Stanford Encyclopedia of Philosophy

When we announced the creation of the Nutrition Science Initiative (NuSI) in September, I e-mailed  information packets and press releases to many of my colleagues in science and health journalism. I included a few who had vehemently and even publicly disagreed with the arguments I’ve made in my books and articles. I was hoping to see them embrace the idea that our conflicting ideas should and could be tested, and any organization that could make this happen would be a good thing and worthy of public support. They all expressed their admiration for the effort in private, although none of them perceived it as worth writing up publicly, at least not until we have experimental results to discuss.

A couple of these journalists took the opportunity to insist that we didn’t really disagree all that much on what we had argued over the years, and they explained why. Here’s how one put it in an e-mail back to me (and I’ve made a few minor changes so that this writer can remain suitably anonymous):

I think in our hearts we basically agree–I know you believe in calories because you are a scientist. And I know you know that many cultures that consume large amounts of carbohydrate but low fat and relatively low sugar are not fat.  (Japan, for example–and yes, I know they are not all slim, but we both know they have a very low obesity rate.)  Do I think some people get “hooked” on sweets and simple carbs?  Yes, I do.  Do I think that can be a problem for some people, hence steering clear of simple carbs is a good idea for them?  I do, and said so on national radio.  Do I think there is strong evidence that carbs per se (in absence of excess calories) will result in excess weight gain?  No I do not.

As I wrote in [...]–sugar ain’t great, nor is excess fat.  Sugar and fat infused foods packaged and “ready to eat” are what has made America obese–sitting around doesn’t help, either.  When you cut out carbs, you cut out most of the fat–yes, you can eat pure whipped cream, but not the cake or ice cream you’d normally put it on–so why bother?  Yes, you can eat fried porkchops, but not breaded fried porkchops–or clams or chicken–so that cuts calories a bunch.  You eat a lot less butter when you can’t eat the bread or potato or pasta or cake or pie…etc, etc.  There is only so much “whole chicken and steak” and even full fat yogurt with berries that most of us can stomach. So over time, on average we eat fewer calories when we cut out simple and most complex carbs.

I thought a lot about this e-mail after I received it, for two reasons. One is the tendency we all have (or at least I have and I do see it in others) to assume that just because we’ve written something, other people have read it or certainly should have read it. If only…

In this case, I addressed the no-bread-no-butter, only-so-much-steak-you-can-eat position at length in Good Calories, Bad Calories (The Diet Delusion, in the UK). I even quoted Jane Brody making this point in the epigraph to Chapter 20 — “Unconventional Diets” — and set it off against a quote from the DuPont physician Alfred Pennington to contrast how the same observation — weight loss and absence of hunger on a diet unrestricted in calories  – could be perceived by one person, Pennington, as a “mighty stimulant to thought on the matter,” and by another, Brody, as a triviality to be dismissed without much concious thought (as my friend did in the e-mail).

Here’s Pennington in 1954:

Here was a treatment, that, in its encouragement to eat plentifully, to the full satisfaction of the appetite, seemed to oppose not only the prevailing theory of obesity but, in addition, principles basic to the biological sciences and other sciences as well. It produced a sense of puzzlement that was a mighty stimulant to thought on the matter.

And here’s Brody nearly half a century later:

Does it help people lose weight? Of course it does. If you cannot eat bread, bagels, cake, cookies, ice cream, candy, crackers, muffins, sugary soft drinks, pasta, rice, most fruits and many vegetables, you will almost certainly consume fewer calories. Any diet will result in weight loss if it eliminates calories that previously were overconsumed.

So rather than get upset at my journalist friend who either hadn’t bothered to read GC,BC after all these years, or read it and found it thoroughly forgettable, I got to thinking about the other point made in the e-mail: “I know you believe in calories because you are a scientist.”

Ignoring the possible mischaracterization of me as a scientist, this statement is a little ambiguous. I obviously believe in calories as a measure of energy, whatever that means to believe in such a thing. (It’s like believing in miles as a measure of distance.) So that’s probably not what my friend meant. What I don’t believe in is that discussions of caloric consumption and expenditure tell us anything meaningful about why we get fat or why we lose fat, and I believe that the mantra that ‘a calorie is a calorie is a calorie” serves only to direct attention away from the meaningful characteristics of the macronutrients in our diets.

I’ve been arguing that the original sin in obesity research is this belief that our body fat is regulated by the amount of energy we consume and expend. I think this is simply the wrong way to think about obesity and the chronic diseases with which it associates, and it’s because this is the fundamental assumption underlying most obesity research, it’s the reason why we’ve made so little progress. (And to those who think we have made real progress, I suggest they take a look around at the people walking by and reconsider.)

Another way to put this is that I think this energy balance hypothesis of obesity is an incorrect paradigm and it has to be replaced with a correct paradigm before progress will be made. Obviously my friend doesn’t understand this, so it got me thinking about yet other ways to explain it that might get the point across. This led me to a series of thought experiments, or gedankenexperiments as I learned to call them back in the days when I was writing about physics rather than nutrition and health.

The great thing about thought experiments is that they come unfettered by financial, ethical or real world constraints. We can do virtually anything in these experiments and see what we think is likely to happen. And we can do it quickly. So here goes.

Let’s imagine we have a pair of identical twins, unimaginatively named A and B. They’re males, say, and 20-years-old. They’ve stopped growing and both are weight stable when this experiment begins. (Remember we can do anything we like in a thought experiment, so if we say they’re weight stable initially, then they are.) Now we measure their daily energy expenditure. Let’s say they both expend 3000 calories a day. Every day, day in and day out — 3000 calories. Again, this is true of both of them because they’re identical in all respects.

Now the experiment: We’re going to house A and B in our splendidly livable metabolic ward and keep them there for the next 20 years. (Thought experiments also come free of  Institutional Review Boards. We don’t have to worry about whether this is ethical or not. Our imaginary twins will be perfectly happy anyway because we say so.) We’re going to feed them almost identical diets. Each one is going to get exactly 3000 calories a day so that their intake matches their initial expenditure. If we believe in calories, as my friend might have put it, the fact that we’re matching intake to expenditure and both twins are getting the same intake suggests they will both maintain a stable weight for the duration of the experiment.

But here’s the experimental twist: the diets are not identical, they’re only almost identical. They differ in the macronutrient content of ten percent of the calories. So 2700 calories of the two diets are identical. The other 300 calories of A’s diet will come from sugar — sucrose, to be precise,  molecules of glucose bonded to molecules of fructose. In B’s diet, these 300 calories will come from glucose alone. So A will get 150 calories of fructose that B won’t get, and B will get 150 calories more glucose than A. Other than that the diets are indeed identical with all the macro and micronutrients necessary for the twins to flourish.

Now we run the experiment for 20 years. What happens? Care to guess?  Will A and B still be identical after 20 years of A eating 300 calories of sugar every day that B does not eat?

We know sugar is metabolized differently from the glucose in starch because of the fructose component. Glucose is metabolized by cells throughout the body; fructose is metabolized primarily in the liver. We know the liver will turn some of this fructose into fat and if the fructose is delivered quickly enough (say in liquid form as sugar water), it  likely to cause insulin resistance in the liver, which in turn might cause systemic insulin resistance.  The extra 150 calories of glucose in B’s diet will stimulate more insulin secretion, although for B this will come in the absence of any fructose-induced effects in the liver. One way or the other, A and B will experience different metabolic and hormonal effects, despite eating precisely the same amount of calories in diets that are otherwise 90 percent identical. Their fat cells, for instance, will be on the receiving end of different hormonal and metabolic signals. As Claude Bernard would say, the fat cells would be living in a different  milieu intérieur and this will effect how they change over time.

Another way of asking what happens in this experiment is to ask whether this difference in hormonal and metabolic responses to diets of equal caloric content will have a meaningful effect on, say, fat accumulation and risk factors of chronic disease. Indeed, why would we expect our twins to end up with identical body compositions, weights, and chronic diseases risk factors, when their hormonal and metabolic experiences over those 20 years are indeed different? The question, of course, is are these differences meaningful.

If we’re wedded to the energy balance way of thinking — if we believe in calories, as my friend said — we’re supposed to predict that the twins will end up identical. (That’s certainly what the sugar industry would like us to think. Although the industry might even argue, based on observational studies from the 1980s, that the twin eating sugar will end up leaner and healthier than the twin getting only glucose.)But we’re also likely to maybe hedge a little bit. Okay, maybe the twins will have slightly different body compositions after 20 years. Maybe they’ll even have slightly different chronic disease risks, depending on how this sugar-starch/fructose-glucose trade-off plays out. Surely, though, they’ll weigh the same.  After all, they’re consuming identical calories and these calories are exactly matched to their initial expenditure. So they should end up weighing what they weighed at the get-go and they should both weigh the same. Yes? (And by the way, this thought experiment also negates any effects of food reward or the addictive nature of sugars, because we’re limiting consumption and so even if sugar is addictive and A wants to eat more of it, he’s not getting the opportunity.)

But the fact is even their weight’s can differ, because we’ve only fixed caloric intake. We haven’t fixed their expenditure. Both will get exactly the same number of calories. That’s a condition of the thought experiment. But the different macronutrient composition of A’s diet vs. B’s, can have an effect on fat accumulation and so weight despite caloric consumption being equivalent.

Here’s how: let’s assume two things for the sake of argument. First, the sugar in A’s diet causes A to become insulin resistant. And second, insulin works to put fat in fat tissue. There’s some evidence for our first assumption and the second assumption is in the textbooks; there’s a lot of evidence for that.

Now as A becomes insulin resistant, his pancreas has to secrete more insulin than B’s to handle the equivalent carbohydrate load. So A now might have higher circulating levels of insulin than B. If he does, this means more calories might be fixed in A’s fat tissue than in B’s. Put simply, A might now be getting fatter than B. And as A gets fatter, his body has to compensate for the calories that are being locked away in the fat tissue and for the greater metabolic demands of a heavier body. What does A do?

What A can’t do is eat more, because we’ve fixed his caloric consumption at 3000 calories per day. One option is he could cannibalize his lean mass to feed his growing fat tissue. This can certainly be done without violating any thermodynamic laws. Now A gets fatter while simultaneously losing muscle mass and his weight remains more or less the same. A second option is that A’s body merely expends less energy to make up for the calories being locked away in fat tissue and the greater caloric requirement that comes from being heavier.

Now A gets fatter while his energy expenditure goes down. While B remains in energy balance throughout the experiment — eating 3000 calories a day to match the 3000 calories he expends — A moves into positive energy balance. He’s still consuming 3000 calories every day, but he’s expending less. And the reason he’s in positive energy balance is because he’s amassing fat in his fat tissue and getting heavier. (Although a naive observer, wedded to the energy balance, might decide that A has become a couch potato and that’s why he’s getting fatter. In this case, the direct effect of the sugar is to make A expend less energy and this in turn causes the energy imbalance that makes A fat. The causality is reversed.)

So here’s a possible chain of events in our thought experiment that’s perfectly consistent with the laws of thermodynamics but is inconsistent with the notion that a calorie is a calorie is a calorie: first,  the sugar causes A to be insulin resistant; second, the insulin resistance serves to cause a compensatory elevation of serum insulin levels; third, the elevated insulin causes A to store calories in his fat tissue every day and grow fatter.

If this effect is tiny, say, five calories worth of fat get trapped in A’s fat cells every day, he’ll still put on ten pounds of fat over the 20 years of the experiment and weigh 10 pounds more than his genetically-identical brother eating his almost identical diet. If this fat-trapping amounts to 20 calories a day — still less than one percent of the calories A is consuming — that would amount to forty pounds of excess fat over the course of the experiment. It would still be too subtle of an effect to be observable in the relatively short-term experiments done to date on sucrose consumption.

Now, assuming this did happen, or at least could happen, it would lead us to some other interesting observations as well. For instance, if A puts on this fat above the waist, it will increase his heart disease risk. The more fat he gains, the greater his risk of diabetes. In fact, depending on the size of the effect, he might become diabetic over the course of the study. His brother might not. A’s cancer risk goes up, as well, with his adiposity. So does his risk of getting Alzheimer’s. All without consuming a single calorie more than his twin brother did. In fact, if we run the experiment long enough, the brothers might die of different diseases and one might out live the other by a significant amount.

