1. Weight gain does not depend on calories in versus calories out, even in the loosest sense.

2. Weight gain is entirely a function of calories in versus calories out, but calories may move in unexpected ways not linked to the classic “eat” and “exercise” dichotomy. For example, some people may have “fast metabolisms” which burn calories even when they are not exercising. These people may stay very thin even if they eat and exercise as much as much more obese people.

3. Weight gain is entirely a function of calories in versus calories out, and therefore of how much you eat and exercise. However, these are in turn mostly dependent on the set points of a biologically-based drive. For example, some people may have overactive appetites, and feel starving unless they eat an amount of food that will make them fat. Other people will have very strong exercise drives and feel fidgety unless they get enough exercise to keep them very thin. These things can be altered in various ways which cause weight gain or loss, without the subject exerting willpower. For example, sleep may cause weight loss because people who get a good night sleep have decreased appetite and lower levels of appetite-related hormones.

4. Weight gain is entirely a function of calories in versus calories out, and therefore of how much you eat and exercise. That means diet is entirely a function of willpower and any claim that factors other than amount of food eaten and amount of exercise performed can affect weight gain is ipso facto ridiculous. For example, we can dismiss claims that getting a good night’s sleep helps weight loss, because that would violate the laws of thermodynamics.

1 and 4 are kind of dumb. 1 is dumb because…well, to steal an Eddington quote originally supposed apply to the second law of thermodynamics:

If someone points out to you that your pet theory of the universe is in disagreement with Maxwell’s equations — then so much the worse for Maxwell’s equations. If it is found to be contradicted by observation — well, these experimentalists do bungle things sometimes. But if your theory is found to be against…thermodynamics I can give you no hope; there is nothing for it but to collapse in deepest humiliation.

But 4 is also dumb. We have a long list of things that affect weight gain – for example, patients on the powerful psychiatric medication clozapine usually gain a lot of weight – fifteen pounds more on average than people on safer antipsychotics. Other medications are known to increase weight to a lesser degree, and some medications even decrease weight, though you wouldn’t like the side effects of most of them. Certain genetic diseases are also known to cause increased weight – Prader-Willi syndrome, for example.

One could try to rescue 4 by saying that people with rare genetic diseases or taking powerful prescription-only medications are a different story and in normal people it’s entirely controlled by willpower. But first, this is an area where possibility proofs are half the battle, and we have a possibility proof. And second, there are more than enough studies about genetics, microbiome, and, yes, sleep showing that all of these things can have effects in normal people.

So 1 and 4 are out. And although I do sometimes see people pushing them, they mostly seem to do a thriving business as straw men for people who want to accuse their opponents of saying something absurd.

The most interesting debate to be had is between 2 and 3. 3 says that all of the interventions that we know affect weight – certain pills, certain recreational drugs, changes in gut bacteria, whatever – do it by affecting appetite and exercise drive. 2 says that basal metabolism is also involved. 3 seems to at least leave open the possibility of just starving yourself even when your body is telling you really hard to eat. 2 says even that won’t work.

There’s room for a little bit of gradation between 2 and 3. A lot of people suggest that one way “fast metabolism” presents is by people fidgeting a lot, which is sort of the same as “your body increases its exercise drive”.

But in general, I think 2 is an important issue that does cause at least some interpersonal weight differences.

We’ll start with the “possibility proof” again. MRAP2. It’s a gene. Scientists can delete it in mice. These mice will eventually develop excessive appetites. But when they are young, they eat the same amount as any other mouse, but still get fatter.

Likewise, 2,4-dinitrophenol is a cellular uncoupling agent which increases metabolic rate and consistently produces weight loss of 2-3 pounds per week. It would be an excellent solution to all of our obesity-related problems if the papers on it didn’t keep having names like 2,4-Dinitrophenol: A Weight Loss Agent With Significant Acute Toxicity And Risk Of Death.

So what about everyday life?

A study of individual variation in basal metabolic rate found very significant interpersonal differences. A lot of that was just “some people are bigger than others”, but some of it wasn’t – they state that “twenty-six percent of the variance remained unexplained”. The Wikipedia article puts this in context: “One study reported an extreme case where two individuals with the same lean body mass of 43 kg had BMRs of 1075 kcal/day (4.5 MJ/day) and 1790 kcal/day (7.5 MJ/day). This difference of 715 kcal/day (67%) is equivalent to one of the individuals completing a 10 kilometer run every day”

Dr. Claude Bouchard and his team stuck 12 pairs of male identical twins in isolation chambers where their caloric intake and exercise could be carefully controlled, then fed them more calories than their bodies needed. All sets of twins gained weight, and in all twin groups both twins gained about the same amount of weight as each other, but the amount of weight gained varied between twin pairs by a factor of 3 (from 4 to 13 kg).

A lot of the sites that talk about this thing are careful to say that people “can’t blame” genes for their obesity, because obesity levels have been rising for decades and genes can’t change that quickly. I think this is wrong-headed. True, genes are not the source of the modern rise in obesity levels. But it’s entirely possible that a globally rising tide of obesity has disproportionately affected the people with the wrong genes. Just as Bouchard fed the same amount extra to all his study participants but some of them gained more weight than others, so if you put an entire civilization worth of people in an obesogenic environment, some of them might be genetically predisposed to do worse than the rest.

A more practical question – can individual people’s metabolism change?

