Interview on Super Human Radio

Today, I did an audio interview with Carl Lanore of Super Human Radio.  Carl seems like a sharp guy who focuses on physical fitness, nutrition, health and aging.  We talked mostly about food reward and body fatness-- I think it went well.  Carl went from obese to fit, and his fat loss experience lines up well with the food reward concept.  As he was losing fat rapidly, he told friends that he had "divorced from flavor", eating plain chicken, sweet potatoes and oatmeal, yet he grew to enjoy simple food over time.

The interview is here.  It also includes an interview of Dr. Matthew Andry about Dr. Loren Cordain's position on dairy; my interview starts at about 57 minutes.  Just to warn you, the website and podcast are both full of ads.

Weight Gain and Weight Loss in a Traditional African Society

The Massas is an ethnic group in Northern Cameroon that subsists mostly on plain sorghum loaves and porridge, along with a small amount of milk, fish and vegetables (1, 2).  They have a peculiar tradition called Guru Walla that is only undertaken by men (2, 1):
Read more »

Food Reward: a Dominant Factor in Obesity, Part VIII

Further reading

I didn't come up with the idea that excessive food reward increases calorie intake and can lead to obesity, far from it.  The idea has been floating around the scientific literature for decades.  In 1976, after conducting an interesting diet study in humans, Dr. Michel Cabanac stated that the "palatability of the diet influences the set point of the ponderostat [system that regulates body fatness]" (1).  

Currently there is a growing consensus that food reward/palatability is a major contributor to obesity. This is reflected by the proliferation of review articles appearing in high-profile journals.  For the scientists in the audience who want more detail than I provide on my blog, here are some of the reviews I've read and enjoyed.  These were written by some of the leading scientists in the study of food reward and hedonics:

Palatability of food and the ponderostat.  Michel Cabanac, 1989.
Food reward, hyperphagia and obesity.  Hans-Rudolf Berthoud et al., 2011.
Reward mechanisms in obesity: new insights and future directions.  Paul J. Kenny, 2011.
Relation of obesity to consummatory and anticipatory food reward.  Eric Stice, 2009.
Hedonic and incentive signals for body weight control.  Emil Egecioglu et al., 2011.
Homeostatic and hedonic signals interact in the control of food intake.  Michael Lutter and Eric J. Nestler, 2009.
Opioids as agents of reward-related feeding: a consideration of the evidence.  Allen S. Levine and Charles J. Billington, 2004.
Central opioids and consumption of sweet tastants: when reward outweighs homeostasis.  Pawel K. Olszewski and Allen S. Levine, 2007.
Oral and postoral determinants of food reward.  Anthony Sclafani, 2004.
Reduced dopaminergic tone in hypothalamic neural circuits: expression of a "thrifty" genotype underlying the metabolic syndrome?  Hanno Pijl, 2003.

If you can read all these papers and still not believe in the food reward hypothesis... you deserve some kind of award.

Food Reward: a Dominant Factor in Obesity, Part VII

Now that I've explained the importance of food reward to obesity, and you're tired of reading about it, it's time to share my ideas on how to prevent and perhaps reverse fat gain.  First, I want to point out that although food reward is important, it's not the only factor.  Heritable factors (genetics and epigenetics), developmental factors (uterine environment, childhood diet), lifestyle factors (exercise, sleep, stress) and dietary factors besides reward also play a role.  That's why I called this series "a dominant factor in obesity", rather than "the dominant factor in obesity".
Read more »

Drug Cessation and Weight Gain

Commenter "mem", who has been practicing healthcare for 30+ years, made an interesting remark that I think is relevant to this discussion:

Recovering substance dependent people often put on lots of weight and it is not uncommon for them to become obese or morbidly obese.

This relates to the question that commenter "Gunther Gatherer" and I have been pondering in the comments: can stimulating reward pathways through non-food stimuli influence body fatness?  

It's clear that smoking cigarettes, taking cocaine and certain other pleasure drugs suppress appetite and can prevent weight gain.  These drugs all activate dopamine-dependent reward centers, which is why they're addictive.  Cocaine in particular directly inhibits dopamine clearance from the synapse (neuron-neuron junction), increasing its availability for signaling.
Read more »

Healthy Skeptic Podcast

Chris Kresser has just posted our recent interview/discussion on his blog The Healthy Skeptic.  You can listen to it on Chris's blog here.  The discussion mostly centered around body fat and food reward.  I also answered a few reader questions.  Here are some highlights:

• How does the food reward system work? Why did it evolve?
• Why do certain flavors we don’t initially like become appealing over time?
• How does industrially processed food affect the food reward system?
• What’s the most effective diet used to make rats obese in a research setting? What does this tell us about human diet and weight regulation?
• Do we know why highly rewarding food increases the set point in some people but not in others?
• How does the food reward theory explain the effectiveness of popular fat loss diets?
• Does the food reward theory tell us anything about why traditional cultures are generally lean?
• What does cooking temperature have to do with health?
• Reader question: How does one lose fat?
• Reader question: What do I (Stephan) eat?
• Reader question: Why do many people gain fat with age, especially postmenopausal women?

The podcast is a sneak preview of some of the things I'll be discussing in the near future.  Enjoy!

Food Reward: a Dominant Factor in Obesity, Part III

Low-Fat Diets

In 2000, the International Journal of Obesity published a nice review article of low-fat diet trials.  It included data from 16 controlled trials lasting from 2-12 months and enrolling 1,910 participants (1).  What sets this review apart is it only covered studies that did not include instructions to restrict calorie intake (ad libitum diets).  On average, low-fat dieters reduced their fat intake from 37.7 to 27.5 percent of calories.  Here's what they found:
Read more »

Food Reward: a Dominant Factor in Obesity, Part II

How to Make a Rat Obese

Rodents are an important model organism for the study of human obesity. To study obesity in rodents, you have to make them fat first. There are many ways to do this, from genetic mutations, to brain lesions, to various diets. However, the most rapid and effective way to make a normal (non-mutant, non-lesioned) rodent obese is the "cafeteria diet." The cafeteria diet first appeared in the medical literature in 1976 (1), and was quickly adopted by other investigators. Here's a description from a recent paper (2):

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.