If you believe this scenario is a possibility, even a likely possibility, as I do, you still believe in the laws of thermodynamics. You’re still thinking like a scientist (as my friend might say). But now, I hope, you can see what I mean by calories being the wrong paradigm. If we believe in calories,as my friend put it, then we believe that the twins end up identical, just as they started, because the quantity of calories consumed in the two diets was identical and it’s quantity that matters, not quality.  What do  you believe?

We’ve started with identical twins, hence the very same genetic make-up. We’ve fed them diets of identical calories. We’ve made a relatively subtle change in macronutrient composition. Do we end up with twins that are still identical; or do we end up with one twin fatter and perhaps sicker than the other? And, keep in mind, as I said, that both twins are limited to 3000 calories a day, and we’re making them eat all 3000, so any addictive effects of the sugar, say, are not relevant. (And if food reward characteristics are meaningful, they have to manifest themselves via the periphery — increasing fat mass, for instance, through central nervous system stimulation of adipocytes — not merely by making us want to eat more. )

Now we can do a host of variations on this experiment. For instance, we can start off with two villages — A and B. Each village has one of each pair of 5000 identical twins. So each sibling pair is identical, but the 5000 sets of twins are as genetically diverse as any 5000 individuals chosen at random. We put  5000 siblings in village A and their 5000 twins  in village B. Now we do the same experiment on this population scale. We measure their energy expenditures. We match intake to expenditure for each pair of twins. Then the twins in village A all get ten percent of their calories as  a sucrose-sweetened beverage. The twins in village B get ten percent of theirs as a beverage with  glucose, not the glucose-fructose mixture that village A is getting.

Now let’s run it out for 20 years. Do villages A and B end up with exactly the same number of obese villagers, exactly the same incidence of diabetes? Heart disease? Cancer? If we run it out for decades, do the two villages have the same mortality rates? The only difference in their diets is the type of carbohydrate that’s sweetening their daily drinks. (And remember, this is a thought experiment: each villager is eating and drinking precisely what we say they’re eating and drinking because they’re under our imaginary control. No propositions need be voted on. We get perfect compliance to our interventions.)

If you believe in the primacy of calories, or you’re a sugar industry spokesperson, then you believe that the two villages start off identical and they end up identical. (Or, for the sugar industry spokespeople, maybe Village A ends up healthier.) If you believe that one village is going to end up fatter and sicker than the other because they’re experiencing different metabolic and hormonal experiences for 20 years, then you’re thinking as I now think and Robert Lustig has argued so publicly. It’s not about the calories; it’s about what those macronutrients do metabolically and hormonally. And who knows what else, maybe the sucrose has an effect on gut biota that the glucose alone does not,or vice versa, and if the two twins develop different bacterial populations in their guts, then this might induce a whole host of other downstream differences that could effect their weight and health.

We can play these thought experiments all day long. That’s the joy of gedankenexperiments. They’re ridiculously inexpensive and we can do them fast. Twenty years in a gedankenexperiment can be instantaneous in reality.

How about this one: instead of feeding twin B (or village B) glucose instead of sucrose, what if he (or it) got dietary fat. So now twin A gets ten percent of his calories as sugar water — pick your poison, so to speak, soda or fruit juice. Now we’re accelerating the delivery method in this thought experiment by making sure these calories are digested quickly. Twin B gets liquid fat, say heavy cream. watered down so that the energy density is effectively identical to the sugar water. So we control for energy density, a factor that the authorities think is key to weight gain. But we dramatically change the macronutrient content of these 300 calories — glucose and fructose calories for A, fat calories for B. Now the hormonal and metabolic responses to these 300 calories are entirely different. Nothing subtle about it. What happens over 20 years? Same body composition, same disease status because the calories are identical? Yes or no?

What if we play more extreme variations with the diets. Rather than play with just ten percent of the calories they consume, let’s play with 50 percent of them.  Twin A (or village A) gets his (or its) calories as a standard American diet, replete with 50 percent carbohydrates, of which, say, a fourth is sugar or high fructose corn syrup as is about the case today in the standard American diet. Twin B (or village B) gets a paleo diet or even a ketogenic diet, same amount of calories, far fewer to almost no carbohydrates, far more fat. What happens? Both twins (or villages) eat precisely the same amount of calories (each or per capita) every day for 20 years. Do they end up identical. Is village A healthier and leaner or village B or neither?

Now let’s change it up entirely, and this will be the last experiment I’ll suggest for the moment. Rather than start with genetically identical twins eating different diets and so generating different hormonal/metabolic responses that way, let’s start with subjects who are not genetically identical, and give them the identical diet. So we can use fraternal twins or siblings, or total strangers as our subjects, but now feed them the exact same diet. We’ll choose our subjects so that they’re the same age, to the  day, the same height and weight and they expend the exact same amount of energy every day (at least when the experiment begins). They’re both equally healthy. And now we feed them the same diet — intake matched perfectly to expenditure — with 10 percent of the calories coming as sucrose. What happens? How do they change over the twenty years of the experiment, given the exact same diet, precisely the same calories, precisely the same physical characteristics, but different genetic make-ups?

The differences in their DNA means they’ll almost assuredly have different hormonal and metabolic responses to the diet. Maybe one does a slightly better job of metabolizing fructose in the liver than the other does. Maybe one secretes a little more insulin in response to the glucose, or is a little more sensitive or resistant to the insulin secreted. Maybe the gut biota in one responds differently. Maybe leptin resistance develops in one but not the other. Anything can happen right, because genes ultimately determine all these responses and their genes are different.

So we’re feeding them exactly the same diet — same quality and quantity — but the hormonal and metabolic responses are going to be different. Their milieu intérieur is going to be different. Maybe a little different as the years go by; maybe a lot. We don’t know.  They may start out relatively identical in relevant physical characteristics, but little by little, they’re going to diverge. Why would we expect them to end up with with the same weight, same fat mass, and even the same chronic disease risk profile?

And if all these things do end up different, would our belief in calories have led us to the same understanding of what happened and why?

If we could do this experiment in real life, it wouldn’t really matter what we believe. Right? We just do the experiment and see how it comes out. (And this is what NuSI hopes to achieve, albeit in far more realistic experiments.) Because we can’t do the experiments, we can do these thought experiments instead and inform our understanding. Time permitting, more will be coming in later posts.

One last note before I conclude here. Let’s go back to our original experiment with twins A and B and their almost identical diets. Imagine, now, as I suggested, that A gets fatter than B and even heavier, because of the effect of the sugar in A’s diet on hepatic metabolism and insulin sensitivity and so serum insulin levels and fat accumulation in fat cells (and maybe all those other factors like gut biota). But A is never able to eat more to compensate for this loss of calories into his fat tissue and his increasing weight, because we don’t let him: we’ve fixed his caloric intake. As a result it’s safe to assume that A would be  hungrier than B is for the entire 20 years. B can eat to satisfy the metabolic requirements of his body; A cannot. How would that manifest itself? Would A at least feel like binging on occasions? Could we create a binge eating disorder that never gets to manifest itself in this particular thought experiment, just by changing the macronutrient composition of the diet?

You can see how thought experiments can lead us to all kinds of conclusions and (at least hypothetical) observations that might not be intuitively obvious otherwise.

I could go on. I’m hoping the point is clear.

Getting back to my friend’s e-mail: Yes, I believe that calories are a useful measure of the energy contained in the foods we consume and a useful measure of the energy our bodies expend. (Just as I believe miles are a useful measure of how far I have to travel to get, say, from Oakland to Los Angeles.) Yes, I believe in the laws of thermodynamics and I believe, as I say in both my books, they always hold true. That’s why we call them laws. But, no, I do not believe that we can learn anything useful about why people get fat or why they get the diseases that associate with getting fat, by focusing on the calories they consume and expend. It’s not about the calories.

This morning, at 9am Eastern Time, we officially launched The Nutrition Science Initiative — NuSI (pronounced “new see”). NuSI is a non-profit organization, technically a 501(c)(3). Its purpose  is to facilitate and fund rigorous, well-controlled experiments targeted at resolving unambiguously many of the outstanding nutrition controversies — to answer the question definitively of what constitutes a healthy diet.

Our conventional dietary wisdom, as I’ve described in my books, is based on science that was simply not adequate to the task of establishing reliable knowledge — poorly-controlled human experiments, observational studies incapable of establishing cause and effect, and animal studies that may or may not say anything meaningful about what happens in humans. NuSI was founded to address this issue and by doing so, we hope, reduce the social and economic burden of obesity and its related diseases. NuSi’s co-founder, and my collaborator in this endeavor, is Peter Attia, who will serve as NuSI’s president.

Peter and I started NuSI as a nights and weekends endeavor with the hope of raising the necessary money to keep the organization running and fund the necessary experiments using crowdfunding  techniques on the internet. Last November, however, we heard from the Laura and John Arnold Foundation expressing interest in what we were trying to accomplish. After many meetings and their due diligence, LJAF has provided NuSI with a seed grant to get our organization up and running and a verbal commitment to help fund some of the key studies. For the past six months, we’ve been working days, nights, and weekends to make it happen. We’ve opened an office in San Diego (where Peter lives) and, as mentioned in my previous post, we’re hiring staff, a research associates and eventually a research director as well.

The support we’ve received  for NuSI has already been remarkable. If you go to the NuSI website, you can see our board of directors, our scientific advisory board and our board of advisors just to get a feel for the researchers and individuals who believe in the mission and are dedicated to making this work.

As to the mission itself, Peter and I have already had meetings with researchers from around the country to discuss and begin the design process of the research that NuSI hopes to fund. These researchers are all excellent scientists, and they’re all skeptical of the hypotheses that we hope to test — the ideal combination.  The experiments will be human trials; they’ll all be rigorously well-controlled, and they’ll all be aimed at identifying unambiguously the causes of obesity and type 2 diabetes, elucidating the underlying mechanisms involved. If all goes well, we’ll move later onto studies that look at longterm effectiveness of dietary therapies based on what we’ve learned.

Both Peter and I have our beliefs about what we’re likely to find, as do the researchers we’ve recruited to join the effort. As we say in our founder’s letter on the NuSI site, we’re not invested in particular outcomes, we’re invested in establishing reliable knowledge on the relationship between diet and disease and so scientifically-sound solutions to the health problems that beset us. One of the quotes that we use on the NuSI website and that I’ve taken to using in my lectures is particularly apt. It’s from Robert Burton’s 1893 book, The Anatomy of Melancholy: ”It is in vain to speak of cures, or think of remedies, until such time as we have considered of the causes . . . cures must be imperfect, lame, and to no purpose, wherein the causes have not first been searched.”

NuSI was founded on the premise that the reason we are beset today by epidemics  of obesity and type 2 diabetes, and the reason physicians and researchers think these diseases are so recalcitrant to dietary therapies, is because of our flawed understanding of their causes. We believe that with a concerted effort and the best possible science, this problem can be fixed.  We hope you’ll give your support to NuSi in anyway you can.

I promised in my last post — yes, far too long ago — that I would give an update on the Metabolism, Diet and Disease Conference, which was held at the end of May in Washington, DC. As the months passed, I was waiting to hear from the organizers that they had posted a video of the panel discussion that ended the conference, and now, as of a few days ago, they have.

The conference itself was rather remarkable. The idea was to bring together from all disciplines researchers working on the various pathologies associated with insulin resistance. It was organized by the editors of BioMed Central, who had come upon the idea after reading The Diet Delusion, which is the British edition of Good Calories, Bad Calories. I was enlisted to help organize and suggest and recruit speakers and executive committee members. The conference also provided the opportunity  to get researchers who had worked on carb-restricted diets — Eric Westman and Jeff Volek, in particular — presenting in a non-nutrition venue to researchers who might otherwise never take their work seriously or at least never imagine that it had relevance to their research in insulin resistance, hyperinsulinemia and the related pathologies. Eugene Fine was also there with a poster on his just published pilot study on ketogenic diets and cancer —  ”Targeting insulin inhibition as a metabolic therapy in advanced cancer: A pilot safety and feasibility dietary trial in 10 patients.”

What I found most fascinating about the conference was how beliefs shifted over the course of the three day event, from unconditional faith in the conventional wisdom to openness and scientific curiosity about the kinds of alternative hypotheses put forward by myself and others.  On the first day of the conference I was having arguments/discussions with researchers about the laws of thermodynamics and how they apply to obesity (or don’t, as I believe) only to find myself sitting with them on a panel on day three as they agreed that the role of refined grains and sugars in cancer and cancer therapy had to be taken seriously.