I am personally predisposed to answer in the affirmative. In my early twenties, I ate a crazy amount every day – two bagels with breakfast, cookies with lunch, a big dinner followed by dessert – and I stayed pretty thin throughout. Now I’m thirty, I eat a very restrained diet, and my weight still hovers at just above the range where I am supposed to be. I know that people are famously bad at understanding how much they’re eating and exercising, but seriously if you try to convince me that I’m eating more now than I was then I’m going to start doubting my own sanity, or at least my autobiographical memory.

But there’s not much evidence to back me up. Metabolic rate is well-known to decline with age, but linearly and predictably. And it changes with muscle mass, but only minimally – and I don’t think I used to be any more muscular.

The sites that talk about drastic and unexpected ways to change metabolism seem mostly crackpottish. This isn’t to say their methods don’t work – green tea, for example, has a statistically significant effect – but it’s all so small as to be pretty meaningless in a real-world context.

So my own story seems to be on shaky ground. But as far as I can tell, the people arguing that they’re trying just as hard as anybody else but still unable to lose weight because of their metabolism are very possibly right.

There are two sets of arguments against including wheat in the diet. First, wheat is a carbohydrate, and some people support low carbohydrate diets. Second, something might be especially dangerous about wheat itself.

It was much easier to figure out the state of the evidence on low-carbohydrate diets. They seem to be at least as good and maybe a little better for weight loss than traditional diets, but this might just be because there are lots of carbohydrates that taste very good and when forced to avoid them, people eat less stuff. They may or may not positively affect metabolic parameters and quality of life. (1, 2, 3, 4). They don’t seem to cause either major health benefits or major health risks in the medium term, which is the longest term for which there is good data available – for example, they have no effect on cancer rates. Overall they seem solid but unspectacular. But there’s a long way between “low carbohydrate diet” and “stop eating wheat”.

So I was more interested in figuring out what was going on with wheat in particular.

Wheat contains chemicals [citation needed]. The ones that keep cropping up (no pun intended) in these kinds of discussions are phytates, lectins, gluten, gliadin, and agglutinin, the last three of which for your convenience have been given names that all sound alike.

Various claims have been made about these chemicals’ effects on health. These have some prima facie plausibility. Plants don’t want to be eaten [citation needed] and they sometimes fill their grains with toxins to discourage animals from eating them. Ricin, a lectin in the seeds of the castor oil plant so toxic it gets used in chemical warfare, is a pretty good example. Most toxins are less dramatic, and most animals have enzymes that break down the toxins in their preferred food sources effectively. But if humans are insufficiently good at this, maybe because they didn’t evolve to eat wheat, some of these chemicals could be toxic to humans.

On the other hand, this same argument covers every pretty much every grain and vegetable and a lot of legumes – pretty much every plant-based food source except edible fruits. So we need a lot more evidence to start worrying about wheat.

I found the following claims about negative effects of wheat:

1. Some people without celiac disease are nevertheless sensitive to gluten.
2. Wheat increases intestinal permeability, causing a leaky gut and autoimmune disease.
3. Digestion of wheat produces opiates, which get you addicted to wheat.
4. Wheat something something something autism and schizophrenia.
5. Wheat has been genetically modified recently in ways that make it much worse for you.
6. The lectins in wheat interfere with leptin receptors, making people leptin resistant and therefore obese.

I’ll try to look at each of those and then turn to the positive claims made about wheat to see if they’re strong enough to counteract them.

Some People Without Celiac Disease Are Sensitive To Gluten – Mostly true but of limited significance

Celiac disease is one source of concern. Everybody on all sides of the wheat debate agree about the basic facts of this condition, which affects a little less than 1% of the population. They have severe reactions to the gluten in wheat. Celiac disease is mostly marked by gastroentereological complaints – diarrhea, bloating, abdominal pain – but it is also associated with vitamin deficiencies, anaemia, skin reactions, infertility, and “malaise”. It can be pretty straightforwardly detected by blood tests and gut biopsies and is not subtle.

People start to disagree about the existence of “gluten sensitivity”, which if it existed would be a bad reaction to gluten even in people who don’t test positive for celiac disease. Many people believe they have gastrointestinal (or other) symptoms that go away when they eat gluten-free diets, but science can’t find anything wrong with their intestines that could be causing the problems.

A recent study somewhat vindicated these people. Biesiekierski 2011 describes a double-blind randomized controlled trial: people who said they had “gluten-sensitive” irritable bowel syndrome were put on otherwise gluten-free diets and then randomly given either gluten or a placebo. They found that the patients given gluten reported symptoms (mostly bowel-related and tiredness) much more than those given placebo (p = 0.0001) but did not demonstrate any of the chemical, immunological, or histological markers usually associated with celiac disease. A similar Italian study found the same thing, except that they did find a higher rate of anti-gluten antibodies in their patients. Another study found that non-celiacs with antibodies to gluten had higher rates of mortality. And another study did find a histological change in bowel barrier function on this group of patients with the introduction of gluten. And another study from the same group found that maybe FODMAPs, another component of wheat, are equally or more responsible.

The journal Gastroenterology, which you may not be surprised to learn is the leading journal in the field of gastroenterology, proclaims:

The current working definition of nonceliac gluten sensitivity (NCGS) is the occurrence of irritable bowel syndrome (IBS)-like symptoms after the ingestion of gluten and improvement after gluten withdrawal from the diet after exclusion of celiac disease based on negative celiac serologies and/or normal intestinal architecture and negative immunoglobulin (Ig)E-mediated allergy tests to wheat. Symptoms reported to be consistent with NCGS are both intestinal (diarrhea, abdominal discomfort or pain, bloating, and flatulence) and extra-intestinal (headache, lethargy, poor concentration, ataxia, or recurrent oral ulceration). These criteria strongly and conveniently suggest that NCGS is best understood as a subset of IBS or perhaps a closely related but distinct functional disorder. Although the existence of NCGS has been slowly gaining ground with physicians and scientists, NCGS has enjoyed rapid and widespread adoption by the general public.