Investigators have known for decades that the cafeteria diet is a highly effective way of producing obesity in rodents, but what was interesting about this particular study from my perspective is that it compared the cafeteria diet to three other commonly used rodent diets: 1) standard, unpurified chow; 2) a purified/refined high-fat diet; 3) a purified/refined low-fat diet designed as a comparator for the high-fat diet. All three of these diets were given as homogeneous pellets, and the textures range from hard and fibrous (chow) to soft and oily like cookie dough (high-fat). The low-fat diet contains a lot of sugar, the high-fat diet contains a modest amount of sugar, and the chow diet contains virtually none. The particular high-fat diet in this paper (Research Diets D12451, 45% fat, which is high for a rat) is commonly used to produce obesity in rats, although it's not always very effective. The 60% fat version is more effective.

Consistent with previous findings, rats on every diet consumed the same number of calories over time... except the cafeteria diet-fed rats, which ate 30% more than any of the other groups. Rats on every diet gained fat compared to the unpurified chow group, but the cafeteria diet group gained much more than any of the others. There was no difference in fat gain between the purified high-fat and low-fat diets.

So in this paper, they compared two refined diets with vastly different carb:fat ratios and different sugar contents, and yet neither equaled the cafeteria diet in its ability to increase food intake and cause fat gain. The fat, starch and sugar content of the cafeteria diet was not able to fully explain its effect on fat gain. However, each diets' ability to cause fat gain correlated with its respective food reward qualities. Refined diets high in fat or sugar caused fat gain in rats relative to unpurified chow, but were surpassed by a diet containing a combination of fat, sugar, starch, salt, free glutamate (umami), interesting textures and pleasant and invariant aromas.

Although the cafeteria diet is the most effective at causing obesity in rodents, it's not commonly used because it's a lot more work than feeding pellets, and it introduces a lot of variability into experiments because each rat eats a different combination of foods.

How to Make an Obese Human Lean

In 1965, the Annals of the New York Academy of Sciences published a very unusual paper (3). Here is the stated goal of the investigators:

The study of food intake in man is fraught with difficulties which result from the enormously complex nature of human eating behavior. In man, in contrast to lower animals, the eating process involves an intricate mixture of physiologic, psychologic, cultural and esthetic considerations. People eat not only to assuage hunger, but because of the enjoyment of the meal ceremony, the pleasures of the palate and often to gratify unconscious needs that are hard to identify. Because of inherent difficulties in studying human food intake in the usual setting, we have attempted to develop a system that would minimize the variables involved and thereby improve the chances of obtaining more reliable and reproducible data.

Here's a photo of their "system":
It's a machine that dispenses bland liquid food through a straw, at the push of a button. They don't give any information on the composition of the liquid diet, beyond remarking that "carbohydrate supplied 50 per cent of the calories, protein 20 per cent and fat 30 per cent. the formula contained vitamins and minerals in amount adequate for daily maintenance."

Volunteers were given access to the machine and allowed to consume as much of the liquid diet as they wanted, but no other food. Since they were in a hospital setting, the investigators could be confident that the volunteers ate nothing else.

The first thing they report is what happened when they fed two lean people using the machine, for 16 or 9 days. Both of them maintained their typical calorie intake (~3,075 and ~4,430 kcal per day) and maintained a very stable weight during this period.

Next, the investigators did the same experiment using two "grossly obese" volunteers. Again, they were asked to "obtain food from the machine whenever hungry." Over the course of the first 18 days, the first (male) volunteer consumed a meager 275 calories per day. The second (female) volunteer consumed a ridiculously low 144 calories per day over the course of 12 days, losing 23 pounds. Without showing data, the investigators remarked that an additional three obese volunteers "showed a similar inhibition of calorie intake when fed by machine."

The first volunteer continued eating bland food from the machine for a total of 70 days, losing approximately 70 pounds. After that, he was sent home with the formula and instructed to drink 400 calories of it per day, which he did for an additional 185 days, after which his total weight loss was 200 lbs. The investigators remarked that "during all this time weight was steadily lost and the patient never complained of hunger or gastrointestinal discomfort." This is truly a starvation-level calorie intake, and to eat it continually for 255 days without hunger suggests that something rather interesting was happening in this man's body.

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. This also implies that obesity may not be a disorder, but rather a normal response to the prevailing dietary environment in affluent nations.

A second study by Dr. Michel Cabanac in 1976 confirmed that reducing food reward (by feeding bland food) lowers the fat mass setpoint in humans, using a clever method that I won't discuss for the sake of brevity (4). I learned about both of these studies through the writing of Dr. Seth Roberts, author of The Shangri-La Diet. I'd also like to thank Dr. Stephen Benoit, a researcher in the food reward field, for talking through these ideas with me to make sure I wasn't misinterpreting them.

I'd like to briefly remark that there's an anatomical basis for the idea of two-way communication between brain regions that determine reward and those that control body fatness. It's well known that the latter influence the former (think about your drive to obtain food after you've just eaten a big meal vs. after you've skipped a meal), but there are also connections from the former to the latter via a brain region called the lateral hypothalamus. The point is that it's anatomically plausible that food reward determines in part the amount of body fat a person carries.