With that, I highly recommend reading the BioMed Central blog post on the last day’s panel discussion and then watching the video of the discussion itself to see how it played out. You can see for yourself how beliefs and opinions had shifted so that the outcome of the panel discussion was probably something that few of the researchers going in would have ever imagined. I’m not optimistic enough to think that this is a long term change in thinking, or at least not without other factors, experiments and influential researchers keeping the momentum up — and, of course, the science has to turn out to be right or at least mostly right. But it certainly gave me hope that the kinds of issues we’ve been raising again and again outside the research community will soon be addressed critically (i.e., not in a knee-jerk, dismissive manner)  by researchers within the community.

This brings up item number two in this post, and here I’m going to be cribbing considerably  from what Peter Attia recently posted on his blog — eatingacadmy.com . This is our update on NuSI, the Nutrition Science Initiative, and a job we’re hoping to fill in the near future.

As we’ve both alluded to in previous posts, Peter and I founded NuSI earlier this year. Peter is the president and I’m, well, the co-founder. (We rejected “provocateur-in-residence” on the basis that it only captured part of what I do and didn’t quite work officially for an organization that we, and the foundation supporting us, and the scientists with whom we’re working, all take very, very seriously.) NuSi is a non-profit organization with the mission of reducing the economic and social cost of obesity and its related chronic diseases. We hope to achieve this by facilitating and funding the kind of rigorous, meticulously well-controlled and targeted experimental research that has been conspicuously lacking in nutrition research for the half past century.

We’ll say much more about this when we formally and publicly launch NuSI in early September. The ultimate goal is to create what would ideally become a kind of Manhattan Project of Nutrition: a concerted, directed, well-funded research effort composed of the best scientists in the field — all independent and suitably skeptical — working together to generate the evidence necessary to put to rest, one way or the other, all the major and many of the minor controversies in nutrition research. Peter and I have already enlisted  some of the researchers we’d like to get involved, and we’ve spoken to others about possible experiments that might be done in the future. Our hope is that regardless of any initial biases, the evidence generated in these experiments (and replicated in further experiments) will be suitably unambiguous that in, say, 15 years we’ll have little  left to argue about. And if the evidence still leaves room for argument and controversy, then we’ll do more experiments until it doesn’t.

The best part, as Peter has pointed out, is that all this should be doable for less than the cost of developing just one drug in the United States.

Peter and I have been working obsessively to build a world-class team at NuSI, including our Board of Directors, Scientific Advisory Board, Board of Advisors, scientific consortium, and full-time staff.  We can’t wait until we can formally introduce you to our team and collaborators.

We have already hired several positions within NuSI through standard recruiting channels and referrals, but there is one position, in particular, Peter and I thought would be worth bringing directly to the attention of our readers – our Research Associate. We’ve already received a few dozen tremendous applications from individuals with great credentials, but we’re wondering if one critical attribute may be missing or under-represented in our applicants so far. Beyond the tangible skills necessary for this particular role – outlined in the downloadable job posting (below) – this role requires an almost maniacal obsession with nutrition science and a passion for answering the kinds of questions we’ve all been debating in print and in our blogs.  We think there’s a reasonable chance that our future Research Associate is one of you out there reading this right now.

For the full list of job responsibilities and requirements, please download the job posting, which also explains exactly how to apply.  Please do not send any of the application materials to me or Peter directly. You can consider this the first test of the ability to follow simple instructions.

This position should prove to be both extremely challenging and highly rewarding.  We think that we have the opportunity with this organization to change the world, and that the odds are pretty good that we can pull it off. Such opportunities don’t come along frequently in life. And as I said, we’ve already enlisted some of the best scientists in nutrition and obesity research to design and conduct the studies we’ll be funding; you’ll get the opportunity to support them day in and day out.

One very important disclosure: This role will make the proverbial “drinking from a fire hose” seem manageable.  We consider this role (as we do our own)  more of a calling than a job.  If you’re interested, if you feel you meet the necessary requirements, and if we haven’t scared you off yet, please consider applying for this position.

Thank you, and we’re looking forward to sharing the progress of NuSI with all of you.

There’s an interesting conference starting Tuesday the 29^th in Washington – today –and I should have written about it months ago. It’s the Metabolism, Diet and Disease conference being held at the Georgetown University Conference Center. The editors of BioMed Central, a British open-access science publishing company, are the organizers. They contacted me in October 2010 to tell me they had read the British version of Good Calories, Bad Calories – The Diet Delusion – and found it compelling. They were particularly struck by the notion that there are many disciplines involved in the science of obesity, diabetes and their associated chronic diseases, but they don’t read the same journals and they don’t tend to interact in conferences. So their idea was to put together a conference that would solve this problem. Between us, we recruited a first-rate executive board – including the two Nobel Laureates, Michael Brown and Joseph Goldstein – and put together a conference that, as I see it, includes most of the major issues revolving around and feeding into insulin resistance and metabolic syndrome.

The original idea – if you’ll pardon the cliché – is that insulin resistance and metabolic syndrome are the elephant, and we would get all these people in one room, who were studying the legs, the trunks, the tail, the ears, etc., and may not have realized quite what the whole elephant itself looked like. As the recruiting of speakers started, the conference evolved and took on a life of its own. Much of the original idea still exists, but there’s a fair share of the latest esoteric research ideas, for good or for bad (sirtuins, irisin, etc.), which may or may not have anything to do with the elephant itself. But we still have some of the leading researchers in the world talking about everything from the epidemiology of insulin resistance to the pathologies that associate with it – diabetes, heart disease, cancer, aging, etc. – and the pathways and mechanisms that link them all together.

One session, which I’m chairing on Thursday morning, is called “Dietary factors in metabolic diseases,” and the speakers are all addressing the carbohydrate issue.  We have Luc Tappy, the leading fructose biochemist in the world, talking about the role of fructose in metabolic disorders, and then Jeff Volek and Eric Westman talking about the effects of carb-restricted diets, suggesting that the carbohydrates (refined grains, of course, and sugars in particular) might be the fundamental problem. It will be interesting to see how the mainstream researchers take this, as they’re used to thinking about carb-restriction as quackery and now it will be presented as potentially mainstream itself.

On Saturday, I’m in Scottsdale Arizona talking at the National Lipid Associations annual convention. This, too, should be interesting, as I’ll be presenting my Why We Get Fat lecture an hour after Robert Eckel speaks. Eckel is a former president of the American Heart Association who is on record saying that he doesn’t even think low-carb-high-fat diets should ever be tested, that it’s unethical, because they’re so dangerous. After I speak, I’ll get to hear Rob Lustig and Peter Havel talk about sugar and fructose. As I said, it should be an interesting day and an interesting week.

I’m writing this post with a little more haste than is my wont. I’ve received dozens of e-mails asking me to comment on the recent news — ala the the New York Times — that meat-eating apparently causes premature death and disease. So this post is likely to contain more than my usual number of typos, egregious spelling mistakes, grammatical errors, etc. Bear with me. Rather than spend a week rewriting and editing, as I usually do, I’m going to do my best to get this up and out in a few hours.

Back in 2007 when I first published Good Calories, Bad Calories I also wrote a cover story in the New York Times Magazine on the problems with observational epidemiology. The article was called “Do We Really Know What Makes Us Healthy?” and I made the argument that even the better epidemiologists in the world consider this stuff closer to a pseudoscience than a real science. I used as a case study the researchers from the Harvard School of Public Health, led by Walter Willett, who runs the Nurses’ Health Study. In doing so, I wanted to point out one of the main reasons why nutritionists and public health authorities have gone off the rails in their advice about what constitutes a healthy diet. The article itself pointed out that every time in the past that these researchers had claimed that an association observed in their observational trials was a causal relationship, and that causal relationship had then been tested in experiment, the experiment had failed to confirm the causal interpretation — i.e., the folks from Harvard got it wrong. Not most times, but every time. No exception. Their batting average circa 2007, at least, was .000.

Now it’s these very same Harvard researchers — Walter Willett and his colleagues — who have authored this new article claiming that red meat and processed meat consumption is deadly; that eating it regularly raises our risk of dying prematurely and contracting a host of chronic diseases. Zoe Harcombe has done a wonderful job dissecting the paper at her site. I want to talk about the bigger picture (in a less concise way).

This is an issue about science itself and the quality of research done in nutrition. Those of you who have read Good Calories, Bad Calories (The Diet Delusion in the UK) know that in the epilogue I make a point to say that I never used the word scientist to describe the people doing nutrition and obesity research, except in very rare and specific cases. Simply put, I don’t believe these people do science as it needs to be done; it would not be recognized as science by scientists in any functioning discipline.

Science is ultimately about establishing cause and effect. It’s not about guessing. You come up with a hypothesis — force x causes observation y — and then you do your best to prove that it’s wrong. If you can’t, you tentatively accept the possibility that your hypothesis was right. Peter Medawar, the Nobel Laureate immunologist, described this proving-it’s-wrong step as the  ”the critical or rectifying episode in scientific reasoning.” Here’s Karl Popper saying the same thing: “The method of science is the method of bold conjectures and ingenious and severe attempts to refute them.” The bold conjectures, the hypotheses, making the observations that lead to your conjectures… that’s the easy part. The critical or rectifying episode, which is to say, the ingenious and severe attempts to refute your conjectures, is the hard part. Anyone can make a bold conjecture. (Here’s one: space aliens cause heart disease.) Making the observations and crafting them into a hypothesis is easy. Testing them ingeniously and severely to see if they’re right is the rest of the job — say 99 percent of the job of doing science, of being a scientist.

The problem with observational studies like those run by Willett and his colleagues is that they do none of this. That’s why it’s so frustrating. The hard part of science is left out  and they skip straight to the endpoint, insisting that their interpretation of the association is the correct one and we should all change our diets accordingly.

In these observational studies, the epidemiologists establish a cohort of subjects to follow (tens of thousands of nurses and physicians, in this case) and then ask them about what they eat. The fact that they use questionnaires that are notoriously fallible is almost irrelevant here because the rest of the science is so flawed. Then they follow the subjects for decades — 28 years in this case. Now they have a database of diseases, deaths and foods consumed, and they can draw associations between what these people were eating and the diseases and deaths.

The end result is an association. In the latest report, eating a lot of red meat and processed meat is associated with premature death and increased risk of chronic disease. That’s what they observed in the cohorts — the observation.  The subjects who ate the most meat (the top quintile) had a 20 percent greater risk of dying over the course of the study than the subjects who ate the least meat (the bottom quintile). This association then generates a hypothesis, which is why these associations used to be known as “hypothesis-generating data” (before Willett and his colleagues and others like them decided they got tired of their hypotheses being shot down by experiments and they’d skip this step). Because of the association that we’ve observed, so this thinking goes, we now hypothesize that eating red meat and particularly processed meat is bad for our health and we will live longer and prosper more if we don’t do it. We hypothesize that the cause of the association we’ve observed is that red and processed meat is unhealthy stuff.

Terrific. We have our bold conjecture. What should we do next?

Well, because this is supposed to be a science, we ask the question whether we can imagine other less newsworthy explanations for the association we’ve observed. What else might cause it? An association by itself contains no causal information. There are an infinite number of associations that are not causally related for every association that is, so the fact of the association itself doesn’t tell us much.

Moreover, this meat-eating association with disease is a tiny association. Tiny. It’s not the 20-fold increased risk of lung cancer that pack-a-day smokers have compared to non-smokers. It’s a 0.2-fold increased risk — 1/100th the size. So with lung cancer we could buy as a society the observation that cigarettes cause lung cancer because it was and remains virtually impossible to imagine what other factor could explain an association so huge and dramatic. Experiments didn’t need to be done to test the hypothesis because, well, the signal was just so big that the epidemiologists of the time could safely believe it was real. And then experiments were, in effect, done anyway. People quit smoking and lung cancer rates came down, or at least I assume they did. (If not, we’re in trouble here.) When I first wrote about the pseudoscience of epidemiology in Science back in 1995,  “Epidemiology Faces It’s Limits”, I noted that very few epidemiologists would ever take seriously an association smaller than a 3- or 4-fold increase in risk. These Harvard people are discussing, and getting an extraordinary amount of media attention, over a 0.2-fold increased risk. (Horn-blowing alert: my Science article has since been cited by over 400 articles in the peer-reviewed medical literature, according to Thomson Reuter’s Web of Knowledge.)

