REVIEW: Burn, by Herman Pontzer

Burn: New Research Blows the Lid Off How We Really Burn Calories, Stay Healthy, and Lose Weight, Herman Pontzer (Avery, 2022).

Compare these two covers. The one on the left, with the design that says “this is pop science to shelve with your Malcolm Gladwell,” was the original. A year later, a different press put out a second edition with a new design that screams DIET BOOK (and which probably sold an order of magnitude more copies). And now ignore the cover. Ignore the subtitle. Ignore the little “PhD.” after the author’s name. (I mean, obviously he has one, but this is not the kind of book that usually points it out on the cover; they assume you know an evolutionary anthropologist at Duke can, in fact, be relied upon to have a doctorate.) Ignore the marketing altogether: this is not a diet book, and it has few if any new lessons for how you, personally, should lose weight. What it does have is an intriguing new model and a hefty serving of food for thought when it comes to society at large.

Let’s begin, as Pontzer does, with the Hadza. One of the world’s few remaining hunter-gatherer peoples, the thousand or so Hadza live in small bands on the savannas of northern Tanzania. Their language is full of click consonants but apparently unrelated to other click-heavy languages of southern Africa, and aside from a bit of Bantu admixture (because of course there’s Bantu admixture), they split from their closest relatives — the agriculturalist Sandawe, who live about a hundred miles away — more than 15,000 years ago. In short, as far as we can tell, the Hadza live more or less the way their ancestors did thousands of years ago, which makes them — like all hunter-gatherers — a source of unending fascination for anthropologists who look to them for clues about how all our ancestors lived in the deep past. (In fact “Hadzaland” is only about thirty miles from Olduvai Gorge.) You may also remember the Hadza from Slime Mold Time Mold’s series on human obesity: they’re the people who get about 15-20% of all their calories from honey, much of it located with the active assistance of the wild honeyguide bird.¹

As you might expect, the Hadza are tremendously more physically active than we are. Hadza women routinely walk five miles or more, carrying their youngest children on their backs, to dig tubers out of the hard ground using only a pointed stick. (Unlike the domesticated potatoes and yams we’re used to, the wild tubers that make up much of the Hadza diet are so fibrous that once they’re cooked you have to chew them up, suck out the starch, and spit out what’s left.) A Hadza man’s daily hunting routine is even more physically intense: he’ll walk almost twice as far as a woman in his search for large game or wild honey, and if he’s successful he’ll have to carry it all back to camp. Even children routinely hike for several miles to fetch water for their families; Pontzer describes a group of eight-year-old Hadza boys who decided they didn’t like the boarding school where their parents had sent them and thought nothing of making the three-day trek home. The Hadza “get more physical activity in a day,” Pontzer writes, “than the typical Westerner gets in a week.”

Given all this, no one was surprised to learn that obesity and cardiometabolic disease (Type 2 diabetes, heart disease, etc.) are virtually unknown among the Hadza. General medical wisdom asserts that these “lifestyle diseases” are a consequence of our sedentary lives: we don’t move around much, the theory goes, so we don’t burn many calories. Our bodies store what we don’t burn, which makes us fat and, eventually, sick. Obviously the thin, healthy Hadza — and by implication our hunter-gatherer ancestors — were burning way more calories with their active lifestyles!

Except they’re not.

Every story of scientific discovery has its exciting new tool: the saga of aDNA had next-gen sequencing; this one has doubly labeled water, a simple but ingenious technique that exploits our knowledge of the Krebs cycle and our industrial prowess. It works like this: when you use energy, you make and then respire carbon dioxide. If you can measure how much carbon dioxide a person is breathing out, you know how many calories their body is burning. So for centuries,² metabolic research depended on capturing the exhalations of people and animals as they engaged in a variety of activities, then analyzing the ratio of oxygen to carbon dioxide in those exhalations. All those charts you can find online purporting to tell you how many calories you burn in your daily life (food shopping with cart: 85 calories; playing with kids, moderate effort: 114 calories) are based on some guy in a lab gamely plugging away while wearing an astronaut helmet that funnels his respiration into a oxygen analyzer.³