Some people may be inclined to think "well, if food tastes bad, you eat less of it; so what!" Although that may be true to some extent, I don't think it can explain the fact that bland diets affect the calorie intake of lean and obese people differently. To me, that implies that highly rewarding food increases the body fat setpoint in susceptible people, and that food with few rewarding properties allows them to return to a lean state.

In the next few posts, I'll describe how food reward explains the effectiveness of many popular fat loss diets, I'll describe how this hypothesis fits in with the diets and health of non-industrial cultures, and I'll outline new dietary strategies for preventing and treating obesity and certain forms of metabolic dysfunction.

Food Reward: a Dominant Factor in Obesity, Part I

A Curious Finding

It all started with one little sentence buried in a paper about obese rats. I was reading about how rats become obese when they're given chocolate Ensure, the "meal replacement drink", when I came across this:

...neither [obesity-prone] nor [obesity-resistant] rats will overeat on either vanilla- or strawberry-flavored Ensure.

The only meaningful difference between chocolate, vanilla and strawberry Ensure is the flavor, yet rats eating the chocolate variety overate, rapidly gained fat and became metabolically ill, while rats eating the other flavors didn't (1). Furthermore, the study suggested that the food's flavor determined, in part, what amount of fatness the rats' bodies "defended."

As I explained in previous posts, the human (and rodent) brain regulates the amount of fat the body carries, in a manner similar to how the brain regulates blood pressure, body temperature, blood oxygenation and blood pH (2). That fact, in addition to several other lines of evidence, suggests that obesity probably results from a change in this regulatory system. I refer to the amount of body fat that the brain defends as the "body fat setpoint", however it's clear that the setpoint is dependent on diet and lifestyle factors. The implication of this paper that I could not escape is that a food's flavor influences body fatness and probably the body fat setpoint.

An Introduction to Food Reward

The brain contains a sophisticated system that assigns a value judgment to everything we experience, integrating a vast amount of information into a one-dimensional rating system that labels things from awesome to terrible. This is the system that decides whether we should seek out a particular experience, or avoid it. For example, if you burn yourself each time you touch the burner on your stove, your brain will label that action as bad and it will discourage you from touching it again. On the other hand, if you feel good every time you're cold and put on a sweater, your brain will encourage that behavior. In the psychology literature, this phenomenon is called "reward," and it's critical to survival.

The brain assigns reward to, and seeks out, experiences that it perceives as positive, and discourages behaviors that it views as threatening. Drugs of abuse plug directly into reward pathways, bypassing the external routes that would typically trigger reward. Although this system has been studied most in the context of drug addiction, it evolved to deal with natural environmental stimuli, not drugs.

As food is one of the most important elements of survival, the brain's reward system is highly attuned to food's rewarding properties. The brain uses input from smell, taste, touch, social cues, and numerous signals from the digestive tract* to assign a reward value to foods. Experiments in rats and humans have outlined some of the qualities of food that are inherently rewarding:

• Fat
• Starch
• Sugar
• Salt
• Meatiness (glutamate)
• The absence of bitterness
  
• Certain textures (e.g., soft or liquid calories, crunchy foods)
• Certain aromas (e.g., esters found in many fruits)
• Calorie density ("heavy" food)
  

We are generally born liking the qualities listed above, and aromas and flavors that are associated with these qualities become rewarding over time. For example, beer tastes terrible the first time you drink it because it's bitter, but after you drink it a few times and your brain catches wind that there are calories and a drug in there, it often begins tasting good. The same applies to many vegetables. Children are generally not fond of vegetables, but if you serve them spinach smothered in butter enough times, they'll learn to like it by the time they're adults.

The human brain evolved to deal with a certain range of rewarding experiences. It didn't evolve to constructively manage strong drugs of abuse such as heroin and crack cocaine, which overstimulate reward pathways, leading to the pathological drug seeking behaviors that characterize addiction. These drugs are "superstimuli" that exceed our reward system's normal operating parameters. Over the next few posts, I'll try to convince you that in a similar manner, industrially processed food, which has been professionally crafted to maximize its rewarding properties, is a superstimulus that exceeds the brain's normal operating parameters, leading to an increase in body fatness and other negative consequences.


* Nerves measure stomach distension. A number of of gut-derived paracrine and endocrine signals, including CCK, PYY, ghrelin, GLP-1 and many others potentially participate in food reward sensing, some by acting directly on the brain via the circulation, and others by signaling indirectly via the vagus nerve. More on this later.

Paleolithic Diet Clinical Trials, Part V

Dr. Staffan Lindeberg's group has published a new paleolithic diet paper in the journal Nutrition and Metabolism, titled "A Paleolithic Diet is More Satiating per Calorie than a Mediterranean-like Diet in Individuals with Ischemic Heart Disease" (1).

The data in this paper are from the same intervention as his group's 2007 paper in Diabetologia (2). To review the results of this paper, 12 weeks of a Paleolithic-style diet caused impressive fat loss and improvement in glucose tolerance, compared to 12 weeks of a Mediterranean-style diet, in volunteers with pre-diabetes or diabetes and ischemic heart disease. Participants who started off with diabetes ended up without it. A Paleolithic diet excludes grains, dairy, legumes and any other category of food that was not a major human food source prior to agriculture. I commented on this study a while back (3, 4).

One of the most intriguing findings in his 2007 study was the low calorie intake of the Paleolithic group. Despite receiving no instruction to reduce calorie intake, the Paleolithic group only ate 1,388 calories per day, compared to 1,823 calories per day for the Mediterranean group*. That's a remarkably low ad libitum calorie intake in the former (and a fairly low intake in the latter as well).