So how can we explain this tiny association between the risk of eating a lot of red and processed meat — the 1/100th-the-size-of-the-lung-cancer-cigarette effect–compared to eating virtually none? Again, we have an observation — or an association, two or more things happening in concert; let’s think of all the possible reasons that might explain why these two variables, meat-eating and disease, associate together in our cohorts of nurses and physicians.  Here’s how the great German pathologist Rudolph Virchow phrased this in 1849: How, he said, can we “with certainty decide which of two coexistent phenomena is the cause and which the effect, whether one of them is the cause at all instead of both being effects of a third cause, or even whether both are effects of two entirely unrelated causes?” This is the hard part.

The answer ultimately is that we do experiments, which is what Virchow went on to discuss. But we’ll get back to this in a minute. Before we get around to doing the experiments, we must rack our brains to figure out if there are other causal explanations for this association beside the the meat-eating one. Another way to think of this is that we’re looking for all the myriad possible ways our methodology and equipment might have fooled us.  The first principle of good science, as Richard Feynman liked to say, is that you must not fool yourself because you’re the easiest person to fool. And so before we go public and commit ourselves to believing this association is meaningful and causal, let’s think of all the ways we might be fooled. Once we’ve thought up every possible, reasonable alternative hypotheses (space aliens are out on this account), we can then go about testing them to see which ones survive the tests: our preferred hypothesis (meat-eating causes disease, in this case) or one of the many others we’ve considered.

So let’s think of reasonable ways in which people who eat a lot of meat might be different from people who don’t, looking specifically for differences that might also explain some or all of the association we observed between meat-eating, disease and premature death. What else can explain this association, which might have nothing to do with whatever happens when we consume meat or processed meat?

Zoe Harcombe made this point beautifully using the Harvard data. The obvious clue is that as we move from the bottom quintile of meat-eaters (those who are effectively vegetarians) to the top quintile of meat-eaters we see an increase in virtually every accepted unhealthy behavior — smoking goes up, drinking goes up, sedentary behavior (or lack of physical activity) goes up — and we also see an increase in markers for unhealthy behaviors — BMI goes up, blood pressure, etc. So what could be happening here?

If you go back and read my New York Times Magazine article on this research, you’ll see that I discussed a whole host of effects, known technically as confounders  — they confound the interpretation of the association — that could explain associations between two variables but have nothing to do biologically with the variables themselves. One of these confounders is called the compliance or adherer effect. Heres’ what I said about it in the article:

The Bias of Compliance

A still more subtle component of healthy-user bias has to be confronted. This is the compliance or adherer effect. Quite simply, people who comply with their doctors’ orders when given a prescription are different and healthier than people who don’t. This difference may be ultimately unquantifiable. The compliance effect is another plausible explanation for many of the beneficial associations that epidemiologists commonly report, which means this alone is a reason to wonder if much of what we hear about what constitutes a healthful diet and lifestyle is misconceived.

The lesson comes from an ambitious clinical trial called the Coronary Drug Project that set out in the 1970s to test whether any of five different drugs might prevent heart attacks. The subjects were some 8,500 middle-aged men with established heart problems. Two-thirds of them were randomly assigned to take one of the five drugs and the other third a placebo. Because one of the drugs, clofibrate, lowered cholesterol levels, the researchers had high hopes that it would ward off heart disease. But when the results were tabulated after five years, clofibrate showed no beneficial effect. The researchers then considered the possibility that clofibrate appeared to fail only because the subjects failed to faithfully take their prescriptions.

As it turned out, those men who said they took more than 80 percent of the pills prescribed fared substantially better than those who didn’t. Only 15 percent of these faithful “adherers” died, compared with almost 25 percent of what the project researchers called “poor adherers.” This might have been taken as reason to believe that clofibrate actually did cut heart-disease deaths almost by half, but then the researchers looked at those men who faithfully took their placebos. And those men, too, seemed to benefit from adhering closely to their prescription: only 15 percent of them died compared with 28 percent who were less conscientious. “So faithfully taking the placebo cuts the death rate by a factor of two,” says David Freedman, a professor of statistics at the University of California, Berkeley [who passed away, regrettably, in 2008]. “How can this be? Well, people who take their placebo regularly are just different than the others. The rest is a little speculative. Maybe they take better care of themselves in general. But this compliance effect is quite a big effect.”

The moral of the story, says Freedman, is that whenever epidemiologists compare people who faithfully engage in some activity with those who don’t — whether taking prescription pills or vitamins or exercising regularly or eating what they consider a healthful diet — the researchers need to account for this compliance effect or they will most likely infer the wrong answer. They’ll conclude that this behavior, whatever it is, prevents disease and saves lives, when all they’re really doing is comparing two different types of people who are, in effect, incomparable.

This phenomenon is a particularly compelling explanation for why the Nurses’ Health Study and other cohort studies saw a benefit of H.R.T. [hormone replacement therapy, one subject of the article] in current users of the drugs, but not necessarily in past users. By distinguishing among women who never used H.R.T., those who used it but then stopped and current users (who were the only ones for which a consistent benefit appeared), these observational studies may have inadvertently focused their attention specifically on, as Jerry Avorn says, the “Girl Scouts in the group, the compliant ongoing users, who are probably doing a lot of other preventive things as well.”

It’s this compliance effect that makes these observational studies the equivalent of conventional wisdom-confirmation machines. Our public health authorities were doling out pretty much the same dietary advice  in the 1970s and 1980s, when these observational studies were starting up, as they are now. The conventional health-conscious wisdom of the era had it that we should eat less fat and saturated fat, and so less red meat, which also was supposed to cause colon cancer, less processed meat (those damn nitrates) and more fruits and vegetables and whole grains, etc. And so the people who are studied in the cohorts could be divided into two groups: those who complied with this advice — the Girl Scouts, as Avorn put it — and those who didn’t.

Now when we’re looking at the subjects who avoided red meat and processed meat and comparing them to the subjects who ate them in quantity, we can think of it as  effectively comparing the Girl Scouts to the non-Girl Scouts, the compliers to the conventional wisdom to the non-compliers. And the compliance effect tells us right there that we should see an association — that the Girl Scouts should appear to be healthier. Significantly healthier. Actually they should be even healthier than Willet et al. are now reporting, which suggests that there’s something else working against them (not eating enough red meat?). In other words, the people who avoided red meat and processed meats were the ones who fundamentally cared about their health and had the energy (and maybe the health) to act on it. And the people who ate a lot of red meat and processed meat in the 1980s and 1990s were the ones who didn’t.

Here’s another way to look at it: let’s say we wanted to identify markers of people who were too poor or too ignorant to behave in a health conscious manner in the 1980s and 1990s or just didn’t, if you’ll pardon the scatological terminology, give a sh*t. Well, we might look at people who continued to eat a lot of bacon and red meat after Time magazine ran this cover image in 1984 — “Cholesterol, and now the bad news”. I’m going to use myself as an example here, realizing it’s always dangerous and I’m probably an extreme case. But I lived in LA in the 1990s where health conscious behavior was and is the norm, and I’d bet that I didn’t have more than half a dozen servings of bacon or more than two steaks a year through the 1990s. It was all skinless chicken breasts and fish and way too much pasta and cereal (oatmeal or some other non-fat grain) and thousands upon thousands of egg whites without the yolks. Because that’s what I thought was healthy.

So when we compare people who ate a lot of meat and processed meat in this period to those who were effectively vegetarians, we’re comparing people who are inherently incomparable. We’re comparing health conscious compliers to non-compliers; people who cared about their health and had the income and energy to do something about it and people who didn’t.  And the compliers will almost always appear to be healthier in these cohorts because of the compliance effect if nothing else. No amount of “correcting” for BMI and blood pressure, smoking status, etc. can correct for this compliance effect, which is the product of all these health conscious behaviors that can’t be measured, or just haven’t been measured. And we know this because they’re even present in randomized controlled trials. When the Harvard people insist they can “correct” for this, or that it’s not a factor, they’re fooling themselves. And we know they’re fooling themselves because the experimental trials keep confirming that.

That was the message of my 2007 article. As one friend put it years ago to me (and I wish I could remember who so I could credit him or her properly), when these cohort studies compare mostly health advice compliers to non-compliers,  they might as well be comparing Berkeley vegetarians who eat at Alice Water’s famous Chez Panisse restaurant once a week after their yoga practice to redneck truck drivers from West Virginia whose idea of a night on the town is chicken-fried steak (and potatoes and beer and who knows what else) at the local truck stop. The researchers can imply, as Willett and his colleagues do, that the most likely reason these people have different levels of morbidity and mortality is the amount of meat they eat; but that’s only because that’s what Willett and his colleagues have to believe to justify the decades of work and tens, if not hundreds, of millions of dollars that have been spent on these trials. Not because it’s the most likely explanation. It’s far more likely that the difference is caused by all the behaviors that associate with meat-eating or effective vegetarianism — whether you are, in effect, a Girl Scout or not.

This is why the best epidemiologists — the one’s I quote in the NYT Magazine article — think this nutritional epidemiology business is a pseudoscience at best. Observational studies like the Nurses’ Health Study can come up with the right hypothesis of causality about as often as a stopped clock gives you the right time. It’s bound to happen on occasion, but there’s no way to tell when that is without doing experiments to test all your competing hypotheses. And what makes this all so frustrating is that the Harvard people don’t see the need to look for alternative explanations of the data — for all the possible confounders — and to test them rigorously, which means they don’t actually see the need to do real science.

As I said, it’s a sad state of affairs.

Now we’re back to doing experiments — i.e., how we ultimately settle this difference of opinion. This is science.  Do the experiments.  We have alternative causal explanations for the tiny association between meat-eating and morbidity and mortality. One is that it’s the meat itself. The other is that it’s the behaviors that associate with meat-eating. So do an experiment to see which is right. How do we do it? Well you can do it with an N of 1. Switch your diet, see what happens. Or we can get more meaningful information by starting with your cohort of subjects and assigning them at random either to a diet rich red meat and processed meat, or to a diet that’s not — a mostly vegetarian diet. By assigning subjects at random to one of these two interventions, we mostly get rid of the behavioral (and socio-economic and educational…) factors that might associate with choosing of your own free will whether to be a vegetarian (or a mostly-vegetarian) or a meat-eater.

So we do a randomized-controlled trial. Take as many people as we can afford, randomize them into two groups — one that eats a lot of red meat and bacon, one that eats a lot of vegetables and whole grains and pulses-and very little red meat and bacon — and see what happens. These experiments have effectively been done. They’re the trials that compare Atkins-like diets to other more conventional weight loss diets — AHA Step 1 diets, Mediterranean diets, Zone diets, Ornish diets, etc. These conventional weight loss diets tend to restrict meat consumption to different extents because they restrict fat and/or saturated fat consumption and meat has a lot of fat and saturated fat in it. Ornish’s diet is the extreme example. And when these experiments have been done, the meat-rich, bacon-rich Atkins diet almost invariably comes out ahead, not just in weight loss but also in heart disease and diabetes risk factors. I discuss this in detail in chapter 18 of Why We Get Fat, ”The Nature of a Healthy Diet.” The Stanford A TO Z Study is a good example of these experiments. Over the course of the experiment — two years in this case — the subjects randomized to the Atkins-like meat- and bacon-heavy diet were healthier. That’s what we want to know.

Now Willett and his colleagues at Harvard would challenge this by saying somewhere along the line, as we go from two years out to decades, this health benefit must turn into a health detriment. How else can they explain why their associations are the opposite of what the experimental trials conclude? And if they don’t explain this away somehow, they might have to acknowledge that they’ve been doing pseudoscience for their entire careers. And maybe they’re right, but I certainly wouldn’t bet my life on it.

Ultimately we’re left with a decision about what we’re going to believe: the observations, or the experiments designed to test those observations. Good scientists will always tell you to believe the experiments. That’s why they do them.