But even this isn’t really that interesting. Maybe some people with irritable bowel syndrome or certain positive antibodies should try avoiding gluten to see if it helps their specific and very real symptoms. At most ten percent of people are positive antibody testing, and not all of those even have symptoms. That’s still a far cry from saying no one should eat wheat.

But the anti-wheat crowd says an alternative more sensitive antibody test could raise sensitivity as high as a third of the population. The test seems to have been developed by a well-respected and legitimate doctor, but it hasn’t as far as I can tell been submitted for peer review or been confirmed by any other source. Meh.

That’s boring anyway. The real excitement comes from sweeping declarations that the entire population is sensitive to wheat.

Wheat Increases Intestinal Permeability Causing A Leaky Gut – Probably true, of uncertain significance

There are gluten-induced mucosal changes in subjects without small bowel disease. And gliadin increases intestinal permeability in the test tube, which should be extremely concerning to any test tubes reading this.

But probably the bigger worry here are lectins, which include wheat germ agglutinin. WGA affects the intestinal permeability of rats, which should be extremely concerning to any rats reading this. The same substance has been found to produce pro-inflammatory cytokines and interfere with the growth of various organs including the gut.

So there’s pretty good evidence that chemicals in wheat can increase intestinal permeability. Who cares?

For years, “leaky gut syndrome” was an alternative medicine diagnosis that was soundly mocked by the mainstream medical establishment. Then the mainstream medical establishment confirmed it existed and did that thing where they totally excused their own mocking of it but were ABSOLUTELY OUTRAGED that the alternative medicine community might have in some cases been overenthusiastic about it.

Maybe I’m being too harsh. The alternative medicine community often does take “leaky gut syndrome” way too far.

On the other hand, it’s probably real and Nature Clinical Practice is now publishing papers saying it is “a key ingredient in the pathogenesis of autoimmune diseases” and “offers innovative, unexplored approaches for the treatment of these devastating diseases” and gut health has been deemed “a new objective in medicine”. Preliminary changes to intestinal permeability have been found in asthma, in diabetes, and even in depression.

But it’s not yet clear if this is cause and effect. Maybe the stress of having asthma increases intestinal permeability somehow. Or maybe high intestinal permeability causes asthma somehow. It sure seems like the latter might work – all sorts of weird antigens and stuff from food can make it into the bloodstream and alarm the immune system – but right now this is all speculative.

So what we have is some preliminary evidence that wheat increases intestinal permeability, and some preliminary evidence that increased intestinal permeability is bad for you in a variety of ways.

And I don’t doubt that those two facts are true, but my knowledge of this whole area is so weak that I wonder how much to worry.

What other foods increase intestinal permeability? Do they do it more or less than wheat? Has anyone been investigating this? Are there common things that affect intestinal permeability a thousand times more than wheat does, such that everything done by wheat is totally irrelevant in comparison?

Do people without autoimmune diseases suffer any danger from increased intestinal permeability? How much? Is it enough to offset the many known benefits of eating wheat (to be discussed later?) Fiber seems to decrease intestinal permeability and most people get their fiber from bread; would decreasing bread consumption make leaky gut even worse?

I find this topic really interesting, but in a “I hope they do more research” sort of way, not an “I shall never eat bread ever again” sort of way.

Digestion Of Wheat Produces Opiates, Which Get You Addicted To Wheat – Probably false, but just true enough to be weird

Dr. William Davis, a cardiologist, most famously makes this claim in his book Wheat Belly. He says that gliadin (a component of gluten) gets digested into opiates, chemicals similar to morphine and heroin with a variety of bioactive effects. This makes you addicted to food in general and wheat in particular, the same way you would get addicted to morphine or heroin. This is why people are getting fat nowadays – they’re eating not because they’re hungry, but because they’re addicted. He notes that drugs that block opiates make people want wheat less.

Does Wheat Make Us Fat And Sick, a review published in the Journal of Cereal Science (they have journals for everything nowadays) is a good rebuttal to some of Davis’ claims and a good pro-wheat resource in general.

They say that although gliadin does digest into opiates, those opiates are seven unit peptides and so too big to be absorbed from the gut to the bloodstream.

(note that having opiates in your gut isn’t a great idea either since there are lots of nerves there controlling digestion that can be affected by these drugs)

But I’m not sure this statement about absorption is even true. First, large proteins can sometimes make it into the gut. Second, if all that leaky gut syndrome stuff above is right, maybe the gut is unusually permeable after wheat consumption. Third, there have been sporadically reported cases of gliadin-derived opiates found in the urine, which implied they got absorbed somehow.

There’s a better counterargument on the blog The Curious Coconut. She notes that there’s no evidence these peptides can cross the blood-brain barrier, a precondition for having any psychological effects. And although the opiate-blocker naloxone does decrease appetite, this effect is not preferential for wheat, and probably more related to the fact that opiates are the way the brain reminds itself it’s enjoying itself (so that opiate-blocked people can’t enjoy eating as much).

And then there’s the usual absence of qualifiers. Lots of things are “chemically related” to other chemicals without having the same effect; are gliadin-derived opiates addictive? Are they produced in quantities high enough to be relevant in real life? Corn, spinach, and maybe meat can all get digested into opiates – is there any evidence wheat-derived opiates are worse? This is really sketchy.