The doubly labeled water technique, however, rests on the fact that when your body makes that new carbon dioxide molecule in the process of using energy, one of the oxygen atoms comes from the water that was already in your body. Hydrogen atoms in your body’s internal “water pool” only leave as water; oxygen atoms can leave as water or as CO2. If you can track the relative rates of hydrogen and oxygen flow, then, you can calculate how much CO2 has been produced, and thus how much energy has been expended. And now, by using water enriched with the isotopes oxygen-18 (two extra neutrons) and deuterium (hydrogen with one neutron) — hence “doubly labeled” — you can, in fact, track the relative rates. If on Monday 10% of the hydrogen atoms in my body are deuterium, but by Friday it’s only 5%, you know that half the water in my body has been flushed out and replaced in that time. Any difference between the rate of hydrogen replacement and the rate of oxygen replacement must therefore reflect the rate of CO2 production…and my energy expenditure. All you need is a sample of my precious bodily fluids shortly after the doubly labeled water has had time to disperse through my system, and then another one a few days later. No astronaut helmet required; you can even bring the technique to Hadzaland as long as you can freeze the samples until they can be shipped home to analyze.

The doubly labeled water method was developed in the 1950s, but it wasn’t until the 1980s, when improved mass spectrometry and more efficient isotope production combined to lower the price by several orders of magnitude, that it became practical to use it on large animals like humans. Since then it’s been the fundamental tool of metabolic research, and with it we’ve begun to discover some very interesting things about human metabolism — among them, the fact that the very active Hadza expend about the same amount of energy each day as a far more sessile inhabitant of an industrialized country.⁴

But that’s actually a really weird thing to discover! We are, after all, still bound by the laws of physics here: a Hadza woman’s five mile foraging trip with her toddler on her back takes energy, which is energy I just don’t expend since my toddler can climb into his own car seat for our drive to the grocery store. How can we both be burning roughly 2000 calories a day? It’s tempting to think that Pontzer is just wrong, but he has plenty of evidence on his side. For example, Shuar children in the Amazonian rainforest seem to expend as much energy as children in industrialized countries despite the fact that they live more active lives and their metabolically-demanding immune systems are in constant overdrive in response to parasites and bacterial infections. Women in rural western Nigeria and black women from the Chicago suburbs, despite their very different sizes and lifestyles (mean BMI among the Nigerians, most of whom were farmers: 22.6; mean BMI in Chicago, where 40% of the participants were unemployed: 30.8) expended essentially the same amount of energy even before adjusting for participants’ fat-free mass.⁵ A meta-analysis of 98 doubly labeled water studies agreed that physically active people don’t seem to expend more energy. And it’s not just humans: zoo animals burn the same number of calories as their more active wild counterparts. So what’s going on here?

Pontzer proposes a model that he terms “constrained total energy expenditure”: as the metabolic demands of physical activity ramp up, the body gradually compensates by scaling back other functions. This isn’t the whole story — there is a point on the curve where increased activity means increased energy expenditure, but it comes very early.⁶ Beyond that, total energy expenditure doesn’t increase until you hit “serious athlete” levels. (Pontzer dedicates a whole chapter and a half to really heavy physical activity, including answering the universal “but what about Michael Phelps?” question, but I won’t get into it here. Read the book.)

He cites a number of studies to back up his model. A 1990 study that trained two dozen Dutch people to run a half marathon examined their weight, muscle mass, and metabolic rates over the course of the experiment, and found that by the end of it the women (who had gone from never exercising to running 25 miles a week) were burning about 360 calories more per day from running, but only increased their total daily energy expenditure by 120 calories. In the famous⁷ Midwest Exercise Trial, sedentary overweight young adults worked their way up to 2000 calories of supervised exercise each week (equivalent to running 20 miles), which “should” have burned forty pounds over the course of the 16-month experiment. Instead, the men lost ten pounds within the first nine months, then stopped, and the women didn’t lose any weight at all. The followup, Midwest 2, had an even more ambitious workout regimen of up to 3000 calories of exercise per week, but found that over the ten months of the trial total daily energy expenditures only increased by 220 calories (when you would naively expect more like 430) and subjects lost less than ten pounds. Half the subjects lost no weight at all.

