In a post titled “Ancestral Health Symposium Drama”, Stephan Guyenet begins to expound his scientific differences with Gary Taubes.

Since my views differ a bit from both Stephan and Gary, I thought readers might enjoy a third view.

My General Perspective on Obesity

My view is that obesity is caused in the first place by malnutrition, toxins, and infections. Each can contribute in multiple ways:

• Malnutrition can affect appetite and energy utilization. Micronutrient deficiencies will increase appetite, regardless of energy balance. Macronutrient deficiencies may also do this. The resulting increased calorie intake may be only partially balanced by increased activity and thermogenesis; fat gain in caloric surplus tends to be more weakly opposed by brain regulatory circuits than muscle loss during caloric deficit. Malnutrition can impair energy utilization by several pathways: for instance, loss of mitochondrial antioxidants may lead to oxidative damage that impairs mitochondrial health. Choline deficiency induces metabolic syndrome and obesity (see Choline Deficiency and Plant Oil Induced Diabetes, Nov 12, 2010). Long-term, malnutrition may induce methylation defects which affect epigenetic regulation of metabolism. These can be passed on from mother to child.
• Toxins also have multiple pathways by which they induce obesity. For example, diets that combine fructose or alcohol with polyunsaturated fats are very effective at producing metabolic syndrome and obesity in animals, and food opioids affect the endocannibinoid pathways which can be important in obesity and appetite regulation. See Why We Get Fat: Food Toxins (Jan 20, 2011) and Wheat and Obesity: More from the China Study (Sep 4, 2010) for more.
• Infections have also been linked to obesity. I’ve blogged about how adenovirus infections of adipose cells promote obesity (Obesity: Often An Infectious Disease, Sep 22, 2010), but another very important pathway is from gut infections to obesity. Briefly, gut pathogens release fat-soluble toxins which can enter systemic circulation, and also modulate immune function. Toxins from pathogens have been shown to induce metabolic syndrome in the liver, promoting obesity. Via the immune system, gut flora can promote obesity. I’ve briefly mentioned one pathway (in Thoughts on Obesity Inspired by Stephan, Jun 2, 2011): gut immune modulation in the gut has been shown to determine whether adipose tissue macrophages are in a pro-inflammatory or anti-inflammatory state. A pro-inflammatory state promotes obesity. Research into the many ways gut flora influence obesity is in early stages, but it’s clearly important.

Due to the diversity of factors which conspire to cause obesity, it is a rather heterogeneous disease. Its unifying character is that some combination of causal factors induces “metabolic damage,” such as leptin resistance, in a variety of organs, including the brain. Metabolic damage can affect both appetite regulation and energy homeostasis.

I’ve discussed Stephan’s views and food reward theory (Thoughts on Obesity Inspired by Stephan, Jun 2, 2011). Food reward theory offers a plausible explanation for many aspects of obesity. I agree that food reward is an important factor in obesity, but consider it one among several factors, and believe that different factors may dominate in different people. Also, it seems likely to me that food reward becomes a dominant factor in obesity only after some form of metabolic damage from malnutrition, toxins, or infections begins to affect the brain’s regulatory systems. In a healthy person a highly palatable diet might have little effect on weight for quite some time. Nor am I convinced that low food reward diets are necessarily the best approach for long term weight loss or for the health of the obese, though I do believe they are great for short-term weight loss.

Distinguishing my view from Stephan’s is difficult because the obesity-inducing diets used in animal studies are generally both toxic and malnourishing and highly palatable. The “cafeteria diet” of Cheetos and such – rich in wheat, sugar, and vegetable oil – is an example.

I haven’t previously blogged about Gary’s views, but I consider very low carb dieting to be an imperfect solution for good health generally. (NB: Low-carb, which I endorse, is for me 400-600 carb calories, very low-carb, which I deprecate, is <200 calories.) Ketogenic diets may be beneficial in some cases of obesity, but I believe they should still include some starchy carbohydrates.

The Exchange

Stephan has transcribed the Q&A between Gary and himself and offers revised answers. I’ll insert my thoughts:

GT: How does your food reward hypothesis hypothesis explain a culture in which mothers are obese and their children are starving?  Are the mothers eating Snickers bars and not sharing them with their children?

SG: The food reward/palatability hypothesis of obesity is not mine, it’s a hypothesis that originated in the 1970s, perhaps earlier, and is a major subject of ongoing obesity research.  I don’t expect it to explain every instance of obesity.  Obesity involves multiple factors, an important one of which is food reward and palatability.  That being said, you have to examine a culture’s food habits in some detail, both before and after a change in obesity prevalence, to determine if reward/palatability may have played a role.  I don’t know enough about that specific culture to judge whether food reward would have played a role there.