With such a low calorie intake over 12 weeks, you might think the Paleolithic group was starving. Fortunately, the authors had the foresight to measure satiety, or fullness, in both groups during the intervention. They found that satiety was almost identical in the two groups, despite the 24% lower calorie intake of the Paleolithic group. In other words, the Paleolithic group was just as full as the Mediterranean group, despite a considerably lower intake of calories. This implies to me that the body fat "set point" decreased, allowing a reduced calorie intake while body fat stores were burned to make up the calorie deficit. I suspect it also decreased somewhat in the Mediterranean group, although we can't know for sure because we don't have baseline satiety data for comparison.

There are a few possible explanations for this result. The first is that the Paleolithic group was eating more protein, a highly satiating macronutrient. However, given the fact that absolute protein intake was scarcely different between groups, I think this is unlikely to explain the reduced calorie intake.

A second possibility is that certain potentially damaging Neolithic foods (e.g., wheat and refined sugar) interfere with leptin signaling**, and removing them lowers fat mass by allowing leptin to function correctly. Dr. Lindeberg and colleagues authored a hypothesis paper on this topic in 2005 (5).

A third possibility is that a major dietary change of any kind lowers the body fat setpoint and reduces calorie intake for a certain period of time. In support of this hypothesis, both low-carbohydrate and low-fat diet trials show that overweight people spontaneously eat fewer calories when instructed to modify their diets in either direction (6, 7). More extreme changes may cause a larger decrease in calorie intake and fat mass, as evidenced by the results of low-fat vegan diet trials (8, 9). Chris Voigt's potato diet also falls into this category (10, 11). I think there may be something about changing food-related sensory cues that alters the defended level of fat mass. A similar idea is the basis of Seth Roberts' book The Shangri-La Diet.

If I had to guess, I would think the second and third possibilities contributed to the finding that Paleolithic dieters lost more fat without feeling hungry over the 12 week diet period.


*Intakes were determined using 4-day weighed food records.

**Leptin is a hormone produced by body fat that reduces food intake and increases energy expenditure by acting in the brain. The more fat a person carries, the more leptin they produce, and hypothetically this should keep body fat in a narrow window by this form of "negative feedback". Clearly, that's not the whole story, otherwise obesity wouldn't exist. A leading hypothesis is that resistance to the hormone leptin causes this feedback loop to defend a higher level of fat mass.

Potato Diet Interpretation

If you read my post on December 16th, you know that Chris Voigt saw remarkable fat loss and improvements in health markers as a result of two months of eating almost nothing but potatoes. This has left many people scratching their heads, because potatoes are not generally viewed as a healthy food. This is partially due to the fact that potatoes are very rich in carbohydrate, which also happens to be a quickly digested type, resulting in a high glycemic index. The glycemic index refers to the degree to which a particular food increases blood glucose when it's eaten, and I've questioned the relevance of this concept to health outcomes in the past (1, 2, 3). I think Mr. Voigt's results once again argue against the importance of the glycemic index as a diet-health concept.

It's often pointed out that potatoes are low in vitamins and minerals compared to vegetables on a per-calorie basis, but I think it's a misleading comparison because potatoes are much more calorie-dense than most vegetables. Potatoes compare favorably to other starchy staples such as bread, rice and taro.

Over the course of two months, Mr. Voigt lost 21 pounds. No one knows exactly how much of that weight came out of fat and how much out of lean mass, but the fact that he reported a decrease in waist and neck circumference indicates that most of it probably came out of fat. Previous long-term potato feeding experiments have indicated that it's possible to maintain an athletic muscle mass on the amount of protein in whole potatoes alone (4). So yes, Mr. Voigt lost fat on a very high-carbohydrate diet (75-80% carbohydrate, up to 440g per day).

On to the most interesting question: why did he lose fat? Losing fat requires that energy leaving the body exceed energy entering the body. But as Gary Taubes would say, that's obvious but it doesn't get us anywhere. In the first three weeks of his diet, Mr. Voigt estimates that he was only eating 1,600 calories per day. Aha! That's why he lost weight! Well, yes. But let's look into this more deeply. Mr. Voigt was not deliberately restricting his calorie intake at all, and he did not intend this as a weight loss diet. In my interview, I asked him if he was hungry during the diet. He said that he was not hungry, and that he ate to appetite during this period, realizing only after three weeks that he was not eating nearly enough calories to maintain his weight*. I also asked him how his energy level was, and he said repeatedly that it was very good, perhaps even better than usual. Those were not idle questions.

Calorie restriction causes a predictable physiological response in humans that includes hunger and decreased energy. It's the starvation response, and it's powerful in both lean and overweight people, as anyone knows who has tried to lose fat by decreasing calorie intake alone. The fact that he didn't experience hunger or fatigue implies that his body did not think it was starving. Why would that be?

I believe Mr. Voigt's diet lowered his fat mass 'setpoint'. In other words, for whatever reason, the diet made his body 'want' to be leaner that it already was. His body began releasing stored fat that it considered excess, and therefore he had to eat less food to complete his energy needs. You see this same phenomenon very clearly in rodent feeding studies. Changes in diet composition/quality can cause dramatic shifts in the fat mass setpoint (5, 6). Mr. Voigt's appetite would eventually have returned to normal once he had stabilized at a lower body fat mass, just as rodents do.

Rodent studies have made it clear that diet composition has a massive effect on the level of fat mass that the body will 'defend' against changes in calorie intake (5, 6). Human studies have shown similar effects from changes in diet composition/quality. For example, in controlled diet trials, low-carbohydrate dieters spontaneously reduce their calorie intake quite significantly and lose body fat, without being asked to restrict calories (7). In Dr. Staffan Lindeberg's Paleolithic diet trials, participants lost a remarkable amount of fat, yet a recent publication from his group shows that the satiety (fullness) level of the Paleolithic group was not different from a non-Paleolithic comparison group despite a considerably lower calorie intake over 12 weeks (8, 9). I'll discuss this important new paper soon. Together, this suggests that diet composition/quality can have a dominant impact on the fat mass setpoint.