There are two obvious responses here. First: well, a few hundred calories is better than nothing, right? Even if you’re only expending an extra 100 calories per day, eventually that adds up (to about a pound a month). But a period of ten, twelve, or even sixteen months, while “long term” for scientific studies, is actually pretty short relative to the human lifespan and the body’s ability to self-regulate (as any woman who’s waited for her body to go “back to normal” after having a baby can tell you). The data from populations who have been very active their whole lives suggests that eventually, even that additional 100 calorie expenditure will eventually vanish back into your constrained total energy budget.

And second: if the people in these studies were burning more calories, even a few more, but weren’t losing weight, they must have been eating more — a new exercise program, holding caloric intake constant, will lead to some weight loss. But your metabolism’s first response to unusual new activity is unusual new hunger, and an extra 100 calories is easily balanced out by a small handful of almonds. And anyway, the studies cited above measured people’s activity levels rather than their diets, so what if in addition to increasing exercise, we reduced caloric intake? This is hard to do in a study, because people are really bad at accurately reporting what they eat and it’s expensive and inconvenient to surveil them all day to report it properly (plus IRB approval is probably hard to obtain). Luckily (?), we have a natural experiment in the form of contestants on The Biggest Loser and the producers who allowed endocrinologists to study them. It turns out that while the people who completed the show’s intense exercise and diet regimen lost an average of 127 pounds over thirty weeks and saw a number of cardiometabolic benefits (fasting glucose level, insulin resistance, and triglycerides all dropped), their basal metabolic rate had also plummeted 25%. Their bodies were burning fewer calories at rest because, as far as millions of years of evolutionary fine-tuning could tell, they were starving and needed to conserve energy. And at a follow-up six years later, their BMRs were still low, despite the fact that all but one of the contestants had regained substantial weight. But the really fascinating thing is who gained the weight: it seems intuitive that the former contestants with the most sluggish metabolism would have gotten fattest, but in fact it’s exactly the opposite. The people who regained the most weight were the ones whose BMR went back up the most. In short, Pontzer says, your basal metabolic rate and your daily energy expenditure don’t dictate your weight change, they respond to it.

So when your body responds to increased physical activity by scaling back energy expenditure on other functions, what are you missing out on? In children, it’s mostly growth: those Shuar children, with their high levels of physical activity and their busy immune response to a heavy infectious disease burden, are substantially shorter than children in industrialized countries. In adults, it’s less visible: some researchers have suggested that when we become more active we unconsciously reduce non-exercise physical activity like standing or fidgeting, but Pontzer thinks most of the cutbacks come from the energy that would otherwise be spent on things like stress reactivity, the reproductive system, and inflammatory response. Endurance athletes, for instance, produce less adrenaline and cortisol in stressful situations than sedentary non-exercisers, and exercise is well known to help with anxiety and depression. A reduced metabolic investment in stress response might well explain the mechanism behind this. Similarly, most people in subsistence societies have dramatically lower levels of circulating reproductive hormones than those in industrialized societies: the average testosterone level among Hadza men is about half that of an American man. Endurance training, too, seems to reduce testosterone levels, although gradually — training for a year reduces it by 10%, and five years or more by 30%.⁸ Pontzer suggests that this is good for you: high hormone levels are linked to various cancers of the reproductive system. Finally, exercise suppresses the inflammatory response, which may be part of why it reduces the risk of cardiometabolic issues. And this all makes perfect sense if we’re working with constrained total energy expenditure: if your body is fine-tuning how much it spends on various functions to stay in the sweet spot of (roughly) 2500 calories/day, then making it burn an extra couple hundred on chasing down a gazelle and schlepping it back to camp leaves that many fewer calories to spend on cortisol, testosterone, and cytokines.⁹

So “missing out” is the wrong framing for this: we don’t actually want such sensitive stress responses, high rates of breast and prostate cancer, and widespread inflammation. They’re bad for us. And more active populations don’t have to worry about them.

Pontzer never really comes quite out and says so in the book (he’s more explicit in this article, which is where I stole the graphic above), but the obvious implication here is that many of the public health problems that plague industrialized societies — our unusually high levels of mental illness, autoimmune diseases, and many cancers — are due to our low levels of physical activity. It’s not that the Hadza and groups like them have low testosterone, stress response, or inflammation — it’s that ours are high, because we don’t get enough exercise to keep them in check. Millions of years of evolution fine-tuned the complex feedback loops our metabolism uses to keep everything balanced, and physical activity is an important element of that system. If industrial civilization’s historically unprecedented leisure has knocked out one of the key props, is it any wonder the whole structure is listing slowly to the side?