Egregious (and embarrassing) error correction: In an early version of the post, I suggested that if you read the chapter on nutritional epidemiology in the textbook Modern Epidemiology, you’d see that the best epidemiologists agree that this pursuit is pathological. A reader from my institution — a UC Berkeley grad student — pointed out that the chapter on nutritional epi in the textbook was actually written by Walter Willett and that, not surprisingly, it does not agree with this position. Here’s how Willett ends that chapter:

The last two decades have seen enormous progress in the development of nutritional epidemiology methods. Work by many investigators has provided clear support for the essential underpinnings of this field. Substantial between-person variation in consumption of most dietary factors in populations has been demonstrated, methods to measure diet applicable to epidemiologic studies have been developed, and their validity has been documented. Based on this evidence, many large prospective cohort studies have been established that are providing a wealth of data on many outcomes that will be reported during the next decade. In addition, methods to account for errors in measurement of dietary intake have been developed and are beginning to be applied in reporting findings from studies of diet and disease.

Nutritional epidemiology has contributed importantly to understanding the etiology of many diseases. Low intake of fruits and vegetables has been shown to be related to increased risk of cardiovascular disease. Also, a substantial amount of epidemiologic evidence has accumulated indicating that replacing saturated and trans fats with unsaturated fats can play an important role in the prevention of coronary heart disease and type 2 diabetes. Many diseases—as diverse as cataracts, neural-tube defects, and macular degeneration—that were not thought to be nutritionally related have been found to have important dietary determinants. Nonetheless, much more needs to be learned regarding other diet and disease relations, and the dimensions of time and ranges of dietary intakes need to be expanded further. Furthermore, new products are constantly being introduced into the food supply, which will require continued epidemiologic vigilance.

The development and evaluation of additional methods to measure dietary factors, particularly those using biochemical methods to assess long-term intake, can contribute substantially to improvements in the capacity to assess diet and disease relations. Also, the capacity to identify those persons at genetically increased risk of disease will allow the study of gene–nutrient interactions that are almost sure to exist. The challenges posed by the complexities of nutritional exposures are likely to spur methodologic developments. Such developments have already occurred with respect to measurement error. The insights gained will have benefits throughout the field of epidemiology.

Now the reason I made this mistake is because I was rushing (no excuse, despite the warning up front) and so working from memory about a chapter that the UCLA epidemiologist Sander Greenland, one of the editor/authors of Modern Epidemiology, sent me when I was writing the New York Times Magazine article in 2007. The chapter Greenland was discussing and that he had sent me at the time was one he had authored, chapter 19 — “Bias Analysis” — and it was discussing observational epidemiology in general.

Here’s Greenland on the problem with all these studies — nutritional epi included — and how they’re interpreted:

Conventional methods assume all errors are random and that any modeling assumptions (such as homogeneity) are correct. With these assumptions, all uncertainty about the impact of errors on estimates is subsumed within conventional standard deviations for the estimates (standard errors), such as those given in earlier chapters (which assume no measurement error), and any discrepancy between an observed association and the target effect may be attributed to chance alone. When the assumptions are incorrect, however, the logical foundation for conventional statistical methods is absent, and those methods may yield highly misleading inferences. Epidemiologists recognize the possibility of incorrect assumptions in conventional analyses when they talk of residual confounding (from nonrandom exposure assignment), selection bias (from nonrandom subject selection), and information bias (from imperfect measurement). These biases rarely receive quantitative analysis, a situation that is understandable given that the analysis requires specifying values (such as amount of selection bias) for which little or no data may be available. An unfortunate consequence of this lack of quantification is the switch in focus to those aspects of error that are more readily quantified, namely the random components.

Systematic errors can be and often are larger than random errors, and failure to appreciate their impact is potentially disastrous. The problem is magnified in large studies and pooling projects, for in those studies the large size reduces the amount of random error, and as a result the random error may be but a small component of total error. In such studies, a focus on “statistical significance” or even on confidence limits may amount to nothing more than a decision to focus on artifacts of systematic error as if they reflected a real causal effect.

Download NEJM Image Challenge (PDF)

Checking in after a long absence (working too hard, and blogging too little), I have news and updates for 2012.

The first order of business is a letter to the editor of the New York Times in response to Tara Parker-Pope’s “The Fat Trap” article that ran on the cover of the January 1st  NYT Magazine. I wrote the letter with my colleague Peter Attia, more on whom shortly, and we posted it online at ipetitions.com. We tried to get it signed by as many MDs and PhDs as possible, to make the case to the editors at the Times and to Tara herself that a significant number of medical professionals and researchers take the alternative hypothesis of obesity seriously and so should they. As it is, we were able to get over 250 such degreed cosigners, which was more than we expected and more than we hoped. The Times is running a 150-word summary of the letter as a letter to the editor in this week’s magazine with a link to the full letter on line, which you should all feel free to sign. I also recommend you click on the signatures link to see who’s signed it and read the comments.

“The Fat Trap” made the point that obesity is effectively incurable. The letter argues that it only appears to be incurable because the wrong treatment is being used, and the wrong treatment  is being used because the people studying the disorder don’t understand what causes it in the first place  – like trying to treat a bacterial infection with an anti-viral medication and then throwing up your hands and saying it’s hopeless when the treatment doesn’t work. Should they ever get the cause right, then the correct treatment becomes obvious.

Second order of business is my colleague Peter Attia. Peter and I started working together last April after he came to San Francisco to meet me. He had recently read Why We Get Fat and Good Calories, Bad Calories and my sugar article in The New York Times, and he had 27-pages of questions he wanted to ask. I was impressed, if not awed, and we hit it off. I suggested to Peter that he should take over the insurgency and I’d be the figurehead, as I’ve been burnt out and overworked for a decade. The good news is he took me up on it, except that now he has me working twice as hard as ever to help.

Among the projects we have in the works is a non-profit, the Nutrition Science Initiative (NuSI), to raise money for the kind of research we think is necessary to clarify the relationship between dietary nutrients, obesity. diabetes and their related chronic diseases. We have a specific plan of research to pursue (or rather to fund so that established, unbiased researchers can then do the studies) and have already recruited a world class scientific advisory board and executive board. I’ll fill in the details in the next couple of weeks as we get closer to going on-line.

Peter’s blog, The War on Insulin.com, is up and running. Peter takes many of my ideas and expands on them from his own unique perspective. He also blogs about his own personal experience on a conventional healthy diet and then a ketogenic diet. What makes this particularly interesting is that Peter is a fanatic endurance athlete and an obsessive self-experimenter, and he comes at his experience and his blog with a significant amount of medical training and acumen. I highly recommend that he be read.   (And if anyone can figure out how he manages to workout 23 hours a week, function in a full-time job, blog regularly on nutrition and physical activity, and be there for his family and not collapse in a puddle of exhaustion, please let me know.)

The next order of business is an exciting and promising project. My friend Larry Istrail is a medical student at Virginia Commonwealth University. He’s recently created the Ancestral Weight Loss Registry, to collect and publish data on individuals who have tried to lose weight with a paleo/carbohydrate-restricted diet. He’s also spent much of the last few years compiling clinical data on many aspects of carb-restricted eating in the related science section, such as the efficacy of such diets for weight loss or the effects of saturated fats and cholesterol intake on heart disease. Tara Parker-Pope’s article in The New York Times claimed that the National Weight Control Registry (about which I could easily fill up a few blog posts with criticisms) has some 10,000 people registered in over 15 years. We’re hoping that the  Ancestral Weight Loss Registry will beat that in a few months. Using this kind of self-selected data to do good science is tricky if not perhaps impossible, but it will be interesting to see what happens.

If you’re reading this and you’ve lost significant weight on a carb-restricted/paleo diet (or if you haven’t), please check out the  Ancestral Weight Loss Registry and enter your details. I’m also hoping you’ll pass this on to your friends and if you have a blog or a podcast, to your readers and/or listeners.

The last order of nutrition/obesity-related interest is that I have some speaking engagements coming up in the next few months  and I thought I’d mention them, which I’ve been lax in doing in the past. I’m giving a couple of readings in the Berkeley/Oakland area, a lecture at the Seattle Town Hall and the University of Texas San Antonio, an after dinner talk at a diabetes conference at Auburn University, a talk at an integrative health conference in Los Angeles and a lecture in Santa Cruz, courtesy of the local office of education. There’s also a couple of talks coming up in Europe — in Amsterdam and Brussels — and a slew of conferences and keynote addresses into the summer and fall, details of which will go up shortly. The dates and locations and links through April (Amsterdam not yet included)  are up on my calendar, and I promise I’ll do a much better job in the future of keeping the calendar updated far in advance.

Along these lines (okay, this is the last order of nutrition/obesity-related interest), the email subscription to my blog now works as it should, and will NOT subscribe readers to “Two and the Zoo” like it used to. Feel free to sign up if you’re interested.

Finally, my best friend Marion Smith recently did a post on her website about our shared history with one of the more distasteful but apt terms in journalism and writing. Marion credits me for the terminology, although I’m going to give credit here to Calvin Trillin, who came up with it. It was from reading Trillin in my salad days — back when I had time to read — that I learned of the term and the technique and embraced it. Either way, Marion is a wonderful writer and nutrition and weight is not the subject. Both reasons to read her blog.

And with that, a belated Happy New Years.

In our last post, we discussed, among other things, an experiment from the 1960s that Dr. Stephan Guyenet of wholehealthsource.org has evoked to support the food reward/palatability hypothesis of obesity. This was an experiment by Sami Hashim and Ted Van Itallie published in 1965. Four subjects, two lean and two obese, consumed a formula diet from a machine that dispensed it through a tube. The lean subjects consumed as many calories as they always did and maintained their weight. The obese subjects responded by dramatically reducing their caloric intake – to 400 calories a day or less. The male subject continued this regimen for the better part of a year (although dispensing with the tube-feeding) and lost 200 pounds, half his body weight, doing so.

When I interviewed Dr. Van Itallie for GC,BC he considered this tube-feeding device to be his primary contribution to the field of obesity research. I would question whether this is a significant contribution to anything, but his experiments were interesting, nonetheless, and I’d like to take a little more space here to discuss them.

The questions we want to answer ultimately about this experiment are simple ones: why did the obese subjects consume so few calories on the feeding tube, and why was the one guy able to keep this up with the formula, eating 400 calories a day, lose so much weight, and do so without apparently being particularly hungry. Dr. Guyenet evokes the low food reward value of the formula to explain this (nonsensical as it seems when the formula itself, as discussed in the last post, was loaded with sugar). But do we need the food reward/palatability hypothesis to explain this particular n=1 anecdotal observation about extreme weight loss?

One flaw I find in Dr. Guyenet’s interpretation of this experiment may have to do with the fact that he’s a laboratory researcher not a clinician working with human subjects or patients. As such, it may be hard for him to imagine all the various ways that humans will screw up an experiment, ideal as they might be otherwise as the experimental subject of interest.

One obvious problem with any such study like Hashim and Van Itallie’s – non-randomized, only a few subjects (4 in this case) – is that there are an almost infinite number of reasons to explain what they observed. Here’s Hashim and Van Itallie mulling this issue themselves:

Whether the inhibition of food intake exhibited by obese patients represents a physiologic effect of massive stores of fat, or whether it results from psychologic factors relating to guilt about the obesity, fear of the feeding device, inability to adjust to the formula, or some other cause, is unknown. The fact that such a striking difference does seem to be present merits further investigation.

Oddly enough, Hashim and Van Itallie neglect the most obvious explanation for what they witnessed, and I suspect it’s because they, too, had little experience until then with obesity research on human subjects. Maybe their two obese subjects reduced their food intake so dramatically because they had reason to do so. They wanted to lose weight and had significant excess weight to lose. The lean individuals didn’t, and so they didn’t bother to reduce their intake. That simple. The obese man, for instance, was pathologically obese – 400 pounds – and he volunteered for the experiment.  So maybe he was using this as an opportunity to lose weight, consciously or unconsciously. Why else volunteer?

This is a very common problem with experiments that use obese or overweight human subjects and effectively any kind of diet or lifestyle intervention. People who are overweight have a motivation to lose weight. We could argue that the fatter they are, the greater their motivation. Even if they’re instructed specifically not to eat less, for instance, because the researcher want to study the efficacy of exercise or of ad libitum eating on, say, a low-fat or low-carb diet, the researchers can never be sure that the subjects didn’t ignore their instructions because, well, they want to lose weight and they’re in the kind of institutional setting that will help motivate them.