The most convincing counterargument is that as far as anyone can tell, wheat makes people eat less, not more:

Prospective studies suggest that weight gain and increases in abdominal adiposity over time are lower in people who consume more whole grains. Analyses of the Physicians’ Health Study (27) and the Nurses’ Health Study (26) showed that those who consumed more whole grain foods consistently weighed less than those who consumed fewer whole grain foods at each follow-up period of the study. Koh-Banerjee et al. (27) estimated that for every 40-g increase in daily whole grain intake, the 8-y weight gain was lower by 1.1 kg.

I’ll discuss this in more detail later, but it does seem like a nail in the coffin for the “people eat too much because they’re addicted to wheat” theory.

Still, who would have thought that wheat being digested into opiates was even a little true?

Wheat Something Something Something Autism And Schizophrenia – Definitely weird

Since gluten-free diets get tried for everything, and everything gets tried for autism, it was overdetermined that people would try gluten-free diets for autism.

All three of the issues mentioned above – immune reactivity to gluten, leaky guts, and gliadin-derived opiates – have been suggested as mechanisms for why gluten free diets might be useful in autism.

Of studies that have investigated, a review found that seven reported positive results, four negative results, and two mixed results – but that all of the studies involved were terrible and the ones that were slightly less terrible seemed to be more negative. The authors described this as evidence against gluten-free diets for autism, although someone with the opposite bias could have equally well looked at the same review and described it as supportive.

However, a very large epidemiological study found (popular article, study abstract) that people with antibodies to gluten had three times the incidence of autism spectrum disease than people without, and that the antibodies preceded the development of the condition.

Also, those wheat-derived opioids from the last section – as well as milk-derived opioids called casomorphins – seem to be detected at much higher rates in autistic people.

Both of these factors may have less to do with wheat in particular and more to do with some general dysregulation of peptide metabolism in autism. If for some reason the gut kept throwing peptides into the body inappropriately, this would disrupt neurodevelopment, lead to more peptides in the urine, and give the immune system more chance to react to gluten.

The most important thing to remember here is that it would be really wrong to say wheat might be “the cause” of autism. Most likely people do not improve on gluten-free diets. While there’s room to argue that people might have picked up a small signal of them improving a little, the idea that this totally removes the condition is right out. If we were doing this same study with celiac disease, we wouldn’t be wasting our time with marginally significant results. Besides, we know autism is multifactorial, and we know it probably begins in utero.

Schizophrenia right now is in a similar place. Schizophrenics are five to seven times more likely to have anti-gliadin antibodies as the general population. We can come up with all sorts of weird confounders – maybe antipsychotic medications increase gut permeability? – but that’s a really strong result. And schizophrenics have frank celiac disease at five to ten times the rate of the general population. Furthermore, a certain subset of schizophrenics sees a dramatic reduction in symptoms when put on a strict gluten-free diet (this is psychiatrically useless, both because we don’t know which subset, and because given how much trouble we have getting schizophrenics to swallow one lousy pill every morning, the chance we can get them to stick to a gluten-free diet is basically nil). And like those with autism, schizophrenics show increased levels of weird peptides in their urine.

But a lot of patients with schizophrenia don’t have reactions to gluten, a lot don’t improve on a gluten free diet, and other studies question the research showing that any of them at all do.

The situation here looks a lot like autism – a complex multifactorial process that probably isn’t caused by gluten but where we see interesting things going on in the vague territory of gluten/celiac/immune response/gut permeability/peptides, with goodness only knows which ones come first and which are causal.

Wheat Has Been Genetically Modified Recently In Ways That Make It Much Worse For You – Probably true, especially if genetically modified means “not genetically modified” and “recently” means “nine thousand years ago”

If you want to blame the “obesity epidemic” or “autism epidemic” or any other epidemic on wheat, at some point you have to deal with people eating wheat for nine thousand years and not getting epidemics of these things. Dr. Davis and other wheat opponents have turned to claims that wheat has been “genetically modified” in ways that improve crop yield but also make it more dangerous. Is this true?

Wheat has not been genetically modified in the classic sense, the one where mad scientists with a god complex inject genes from jellyfish into wheat and all of a sudden your bread has tentacles and every time you try to eat it it stings you. But it has been modified in the same way as all of our livestock, crops, and domestic pets – by selective breeding. Modern agricultural wheat doesn’t look much like its ancient wild ancestors.

The Journal Of Cereal Science folk don’t seem to think this is terribly relevant. They say:

Gliadins are present in all wheat lines and in related wild species. In addition, seeds of certain ancient types of tetraploid wheat have even greater amounts of total gliadin than modern accessions…There is no evidence that selective breeding has resulted in detrimental effects on the nutritional properties or health benefits of the wheat grain, with the exception that the dilution of other components with starch occurs in modern high yielding lines (starch comprising about 80% of the grain dry weight). Selection for high protein content has been carried out for bread making, with modern bread making varieties generally containing about 1–2% more protein (on a grain dry weight basis) than varieties bred for livestock feed when grown under the same conditions. However, this genetically determined difference in protein content is less than can be achieved by application of nitrogen fertilizer. We consider that statements made in the book of Davis, as well as in related interviews, cannot be substantiated based on published scientific studies.

In support of this proposition, in the test tube ancient grains were just as bad for celiac patients’ immune systems as modern ones.

And yet in one double-blind randomized-controlled trial, people with irritable bowel syndrome felt better on a diet of ancient grains than modern ones (p < 0.0001); and in another, people on an ancient grain diet had lower inflammatory markers and generally better nutritional parameters than people on a modern grain one. Isn’t that interesting?

