"Wheat Belly is an anti-scientific" - discussion about gluten, phytic acid and pseudoscience

(This is an archive of discussion on Reddit.)

You can think of Wheat Belly as an anti-scientific source. I once had to go through the book page-by-page because a friend of mine adopted it like religion. Every single thing in the book, if you actually look up and read the cited papers, is either for tangential, irrelevant, or directly contradictory material. I generally take it as a rule of thumb that when someone takes the effort to write a book chock-full of citations and can't find a single one to support their argument, it's because there aren't any. DrArkades

How about this?

http://authoritynutrition.com/6-ways-wheat-can-destroy-your-health/

http://authoritynutrition.com/why-is-bread-bad-for-you/

"The phytic acid blocks absorption of minerals like iron, zinc and calcium"

"Wheat fiber may cause your body to burn through its Vitamin D stores much faster and contribute to vitamin d deficiency (20), which is associated with cancer, diabetes and death (21, 22, 23)."

"Bread is High in Carbs and Can Spike Blood Sugar Levels"

"Wheat Consumption is Associated With Several Brain Diseases"

"The digestive products of gluten may be able to stimulate opioid receptors and contribute to addiction. However, this needs to be studied more before we can make any conclusions."

It's not a scientific source when you click on their citations and their citations read as the opposite of what they claim. For instance, the first thing I went looking for was the phytic acid line.

From the page you linked:

Bread contains other harmful substances

Most grains also include the “anti nutrient” phytic acid.

Phytic acid is a molecule that strongly binds essential minerals like >calcium, iron and zinc, preventing them from being absorbed (16).

Soaking grains before baking can degrade the phytic acid, which should improve the availability of minerals.

Wow! It makes it sound as though phytic acid is harmful, and as though we really need those "essential" minerals like calcium, iron, and zinc in the bread.

Of course, first-run pass through that sentence my brain says, wait a second, since when do I use bread as a primary source of iron, calcium, or zinc? I don't, actually.

Then I click on the citation link.

We find an abstract noting that phytic acid is, indeed, a chelating agent. Specifically, a chelating agent that keeps us from absorbing zinc, calcium, and iron. Is that a bad thing? Well, we don't really get those minerals primarily from bread, and the first thing the actual article notes is (A) the probable protective effects (lowering serum cholesterol, reducing exposure of colon tissue to iron), and (B) the fact that actual consequences in humans haven't been assessed.

So, what does the link actually show? Phytic acid may be beneficial, we don't really know for sure, but its primary effect is keeping you from fully absorbing nutrients that you don't really need (we get iron and zinc from meats; calcium from dairy - it's in bread, generally, because the bread has been fortified).

As opposed to its presentation on that page as a "Harmful Substance" whose effect is depriving you of something "essential."

I don't trust authors who use this as their version of "citing a source." It's just on this line of outright lying.DrArkades

Of course humans get their calcium, iron and zinc from other sources (not bread). But they usually consume vegetables and meat along with bread or other food that contains wheat (and other sources of Phytic acid).

Therefore, in the stomach (and intestines) vegetables and meat are mixed with phytic acid.

So, you claim that phytic acid is not interfering with absorbtion of nutrients from non-wheat sources (when we consume them with bread or other food that cointan wheat - like pasta)?

Also:

Phytic acid’s “chelating” ability is considered by some to be a detriment to one’s health. On the other hand, many researches embrace this ability to bind with minerals as its most powerful asset. In her book, Diet for the Atomic Age, Sara Shannon, lists 11 nutrients in particular that protect against heavy metal toxicity and radiation damage. Phytates bind with radioactive and toxic substances and carry them out of the body.

The author claim that phytic acid is protecting against "heavy metal toxicity and radiation damage". Obviously "heavy metals" are not in wheat. They are from other soruces that humans eat along with bread or other sources of phytic acid.

Of course humans get their calcium, iron and zinc from other sources (not bread). But they usually consume vegetables and meat along with bread or other food that contains wheat (and other sources of Phytic acid). Therefore, in the stomach (and intestines) vegetables and meat are mixed with phytic acid. So, you claim that phytic acid is not interfering with absorbtion of nutrients from non-wheat sources (when we consume them with bread or other food that cointan wheat - like pasta)?

Most of this is way off-base. I'm going to address this one directly, though, because it's worth mentioning on general principle:

Things don't exist as essences. They exist in quantities. For example, when discussing phytic acid, there's usually about 12 mg phytic acid / gram of bread after baking.

In your average serving size of bread - 50 grams - that's 60 mg phytate. For chemical stoichiometry, let's convert that into moles: .09 moles of phytate.