The naively mechanistic approach to human metabolism — the idea that increased energy expenditure on physical activity will lead to increased total expenditure, like driving a few extra miles burns that much more gas in your car’s engine — has turned out to be totally wrong, but it’s just a special case of a more general problem. We like to think about everything as if it were a machine, because there are a lot of machines in our lives and they do all sorts of wonderful things, but complex evolved systems — like the human body, or, say, culture — don’t work that way. They’re made up of feedback loops and inhibitors, uncountable tiny things that all work together in unpredictable ways, more like a large language model than an internal combustion engine. You can think of evolved systems as the world’s worst spaghetti code: they might work right now, but change something and there’s no telling what snowballing consequences will wreck everything.

So yes, Pontzer says, our lack of physical activity is making us sick. But it’s not making us fat.¹⁰ What is?

There’s only one possible answer, because we’re still dealing with the laws of physics here. Weight is — weight must be — a matter of calories in vs. calories out. Constrained total energy expenditure just means it’s hard to meaningfully increase the out; to lose weight, you have to decrease the in. Of course, eating less is difficult, especially when you’re surrounded by such a variety of delicious, energy-dense food. There’s a whole pile of things going on here, beginning with a phenomenon called sensory specific satiety (also known as “I’m full of dinner but there’s still room in my dessert stomach”) and moving on to entire stores full of foods designed to hit all your palatability triggers without any of the satiety ones (which is why you can’t eat just one), but even beyond that, we saw another problem with the Biggest Loser contestants: if you decrease your energy intake by too much, your metabolism throttles down to keep you alive through this “lean season.” It takes slow, gradual, consistent change to your diet, and this is where exercise does come in: for people who have lost significant quantities of weight, regular exercise seems to help keep the weight off.

Anyway, the idea that weight loss happens in the kitchen rather than the gym is not actually news to anyone who’s tried to lose weight, but it has some interesting implications for the obesity crisis. One of Pontzer’s main targets throughout his work has been a public health establishment that encourages exercise as a route to weight loss; it just doesn’t work that way, he insists. Which is too bad, because it’s a lot easier to convince someone to go for a walk after dinner than to eat less while they’re at the table! But exercise is still a good thing, even if it doesn’t do much for obesity and its associated cardiometabolic issues; in the constrained total energy expenditure model, every calorie you burn in physical activity is a calorie your body can’t use to fuel mental illness, autoimmune disease, and cancer. Is that enough motivation to get people moving? Pontzer quotes an anonymous obesity researcher who advised him, “Once people find out exercise won’t help them lose weight, they’ll stop doing it. Avoiding death isn’t a big enough incentive. The only reliable motivation to exercise is vanity.” Now I think we’ve all had enough of public health authorities lying to us “for our own good,” but there’s another problem with lying: if we promote exercise for weight loss, which it doesn’t do, then we risk losing it and all it does do when something else (say semaglutide, which really does seem to help people eat less) comes along and shows actual effects against obesity.

Then again, I don’t think these “public health interventions” actually do much of anything besides giving someone with her MPH in her Twitter handle something to do. Who goes to the gym because of what the WHO says? Who doesn’t know that exercise is good for you and that eating too much junk food makes you fat? No, what’s really novel in the constrained total energy expenditure model is its implications about everything else. We’ve known for a long time that physical activity has all sorts of benefits, from physical strength to cardiopulmonary health to improving mood and reducing stress (Pontzer suggests the book Exercised: Why Something We Never Evolved to Do Is Healthy and Rewarding, by his PhD. advisor,¹¹ for more), but this vision of metabolic trade-offs helps to explain why. Our world is simply teeming with problems we already know regular exercise can help us avoid: anxiety! ADHD! prostate cancer! diabetes! If Pontzer’s model is correct, though — if a sedentary lifestyle leads the body to overinvest in inflammatory response and various stress and reproductive hormones — there are dozens more it might be worth investigating. Is there a relationship between physical activity and menstrual cramps? Acne? PTSD? Celiac? It’s a fascinating suggestion.