Even though it’s a little bit weird and I don’t think anyone understands the exact nutrients at work, sure, let’s give this one to the ancient grain people.

The Lectins In Wheat Interfere With Leptin Receptors, Making People Leptin Resistant And Therefore Obese – Currently at “mere assertion” level until I hear some evidence

So here’s the argument. Your brain has receptors for the hormone leptin, which tells you when to stop eating. But “lectin” sounds a lot like “leptin”, and this confuses the receptors, so they give up and tell you to just eat as much as you want.

Okay, this probably isn’t the real argument. But even though a lot of wheat opponents cite the heck out of this theory, the only presentation of evidence I can find is Jonsson et al (2005), which points out that there are a lot of diseases of civilization, they seem to revolve around leptin, something common to civilization must be causing them, and maybe that thing could be lectin.

But civilization actually contains more things than a certain class of proteins found in grains! There’s poor evidence of lectin actually interfering with the leptin receptor in humans. The only piece of evidence they provide is a nonsignificant trend toward more cardiovascular disease in people who eat more whole grains in one study, and as we will see, that is wildly contradicted by all other studies.

This one does not impress me much.

Wheat Is Actually Super Good For You And You Should Have It All The Time – Probably more evidence than the other claims on this list

Before I mention any evidence, let me tell you what we’re going to find.

We’re going to find very, very many large studies finding conclusively that whole grains are great in a lot of different ways.

And we’re not going to know whether it’s at all applicable to the current question.

Pretty much all these studies show that people with some high level of “whole grain consumption” are much healthier than people with some lower level of same. That sounds impressive.

But what none of these studies are going to do a good job ruling out is that whole grain is just funging against refined grain which is even worse. Like maybe the people who report low whole grain consumption are eating lots of refined grain, and so more total grain, and the high-whole-grain-consumption people are actually eating less grain total.

They’re also not going to rule out the universal problem that if something is widely known to be healthy (like eating whole grains) then the same health-conscious people who exercise and eat lots of vegetables will start doing it, so when we find that the people doing it are healthier, for all we know it’s just that the people doing it are exercising and eating vegetables.

That having been said, eating lots of whole grain decreases BMI, metabolic risk factors, fasting insulin, and body weight (1, 2, 3, 4,5.)

The American Society For Nutrition Symposium says:

Several mechanisms have been suggested to explain why whole grain intake may play a role in body weight management. Fiber content of whole grain foods may influence food volume and energy density, gastric emptying, and glycemic response. Whole grains has also been proposed to play an important role in promoting satiety; individuals who eat more whole grain foods may eat less because they feel satisfied with less food. Some studies comparing feelings of fullness or actual food intake after ingestion of certain whole grains, such as barley, oats, buckwheat, or quinoa, compared with refined grain controls indicated a trend toward increased satiety with whole grains. These data are in accordance with analyses determining the satiety index of a large number of foods, which showed that the satiety index of traditional white bread was lower than that of whole grain breads. However, in general, these satiety studies have not observed a reduction in energy intake; hence, further research is needed to better understand the satiety effects of whole grains and their impact on weight management.

Whole grains, in some studies, have also been observed to lower the glycemic and insulin responses, affect hunger hormones, and reduce subsequent food intake in adults. Ingestion of specific whole grains has been shown to influence hormones that affect appetite and fullness, such as ghrelin, peptide YY, glucose-dependent insulinotropic polypeptide, glucagon-like peptide 1, and cholecystokinin. Whole grain foods with fiber, such as wheat bran or functional doses of high molecular weight β-glucans, compared with lower fiber or refined counterparts have been observed to alter gastric emptying rates. Although it is likely that whole grains and dietary fiber may have similar effects on satiety, fullness, and energy intake, further research is needed to elucidate how, and to what degree, short-term satiety influences body weight in all age groups.

Differences in particle size of whole grain foods may have an effect on satiety, glycemic response, and other metabolic and biochemical (leptin, insulin, etc.) responses. Additionally, whole grains have been suggested to have prebiotic effects. For example, the presence of oligosaccharides, RS, and other fermentable carbohydrates may increase the number of fecal bifidobacteria and lactobacilli (49), thus potentially increasing the SCFA production and thereby potentially altering the metabolic and physiological responses that affect body weight regulation.

In summary, the current evidence among a predominantly Caucasian population suggests that consuming 3 or more servings of whole grains per day is associated with lower BMI, lower abdominal adiposity, and trends toward lower weight gain over time. However, intervention studies have been inconsistent regarding weight loss

The studies that combined whole and refined grains are notably fewer. But Dietary Intake Of Whole And Refined Grain Breakfast Cereals And Weight Gain In Men finds that among 18,000 male doctors, those who ate breakfast cereal (regardless of whether it was whole and refined) were less likely to become overweight several years later than those who did not (p = 0.01). A book with many international studies report several that find a health benefit of whole grains, several that find a health benefit of all grains (Swedes who ate more grains had lower abdominal obesity; Greeks who ate a grain-rich diet were less likely to become obese; Koreans who ate a “Westernized” bread-and-dairy diet were less likely to have abdominal obesity) and no studies that showed any positive association between grains and obesity, whether whole or refined.

I cannot find good interventional trials on what happens when a population replaces non-grain with grain.

On the other hand, Dr. Davis and his book Wheat Belly claim:

Typically, people who say goodbye to wheat lose a pound a day for the first 10 days. Weight loss then slows to yield 25-30 pounds over the subsequent 3-6 months (differing depending on body size, quality of diet at the start, male vs. female, etc.)

