This is a summary of a paper that I found open in a browser tab; I don't recall where I came across it. I think it's a nice paper, but it's also 63 pages long and seemed worth a synopsis for those who wouldn't otherwise tackle it.
Scott concluded in For, then Against, High-saturated-fat diets that the obesity crisis seemed to imply one of three answers:
- That weight less is really hard and people in previous centuries had really hard lives and that's why there was so little obesity back then.
- That it's “being caused by plastics or antibiotics affecting the microbiome or something like that”.
- That there is hysteresis—once you become overweight it's semi-permanent.
This paper argues for the second answer, and against the other two.
At the outset, there are reasons to be wary of this paper: neither author (who share a family name) appear to have expertise in applicable fields, and it appears to be set in Computer Modern Roman, hardly the style of a journal submission. So it's coming from outside of traditional expertise. (I don't have any expertise here either.)
With that in mind, the paper starts by posing a series of challenging facts about obesity. (References in the original:)
- It's new. One hundred years ago obesity was very rare (~1% of the population) but there were plenty of people who had enough to eat and, from our point of view, ate a lot of fattening foods.
- It's not just new, it seemed to suddenly kick off around 1980. “Today the rate of obesity in Italy, France, and Sweden is around 20%. In 1975, there was no country in the world that had an obesity rate higher than 15%”.
- It's still getting worse. It's less in the news but if anything it's accelerating in the US. This is despite Americans significantly cutting back on sugars and carbs since 2000.
- It's not just humans: lab animals and wild animals appear to be getting fatter over time too. (A surprise to me, but casual inspection seems to confirm that this is really a thing that reviewed papers are noting.)
- Junk food from a supermarket fattens rats far more than giving them more of any macro-nutrient does. Somehow junk food is more than the sum of its sugars, proteins, and fats.
- Across several countries, living at sea-level seems to increase obesity.
- Diets produce modest reductions in weight over the span of weeks or months, but the weight comes back over time. There's been a lot of searching for effective diets, but they're all about the same in large populations.
The next section answers some of the competing explanations for obesity:
“It's from overeating!”, they cry. But controlled overfeeding studies (from the 1970's—pre-explosion) struggle to make people gain weight and they loose it quickly once the overfeeding stops. (Which is evidence against a hysteresis theory.)
“It's lack of exercise”, they yell. But making people exercise doesn't seem to produce significant weight loss, and obesity is still spreading despite lots of money and effort being put into exercise.
“It's from eating too much fat”, rings out. But Americans reduced their fat intake in response to messaging about the evils of fat some decades ago and it didn't help. Nor are low-fat diets very effective.
“It's too much sugar / carbs”, you hear. But Americans reduced their sugar and (more generally) carb intakes over recent years and that didn't help either. Gary Taubes's study was a bit of a damp squib.
In this section there's a hunter-gatherer tribe for everything. I'm a little suspicious of this line of evidence because these small human populations could plausibly have evolved to tolerate their specific environment but, if you want a group of humans with zero-percent obesity who eat 60%+ carbs, or 60%+ fat, this paper has one for you. They have plenty of food, they just live happily and remain thin.
Next the paper establishes that there is clearly some degree of homeostatic regulation of weight by the brain. You can damage a specific part of the brain and cause obesity. Or you can have a genetic flaw that results in fat cells not producing leptin, which results in an insatiable appetite. (But adding leptin to overweight people doesn't work.)
Now the paper presents its thesis: it's all caused by a subtle poison! Manufacture of which really took off slightly before 1980, and is increasing or is bio-accumulative. Diets don't work because it's not a diet problem. Supermarket food fattens rats much more than any macro-nutrient chow because supermarket food contains more of the contamination. Wild animals are getting fatter because they're consuming it too. Living at sea-level means that your water supply has traveled much further and picked up more of it, which is why altitude is anti-correlated with obesity.
There are many drugs that cause weight gain and that appear to do so by acting on the brain, so these things can exist. This is hardly the first paper to suggest that certain chemicals contribute to the problem, but this paper is distinguishing itself by saying that it's the dominant factor.
Three specific families of chemicals are detailed for consideration: antibiotics; per-, and poly-fluoroalkyls (PFAS); and lithium. All have ambiguous evidence.
Antibiotics certainly make animals fatter, but do they do the same to humans in the amounts consumed? If so, why aren't places that use a lot in livestock fatter than those which use less? Why aren't vegan diets magic for weight loss?
PFAS is a family of thousands of under-studied chemicals, but they doesn't clearly cause weight gain in humans at plausible levels. But they are certainly getting everywhere.
Lithium clearly does cause weight gain in humans, but are the amounts that people are exposed to increasing, and they are large enough to cause weight gain?
The paper also notes that things are even worse to think about because chemicals do complex things in the environment, and in animals. You don't just have to think about the chemicals that are made, you have to worry about everything they can become. The paper includes the example of a factory in Colorado that made “war materials” and released some chemicals into the ground around the factory. It took several years for the chemicals to travel through the ground water to farms several miles away. During that time they had reacted to form 2,4-D, a herbicide, which killed crops on those farms. (The unreacted chemicals were also pretty nasty.)
Switching to speculation that should be blamed entirely on me, not the paper: it seems that there might be a tendency for any chemical that affects the regulation of adiposity to do so in the direction of obesity. There are several drugs that target the brain and cause weight gain, but fewer safe drugs that cause significant down regulation of weight. (If there were several such drugs, lots of people would be taking them.) Rather, drugs that cause weight loss often cause energy to be wasted from the body rather than changing the regulation of weight: DNP causing heat-loss, or SGLT2 inhibitors causing glucose excretion. Thus we might be facing a situation where multiple minor factors affect adipose regulation, but the overall effect is towards obesity because any effect tends to be in that direction.
The obvious example of something that causes down-regulation of weight is smoking. (We wouldn't call it “safe” though.) I wonder whether the paper is overly focused on something that had a step change in prevalence shortly before 1980. It might have been building steadily in the prior decades but the 40%ish of American adults who smoked in the 60s/70s hid it for a while.
If we were to hypothesise that some environmental factor is causing a significant fraction of the obesity problem then how would we test it? It could well be the sum of multiple factors, some of which may be carried in water given the correlation with elevation. It seems that one would need groups of overweight people willing to consume exclusively provided water (from distillation) and a source of food that is somehow pristine. The half-life of PFAS, at least, is measured in years in humans, so the subjects would have to remain compliant with this proscribed diet for extended periods of time. In order to have control groups we would have to (double-blind) contaminate the pristine food with environmentally-plausible levels of candidate chemicals, I guess? Would that get past any IRB? The paper contains easier experiments, like having Kuwait change its desalination process or, more reasonably, have a car-mechanic company change grease, but these would only produce partial answers unless we got lucky and one factor dominates.
The environmental hypothesis is primarily one of exclusion, and this paper makes a good case. (Although one should have significant epistemological humility about any complex technical argument outside of one's expertise.) And I haven't even covered it all! The paper continues with arguments about paradoxical reactions and the occurrence of anorexia! There is much more within if this summary piqued an interest.