One possibility is that cutting the wheat, sugar, most vegetable oil and other processed food out of Mr. Voigt's diet was responsible for the fat loss. I think that's likely to have contributed. Many people find, for example, that they lose fat simply by eliminating wheat from their diet.

Another possibility that I've been exploring recently is that changes in palatability (pleasantness of flavor) influence the fat mass setpoint. There is evidence in rodents that it does, although it's not entirely consistent. For example, rats will become massively obese if you provide them with chocolate flavored Ensure (a meal replacement drink), but not with vanilla or strawberry Ensure (10). They will defend their elevated fat mass against calorie restriction (i.e. they show a physiological starvation response when you try to bring them down to a lower weight by feeding them less chocolate Ensure) while they're eating chocolate Ensure, but as soon as you put them back on unpurified rodent pellets, they will lose fat and defend the lower fat mass. Giving them food in liquid or paste form often causes obesity, while the same food in solid pellet form will not. Eating nothing but potatoes is obviously a diet with a low overall palatability.

So I think that both a change in diet composition/quality and a decrease in palatability probably contributed to a decrease in Mr. Voigt's fat mass setpoint, which allowed him to lose fat mass without triggering a starvation response (hunger, fatigue).

The rest of his improvements in health markers were partially due to the fat loss, including his decreased fasting glucose, decreased triglycerides, and presumably increased insulin sensitivity. They may also have been partially due to a lack of industrial food and increased intake of certain micronutrients such as magnesium.

One of the most striking changes was in his calculated LDL cholesterol ("bad" cholesterol), which decreased by 41%, putting him in a range that's more typical of healthy non-industrial cultures including hunter-gatherers. Yet hunter-gatherers didn't eat nothing but potatoes, often didn't eat much starch, and in some cases had a high intake of fat and saturated fat, so what gives? It's possible that a reduced saturated fat intake had an impact on his LDL, given the relatively short timescale of the diet. But I think there's something mysterious about this setpoint mechanism that has a much broader impact on metabolism than is generally appreciated. For example, calorie restriction in humans has a massive impact on LDL, much larger than the impact of saturated fat (11). And in any case, the latter appears to be a short-term phenomenon (12). It's just beginning to be appreciated that energy balance control systems in the brain influence cholesterol metabolism.

Mr. Voigt's digestion appeared to be just fine on his potato diet, even though he generally ate the skins. This makes me even more skeptical of the idea that potato glycoalkaloids in common potato varieties are a health concern, especially if you were to eliminate most of the glycoalkaloids by peeling.

I asked Mr. Voigt about what foods he was craving during the diet to get an idea of whether he was experiencing any major deficiencies. The fact that Mr. Voigt did not mention craving meat or other high-protein foods reinforces the fact that potatoes are a reasonable source of complete protein. The only thing he craved was crunchy/juicy food, which I'm not sure how to interpret.

He also stopped snoring during the diet, and began again immediately upon resuming his normal diet, perhaps indicating that his potato diet reduced airway inflammation. This could be due to avoiding food allergies and irritants (wheat anyone?) and also fat loss.

Overall, a very informative experiment! Enjoy your potatoes.


*Until the last 5.5 weeks, when he deliberately stuffed himself beyond his appetite because his rapid weight loss worried him. Yet, even with deliberate overfeeding up to his estimated calorie requirement of 2,200 calories per day, he continued to lose weight. He probably was not quite reaching his calorie goal, or his requirement is higher than he thought.

Paleolithic Diet Clinical Trials Part II

There were a number of remarkable changes in both trials. I'll focus mostly on Dr. Lindeberg's trial because it was longer and better designed. The first thing I noticed is that caloric intake dropped dramatically in both trials, -36% in the first trial and a large but undetermined amount in Dr Lindeberg's. The Mediterranean diet group ended up eating 1,795 calories per day, while the paleolithic dieters ate 1,344. In both studies, participants were allowed to eat as much as they wanted, so those reductions were purely voluntary.

This again agrees with the theory that certain grains (wheat) promote hyperphagia, or excessive eating. It's the same thing you see in low-carbohydrate diet trials, such as this one, which also reduce grain intake. The participants in Lindeberg's study were borderline obese. When you're overweight and your body resets its fat mass set-point due to an improved diet, fatty acids come pouring out of fat tissue and you don't need as many calories to feel satisfied. Your diet is supplemented by generous quantities of lard. Your brain decreases your calorie intake until you approach your new set-point.

That's what I believe happened here. The paleolithic group supplemented their diet with 3.9 kg of their own rump fat over the course of 12 weeks, coming out to 30,000 additional calories, or 357 calories a day. Not quite so spartan when you think about it like that.

The most remarkable thing about Lindeberg's trial was the fact that the 14 people in the paleolithic group, 2 of which had moderately elevated fasting blood glucose and 10 of which had diabetic fasting glucose, all ended up with normal fasting glucose after 12 weeks. That is truly amazing. The mediterranean diet worked also, but only in half as many participants.

If you look at their glucose tolerance by an oral glocose tolerance test (OGTT), the paleolithic diet group improved dramatically. Their rise in blood sugar after the OGTT (fasting BG subtracted out) was 76% less at 2 hours. If you look at the graph, they were basically back to fasting glucose levels at 2 hours, whereas before the trial they had only dropped slightly from the peak at that timepoint. The mediterranean diet group saw no significant improvement in fasting blood glucose or the OGTT. Lindeberg is pretty modest about this finding, but he essentially cured type II diabetes and glucose intolerance in 100% of the paleolithic group.