When I was researching GC,BC, quite a few obesity researchers made note of this problem. You could give obese subjects virtually any dietary or lifestyle instructions in an institutional setting, I was told, and they’ll promptly lose at least a little weight, because they want to lose weight. That’s why they’re there; that’s why they volunteered. And they are likely to eat less (calories and/or carbs or fat) no matter what instructions they get. That’s why when you do studies with obese subjects you randomize them into two or more arms and try to make sure that each arm has an equal intervention – a diet or lifestyle change that requires an equal amount of counseling and effort to get right. This should do a pretty good job of dealing with this motivation confounder, although you can never be sure.

In 1971 Hashim and Van Itallie returned to their food dispensing machine experiments in a paper in the New England Journal of Medicine. This time they reported that female obese subjects consumed less and lost weight on the tube-fed formula diet, while male lean subjects did neither. The NEJM published a letter in response, by a Virginia physician, making this precise point:

“All that one has to postulate to explain the differences in the behavior of the two groups is that the male group was at acceptable weight, and knew it, and that the other group was obese, and knew it, and that the obese females were trying to lose weight and did so by the very simple expedient of reducing the volume of food intake.”

In their response to the letter, Hashim and Van Itallie argued that this was unlikely to be the case based on the fact that all the obese subjects, men and women, behaved the same way: “It overstrains credulity to interpret such uniform behavior as being entirely the product of a systematic attempt to lose weight.”  Well, not my credulity, but it’s open to investigation. In a real science, Hashim and Van Itallie would have tried to figure out a way to rigorously test the hypothesis, not leave it up to us to judge whether their defense was adequate.

(That van Itallie and Hashim defended their decision to compare male lean subjects with female obese subjects on the basis that this was the best they could do – “Unfortunately, our choice of subjects was limited; thus, this `fault’ in design was thrust upon us,” they say– boggles the mind. It may be all you need to know to understand why I consider most obesity research to border on pseudoscience. And this, in a paper published in what’s considered to be one of the two best medical journals in the world. Sigh.)

Okay, so we still haven’t answered the question of why even one obese subject – male or female – would be able to maintain a 400 calorie-a-day regimen of this formula for months and lose significant weight without apparent hunger. Is it because the formula has a low-food reward value (sugar or not) or because, well, it lowers insulin levels and so this man is happily consuming his own fat, just as if he was on a protein-sparing modified fast (as many as 600 or 800 calories a day, but no carbohydrates) or starving himself entirely, or even, perhaps, eating an ad libidum ketogenic diet of virtually all protein and fat, in which hunger is also apparently suppressed in association with weight loss?

This brings us back to the question of calories vs. carbohydrates and why diets work when they do, as I discussed in this post. And it always discourages me when this issue is missed by otherwise intelligent commentators like Dr. Guyenet.

Here’s the composition of the formula, according to Dr. Guyenet: “carbohydrate supplied 50 percent of the calories, protein 20 per cent and fat 30 percent.” So, regardless of the food reward value, the 400-pound subject was eating a diet of roughly 200 calories a day of carbohydrates. Hashim and Van Italie tell us that he was eating about 3000 calories a day prior to going on the feeding tube, and if his pre-experiment diet was half carbs, than our subject cut his carbohydrate consumption on the feeding tube from 1500 calories a day or so down to 200.

Two hundred calories of carbohydrates – 50 grams worth – was low enough to be ketogenic. “Ketonuria was always present,” Hashim and Van Itallie write, “ and the blood ketone levels on several occasions was 15 mg. per 100 ml.” This means insulin levels on the diet were extremely low, despite the 50 percent carbs.

Why wouldn’t he be hungry? Well, here’s a guy who lost 200 pounds in 265 days. Assuming three quarters of that weight loss is fat and the rest lean tissue, that’s 150 pounds of fat that he burnt in 265 days, and this was fat he wasn’t burning back when he was weight stable and eating 3000 calories a day to run his body. That’s an average of about 1980 calories of his own fat he was consuming every day. The calculation is simple:

150 pounds of fat x 3500 kcal per pound of fat/265 days = 1981 kcal/day

Add to this the fifth of a pound or so of lean tissue he’s also consuming daily (50 pounds of lean tissue/265 days), and we can easily add another few hundred calories a day from the protein. So the fact that he could reduce his food consumption from 3000 calories a day down to 400 without feeling hunger, is not something that requires a lot of extra thought if we realize that in doing so, and lowering his insulin levels, he was consuming far more than 2000 calories a day of his own body – fat and protein — to make up the difference. We don’t have to evoke somewhat vague properties about the hyper-rewarding nature of the food to explain this.

There’s one other aspect of Hashim and Van Itallie’s crazy exercise in dispensing formula diets from a tube that I find relevant. The goal of  the experiment they published in the New England Journal of Medicine in 1971 was to examine whether obese and lean subjects would respond to surreptitious changes in the nutrient density of the formula they were consuming. Hashim and Van Itallie diluted the formula with different amounts of water, fed it through the tube, and measured whether the subjects responded to the dilution by taking in more formula.

In this case, we can assume that the formula might indeed have been bland and so low in food reward value because Van Itallie and his researchers made it themselves without sugar. The carbohydrates in it were dextrose (glucose) and lactose.

Now the obese female subjects all “lost weight as long as they fed themselves by machine.” The lean males seemed to adjust to dilution of the diet by drinking more of it. Not so the obese women. They just kept drinking little of the stuff and losing weight, regardless of the dilution.

So if we buy into the food reward hypothesis (and we don’t buy the notion that the “obese females were trying to lose weight and did so by the very simple expedient of reducing the volume of food intake”), then we can conclude that non-rewarding food serves to lower the set-point of women and so they eat less and lose weight. All well and good, enormous assumption aside.

But Hashim and Van Italie also studied two obese adolescents under the same circumstances. The obese adolescents (boys, 13 and 15) ate as much as ever of the formula diet, although one of them actually refused to use the feeding machine after a week. Smart kid. One way or the other, “these two obese juvenile subjects differed from the adult subjects in that they either maintained or gained weight while receiving the machine-dispensing formula.”

Hmmmm…. Does this mean the reward value if a food is dependent on age? On sex? Men find bland formulas rewarding and get fat on them, women don’t? Or children do and adults don’t? Does it mean that Hashim and Van Italie’s food reward studies are so poorly controlled that, to steal a phrase, it overstrains credulity to consider any interpretation meaningful? My vote is with the latter, but once again I’m biased.

When last I left off, the subject of discussion was the critical question about the food reward/palatability hypothesis of obesity: Can palatability and reward value of foods be disassociated from the metabolic and hormonal effects of the individual nutrients being consumed and, in particular, the sugar and refined grains that “hyper-rewarding” foods seem to invariably contain?

Let’s start with the experiment in humans that Dr. Stephan Guyenet of wholehealthsource.org finds such compelling support of the food reward hypothesis. This was work done by Ted Van Itallie and Sami Hashim back in the 1960s. (For an idea of the simplistic notions held by Dr. Hashim about obesity and its cause and prevention, I highly recommend this video here. I discuss Dr. Van Itallie’s critical role in shaping the current thinking about obesity — i.e., the mess we’re in today — in chapter 23 of Good Calories, Bad Calories.)

In his “Case for the Food Reward Hypothesis of Obesity, Part II” post, Dr. Guyenet argues that this experiment is important because it demonstrates what he considers one of several critical requirements for the validity of the food reward hypothesis: “Decreasing the reward/palatability of the diet should cause fat loss in animals and humans that carry excess fat.” Here’s what he says:

One of the most striking weight loss studies I’ve seen was conducted in 1965 and involved feeding a bland liquid diet through a dispensing straw (12).  Lean and obese volunteers were instructed to eat as much of the liquid food as they wanted, but they were permitted no other food.  While lean volunteers ate a normal amount of calories and maintained weight, obese volunteers dramatically reduced their spontaneous calorie intake and lost fat rapidly, with one man losing 200 lbs in 255 days without hunger.  This is exactly what one would expect if unpalatable/unrewarding food lowered the biologically “defended” level of fat mass.  Interestingly, the diet was high in sugar but was otherwise very low in palatability/reward value.

This was Dr. Guyenet’s second discussion of the tube-feeding paper. As he explained in an earlier post on this experiment, the total number of subjects was four: two lean and two obese. The two lean were kept on the feeding machine for 16 and 9 days. They didn’t bother to decrease caloric intake, and so their experiment ended then. The two obese subjects, however, curtailed intake dramatically, to 275 calories per day for the male volunteer and 144 for the female). The man stayed with the feeding machine for another 70 days and was then sent home with the formula and the instruction to drink only 400 calories a day. He kept this up for another half year until he had lost the 200 pounds. Says Dr.  Guyenet (in “Food Reward: a Dominant Factor in Obesity, Part II”, his earlier post):

 This machine-feeding regimen was nearly as close as one can get to a diet with no rewarding properties whatsoever. Although it contained carbohydrate and fat, it did not contain any flavor or texture to associate them with, and thus the reward value of the diet was minimized. As one would expect if food reward influences the body fat setpoint, lean volunteers maintained starting weight and a normal calorie intake, while their obese counterparts rapidly lost a massive amount of fat and reduced calorie intake dramatically without hunger. This suggests that obesity is not entirely due to a “broken” metabolism (although that may still contribute), but also at least in part to a heightened sensitivity to food reward in susceptible people. [The italics are mine.]

So immediately we have a problem, and it strikes me as near-fatal for the food reward hypothesis of obesity. In Dr. Guyenet’s first post on the experiment (the one immediately above), he says that the regimen “was nearly as close as one can get to a diet with no rewarding properties whatsoever…. It did not contain any flavor or texture to associate them with, and thus the reward value of the diet was minimized.” In his second post (the first of the two I quote, just to make life confusing), he notes that the diet was “high in sugar” although he tries to hold onto the food reward hypothesis by stating that it “was otherwise very low in palatability/reward value.”

This is why I asked Dr. Guyenet at the AHS whether the formula diet had sugar in it. If it did, then how could it be bland? And how could it have a low food reward value, which is, more or less, the whole point? It might have a lower food value than what the two obese subjects were eating prior to the experiment, but low?

As Hashim and Van Italie noted in a footnote in their 1965 paper, and as Dr. Guyenet notes in his blog, the formula used was Nutrament. This was a liquid diet formula that went on sale in 1960 (according to Wikipedia), and if the composition then was anything like the composition now, a significant portion of the carbohydrate calories came from sugar.

So was it non-rewarding? Hashim and Van Italie refer to it as “bland” and Guyenet assumes it was as well, hence his description of it as “otherwise very low in palatabily/reward value.” But it had to be sweet if it had significant sugar in it; and indeed in the modern incarnation of Nutrament, which may or may not be the same as the original, there are 47 grams in every 12-ounce serving. This happens to be more sugar than you’d find in a 12-ounce can of Coke.

Frankly, the stuff sounds awful, but low in reward value? Well, only if a Coke is, too, and certainly not if Dr. Guyenet includes “liquid calories, particularly sweetened beverages” among the low-hanging fruit of the food reward hypothesis, which he does.

In fact, the point of a diet formula like Nutrament is not just that it contains enough protein and other nutrients that people can thrive on, say, 400 or 800 calories-a-day of the stuff. But it also has to taste good, so that consumers will continue to buy it and drink it day in and day out, even after they’ve moved into the weight maintenance phase of their lives — i.e., for the rest of their lives. It’s an example from the 1960s of what Dr. Guyenet describes as “the goal of processed food manufacturers… to create a product that maximally reinforces purchase and consumption behaviors—food reward!”

We can try to get around this problem by suggesting that bland and sweet is just not high in food reward value, as Dr. Guyenet tries to do, but we’re going to resort to this kind of, well, blatant contradiction only because we want to salvage this experiment as support for the hypothesis. So this formula must have a low-food reward value because an obese subject consumed less of it and lost weight and because we believe that foods with high reward value cause people to gain weight. Now we’re back to circular-definition land, a place I would prefer to never visit.

Now, how about the idea that a “cafeteria” or “junk food” diet makes humans and animals fat, a concept that was pioneered by Anthony Sclafani. The assumption is that such a diet is fattening because there’s something about eating a variety of foods, mostly junk foods, that is so rewarding or at least so less bland than a plain chow diet that both humans and animals get fat eating it. Here’s how Dr. Guyenet describes it:

In this model, animals are allowed free access to standard chow and water while concurrently offered highly palatable, energy dense, unhealthy human foods ad libitum.