Recall that people who are wheat-free consume, on average, 400 calories less per day and are not driven by the 90-120 minute cycle of hunger that is common to wheat. It means you eat when you are hungry and you eat less. It means a breakfast of 3 eggs with green peppers and sundried tomatoes, olive oil, and mozzarella cheese for breakfast at 7 am and you’re not hungry until 1 pm. That’s an entirely different experience than the shredded wheat cereal in skim milk at 7 am, hungry for a snack at 9 am, hungry again at 11 am, counting the minutes until lunch. Eat lunch at noon, sleepy by 2 pm, etc. All of this goes away by banning wheat from the diet, provided the lost calories are replaced with real healthy foods.”

Needless to say, he has no studies supporting this assertion. But the weird thing is, his message board is full of people who report having exactly this experience, my friends who have gone paleo have reported exactly this experience, and when I experimented with it, I had pretty much exactly this experience. Even the blogger from whom I took some of the strongest evidence criticizing Davis says she had exactly this experience.

The first and most likely explanation is that anecdotal evidence sucks and we should shut the hell up. Are there other, less satisfying explanations?

Maybe completely removing wheat from the diet has a nonlinear effect relative to cutting down on it? For example, in celiac disease there is no such thing as “partially gluten free” – if you have any gluten at all, your disease comes back in full force. This probably wouldn’t explain Dr. Davis’ observation – neither I nor my other wheatless-experimentation friends were as scrupulous as a celiac would have to be. But maybe there’s a nonlinear discrepancy between people who have 75% the wheat of a normal person and 10% the wheat of a normal person?

Maybe there’s an effect where people who like wheat but remove it from the diet are eating things they don’t like, and so eat less of them? But people who don’t like wheat like other stuff, and so eat lots of that?

Maybe wheat in those studies is totally 100% a confounder for whether people are generally healthy and follow their doctor’s advice, and the rest of the doctor’s advice is really good but the wheat itself is terrible?

Maybe cutting out wheat has really positive short-term effects, but neutral to negative long-term effects?

Maybe as usual in these sorts of situations, the simplest explanation is best.

Final Thoughts

Non-celiac gluten sensitivity is clearly a real thing. It seems to produce irritable bowel type symptoms. If you have irritable bowel type symptoms, it might be worth trying a gluten-free diet for a while. But the excellent evidence for its existence doesn’t seem to carry over to the normal population who don’t experience bowel symptoms.

What these people have are vague strands of evidence. Something seems to be going on with autism and schizophrenia – but most people don’t have autism or schizophrenia. The intestinal barrier seems to become more permeable with possible implications for autoimmune diseases – but most people don’t have autoimmune disease. Some bad things seem to happen in rats and test tubes – but most people aren’t rats or test tubes.

You’d have to want to take a position of maximum caution – wheat seems to do all these things, and even though none of them in particular obviously hurt me directly, all of them together make it look like the body just doesn’t do very well with this substance, and probably other ways the body doesn’t do very well with this substance will turn up, and some of them probably affect me.

There’s honor in a position of maximum caution, especially in a field as confusing as nutrition. It would not surprise me if the leaky gut connection turned into something very big that had general implications for, for example, mental health. And then people who ate grain might regret it.

But stack that up against the pro-wheat studies. None of them are great, but they mostly do something the anti-wheat studies don’t: show direct effect on things that are important to you. Most people don’t have autism or schizophrenia, but most people do have to worry about cardiovascular disease. We do have medium-term data that wheat doesn’t cause cancer, or increase obesity, or contribute to diabetes, or any of that stuff, and at this point solely based on the empirical data it seems much more likely to help with those things than hurt.

I hope the role of intestinal permeability in autoimmune disease gets the attention it deserves – and when it does, I might have to change my mind. I hope people stop being jerks about gluten sensitivity, admit it exists, and find better ways to deal with it. And if people find that eliminating bread from their diet makes them feel better or lose weight faster, cool.

But as far as I can tell the best evidence is on the pro-wheat side of things for most people at most times.

[EDIT: An especially good summary of the anti-wheat position is 6 Ways Wheat Can Destroy Your Health. An especially good pro-wheat summary is Does Wheat Make Us Fat And Sick?]

Let’s get the obvious out of the way first. The title, dust jacket, and introduction are all doing their best to make it sound like a sketchy diet fad book. Even its proponents admit it does a very poor job at not sounding like a sketchy diet fad book. I don’t know how much importance to give this. Maybe any diet advice, including very scientifically based and sober advice, inevitably ends up looking like a sketchy fad book because that is the most effective way to market it? Let’s just raise the red flag count to “one” and move on.

The best word for my opinions on the content would be “mixed”.

Let’s start with the good. It clearly tries to be science-based. It mostly succeeds. It is evolutionarily informed without being silly about it. It ties a lot of research together into a consistent and conclusive whole.

The overall diet it recommends is…pretty good, actually. Sort of similar to the paleo diet, with its emphasis on unprocessed whole meats and vegetables, and an acceptance bordering on fetish for saturated fat. But it adds the concept of a “safe starch” – where the classic paleo diet has a zero tolerance policy toward carbohydrates, the Perfect Health Diet allows certain carbs like rice, potatoes, and lots of weird things like “taro” and “sago”.

I really like this change. Hunter gatherers do eat certain roots and tubers, so it’s not like an evolutionary perspective has to exclude them entirely. And the Japanese remain some of the healthiest people in the world while eating rice by the bucketload, so anyone who tells you that rice is bad for you has got something like a hundred million contrary data points to overcome. Pre-contact Polynesians were also exceptionally healthy, so all the weird Polynesian starches they recommend have my seal of approval as well.