Fasting insulin, the insulin response to the OGTT and insulin sensitivity improved in the paleolithic diet whereas only insulin sensitivity improved significantly in the Mediterranean diet. Fasting insulin didn't decrease as much as I would have thought, only 16% in the paleolithic group.

Another interesting thing is that the paleolithic group lost more belly fat than the Mediterranean group, as judged by waist circumference. This is the most dangerous type of fat, which is associated with, and contributes to, insulin resistance and the metabolic syndrome. Guess what food belly fat was associated with when they analyzed the data? The strongest association was with grain consumption (probably mostly wheat), and the association remained even after adjusting for carbohydrate intake. In other words, the carbohydrate content of grains does not explain their association with belly fat because "paleo carbs" didn't associate with it. The effect of the paleolithic diet on glucose tolerance was also not related to carbohydrate intake.

So in summary, the "Mediterranean diet" may be healthier than a typical Swedish diet, while a diet loosely modeled after a paleolithic diet kicks both of their butts around the block. My opinion is that it's probably due to eliminating wheat, substantially reducing refined vegetable oils and dumping the processed junk in favor of real, whole foods. Here's a zinger from the end of the paper that sums it up nicely (emphasis mine):

The larger improvement of glucose tolerance in the Paleolithic group was independent of energy intake and macronutrient composition, which suggests that avoiding Western foods is more important than counting calories, fat, carbohydrate or protein. The study adds to the notion that healthy diets based on whole-grain cereals and low-fat dairy products are only the second best choice in the prevention and treatment of type 2 diabetes.

Paleolithic Diet Clinical Trials

If Dr. Ancel Keys (of diet-heart hypothesis fame) had been a proponent of "paleolithic nutrition", we would have numerous large intervention trials by now either confirming or denying its ability to prevent health problems. In this alternate reality, public health would probably be a lot better than it is today. Sadly, we have to settle for our current reality where the paleolithic diet has only been evaluated in two small trials, and medical research spends its (our) money repeatedly conducting failed attempts to link saturated fat to every ill you can think of. But let's at least take a look at what we have.

Both trials were conducted in Sweden. In the first one, lead by Dr. Per Wändell, 14 healthy participants (5 men, 9 women) completed a 3-week dietary intervention in which they were counseled to eat a "paleolithic diet". Calories were not restricted, only food categories were. Participants were told to eat as much as they wanted of fruit, vegetables, fish, lean meats, nuts, flax and canola oil, coffe and tea (without dairy). They were allowed restricted quantities of dried fruit, potatoes (2 medium/day) salted meat and fish, fat meat and honey. They were told not to eat dairy, grain products, canned food, sugar and salt.

After three weeks, the participants had:

• Decreased their caloric intake from 2,478 to 1,584 kcal
• Increased their percentage protein and fat, while decreasing carbohydrate
• Decreased saturated fat, increased dietary cholesterol, decreased sodium intake, increased potassium
  
• Lost 2.3 kg (5 lb)
• Decreased waist circumference, blood pressure and PAI-1

Not bad for a 3-week intervention on healthy subjects. This study suffered from some serious problems, however. #1 is the lack of a control group as a means for comparison. Ouch. #2 is the small study size and resulting lack of statistical power. I consider this one encouraging but by no means conclusive.

The second study was conducted by the author of the Kitava study, Dr. Staffan Lindeberg. The study design was very interesting. He randomly assigned 29 men with ischemic heart disease, plus type II diabetes or glucose intolerance, to either a "Mediterranean diet" or a "paleolithic diet". Neither diet was calorie-restricted. Here's the beauty of the study design: the Mediterranean diet was the control for the paleo diet. The reason that's so great is it completely eliminates the placebo effect. Both groups were told they were being assigned to a healthy diet to try to improve their health. Each group was educated on the health benefits of their diet but not the other one. It would have been nice to see a regular non-intervention control group as well, but this design was adequate to see some differences.

Participants eating the Mediterranean diet were counseled to focus on whole grains, low-fat dairy, potatoes, legumes, vegetables, fruit, fatty fish and vegetable oils rich in monounsaturated fats and alpha-linolenic acid (omega-3). I'm going to go on a little tangent here. This is truly a bizarre concept of what people eat in the Mediterranean region. It's a fantasy invented in the US to justify the mainstream concept of a healthy diet. My father is French and I spent many summers with my family in southern France. They ate white bread, full-fat dairy at every meal, legumes only if they were smothered in fatty pork, sausages and lamb chops. In fact, full-fat dairy wasn't fat enough sometimes. Many of the yogurts and cheeses we ate were made from milk with extra cream added. Want to get a lecture from Grandmere? Try cutting the fat off your pork chop!

The paleolithic group was counseled to eat lean meat, fish, fruit, leafy and cruciferous vegetables, root vegetables (including moderate amounts of potatoes), eggs and nuts. They were told to avoid dairy, grain products, processed food, sugar and beer.

Both groups were bordering on obese at the beginning of the study. All participants had cardiovascular disease and moderate to severe glucose intolerance (i.e. type II diabetes). After 12 weeks, both groups improved on several parameters. That includes fat mass and waist circumference. But the paleolithic diet trumped the Mediterranean diet in many ways:

• Greater fat loss in the the midsection and a trend toward greater weight loss
  
• Greater voluntary reduction in caloric intake (total intake paleo= 1,344 kcal; Med= 1,795)
  
• A remarkable improvement in glucose tolerance that did not occur significantly in the Mediterranean group
• A decrease in fasting glucose
  
• An increase in insulin sensitivity (HOMA-IR)
  

Overall, the paleolithic diet came out looking very good. But I haven't even gotten to the best part yet. At the beginning of the trial, 12 out of the 14 people in the paleo group had elevated fasting glucose. At the end, every single one had normal fasting glucose. In the Mediterranean group, 13 out of 15 began with elevated glucose and 8 out of 15 ended with it. This clearly shows that a paleolithic diet is an excellent way to restore glucose control to a person who still has beta cells in their pancreas.

