higher weight generally causes you to expend more energy just getting around. So the exact same level of food would indeed cause you to lose weight.
But going back again, this is still more complicated, since wouldn't we naively assume that if you now have a higher baseline energy expenditure, wouldn't you *be hungrier?* And therefore eat more, even if you weren't realizing it?
as an aside, David Wong has a substack now.
So being at a higher weight not only makes you spend more energy getting around, but it also just makes you *spend more energy*. You have to pump more blood, you have to maintain that fat, you have to keep it warm, etc.
In the case of "being hungrier", I don't think it's particularly relevant for the two studies he details, primarily because in one case the food intake of the prisoners was presumably set and because in the other the person doesn't describe what the weight loss process was like (did he eat less? did he exercise?). But it's also sort of a big part of the question SMTM is asking.
SMTM believes really heavily in set-point philosophy, very essentially that your body decides it likes certain weights and tries to stay there by any of a number of means (reducing your energy level and making you sluggish or trying to mind-control you into eating more hormonally, etc.). What you are describing is sort of a rudimentary version of that, basically assuming that your body will amp up hunger in an situation where it might be losing weight or that hunger is exactly keyed into maintaining weight.
I'm not sure any of that's false, but it's sort of the question; I'll get to it later but SMTM is basically saying "whatever set point mechanics might exist, lithium or whatever messes with them and makes that higher".
One thing that I think I definitely need to look into in the next article, or at least mention, is that the weight loss the people experienced was sort of *faster* than you'd expect. Dude said he lost 22 pounds over 6 weeks, which is about 1800 k/cal a day of deficit in terms of forcing the weight loss with pure thermodynamic-logic. That's a ton, and it's more than we'd expect dude to save from moving around at a lighter weight, or from even eating a little less if he grossed himself out on food or something.
It is more than one would expect from a pure energy-storage model, but from my experience I would expect that the weight loss would hit a brick wall from that point on until the guy resets his metabolism by eating at maintenance level for a week or two; 6-8 weeks is (IIRC) about as long as a cut can usefully go before the body down-regulates some autonomic functions to maintain status in what it assumes is a low-calorie environment. I experienced this myself last year, losing about 10 kilos in two months, then remaining at that same level (despite lowering my calorie intake to be at a ~500 kcal deficit for my new weight) for another month or so.
Another possibility might be that a significant part of that weight loss is water weight, which people in the fitness community will also mention when people ask why the first few weeks of weight loss are so dramatic, but then it seems to taper off.
I think you're exactly right.
I like to say "all diets work if you do them; the question is which one can you do, and for how long," and I think you're right that's what SMTM's argument shows. My wife and I watch Survivor, and you never see a person stay fat once they go on the 600 calorie diets that the show imposes. (I guess SMTM could argue that they lose weight because they're away from chemicals?)
If he wants to argue that chemicals are predisposing us towards diets that lead to weight gain and away from diets that lead to weight loss, well fine, I guess, but my null hypothesis is that the cause of weight gain is nothing but the main:
- Food is cheaper, easier to get, and more delicious than ever before, for a certain insidious view of delicious.
- Modern lifestyles encourage the consumption of said delicious, cheap and weight intensive food. Whether it's desk jobs, the increasing amount of leisure enjoyed by the working class, or welfare, there's something about our lives that make driving through McDonalds or grabbing a bag of Cheese-Its easier than it used to be relative to preparing food. (My kid knows how to cook and he just won't. If I ask him to cook dinner for the family, he will, but left to feed himself, he puts frozen food in the oven. I've never seen him fry an egg or boil pasta of his own volition.)
Good write up, glad to see you digging through this. It's funny how a few different posts around weight loss popped up in the past week or so... maybe people breaking out the summer clothes and thinking "Eh... I don't know if this a good idea."
One thing I have never seen anyone mention in regards to the obesity epidemic framing and the increases in obesity is the increase of life expectancy, particularly around cardiovascular problems. Weight is correlated with age, as are heart attacks. Whether weight and heart attacks are related is apparently pretty questionable, but people are definitely living longer, the proportion of old to young people is going up, and those older people are going to be heavier because people gain weight as they age. So assuming obesity doesn't affect longevity through heart attacks etc., just the fact people live a lot longer should make us expect more obesity.
If obesity does impact health significantly, either through heart attacks or something else, we should expect an even bigger increase in obesity as medical interventions improve. Whereas before people who were obese (for whatever reason) would get sick and die younger, thus taking themselves off the count, now they have a heart attack or whatever and live another 30 years. Population selection due to untreatable health issues used to select against obesity more before the 1980's than it does today.
By way of example, my grandfather died at about 60 right around 1980 of a heart attack, and some twenty years later my oldest uncle had a heart attack at 60 and lived to have some more. Both heavier guys, both not in great shape, but one died before he retired and one is still alive with his great grandkids. No matter why people are overweight, be it diet, exercise, genetics, what have you, the simple fact doctors can keep them alive when things go wrong has got to contribute a good bit to the increasing numbers of them.
And thanks for the plug, RC! The admiration is mutual! :D
I've also written a post criticizing "A Chemical Hunger", although until now I haven't put much effort into sharing it with the community:
https://someflow.substack.com/p/criticisms-of-a-chemical-hunger
Overall I take a similar tack to what you've done here — not criticizing the contamination hypothesis itself, but criticizing where SMTM doesn't provide the best evidence in favor of the standard explanations. Looking forward to your future posts on the subject!
Resetting the game board on SMTM vs self-helpers: For SMTM to be correct, it has to satisfy a few criteria:
1. That a certain chemical (e.g. Lithium) has to tamper with metabolism
2. That such chemicals also may likely cause mental health changes
3. That standard exercise and caloric control regime cannot improve metabolism when compared to "western lifestyle" as a default
4. That religiosity and "willpower" cannot defeat neurochemical and pathological changes without progressive detox (e.g. mineral intake, fluid excretion, treatment)
P.S. Forever chemicals, antibiotics, growth hormones and PUFA/MUFA need their own research on other weird or weak effects such as heart disease, immune disorders, and decreased virility.
So I actually agree with this!
What I'm tackling right now is the *deductive* side of SMTM's argument. So far, all I've touched is the point where they are saying "Hey, no matter what I propose as a cause next, it's *more likely to be true* because all these other things *can't be true*.
So where I get into the actual Lithium arguments they make, the inductive side, I might say "Hey, this is all super plausible and likely, I think it's this even though I said all that other stuff before". Nothing I've said so far keeps that from happening, it just points out that I don't think most of the deductive-side arguments are great.
A good deduction to do this in statistics is effect size of Lithium vs lifestyle change vs interactions of the previous two factors. Then we can deduce that one effect is stronger than the other. Problem also is that mind-altering drugs can warp experimental effectiveness (inability to finish trial is an issue)