This makes their diet much easier than normal paleo diets. With a moderate amount of work, you can find a lot of the breadstuffs you like made with rice or potato or something rather than wheat – my local Asian food store has rice noodles that can almost pass for spaghetti to my unsophisticated palate. And since I have been trained not to feel full without some carbs in me, having a small plate of rice with a meal really helps.

I am slightly skeptical about admitting potatoes to the same pantheon as rice and Weird Polynesian Starches. This comes from my four years studying medicine in Ireland. The traditional Irish diet is beef, lamb, butter, cabbage, and potato, which is about as Perfect Health Diet compliant as it’s possible to get (yeah, okay, they do eat bread and drink beer, but other than that). The Irish can be very insistent on sticking to just these couple of things. I remember working at a rural GP office and staying at the home of an elderly Irish couple. I made a trip into town and brought back some spaghetti to make for dinner, and my hosts were impressed that I was making them “exotic foreign foods”. By which they meant spaghetti. This was 2010.

Anyway, despite this rigid diet elderly Irish people seemed pretty unhealthy and cardiovascular-disease-ridden, maybe not much better than the US. If we’re sticking Polynesians and Japanese on one side, and Americans on the other, the Irish are over with us. I don’t know if the Perfect Health people want to blame all of that on their bread and alcohol and whatever ungodly oils they fry stuff in, but I am going to view potatoes with suspicion until this gets cleared up.

The other thing about the book I really liked was their discussion of “toxins”. Normally whenever any diet book discusses “toxins” I throw it aside with great force, because they mean “mysterious evil spirits in foods I don’t like which can be banished only with juice cleanses.” Perfect Health Diet, despite using the same language, singles out very specific chemicals and cites all the reasons which each chemical can be expected to be harmful.

The part that stuck with me most here was their section on omega-6 fats, which might be literally the Devil. I find their speculation on omega-6s and obesity intriguing, not to mention the connection to crime I have already blogged about. Stephan Guyenet seems to take this seriously as well, which is always a good sign. And it makes evolutionary-historical sense that this random stuff we’ve been adding to foods in massive quantities over the past couple of decades and not before is screwing us up.

Also wheat. So, so much wheat. I haven’t had time to investigate this much so far, but a lot of smart people seem very worried about wheat and corn and all those other grains, and if one tenth of the darts they throw at them stick, we should probably start looking for a new staple food for Western civilization.

So much for the good. Let’s move on to the bad.

This book has a very consistent pattern of making a strong claim, citing one study that supports their claim, and failing to mention that there have been many more much larger studies since then that have come to the opposite conclusion.

For example, they mentioned every single study to say breathlessly excited things about Vitamin D for the past fifty years, but failed to mention that all of them have since been disconfirmed by larger and more careful trials.

And it’s not just Vitamin D. Vitamin K prevents prostate cancer! Vitamin K prevents liver cancer! Vitamin K reduces all kinds of cancer by 75%! Vitamin K heals cancer in people who already have the disease! A recommendation: take lots of Vitamin K!

But look at how the American Cancer Society describes these same trials:

A large European epidemiologic study published in 2008 found higher risk of prostate cancer in men with low intake of some forms of vitamin K. But people who take in plenty of vitamin K usually get it from fruits and vegetables. This means that people who get lots of vitamin K from foods are more likely to take in many other vitamins and phytochemicals that can affect cancer risk. In some cases, vitamin K may be given credit for benefits that are actually due to other compounds that appear in the same foods. So studies like this, which only observe large groups of people, may not be very useful for looking at the effects of vitamin K intake.

A small clinical trial from Japan suggested that vitamin K lowers the risk of developing liver cancer among women with cirrhosis due to hepatitis C. A later study also seemed to show some effect, but it was not statistically significant. Further study is needed.

There have been some studies examining whether menadione (vitamin K3) can help overcome cancer cells’ resistance to certain types of chemotherapy drugs. Results in lab animals and cell cultures are mixed, but there is no evidence available of significant effects in humans yet.

A small Phase I clinical trial in California tested different doses of intravenous vitamin K3 in people with advanced cancer. The patients did not improve. In that study, several patients also had allergic reactions, especially at higher doses.

Look. I know alternative medicine is hard to do. It’s also useful. Sometimes they come up with genuinely interesting stuff the mainstream has missed. So I’m not demanding that you not speculate about Vitamin K just because the mainstream medical community has interpreted the results as unimpressive. Go ahead and try to find reasons why the earlier successful trials were more believable than the later more pessimistic trials. But at least mention that everyone else disagrees with you about this. Don’t present it as some well-established thing which the mainstream community is just completely ignoring.

We repeat this exact same frickin’ pattern with just about every vitamin and mineral ever discovered. Selenium prevents gastric cancer! Selenium reduces tumor incidence! A randomized controlled trial of selenium in China decreased liver cancer! A trial of selenium in America decreased skin cancer! Selenium improves immune function! Selenium decreases all-cause mortality by a quarter! Selenium cuts cardiovascular mortality in half! Selenium cures all diseases! Selenium will make you rich and famous! Women will want to date you! Men will want to be you! You will live in a beautiful house on the beach, with a huge yacht made entirely of selenium! You will sail it over seas of liquid selenium, to selenium islands where selenium-based life-forms play haunting music beautiful beyond anything you have ever imagined upon their selenium flutes. If you take enough selenium you will never die, yet if by chance you should, you will go to Selenium Heaven, which is to regular Heaven what a supernova is to a firefly.