This post is getting long, so I think I'll save the interpretation for the next post.

Rats on Junk Food

If diet composition causes hyperphagia, we should be able to see it in animals. I just came across a great study from the lab of Dr. Neil Stickland that explored this in rats. They took two groups of pregnant rats and fed them two different diets ad libitum, meaning the rats could eat as much as they wanted. Here's what the diets looked like:

The animals were fed two types of diet throughout the study. They were fed either RM3 rodent chow alone ad libitum (SDS Ltd, Betchworth, Surrey, UK) or with a junk food diet, also known as cafeteria diet, which consisted of eight different types of palatable foods, purchased from a British supermarket. The palatable food included biscuits, marshmallows, cheese, jam doughnuts, chocolate chip muffins, butter flapjacks, potato crisps and caramel/chocolate bars.

It's important to note that the junk food-fed rats had access to rat chow as well. Now here's where it gets interesting. Rats with access to junk food in addition to rat chow ate 56% more calories than the chow-only group! Here's what they had to say about it:

These results clearly show that pregnant rats, given ad libitum access to junk food, exhibited hyperphagia characterised by a marked preference for foods rich in fat, sucrose and salt at the expense of protein-rich foods, when compared with rats that only had access to rodent chow. Although the body mass of dams was comparable among all groups at the start of the experiment, the increased energy intake in the junk food group throughout gestation was accompanied by an increase in body mass at G20 [gestational day 20] with the junk food-fed dams being 13 % heavier than those fed chow alone.

Hmm, this is remarkably reminiscent of what's happening to a certain group of humans in North America right now: give them access to food made mostly of refined grains, sugar, and industrially processed vegetable oil. They will prefer it to healthier food, to the point of overeating. The junk food then drives hyperphagia by interfering with the body's feedback loops that normally keep feeding behaviors and body fat within the optimal range. These data support the hypothesis that metabolic damage is the cause of, not the result of, "super-sized" food portions and other similar cultural phenomena.

The rest of the paper is interesting as well. Pups born to mothers who ate junk food while pregnant and lactating had a greater tendency to eat junk than pups born to mothers who ate rat chow during the same period. This underscores the idea that poor nutrition can set a child up for a lifetime of problems.

Hyperphagia

One of the things I didn't mention in the last post is that Americans are eating more calories than ever before. According to Centers for Disease Control NHANES data, in 2000, men ate about 160 more calories per day, and women ate about 340 more than in 1971. That's a change of 7% and 22%, respectively. The extra calories come almost exclusively from refined grains, with the largest single contribution coming from white wheat flour (correction: the largest single contribution comes from corn sweeteners, followed by white wheat flour).

Some people will see those data and decide the increase in calories is the explanation for the expanding American waistline. I don't think that's incorrect, but I do think it misses the point. The relevant question is "why are we eating more calories now than we were in 1971?"

We weren't exactly starving in 1971. And average energy expenditure, if anything, has actually increased. So why are we eating more? I believe that our increased food intake, or hyperphagia, is the result of metabolic disturbances, rather than the cause of them.

Humans, like all animals, have a sophisticated system of hormones and brain regions whose function is to maintain a proper energy balance. Part of the system's job is to keep fat mass at an appropriate level. With a properly functioning system, feedback loops inhibit hunger once fat mass has reached a certain level, and also increase resting metabolic rate to burn excess calories. If the system is working properly, it's very difficult to gain weight. There have been a number of overfeeding studies in which subjects have consumed huge amounts of excess calories. Some people gain weight, many don't.

The fact that fat mass is hormonally regulated can be easily seen in other mammals. When was the last time you saw a fat squirrel in the springtime? When was the last time you saw a thin squirrel in the fall? These events are regulated by hormones. A squirrel in captivity will put on weight in the fall, even if its daily food intake is not changed.

A key hormone in this process is leptin. Leptin levels are proportional to fat mass, and serve to inhibit hunger and eating behaviors. Under normal conditions, the more fat tissue a person has, the more leptin they will produce, and the less they will eat until the fat mass has reached the body's preferred 'set-point'. The problem is that overweight Westerners are almost invariably leptin-resistant, meaning their body doesn't respond to the signal to stop eating!

Leptin resistance leads to hyperphagia, overweight and the metabolic syndrome (a common cluster of symptoms that implies profound metabolic disturbance). It typically precedes insulin resistance during the downward slide towards metabolic syndrome.

I suspect that wheat, sugar and perhaps other processed foods cause hyperphagia. It's the same thing you see when wheat is first introduced to a culture, even if it's replacing another refined carbohydrate. I believe hyperphagia is secondary to a disturbed metabolism. There's something about the combination of refined wheat, sugar, processed vegetable oils and other industrial foods that reached a critical mass in the mid-70s. The shift in diet composition disturbed our normal hormonal profile (even more than it was already disturbed), and sent us into a tailspin of excessive eating and unprecedented weight gain.

New Low-carb Study

I know you’ve all heard the news about the new low-carb study in the New England Journal of Medicine by now, but I have to chime in. I‘m going to try to offer you a different perspective of the study that you may not have found elsewhere. First of all, this is a Rolls Royce of a study. It was large, well-controlled, and two years long. It was partly funded by the Atkins foundation, but it's a peer-reviewed study in a good journal and if anything the study design is slanted against the low-carbohydrate diet.