In other words, they’re given an unlimited amount of human junk food in addition to their whole food-based “standard chow.” In this particular paper, the junk foods included Froot Loops, Cocoa Puffs, peanut butter cookies, Reese’s Pieces, Hostess Blueberry MiniMuffins, Cheez-its, nacho cheese Doritos, hot dogs, cheese, wedding cake, pork rinds, pepperoni slices and other industrial delicacies. Rats exposed to this food almost completely ignored their healthier, more nutritious and less palatable chow, instead gorging on junk food and rapidly attaining an obese state.

Aside from Dr. Guyenet’s description of standard rat chow as “whole-food based,” my major problem with this (which is the same problem Ramirez et al had 20-odd years ago with the existing research then) is that this is an experiment that changes an unholy host of variables, and the results are evoked to make a point about one: food reward value.

One advantage I have in this nutrition business as an arguably ignorant journalist is that I actually get to interview the researchers who do the work. (Technically anyone could do this, but the researchers are certainly more likely to give their time to a journalist, ignorant or not, than to what one of my acquaintances in academia refers to as “just a person.”) I interviewed Sclafani back on January 30, 2003 for GC,BC, and the interview revealed the obvious problem with this interpretation. As Sclafani told me, they started their cafeteria diet (which he was calling the “supermarket” diet at the time) with a variety of different foods (not quite as wide a variety as Dr. Guyenet is discussing above, but wide nonetheless): chocolate chip cookies, salami, cheese, bananas, marshmallows, milk chocolate, peanut butter and sweetened condensed milk, and then they later simplified it to four foods because the rats didn’t eat all the foods they gave them.

Which foods did they ignore? Sclafani said they never did a systematic study, nor had anyone else, as far as he knew (as of January 2003), but cheese, salami and peanut butter—the foods highest in fat and lowest in refined grains and sugar—seemed to be the foods they avoided in favor of the sweeter, starchy options. So the obvious question: are refined grains and/or sugar necessary to impart not just reward value, but reward value that leads to people and animals getting fat?

In fact, Sclafani told me that they had based their selection of foods on a hunch about what rats preferentially like, and so that’s why they included cheese in the list. It seemed like an obvious choice. After all, don’t you stick cheese in mouse traps when you want to rid your house of mice? And yet, cheese was not among the foods the Sclafani’s rats preferentially ate when given all these other refined carbs and sugary foods to eat instead. Maybe because the cheese was unrewarding. Or maybe because it was relatively if not completely refined-carb and sugar free, as were the salami and peanut butter.

This inability to differentiate food reward and/or palatability from the presence of refined carbs and sugars haunts virtually every example of the studies cited to document food reward and/or palatability.

Another example, not one used by Dr. Guyenet, is Kelly Brownell’s Yale Food Addiction Scale . This scale attempts to identify  people who suffer from addiction to certain foods. The scale is based on a survey that gives a series of statements about eating habits. Subjects must say how true each statement is, on a scale from “never” to “four or more times or daily.” This goes along with a list of foods of which food addicted subjects might have “difficulty controlling their intake.” Here are the first four statements to give you an idea of what Brownell is getting at:

1. I find that when I start eating certain foods, I end up eating much more than planned.
2. I find myself continuing to consume certain foods even though I am no longer hungry.
3. I eat to the point where I feel physically ill.
4. Not eating certain types of food or cutting down on certain type of food is something I worry about.

So let’s assume, for the sake of argument at least, that the “certain foods” that illicit addictive behavior is very similar to the list of hyper-rewarding foods, the ones most likely to cause obesity. Here’s Brownell’s list of the foods that are most likely to be addictive:

-       Sweets like ice cream, chocolate, doughnuts, cookies, cake, candy, ice cream [yes, ice cream is listed twice.]

-       Starches like white bread, rolls, pasta, and rice

-       Salty snacks like chips, pretzels, and crackers

-       Fatty foods like steak, bacon, hamburgers, cheeseburgers, pizza, and French fries

-       Sugary drinks like soda pop

With the exception of steak and bacon, all of these foods are high in carbohydrates — refined or otherwise (the French Fries) — and/or sugars, even the foods defined as “fatty” with the aforesaid exceptions.  If these foods are addictive and if they cause obesity, is it because they’re addictive or is it because of the metabolic and hormonal effects of consuming them — their effects on insulin signalling? There’s no way to tell without an exceedingly well-controlled and well-conceived experiment, but you can guess where my vote lies.

What about the steak and bacon, then? Well, if you ate nothing but those—steak and bacon every day, plus, say, the hamburgers or cheeseburgers without the refined grains attached, i.e., the buns—you’d be eating a weight loss diet (a ketogenic diet) and would almost assuredly lose weight doing it. So whether or not you consider steak and bacon addictive, it’s unlikely that they could be defined as foods high in reward value because they would tend to refute the hypothesis that high food reward value causes obesity. And now we’d be back to this problem of having to differentiate between hyper-rewarding foods or at least addictive foods that come with refined/easily digestible carbohydrates and sugars and cause fat accumulation and hyper-rewarding foods that don’t, and well, don’t.

Of the examples I could find in Dr. Guyenet’s discussions of the food reward/palatability hypothesis that held the promise of differentiating food reward value from underlying metabolic effects of the foods themselves (and the presence of refined or easily-digestible carbs and sugars), none of them actually came through with meaningful evidence.

The studies most likely to offer such a differentiation were those mentioned by Dr. Guyenet  in his post, “The Case for the Food Reward Hypothesis of Obesity, Part II.” These were the studies evoked as evidence for the hypothesis because they demonstrate that “Individual sensitivity to food reward should predict future fat gain.” About this evidence he says:

I’m aware of three studies that have investigated this question.  In the first, researchers found that the reinforcing value of food relative to a non-food stimulus predicted fat gain over the next year in 7-10 year old children (19).  In the second, the responsiveness of reward-related brain regions to imagining palatable vs. unpalatable foods (as assessed using fMRI) predicted body mass index (BMI) gains in adolescent girls, and this effect was modified by gene polymorphisms in dopamine receptor genes (20).  The third study also used fMRI to demonstrate that greater activation in reward-related brain regions during exposure to appetizing food cues predicted greater BMI gains over time in adolescent girls (21).

But of those three studies, none of them define what the high reward foods were, which foods were considered palatable and which were unpalatable. For all we know, the palatable foods were the ones rich with refined grains and sugars and the reason reward-related regions of our brain light up when we eat them (or at least when obese people do) is because our brains are responding to what these foods do to our bodies.

Reference 19 doesn’t specify at all which foods are actually high in reward value, nor do references 20 and 21, which are both by the same authors. They do, however, include a “cheeseburger” as an example of a processed food, demonstrating a certain bias against cheeseburgers that may be misplaced.

And reference 21 says that while BMI may have been related to the extent of activation in reward-related brain regions, as Dr. Guyenet points out, this was true regardless of whether the food being imagined was rewarding or palatable or not. Or the authors put it, “BMI [body mass index] was positively correlated with behavioral response to both appetizing and unappetizing food images, implying that food cues in general trigger greater attention in overweight vs. lean individuals.”

One obvious interpretation is that overweight individuals are hungrier than lean individuals, and so they have a greater response to any food in their reward centers. And, in fact, one point Dr. Guyenet’s mentor, Michael Schwartz, made of interest in his 2006 review in Nature “Central nervous system control of food intake and body weight” was that “food deprivation strongly augments the reward value… reduced food availability seems to exert a global, stimulatory effect on reward perception.” And so maybe the greater the BMI, the more likely the subjects were hungry or food deprived—a phenomenon I discuss at length in GC,BC— a state that could be due to increased insulin secretion and chronic hyperinsulinemia. And maybe refined grains and sugars augment reward value because they cause us to secrete more insulin and store calories away as fat and glycogen and make us hungry.

In GC,BC I quote Mark Friedman commenting on this potential carbs-insulin-hunger connection regarding just the cephalic phase of insulin secretion, the one that comes just by thinking about a particular food:

This cephalic release of insulin also serves to clear the circulation of “essentially anything an animal or a person can use for fuel. Not just blood sugar, but fatty acids, as well. All those nutrients just go away.” Hence, the thought of eating makes us hungry, because the insulin secreted in response depletes the bloodstream of the fuel that the peripheral tissues and organs need to survive.

And if this happens more in individuals who are insulin resistant,  as most obese individuals are, we’re now back at a hypothesis that maybe the insulin signaling in the body is running the brain’s response, not vice versa. Yes, it’s the brain that’s stimulating the insulin secretion when we think about food, but what makes us hungry and makes the food then seem so rewarding is the effect of the insulin secreted in the body.

We’ve been discussing the food reward/palatability hypothesis of obesity and whether this idea adds anything meaningful to our understanding of obesity.  Is the evidence for it sufficiently compelling that we should cease to pay attention to the fact that insulin, as Yalow and Berson noted in 1965, is “the principal regulator of fat metabolism?”

One point I’ve been making in my posts and in my books is that it’s possible to find evidence in favor of virtually any idea – including the Flying Spaghetti Monster as the ruling force in the universe. More important to the validation of an idea or a hypothesis is the strength of the evidence that seems to refute it. Can the hypothesis survive more or less intact our best attempts to refute it?

This is one of the points I was trying to get across at the Ancestral Health Symposium: that the foods we eat today during our current obesity epidemic might have a high reward value, and that diets consumed by lean populations in faraway locales might not, isn’t particularly interesting. Yes, it supports the hypothesis, but how do we explain epidemics of obesity in populations that  eat diets that don’t appear to have a high reward value? Do we need an entirely different hypothesis for them? That would be unfortunate.

“Here’s the fundamental concept that I think explains a lot of obesity in industrialized nations,” writes Dr. Stephen Guyenet of wholehealthsource.org .

We live in a more or less Darwinian economic framework (capitalism). Food manufacturers are in constant competition, and any food that sells poorly will rapidly disappear from stores. How do you get people to buy your product? You produce something that causes them to come back and buy it again. In other words, the goal of processed food manufacturers is to create a product that maximally reinforces purchase and consumption behaviors – food reward! If the product is not extremely rewarding, it won’t sell because it’s competing against other products that are extremely rewarding. Only the most rewarding products survive.

This certainly sounds reasonable, but don’t we also want a hypothesis of obesity that explains obesity rates in populations that lack such highly evolved food industries – obesity in non-industrialized nations? This would be a hypothesis that explains obesity-ridden populations in which the local industry isn’t quite so diligent in increasing food reward, if there are food manufacturers to speak of at all?

This is the question I asked in chapter one of Why We Get Fat. It’s why I listed a host of populations in which levels of obesity were reported, in some cases, approaching or exceeding those in the U.S. today, and yet with none of this Darwinian competition between food manufacturers, none of this extremely rewarding food (or at least not extremely-rewarding as we would define it today).

These populations included the Pima in 1902, the Sioux on the Crow Creek Reservation in 1928, the citizens of Naples in the years of extreme poverty following the Second World War and African-Americans in Charleston South Carolina in 1959. They included Zulu in Durban South Africa in 1960, and the citizens of Nauru in the South Pacific in 1961 — “By European standards,” a local physician wrote, “everyone past puberty is grossly overweight.” They included Trinidadians in the early 1960s and Chilean factory workers. They included urban Bantu pensioners “the most indigent of elderly Bantu,” in Johannesburg, South Africa in 1965, and so on.

What these populations had in common was varying degrees of poverty — from very poor to unimaginably poor — and the absence of a Darwinian food industry as Dr. Guyenet and others would describe it. They did have sugar and refined grains, but don’t we want a hypothesis of food reward that can make a claim more meaningful than “rewarding (or hyper-rewarding) foods are foods with sugar and/or refined grains in them?” (And if this ultimately is our definition, as I’ll discuss shortly, then we should be able to establish whether the reason they’re rewarding is or is not due to the peripheral effects of these foods, rather than their ability to influence brain chemistry, set point, etc.)

We also want a concept (or at least I do) that explains how we can have populations in which obesity and malnutrition and under-nutrition co-exist — for example, obese mothers with starving children, a common observation now in the literature.