Meanwhile, look at any respectable medical source like Cochrane Review, and the actual list of effects that they find for selenium is exactly zero. The process is the same as always. Some preliminary exciting studies come out. So we do much bigger, more rigorous studies and the effects go away. I did a really thorough evaluation of the selenium literature for my Quantified Health Prize entry and although I was totally happy to disagree with expert advice in several areas, in the case of selenium I found that the experts were exactly right and it is useless. Perfect Health Diet doesn’t even mention the expert consensus or rigorous studies here.

This leads to this weird pattern where Perfect Health Diet manages to explain about 1500% of human mortality. That is, if you add up all of their claims (“Vitamin Q decreases mortality by 50%!” “Vitamin R decreases mortality by 75%!”) then although I haven’t exactly done the calculations my guess is they would add up to around 1500%. This isn’t quite mathematically impossible – if there are fifteen vitamins, any single one of them prevents cancer completely, and you can just choose which one you want – then added together they would decrease cancer mortality by 1500%. But even though it’s not mathematically impossible, it doesn’t seem right to me and it seems much more likely that they use hugely inflated claims than that fifteen different interventions all cut mortality even more than any of the things we know cut mortality like exercise or quitting smoking.

I’ll skip over the T3/T4 thyroid thing and a few others that will be more of the same, and just spare a second to complain about folic acid: they say supplementing pregnancies with folic acid causes autism. This is not only not proven, but there’s much stronger evidence for the opposite: multiple good studies show folic acid supplemention decreases autism risk by as much as 40%. Perfect Health Diet’s only source for the opposite claim was a correlational study in, of all places, the Journal of Medical Hypotheses, which specifically exists to signal-boost claims that don’t really have any support but need further investigation.

Now, to be fair to PHD, they did briefly mention this counterexample. But their scare study was in bold and exaggerated, and the existence of a counterexample was in parentheses and had the word “slightly” appended to the front even though it was not slight at all.

So. Alternative medicine people? Please listen very closely. STOP SAYING THAT VITAL CHILDRENS’ HEALTH INTERVENTIONS CAUSE AUTISM. THIS. NEVER. HELPS.

Overall I thought their entire section on vitamin and mineral supplementation was a disaster area. This is a little worrying because I know more about vitamins and minerals than I do about macronutrients and diet in general, and the Gell-Mann Amnesia Effect argues that if the parts you know something about are bad, you should assume the parts you don’t know enough to criticize are as well.

I will, however, give praise where praise is due. PHD gets that multivitamins aren’t a great idea and does not recommend them. It gets that calcium is not a great idea and doesn’t recommend it either. And they have praise for iodine and lithium, which are the two minerals I went out on a limb to recommend in my QHP entry and which I really do think deserve more attention than they’re getting.

In conclusion, PHD is not a terrible book and you could do much worse as far as diet books go. It is written by smart people who are familiar with at least some of the literature. It is a good idea-generation source to learn about things you might not have otherwise known (did you know there is some evidence a chronic bacterial infection can be one cause of Alzheimers Disease?) and the diet it recommends will probably not kill you and in fact probably be good for you. If their stuff about wheat and omega-6 and the like pans out, and I would not be surprised if it does, it might be really good for you. I don’t have any complaints there.

(actually, I tried the diet they recommended for about two weeks, just to see what would happen, and I lost seven pounds without really trying. This is a sufficiently large number of pounds that I find this result implausible and am hesitant to pronounce a “This is what happens when you follow Perfect Health Diet”. But still, nice.)

But I also cannot wholeheartedly recommend it, especially for epistemically-minded people who want something correct as opposed to something full of nice diet advice. It has a lot of shoddy research and presents extremely controversial positions as if they are settled science. There are very few mistakes that will kill anyone (they even couched their folic acid advice with an incongruous yet welcome “but you should still take a little folic acid during pregnancy”) but a lot of mistakes that will leave you very confused or have you taking very silly amounts of vitamins.

This book might be suitable for someone who wants to try a paleo-inspired diet for purely practical reasons, but who needs some encouragement and inspiration through a discussion of the science behind it. It might also be good for someone who already knows a lot about health from a mainstream point of view, is sufficiently skeptical to catch the mistakes, and wants to see what sorts of stuff the paleo diet and alternative-medicine-which-is-smarter-and-more-sciency-than-homeopathy people are about so they can grab the useful insights and discard the rest.

I was originally sent this book because I dissed saturated fat, so I should tie up my review there – especially since it makes a good metaphor for the book in general. Perfect Health Diet gives a lot of good evolutionary reasons why saturated fat might be good for you, and reflects the growing consensus that yeah, we kind of went overboard turning it into demon-food and should take a step back from there – something most people don’t realize and that this book could be very helpful in teaching.

But then it goes on to say that the old story that saturated fats led to heart disease has been debunked, citing as the only discussion for this claim a link to a systematic review that did in fact find saturated fats not to correlate with heart disease. What the book doesn’t mention is that this study was funded by the National Dairy Council, or that nine out of eleven other systematic reviews that have investigated the issue did find a saturated fat/heart disease link, or that Cochrane Review, which is the gold standard for systematic reviews of medical evidence, concluded that on balance there is good evidence for a saturated fat-heart disease link. All they do is cite the increasingly-lonely one that found the conclusions they wanted, declare that to be the state of the evidence, and move on.

So that is PHD in a nutshell. Interesting points, broadly reasonable dietary advice, and total denial of the existence of any evidence that doesn’t support their claims.