The study compared the weights and various health parameters of 322 overweight subjects put on one of three diets: a “low-fat diet”, a Mediterranean diet and a “low-carbohydrate diet”. The first two were calorie-restricted while the low-carb diet was not. First of all, the “low-fat” diet was not particularly low in fat. It was 30% fat by calories, only a few percent short of the US average. What they call low-fat in the study is actually a calorie-restricted version of the American Heart Association diet recommendation, which suggests:

“…30% of calories from fat, 10% calories from saturated fat, and an intake of 300 mg of cholesterol per day. The Participants were counseled to consume low-fat grains, vegetables, fruits, and legumes and to limit their consumption of additional fats, sweets and high-fat snacks.”

So henceforth, I’ll refer to it as the AHA diet rather than the low-fat diet.

The “low-carb” diet wasn’t particularly low in carbohydrate either. The low-carb group was only getting 10% fewer calories from carbohydrate than the low-fat or Mediterranean diet groups. Despite these problems, the low-carbohydrate diet was the most effective overall. It caused a weight loss of 5.5 kg (12 lb), compared to 4.6 kg (10 lb) and 3.3 kg (7.4 lb) for the Mediterranean and AHA diets, respectively.

One of the most amazing aspects of the study is that the low-carb diet was the only one that wasn’t calorie-restricted, yet it caused the most weight loss. People in the low-carb group naturally reduced their calorie intake over the course of the study, ending up with an intake similar to the AHA group.

The low-carb diet also came out on top in most of the markers of health they examined. It caused the largest drop in HbA1c, a measure of average blood glucose level. It caused the largest drop in C-reactive protein, a measure of inflammation (the Mediterranean diet also did well). And finally, it caused the biggest improvement in the triglyceride:HDL ratio. This ratio is the best blood lipid predictor of heart disease risk I’m aware of in modern Western populations. The lower, the better. They didn't calculate it in the study so I had to do it myself. Here's a graph of the change in trig:HDL ratio for each group over the course of the study:


Other interesting findings: despite the calorie restriction, diabetic participants on the AHA group actually saw a significant increase in fasting blood glucose.

I've speculated before that wheat and sugar may cause hyperphagia, or excessive eating. We can see from these results that reducing carbohydrate (and probably wheat) reduces overall caloric intake quite significantly. This squares with the findings of the recent Chinese study that showed an increase in calorie intake and weight, correlating with the replacement of rice with wheat as the primary carbohydrate. It also squares with diet trends in the US, where wheat consumption has risen alongside calorie intake and weight.

I'd love to know what the results would have looked like if they had gone on a true low-carbohydrate diet, or even simply eliminated grains and sugar.

Another China Tidbit

A final note about the Chinese study in the previous post: the overweight vegetable-eaters (read: wheat eaters) exercised more than their non-vegetable-eating, thin neighbors. So although their average calorie intake was a bit higher, their expenditure was as well. So much for 'calories in, calories out'...

Although I speculated in the last post that affluent people might be eating more wheat and fresh vegetables, the data don't support that. Participants with the highest income level actually adhered to the wheat and vegetable-rich pattern the least, while low-income participants were most likely to eat this way.

Interestingly, education showed a (weaker) trend in the opposite direction. More educated participants were more likely to eat the wheat-vegetable pattern, while the opposite was true of less educated participants. Thus, it looks like wheat makes people more educated. Just kidding, that's exactly the logic we have to avoid when interpreting this type of study!

Wheat in China

Dr. Michael Eades linked to an interesting study yesterday on his Health and Nutrition blog. It's entitled "Vegetable-Rich Food Pattern is Related to Obesity in China."

It's one of these epidemiological studies where they try to divide subjects into different categories of eating patterns and see how health problems associate with each one. They identified four patterns: the 'macho' diet high in meat and alcohol; the 'traditional' diet high in rice and vegetables; the 'sweet tooth' pattern high in cake, dairy and various drinks; and the 'vegetable rich' diet high in wheat, vegetables, fruit and tofu. The only pattern that associated with obesity was the vegetable-rich diet. The 25% of people eating closest to the vegetable-rich pattern were more than twice as likely to be obese as the 25% adhering the least.

The authors of the paper try to blame the increased obesity on a higher intake of vegetable oil from stir-frying the vegetables, but that explanation is juvenile and misleading. A cursory glance at table 3 reveals that the vegetable-eaters weren't eating any more fat than their thinner neighbors. Dr. Eades suggests that their higher carbohydrate intake (+10%) and higher calorie intake (+120 kcal/day) are responsible for the weight gain, but I wasn't satisfied with that explanation so I took a closer look.

One of the most striking elements of the 'vegetable-rich' food pattern is its replacement of rice with wheat flour. The 25% of the study population that adhered the least to the vegetable-rich food pattern ate 7.3 times more rice than wheat, whereas the 25% sticking most closely to the vegetable-rich pattern ate 1.2 times more wheat than rice! In other words, wheat flour had replaced rice as their single largest source of calories. This association was much stronger than the increase in vegetable consumption itself!

All of a sudden, the data make perfect sense. Wheat seems to destroy the metabolism of cultures wherever it goes. I think the reason we don't see the same type of association in American epidemiological studies is that everyone eats wheat. Only in a culture that has a true diversity of diet can you find a robust association like this. The replacement of rice with wheat may have caused the increase in calorie intake as well, subsequent to metabolic dysfunction. Clinical trials of low-carbohydrate diets as well as 'paleolithic diets' have shown good metabolic outcomes from wheat avoidance, although one can't be sure that wheat is the only culprit from those data.

I don't think the vegetables had anything to do with the weight gain, they were just incidentally associated with wheat consumption. But I do think these data argue against the commonly-held idea that vegetables protect against overweight.