Now, 50g bread has about 1.0 mg zinc, 2 mg iron, 24 mg calcium. Converting that into moles, .01 moles of zinc, .03 moles iron, 0.6 moles calcium. This adds up to .64 moles of minerals - and just these minerals, never mind anything else the bread may contain that might be chelated by phytate. The phytate chelates out .09 moles of these .64: 14% of the iron, magnesium, and calcium, if all of the phytate grabs a ligand.

IOW, there's not enough phytic acid to chelate out the contents of a sub-set of the bread, never mind anything else you're eating. In a diet that's already over-rich in a number of minerals, because you're generally eating fortified foods.

Actual science deals with numbers.DrArkades

Thank you for the explaination.

I understand that there is no enough phytic acid. But what about other claims? Brain damage (because of Vitamin D deficiency), high carbs (can spike blood sugar levels), brain deseases, the digestive products of gluten may be able to stimulate opioid receptors and contribute to addiction...

Is there any real disadvantage of eating food that contain wheat?

As far as I know humans should eat vegetables, meat (not only muscle tissue because this can lead to vitamin B12 deficiency), fish, fruits, eggs.

Is anything wrong with this diet?

Also I am reading that milk (from other animals) is not good for humans and may increase risk of some types of cancer and iron deficiency (and also is very harmful for small children). Is this real?

I found this about milk:
Iron deficiency anemia - children

Cow's milk is a common cause of iron deficiency. It contains less iron than many other foods and also makes it more difficult for the body to absorb iron from other foods. Cow's milk also can cause the intestines to lose small amounts of blood.

Consumption of cow's milk as a cause of iron deficiency in infants and toddlers.
Looks like "real science".
Also:
The Pros and Cons of Milk and Dairy

The Case Against Milk and Dairy
Not everyone thinks that was a good idea. Indeed, experts at the Harvard School of Public Health have labeled the milk recommendations a “step in the wrong direction.” One the most prominent critics is Walter Willett, MD, PhD, professor of epidemiology and head of the nutrition department at the Harvard School of Public Health.
“One of the main arguments for USDA recommendations is that drinking milk or equivalent dairy products will reduce the risk of fractures. But in fact there’s very little evidence that milk consumption is associated with reduced fractures,” Willett tells WebMD.
Indeed, countries in which almost no milk is consumed, such as many Asian countries, have low rates of fractures, he points out.
But Willett’s chief worry is that drinking too much milk may pose dangers. “By now there’s quite a body of data showing a higher risk of fatal prostate cancer associated with milk,” he tells WebMD. “And though the evidence is somewhat mixed, we’ve still seen a slightly higher risk of ovarian cancer associated with drinking three or more servings of milk.”

Hormones in milk can be dangerous

I'm not entirely sure how to answer this in a way that is thorough and meaningful without braindumping my entire undergraduate and graduate educations on you. I'll do my best, though.

Aside from the big picture "eat stuff your ancestors from a century ago would recognize as food; include more fruits and vegetables" most studies on the effect of diet on health are untrustworthy. When we're designing experiments, we need to worry about something called confounding. Here's a hypothetical: imagine a study in the 1970s on smoking + health in women. They followed a group of smokers and non-smokers for a decade and compared death rates. They found that smokers actually tended to live longer than non-smokers, by a wide margin.

But, when you break the two groups down by age grouping, we find non-smokers have a lower death rate. Why? At the time, smoking was getting fashionable among young women. So smoking tended to correlate with youth, non-smoking with old age. Old age is an independent predictor of death. So, here, where we didn't evenly distribute "age" among both groups, we got a biased result.

This is why scientists use randomly assigned interventions for experimentation. If I completely randomly assign people to experimental groups, then on average, all confounders will distribute evenly. This includes known confounders (age, SES) and unknown confounders (some previously unknown gene that is an independent predictor of our outcome of interest.) RCTs often publish a table showing major known confounders being evenly distributed, as an indicator that, indeed, randomization was successful. (This isn't actually logical from a math standpoint, but whatevs. Point is, randomization is super important.)

Not every question lends itself to a randomized intervention, though. I can choose to randomly assign people to Drug A or placebo if I have no real proof Drug A works (and in fact, so many drugs fail in testing that it's a pretty safe bet Drug A doesn't). I can't always assign people to random interventions, though. If I have good, but not complete, evidence to believe a toxin is fatal, for instance, I can't randomly assign half of a group to be exposed to it. It's unethical. Sometimes it's the sort of thing you just can't really control (how far people live from place X; components of diet over very long times, etc.)