Take Jamaica, for instance, where the British-trained diabetologist Rolf Richards, as I have quoted in Why We Get Fat and in my lectures, discussed the situation in 1973:

It is difficult to explain the high frequency of obesity seen in a relatively impecunious [very poor] society such as exists in the West Indies, when compared to the standard of living enjoyed in the more developed countries. Malnutrition and subnutrition are common disorders in the first two years of life in these areas, and account for almost 25 per cent of all admissions to pediatric wards in Jamaica.  Subnutrition continues in early childhood to the early teens.  Obesity begins to manifest itself in the female population from the 25th year of life and reaches enormous proportions from 30 onwards.

Now if we blame the mother’s obesity on the hyper-rewarding nature of the food she’s eating, we have to ask why these foods are rewarding only to the women and not to their children. The children aren’t fat, after all (not yet, anyway). In fact they’re starving. They’re under-nourished. We also have to explain why these foods only become rewarding from “the 25^th year of life” onward? And, perhaps most important, we have to explain why these women don’t fight the hyper-rewarding nature of these foods and remain lean.

After all, the food reward/palatability hypothesis of obesity, as we discussed in the first post on the subject, dictates that these foods cause neurochemical changes in the brain, which then raises the adiposity set point, thus making us eat more and get fat.  Put simply, raising the set point in the brain makes us hungry or at least hungrier. Okay, so if this is right, then we can assume that the reward value of the food eaten in Jamaica made these women hungrier; they ate more, they got fatter. But why couldn’t they control their impulses and remain lean? Why couldn’t they experience the semi-starvation—or at least the perception of not having enough to eat—rather than their children who are indeed semi-starved?

Rather than giving in to the urge, consuming the superfluous calories themselves, and getting fat, why didn’t these mothers fight the urge and give those excess calories to their starving kids? If one of them has to go hungry or at least feel hungry, evolution, it seems, would always favor the mother doing it rather than the child.

We can try to rescue  the food reward/palatability hypothesis of obesity in a case like this by simply making the claim that if these people are fat, then obviously something about their food must have been hyper-rewarding. (Something other than the refined carbs and sugars, as we’ll discuss in the next post.) But now our definition is becoming circular: The women get fat because of the hyper-rewarding nature of the food they’re eating, and we know that the food is hyper-rewarding because they’re fat. We just have to find or identify the particular foods in their diet that are the hyper-rewarding ones, and, as I said, it would be nice if they weren’t just sugar and refined carbs.

In his blogs, Dr. Guyenet suggests that home cooked food has a lower food reward value than processed, restaurant-produced fast food. This is one reason why, he suggests, populations like the Ache of Paraguay, the !Kung San, Polynesians and Melanesians (not counting those on Nauru and other islands that were obese) were lean: They “cooked their food in earth ovens and used no flavorings or salt .“

That we don’t cook our foods by these simple, spice-free, salt-free methods is offered as another explanation for the current obesity epidemic — “the shift from simpler home-cooked food to professionally engineered/processed food designed to maximize palatability and reward.” And this is also an explanation often offered for why carbohydrate restriction and paleo diets (not necessarily two different things) are weight loss diets. It’s not that they’re simply absent refined grains and sugars, as they are, but that the meats, fish, fowl, vegetables, and maybe tubers consumed are home cooked and/or so relatively bland that somehow they are low in food reward value.

But we can be confident that these extremely poor populations with high levels of obesity were also getting by on simple home-cooked food. Without having had the opportunity to visit Trinidad in the early 1960s or the South Dakota Crow Creek Reservation in 1928, I’m going to assume with confidence that a large proportion of the population, if not all, were not frequenting fast food joints and buying hyper-rewarding candy bars and soft drinks. So why were they fat? Certainly the presence or absence of flavorless home cooking is not enough to explain it. Nor can we explain it by claiming that only the affluent were obese, as Dr. Guyenet suggests, because these populations were anything but affluent.

So why were they fat? A familiar question.

Well, maybe it’s the low-hanging fruit of food reward—the refined grains and sugars? Populations that got fat ate significant quantities—particularly, the sugar—and populations that didn’t, well, didn’t. And when obesity suddenly blossomed in populations, it was because sugar and refined grains were new additions to their diets. And so diets that work for weight loss and weight maintenance are those that restrict refined grains and sugars (and maybe easily digestible starchy vegetables, as well, or maybe not) and diets that don’t, well, don’t.

This is what I argued in my books, although I’m arguing that the problem is caused by the metabolic hormonal effects of these foods in the periphery, their effect primarily on insulin signaling and, ultimately, fat accumulation. And the reason we find these foods rewarding and palatable is because of these metabolic hormonal effects.

As I’ve suggested in prior posts, the kinds of observations that are meaningful in situations like this—two competing hypotheses/paradigms—are only those that can differentiate between the two competitors. Evidence or observations that can be explained equally well by either hypothesis might have rhetorical value—good in an argument, in the spur of the moment—but they don’t add much to the scientific question at hand: Which hypothesis/paradigm is the right one?

This is why the observation that the Ache, the !Kung San, the Polynesians and Kitavans and Masai are lean or were lean, for instance, doesn’t tell us anything of significance about which hypothesis is right: Their lack of excess adiposity might be a result of their bland, unrewarding diets or it might be because their diets lack or lacked any significant amount of refined grains and/or sugars.

And even if the foods or diets that are consumed by obese populations and individuals today in the U.S. and elsewhere do seem indisputably rewarding and palatable, we’re still left having to demonstrate that this palatability, this high food reward value, is not due to the nutritional composition of the diet and the peripheral effects of the nutrients—the metabolic and hormonal effects in the body.

This was a point made back in 1989 by Israel Ramirez, Michael Tordoff and Mark Friedman of the Monell Chemical Senses Center in Philadelphia in an article entitled “Dietary Hyperphagia and Obesity: What causes them?”(Friedman is one of the scientists whose thoughts on obesity and over-eating significantly shaped my own. I owe him a debt of gratitude. For those who want to read what I think may be the single most thoughtful article written on obesity and hunger in the post-WW2 era, I’d recommend Friedman’s article with Edward Stricker, The Physiological Psychology of Hunger: A Physiological Perspective.)

The Monell researchers were discussing only the concept of palatability, not the food reward value of a particular food. (The idea that food reward and palatability could be differentiated — that they weren’t precisely the same thing — hadn’t gotten much if any play up until then.) So the question was whether or not palatability (whether a food tastes good) could be legitimately disassociated from nutrient composition and peripheral effects of the food. As Ramirez et al said repeatedly in this article, researchers almost invariably assumed that a food could be defined as palatable if the animals (or humans) ate more of that food than some other food, but this was an inference and nothing more.

It was well known at the time (although it may have been forgotten since then), as I discussed in Good Calories, Bad Calories, that animals can be made to like one food more than another, and so eat more of the one than the other, by interventions that influenced their underlying physiologic/metabolic/hormonal states. Here’s how I illustrated this in GC,BC:

Throughout the first half of the twentieth century, a series of experimental observations, many of them from [Curt] Richter’s laboratory [at Johns Hopkins University], raised questions about what is meant by the concepts of hunger, thirst and palatability, and how they might reflect metabolic and physiological needs. For example, rats in which the adrenal glands are removed cannot retain salt and will die within two weeks on their usual diet from the consequences of salt depletion. If given a supply of salt in their cages, however, or given the choice of drinking salt water or pure water, they will chose to either eat or drink the salt and, by doing so, keep themselves alive indefinitely. These rats will develop a “taste” for salt that did not exist prior to the removal of their adrenal glands. Rats that have had their parathyroid glands removed will die within days of tetany, a disorder of calcium deficiency. If given the opportunity, however, they will drink a solution of calcium lactate rather than water—not the case with healthy rats—and will stay alive because of that choice. They will appear to like the calcium lactate more than water. And rats rendered diabetic voluntarily choose diets devoid of carbohydrates, consuming only protein and fat. “As a result,” Richter said, “they lost their symptoms of diabetes, i.e., their blood sugar fell to its normal level, they gained weight, ate less food and drank only normal amounts of water.

In short, change underlying physiologic/hormonal conditions and it will affect what an animal chooses to eat and so seems to like or find rewarding. The animal’s behavior and perceptions will change in response to a change in homeostasis – in the hormonal milieu of the cells in the body.

It’s quite possible that all those foods we seem to like, or even the ones we find rewarding but don’t particularly like, as Dr. Guyenet argues, and that subsequently cause obesity (not necessarily the same thing) are those foods that somehow satisfy an underlying metabolic and physiological demand. This in turn might induce our brains to register them as more palatable or rewarding, but the initial cause would be the effect in the periphery.  The nutrient composition of the food, in this case, would be the key—what it’s doing in the body, not necessarily the brain.

Here’s how Ramirez, Tordoff and Friedman phrased this issue back in 1989:

In order to demonstrate that diet palatability per se causes hyperphagia [overeating or a voracious appetite], it must be shown that obesity-inducing foods are more palatable than control foods, this greater palatability is not merely a reflection of the postingestive [after entering the digestive tract] consequences of the foods, and altering palatability without altering nutritional composition can cause obesity. This has not been done.… Although various experiments have been cited as supporting the palatability hypothesis, they are not decisive because, in every case, palatability was confounded with changes in nutritional composition.

That an experiment is “not decisive” unless this is done is the critical point. If an experiment that ostensibly changes food reward makes an animal eat more of a particular food and/or get fatter, and it does so by changing nutritional composition—say, the foods that are defined as more rewarding have more sugar in them, or are more refined, or have a greater water or fat content—then the researchers have to demonstrate that it’s not the change in nutritional composition and post-ingestive effects of that change that is causing the overeating and obesity. An observation that one diet produces obesity compared to another because it’s ostensibly more rewarding or palatable has to do the same. Otherwise either hypothesis could be true, and we haven’t learned anything.

Take the idea, as Dr. Guyenet suggests, that people will eat more at a sitting if foods are palatable than if they’re not, which seems kind of obvious. The better a food tastes, the more likely we are to eat more of it. But then Dr. Guyenet adds that this is true even of foods with “little or no nutritional quality.” This is how he phrases it in a recent post:

Many human studies have shown that people eat more food at a sitting if the food is higher palatability than if it is lower palatability (11).  This is true even if palatability is manipulated using substances that have little or no impact on the nutritional quality of the food, including saccharin (sweet), monosodium glutamate (savory) and herbs/spices.

The reference is a review article that actually makes the point that the evidence is ambiguous on the eating more issue when the foods have little or no nutritional quality. “Several studies showed no effect of sweet taste on either hunger ratings or food intake,” the authors write, “when the sweetener was provided in the form of gelatine, corn flakes or fromage blanc or as aspartame- or saccharin-sweetened drinks.” In fact, the authors then go on to suggest this is true of all “sweet taste” whether from caloric sweeteners or non-caloric, which doesn’t seem to do my hypothesis any favors either.

But what I’m arguing is that the key isn’t whether people eat more, but whether the foods stimulate fat accumulation. And if they do make us fatter, how?

The food reward hypothesis suggests that it happens because of the effect of the sweet taste in the brain, not in the body. If the former, then sugar and saccharine might be expected to be equally fattening, so long as we consider sugar and saccharine-sweetened beverages to have equal reward value or to be considered equally palatable by humans.

If the reward value is not the critical factor, then it’s a reasonable assumption that sugar-sweetened beverages will be more fattening than saccharine or aspartame-sweetened beverages. And we could do a clinical trial and see which turns out to be true, although we can also guess what we think such a trial (randomized, well-controlled) would find. Not surprisingly, I’d vote for the sugar-sweetened beverages being more fattening.

This doesn’t mean, by the way, that artificially-sweetened beverages could be absolved of having any fattening properties because we might still secrete insulin in response to these beverages. They may fool us into thinking that they have carbohydrate or sugar calories in them. And this insulin secretion could be cephalic — a kind of Pavlovian response — which would mean that the brain is telling the pancreas to secrete insulin (via the vagus nerve). But it would now be doing so not because the food is rewarding necessarily, but because the body has come to associate sweet taste with the presence of carbohydrates and feedback loops in the brain are working to get the body ready by secreting insulin.

In my next post, I’ll discuss more of the evidence offered in support of the food reward/palatability hypothesis and ask the question that Ramirez et al did: are palatability and food reward confounded with changes in nutritional composition, and if so, what might that confounding be?