Thursday, September 2, 2010

The China Study on Wheat

Denise Minger has just put up another great China Study post that's worth reading if you haven't already. Denise has been busy applying her statistics skills to the mountain of data the study collected. She noted in a previous post that wheat intake was strongly associated with coronary heart disease (CHD), the quintessential modern cardiovascular disease. I, and several other people, requested that she work her mathmagic to see if the association could be due to some other factor. For example, wheat is eaten mostly in the Northern regions of China, and CHD rates are generally higher at higher latitudes (vitamin D insufficiency?). This is true in Europe as well, and may be partly responsible for the purported benefits of the Mediterranean diet. You can mathematically determine if the association between wheat and CHD is simply due to the fact that wheat eaters live further North.

To make a long story short, nothing could explain the association except wheat itself, even latitude. Furthermore, she found a strong association between wheat intake and body mass index, typically a predictor of fat mass although we can't say that for sure. That finding echos a previous study in China where wheat eaters were more likely to be overweight than rice eaters (1, 2). Head over to Denise's post for the full story.

The China Study has major limitations built into its basic design, due to the fact that it was observational and pooled the blood samples of many individuals. Therefore, its findings can never prove anything, they can only suggest or be consistent with hypotheses. However, the study also has some unique advantages, such as a diversity of diets and regions, and the fact that people had presumably been eating a similar diet for a long time. I feel that Denise's efforts are really teasing out some useful information from the study that have been de-emphasized by other investigators.

There has been very little serious investigation into the health effects of wheat in the general population. Researchers studying celiac disease and other forms of gluten allergy, and the efforts of the paleolithic diet community in spreading that information (for example, Loren Cordain and Pedro Bastos), have been major contributors to understanding the health effects of wheat. Denise's analysis is one of the strongest pieces of evidence I've come by so far. I think there's enough indirect evidence that investigators should begin taking the idea seriously that wheat, particularly in the form of industrial flour products, may contribute to chronic disease in more than just a small subset of the population.


Unknown said...

I wonder if history books will ever refer to wheat as one of the biggest mistakes that human civilization has ever made.

Will we ever learn? Will the USDA ever stop trying to convince us to eat 8 servings of grain per day?

Ruth Almon said...

How do other grains like rye compare to wheat and rice?

Michael Barker said...

If we can't kill the lipid hypothesis I doubt if wheat will be going away anytime soon.

Unknown said...


At this point it seems that grains fall into two categories: gluten grains and non-gluten grains.

Rye, like wheat and barley, contain gluten.

Grains without gluten, such as rice, buckwheat, and corn, seem to be less problematic, although still far from ideal food.

Gyan said...

I believe wheat has always been eaten along with some saturated fat.
Perhaps the wheat toxicity emerges when diet lacks sufficient saturated fat?
Also rancid flour could be a factor too?

Can these considerations be evaluated using China Study data?

Emily said...

thanks again for another thought provoking post!

Derek H said...

Is the problem traditional yeast and quick leavening? I tend to agree with Stephen that natural fermentation and the sourdough process with the removal of the bran and germ, tend to make it a less harsh food for healthy individuals. Again, it's probably not the most optimal food.

Innovator said...

Porcupine, BizJ, Gyan, and Derek H,

May I refer everyone to the comments on Stephan's previous post, which point out that the mostly unnoticed villain in wheat and other grains may well be the lethal combination of fructose and fructans.

Gluten is no doubt a factor too, but the far stronger correlation -- cutting across all foods and all populations of industrial "dieters" -- appears to be a) the proportion of fructose to glucose in the food and b) the limit of fructose (25g) consumed per sitting or fructans (7.5-10g) per day.

David Csonka said...

Wheat makes a lot of people rich, so I doubt it will be going away any time soon.

Hoggan said...

In addition to the immunological reactions to wheat caused in those with celiac disease, resulting in higher cancer rates and more early mortality, those with non-celiac gluten sensitivity have been shown to have a greater risk of malignancy and premature death than celiac patients (Anderson et al). One group of patients with schizophrenia was shown to have a dramatically increased rate of immune reactions to gluten that are dissimilar to those seen in celiac disease. Wheat germ agglutinin is a very small molecule that can bypass the blood brain barrier and the intestinal barrier.

Although whole grains may reduce some of the above negative impacts, they are also higher in anti-nutrients, such as phytates, which create other health problems.

Gluten grains are not a very healthful food for humans.

Best Wishes,
Ron Hoggan, Ed. D.
co-author: Dangerous Grains I
author: The Iron Edge
author: Smarten Up! I
editor: Journal of Gluten Sensitivity
editor: Cereal Killers

Stephan Guyenet said...

Hi Porcupine,

In my opinion, rye is a better grain for several reasons, one of which is its lower gluten content. Still, it does contain gluten so it has the potential to be problematic.

Hi JA Deep,

I doubt the fructans are the reason why wheat is problematic. Wheat is relatively low in fructans compared to other grains, and fructans have been shown to have a number of health benefits in rodents and humans.