PJ: Famines occur in impoverished societies with disrupted social institutions. People in these cultures are driven to eat the cheapest calories, which are the toxic grains such as wheat. They also tend to be malnourished, especially during famines. Malnutrition and toxic foods can create the disease of obesity, especially in a suitable infectious disease context.  Once the disease of obesity is induced, periods of caloric availability lead to weight gain which may be defended during subsequent famines. This explains maternal obesity persisting during a period of food scarcity. The slenderness of their children is a result of the disease process not having had enough time to work. It may take decades for malnutrition and food toxicity to induce obesity in the child.

So the element of long-acting causal factors and history eliminates the apparent conflict between an obese mother and a starving slender child.

Because food reward could induce obesity in the mother prior to the famine which is defended later, and food reward may act differently in growing children, food reward theory may be able to explain the situation. But Stephan prudently allows for the possibility that other causes of obesity besides food reward may be at work.

GT: The Pima indians were obese in 1902, following 20-30 years of famine.  How would your theory explain this?

SG: The Pima were first contacted in 1539 by the Spanish, who apparently found them to be lean and healthy.  At the time, they were eating a high-carbohydrate, low-fat diet based on corn, beans, starchy squash, and a modest amount of gathered animal and plant foods from the forest and rivers in the area.  In 1869, the Gila river went dry for the first time, and 1886 was the last year water flowed onto their land, due to upstream river diversion by settlers.  They suffered famine, and were rescued by government rations consisting of white flour, sugar, lard, canned meats, salt and other canned and processed goods.  They subsequently became obese.  Their diet consisted mostly of bread cooked in lard, sweetened beverages and canned goods, and they also suddenly had salt.  I don’t see why that’s incompatible with the food reward hypothesis.  It is, however, difficult to reconcile with the carbohydrate hypothesis.

PJ: The Pima Indian story seems compatible with both Stephan’s and my views, since they ate a nourishing, low-toxicity, low-food reward diet when they were lean but a malnourishing, toxic, high-food reward diet when they became obese. It seems incompatible with Gary’s ideas, since the Pima ate a high-carb diet at all times. Thus it’s a bit surprising Gary is so fond of the Pima story. It weakens, not helps, his case.

GT: There are two possible hypotheses here.  The alternative hypothesis is that sugar and refined carbohydrate consumption changes the regulation of fat tissue, leading to obesity.  The studies you cited in which people lost weight by consuming bland liquid diets would have been low in sugar as well.  “We need an observation that can refute one of the two hypotheses”.

SG: The bland liquid diet in Hashim et al. that caused massive weight loss is called “Nutrament”.  It is 50% carbohydrate, 30% fat and 20% protein.  The primary three sources of carbohydrate in this formulation are lactose (from milk), sucrose (table sugar) and corn syrup.  The bland liquid used in the study by Cabanac et al. (Renutryl), which also caused weight loss, was high in refined glucose and sucrose.  I find this rather difficult to reconcile with the idea that sugar and refined carbohydrate are inherently obesogenic.

PJ:  It’s unclear to me what Gary’s “alternative hypothesis” is. Why are refined carbohydrates different from unrefined carbohydrates? Both may raise blood glucose and insulin levels similarly. If toxic plant foods are the problem, then he should say toxins rather than carbohydrates are the problem. If it’s the macronutrient that’s the problem, why does refining matter?

Stephan scores a point against both Gary and me here, but especially against Gary, since the liquid diets are fairly high in carbs. As there was some sucrose and polyunsaturated fat, this was not a non-toxic diet, and I don’t know if adequately micronutrients were provided – probably not – but on its face the food reward theory seems to work best in explaining this experiment.

GT: “How was it bland then?”

SG: The diet was a liquid formulation that (judging by the ingredients) probably tastes like powdered milk.  The subjects were drinking that for 100% of their calories.  That fits any reasonable definition of a low reward/palatability diet, regardless of the sugar.

GT: What about the Mexican-Americans in Star county, Texas, who were obese despite the fact that there was only one restaurant in the whole town?

SG: Again, you have to examine a culture’s food habits in some detail, both before and after a change in obesity prevalence, to determine if reward/palatability may have played a role.  I don’t know enough about that specific culture to judge whether food reward would have played a role there.