But even if it’s true, it’s not clear what we can do about it at a societal level. Pontzer suggests we tax or ban junk food (or at least stop subsidizing its production), build cities that don’t depend on cars, and generally restructure our built environments to encourage physical activity and discourage overeating, all of which would be great for reasons that go well beyond fighting obesity and none of which seem very likely to happen under our present regime. He also trots out some standard-issue lib suggestions that we should reduce or eliminate socioeconomic inequality and make healthy foods “cheaper and more plentiful,” eliminate “food deserts,” and “make school nutrition a priority.” But food deserts are not actually a thing: people who buy chips and candy instead of kale aren’t doing it because they can’t afford kale or don’t know how to cook it, they’re doing it because they’d rather eat chips and candy. Kids provided nutritious, whole-food-heavy school lunches don’t eat the fruits and vegetables.¹² Behavioral choices on diet and exercise are a product of culture; we can think of a cuisine as a social technology, which can be lost like the aboriginal Tasmanians lost the ability to make fire, and is difficult to recreate from scratch. But fine, say we do pull it off and get a generation that, on the margin, consumes fewer calories and moves more. It still wouldn’t be enough to return us to the levels of physical activity our metabolic balance-sheets have evolved to assume. Short of wholesale deindustrialization — which couldn’t be done without inflicting tremendous suffering, but would at least force the ragged survivors to get in a lot of steps — we’re stuck with at least a fraction of our present problems. The best we can hope for is making it a much smaller fraction. Which is, in the end, probably why Burn got marketed as a diet book: it’s not exactly easy to fix your own life, but it’s a lot easier than fixing society.

1. Frank Marlowe, who is the guy on the Hadza, uses observation of their honey-gathering techniques to suggest that honey could also have formed an important part of the diet of our hominin ancestors as far back as 1.7 million years ago. Someone tell the Stone Age foragers they’re not very paleo.

2. Really: Lavoisier and Laplace were the first ones to figure this out.

3. It raises some questions about claims for how many calories are burned during “sexual activity, general, moderate effort,” but for my own peace of mind I’m just going to assume they made it up.

4. You can find Pontzer’s original paper on the topic here.

5. Fat is not metabolically demanding in the way that muscle or organ function are, which is why women (who have higher body fat than men) tend to burn fewer calories per pound.

6. Pontzer’s paper on this puts the cutoff at which additional physical activity no longer produces greater energy expenditure at 230 average counts per minute per day (CPM/d) on the wearable tri-axial accelerometers that seem to be standard for the field. This table from another paper gives accelerometer counts for various activities (gardening: 1243, descending stairs: 3157, mopping: 680, folding laundry: 148), which implies that 230 CPM/d is exceptionally attainable with even a lightly active lifestyle — if you’re moving for much of the time you’re awake, you’re probably hitting it. But I haven’t done the actual math, so if anyone wants to correct me please do.

7. Well, famous if you read about this stuff much. I always enjoy dipping into a new field and discovering all the new things it turns out “everybody knows.”

8. I would be interested to learn how this interacts with the role of muscle mass in increasing testosterone levels, and whether these effects are the same for resistance training as opposed to endurance training. I suspect endurance training is closer to our environment of evolutionary adaptation, but frankly it sucks, so if you have any papers about reproductive hormone levels in strength training send ‘em my way.

9. Of course it doesn’t happen in a day — it takes years to reach a steady state — but you get the picture.

10. Okay, that’s a little unfair, because there is pretty good evidence that a complete absence of physical activity interferes with metabolic regulation, possibly by causing chronic inflammation in the brain. This may be what we see in the constrained total energy expenditure graphs, where at very low levels of activity an increase really does result in greater energy expenditure. An interesting example comes from a 1956 study in a Bengali jute factory (sadly not on Sci-Hub) where it turned out that all the employees — from clerks to the men who carried the 190-pound bales of jute around the factory — weighed about the same, except for the men who sat in stalls all day, every day, who weighed fifty pounds more.

11. Excuse me, his Harvard PhD. advisor.

12. This isn’t a huge surprise; most American school cafeterias don’t have functional kitchens, so the vegetables are basically whatever glop you can scoop out of a giant can, but even when I cook my kids fresh vegetables from scratch I usually bat about .400 on getting them to do more than taste.