Hi Ron,

Nice to see you on the blog and thanks for the comment. I read your book Dangerous Grains and reviewed it a while back:

Colldén said...

It's funny though that the french, arguably the least unhealthy western population, are also some of the biggest consumers of wheat in the world, while the US, easily among the most unhealthy, has a relatively low wheat consumption.

Innovator said...

Hi Stephan,

I'm urging the community here to look at fructose/fructans with fresh eyes, because the data is relatively recent and -- unlike celiac disease -- is not yet widely known even by the nutritionists here. For example, functional gut disorders commonly exist in association with untreated celiac disease (a common presentation) and inflammatory bowel disease
(IBD) where such symptoms are two-to-threefold more common than in the community at large.Cite at pg. 52, fn 3-4.

For clarity, the theory is not that fructans in isolation are a problem. In fact, primitive societies survived on fructans in deserts when there was little else to eat. Cite. It is the lethal combination of excessive fructose to glucose and the total load of fructans that correlates, in affluent diets, to fat gain and other gastointestinal disorders.

True, grains other than wheat may be higher in fructans. But affluent societies don't eat as much of these other grains as they do wheat, or enough to create the necessary fructans load (12.5g per day).

Finally, any study that purports to show a health benefit to fructans should be reviewed carefully for conflict of interest -- because the sale of fructans is big business, and dangerous to consumers.

Unknown said...

Colldén, the French don't eat more wheat than other Europeans. The highest per capita supply via FAOstat in Europe is Albania, Romania, Italy, Greece, Ukraine, and Malta.

Raphael Gonçalves Nicésio said...

Hello, I'm from Brazil.
I found your site and wanted to ask, please, for that added the link to my blog for disclosure in order to be partners.
Already added your on my list of partners, ok?
My blog is this:
Thanks and good luck!

Colldén said...


According to this table (

France consumes significantly more wheat than all the nations you listed with the exception of Ukraine (and possibly Malta). They consume almost 40% more wheat than the UK (284 000 vs 204 000 tonnes per million), yet have like a third of their obesity rate. They consume more than twice as much as the US, who have even poorer health in comparison.

gallier2 said...

@Colldén just for info, the french consume lot of wheat but much less than the production/import would suggest. French are extremely wasteful with bread as normal white bread (baguette) is very quickly stale. I think that 30% of the bread production is wasted that way.

Unknown said...

Pretty insightful post. Never thought that it was this simple after all. I had spent a good deal of my time looking for someone to explain this subject clearly and you’re the only one that ever did that. Kudos to you! Keep it up

hire limo warsaw

Colldén said...


No doubt, but I would think the same goes for most western countries.

Peter said...

I've heard that it's the opposite in India. In the north, where they eat wheat, the heart attack rate is far smaller than in the south, where they eat rice.

Anonymous said...

I came across this review, , listing trends in food consumption in th 20th century. Grain consumption was much higher in the days before the widespread epidemic diseases of civilization. I couldn't find anything about wheat consumption specifically, but the US was a wheat exporter even in the early 19th century, suggesting that we've always had enough for our own population, plus a surplus.

Clearly something has changed in our health since then, and food trends are probably very important factors, but there are plenty of healthier populations eating a lot of grain. In terms of eating habits, sugar consumption is much higher than it used to be, eggs are lower, seed oils are MUCH higher, butter/lard are lower, milk is lower, cheese is higher, chicken/beef are higher.

Wheat may be very important in the decline in health, but I tend to think for the population at large (not including those with clear pathologic immunological responses), wheat may not be necessary or sufficient to account for the DOC, just one of many factors that may cause problems, alone or in combination.

It seems that eating a pre-industrial diet, and possibly a pre-agricultural diet, makes the most sense--that way there's no need to identify the smoking gun. (And then there are the innumerable ways our lives differ that aren't related to macro food choices: additives, chemicals, altered nutrient profiles in foods, ?stress, etc.)

Anna said...

Wheat has changed both as a grain and how it is transformed into edible food, particularly since the mid 20th century.

There are now many many more hybridized varieties of wheat that conform better to mechanized production and industrial processing requirements that are very different from the wheat our ancestors ate and we don't know if these newer hybrids present risks to health.

The preparation of wheat grain into edible foods has changed profoundly, esp where Western-style diets prevail. Most industrially produced wheat products are quickly produced (in hours) from wheat that may be pre-ground and rancid, with fast acting yeasts, and other additives to increase profitability, palatability, and shelf-life. There's good reason to question if wheat foods prepared with industrial processing affects health differently than traditional preparations (that may take days to complete) such as sprouting, sourdough fermentation, freshly ground flour, fewer additives, etc.

Unknown said...

Regarding India, I have heard before that Indian vegetarians have one of the highest rates of heart disease in the world, whereas Indians who eat meat do not. Is it possible that northern India eats a different diet, including more meat, that may be a confounding factor?

Is vegetarianism more popular in southern India? I haven't been able to find any data.

Unknown said...

"Is vegetarianism more popular in southern India?"