GT: How can we differentiate between altered palatability and altered carbohydrate intake as important factors in the rising obesity prevalence of industrializing nations?

SG: Increased carbohydrate intake is a particularly poor explanation for obesity in industrializing populations, as the majority of them (for example, most of Asia and Africa) are going from a diet very high in carbohydrate, to one that is lower in carbohydrate and higher in fat.  There are also a smaller number of cultures that developed obesity as they went from high-fat to higher carbohydrate, industrialized food.  Therefore, the ideas that carbohydrate or fat are inherently fattening don’t appear consistent with the evidence as a whole.  An alternative explanation whereby both fat and carbohydrate, as well as other factors, are important for reward/palatability, an excess of which contributes to obesity, fits the evidence better.

PJ: It seems to be easiest to induce obesity with a roughly equal mix of carbs and fat; both low-carb and low-fat diets tend to be less obesogenic. This result is compatible with Stephan’s views because carb and fat together are more rewarding than either alone, and with my views because carb-fat combinations can be highly toxic – for instance, a fructose-PUFA combination is more toxic than either alone; or carbs feed gut pathogens while fats carry their toxins into the body.

It is unclear how Gary would explain the evidence from both animal studies and human populations that obesity becomes more likely as high-carb diets shift toward more fat.

Of Glass Houses

Stephan is a model of scholarly virtue, so Gary’s challenge at the end of his talk was a shock. I thought Stephan’s original reply – “Thank you for the advice” – was perfect, but Stephan revises it:

GT: “I would just recommend in the future you should pay attention to populations that might refute your hypothesis rather than just presenting populations that support.  That’s always key in science.”

SG: People who live in glass houses shouldn’t throw stones.

Presumably Stephan is challenging Gary to address some of the populations who seem to refute his hypothesis: Asian populations that have become more obese while dropping carbs from 75% to 50% of diet, or the Pima who remained lean on a high-carb diet for centuries.

In other words, to seek a theory that can explain all phenomena, as a scientist should.

In general, I find Gary’s work rhetorically artful but not very helpful to scientific progress. He often neglects to consider the full implications of his own evidence. This is especially true when he ventures into molecular and cellular biology.

For instance, he uses genetic lipodystrophies to illustrate that fat storage can be a disease of molecular biology, rather than excess food consumption. Now, the mutations in these lipodystrophies are generally not in insulin, the insulin receptor, or even centrally located on insulin pathways. So the lipodystrophies show that other molecules besides insulin can be responsible for fat storage (or negative regulation of fat storage), and may be relevant to obesity.

But when he looks into which molecules might be responsible for obesity, he offers only one candidate: insulin.

More startling is his neglect of perhaps the single most important molecule in obesity, leptin. Stephan writes:

[H]e sent me a manuscript for his book Why We Get Fat and asked for my advice prior to its publication.  I explained to him that he needed to use the word “leptin” in the book, particularly when discussing animal models of obesity that are obese because of defects in leptin signaling (ob/ob mice and Zucker rats, for example).

This is just like his use of lipodystrophies: mice get obese due to mutations in leptin, but he doesn’t discuss the role of leptin, preferring to keep the spotlight on insulin.

I don’t want to sound harsh because I think Gary is on the side of the angels. He has done very beneficial work refuting saturated-fat-phobia and encouraging low-carb diets, which improve the health of nearly all westerners who adopt them (although the reason is probably reduced toxicity from wheat and sugar, rather than reduced carbohydrate calories).

But I think he would do well to be more generous to others. I was excited when he began blogging, but disappointed by his first post:

conventional wisdom … almost incomprehensibly naïve and wrong-headed … nonsensical notion … I’ve been consistently amazed at the ability of researchers … to accept some of the rote ideas … without seemingly giving it any conscious thought whatsoever, or without wanting to ask the kinds of questions that a reasonably smart junior high school student should ask if given the opportunity…. I don’t understand this failure of intellect … nonsensical explanations … he falls short, as he’s working outside his area of expertise … we’re being fed nonsense … we will typically pass that nonsense along … If the experts had ever been open to a little skeptical thinking from others or had they been appropriately skeptical themselves … What’s been needed (and still is) was for someone (a reasonably smart 14-year-old would suffice) to ask the obvious questions and then insist on intelligent answers.

I find such talk ungenerous; and ironic, because in places in that very post Gary’s own reasoning is unsound.

Biology is complex, none of us have all the answers, and a lifetime is too short to acquire all the answers. Since we have no choice but to live in glass houses, we should all be humble, and refrain from casting stones.