In that case, you go with observational trials. You distribute people into groups according to known confounders, as best you can, and then compare them based on exposures. Theoretically, this approach can work. In practice, however, to address every known confounder, and combinations of confounders, requires absurdly enormous trials, and even then doesn't account for unknown confounders. The result is that in observational studies are kind of worthless. On any topic where observational studies preceded eventual clinical trials, it ends up looking something like this: 65% of observational studies say X. 80% of Clinical Trials say Not X.

The effects of these biases swamp true effects from the variables under investigation. Now, not always. What can sometimes hold true is if you have a huge effect. Generally, if the relative risk is less than 3 (increases risk less than 3-fold), chuck it in the trash. The study design just isn't that powerful. If it's greater than 3 - or closer to the lower 20s, like smoking and lung cancer in men - then you might have a real effect.

Dietary studies are usually observational studies with relative risks between 0.5 and 2 (reduce risk of something by 50%, increase risk of something 100%). Which is to say, you can chuck in the bin the vast majority of nutritional studies that manage to become headlines. The bright side about relatively small effect sizes is that they actually have very little impact on your health: that's the effect we're talking about.

Multiple testing. If you have a .18 chance of winning a roll of the dice, but roll it 6 times, you'll likely win once. If a study on, whatever, wheat, starts digging through charts for every possible variable and combination of variables they'll usually find upwards of 100 to test (I know, I've done them). They then don't mention in their publication the number of cast-offs. You're supposed to use a multiple-testing adjustment (if I test 100 variables, I should make my statistical cut-off 100x more strict). You'll notice in the methods section that multiple testing adjustments are rarely done. In proper RCT's, as monitored by the FDA for drug testing, end-points (the variables you're looking at) are supposed to be submitted to the FDA before the trial ever gets underway (so you can't just cherry-pick results), and you'd better believe their statisticians look for multiple testing adjustments.

Fundamental assumptions. Most of these dietary studies ultimately pop these things into a multiple regression. So you'll see something like Beta = .3 on VitD for Cancer ~ VitD + BMI + AGE + SES + etc. Which is to say, for every "unit" of Vitamin D intake (defined by the way the study's built), assuming all else is held constant, chance of cancer rises 30%. Here's the problem: (1) Not all else is held constant. You'd better fucking believe there's a relationship between these variables. (2) By far more important, think of all the variables VitD intake has a relationship to - like, various elements of your physiology - that change with changes in VitD intake. (3) That 30% number only works for the range of values studied: you really can't extrapolate outside the bounds of a linear regression, although people do it constantly.

So, dietary changes - barring the big picture (eat real food and clean water, focus on fruits and veg, and not too much) - tend to have small effect sizes that aren't well studied, and are swamped by confounders in the sorts of studies we tend to run.

In other words: we tend to know a lot less about diet than we like to think, and most of it doesn't matter.
Luckily, when something gets heavily studied, we also study it in the laboratory. Which means we learn about it chemically and biochemically. We can tie it into the web of knowledge we have of existing nutrition, physiology, anatomy, chemistry, etc.

So, we tend to know wheat is OK for most people. It may have some really subtle effects we don't see, but we can not give a shit, because they're so subtle that if you think you see them, you're probably seeing random variation and attributing it to wheat. It certainly doesn't have the effects being claimed because those are major enough to have been ruled out. It's also unlikely to, on first principles, because it's been a major component of human diet for a lot longer than other elements of human diet that are less important, but that we've generally evolved tolerances for (e.g., caffeine). Some things evolve slowly, but central components of human survival (ability to tolerate key nutritional sources) are not one of them. There are some people that get a disease related to wheat. These people are not so common. They're far, far rarer than people in anti-wheat discussions claim. They're far, far rarer than self-diagnosed. These people do have an actual problem with wheat. For that matter, I have an actual problem with fiber. That's not to say that problem generalizes to the rest of the population: you shouldn't be afraid of fiber just because I have an illness that gives me problems with it.

As for milk, I haven't seen any data to indicate it's harmful. I have seen it started to crop up as the next insidious threat on a lot of quack sites, though, so I imagine I'm going to have to start getting as familiar with that as I have with the wheat topic. All I can say is: I don't see it coming up anywhere other than quack sites. If they're not responding to actual science (if they were, I'd've seen that), it's a fair bet it's another thing being pulled out of their asses.

Seriously: the public health literature in the last 50 years has pretty solidly shown the biggest risks to health are low socioeconomic status and lack of exercise. (Bigger picture: fucked up public health infrastructure). Worry about those things. Don't worry about small effects from shitty studies.

Food doesn't matter as much as people think, beyond "have fruits and vegetables, don't eat too much". It tends to be a topic of heated discussion because, historically, it's something we can control, and makes us feel like we're in control of our lives. It's a safety-blanket, not a major health determinant.DrArkades