Ha, yes! Southern Indians are usually vegetarians and love vegetarian food fried in seed oil particularly.

IMHO Seed oil= heart disease and obesity. Wheat = autoimmune disorders. Wheat + seed oil = OMGWTF.

The French might not be as likely to keel over from a heart attack as we are, but incidence of things like allergies is quite similar.

Maureen Diaz said...

In an article I wrote several months ago I attempt examine several potential causes for all of this gluten/wheat/grain sensitivity, from a lay person's point of view. Might this explain some of the wheat connection you and Denise are speculating about:

Alan said...

Don't the French consume most of there wheat as white flour products (baquette, croissant) which is much lower in gluten. Also their diets in general are alot healthier than in America which helps compensate for the wheat consumption. Someone else said, maybe you need to consume good fats with the wheat. The French also tend to eat freshly made bread and not the type that gets preserved to last for two weeks.

Gyan said...

It is not true that South Indian more vegetarians relative to North Indians.
Fish-eating for example is very low in North.

What is true is that healthy North Indians eat a lot of ghee.

Heather Twist said...

Thanks for that! An earlier blog posted something about the same thing, but this is much more recent.

It brings to mind what I read constantly in Price's work: that "white flour and sugar" were the culprits in the diseases he saw. Mind you he thought the problem was the "white" part, but whole grain flour just wasn't shipped around the world back then (or now: it goes rancid way too fast. But whole wheat berries *are* shipped these days, and they are just as devastating as white flour is).

I was reading an encyclopedia from 1909, and there was a whole article on grain. In that article, the authors attributed the "energy" of the West to the fact that the West ate wheat, not rice. Rice, it said, just didn't have the vital energy to get people motivated to do stuff. (my paraphrase: their wording was far more flowery).

I'm fairly sure that when history looks back at all this, wheat will be one of the huge factors that caused the uniqueness (and problems) of Europe and the US. Parts of China do have wheat, but it's nowhere near as high in gluten as the wheat that was invented in the US and Canada.

Fredrik Gyllensten said...

Very interesting, I don't consider wheat to be the devil, but there are definitely better 'staple' foods out there.

Stephan Guyenet said...

Hi Everyone,

Great discussion. I think it's clear from numerous examples that wheat isn't the sole factor in the diseases of civilization. North vs. South India is a good example, and so is the US for that matter (we ate more flour 100 years ago, if you believe the statistics). It's also important to note that the incidence of heart attacks in the China study was low compared to the US even in the wheat-eating areas.

I think wheat is one factor that can contribute to certain diseases of civilization, but probably can't do it single-handedly, at least in the context of an otherwise good diet.

Another thing to keep in mind is that the wheat we eat today is not the same as the wheat we ate even 150 years ago, as Anna mentioned. Our modern bread flour varieties have been selected for a high gluten content, precisely because that's what makes good bread. Whereas wheat used to be fermented with a natural sourdough starter, which partially degrades gluten and makes the flour overall more nutritious, now it's fermented with isolated quick-rise yeast strains that are better than nothing but inferior to sourdough. So bread today has considerably more gluten than it used to.

Flour is also processed differently. Today's flour is both more refined and more processed than it was 150 years ago. Even though they had white flour 150 years ago, it wasn't as thoroughly refined. Today, flour is typically treated with bromine gas ("bromated"), which improves its baking qualities and chemically alters the flour. It's also bleached, another chemical modification that is done for aesthetic purposes but may make it more toxic.

You can buy unbleached, unbromated white flour and make sourdough yourself, which I suspect is a lot healthier than buying packaged white bread at the grocery store.

Unknown said...

This is such a great discussion, and I feel so privileged to even be in contact with such informed and knowledgeable people.

Especially Stephan. Thank you for doing what you do; this website is a gold mine of useful information for people who care about health. You are doing an amazing service.

If you ever decide to write a book, Stephan, I'll buy it.

Dr. Curmudgeon Gee said...

thanks for the update on China Study vs. wheat.

yes, please write a summary of the "evil" wheat with references. most of my associates still think wheat is wholesome. & i don't have enough "hat" to convince them.

although i suggested one colleague to "eat less wheat" who's struggling with few pounds for years. she happily reported of loosing 3 lb "effortlessly" in a month.



Jeevita said...

Thanks for your blog. I recently discovered it and enjoy reading it. I am originally from South India where people are being encouraged to eat wheat instead of rice by doctors and popular media because it is 'healthier'. We have also replaced sesame oil and peanut oil with sunflower oil based on the same sources. Ironical! I have only now started trying to cook our traditional recipes without so called 'healthier' substitutions.

Anonymous said...

I think some investigation needs to be done on the Italian (and French) thing. The Italians are the 3rd longest lived nation on earth behind Iceland and Japan despite having white flour products (pasta and bread) as their staple starch.

Do they have bad teeth or bones or high levels of other diseases of civilization?

I know they place great value on local and fresh produce and that they are passionate about good food, mostly always making things like sauces etc from scratch unlike here where people buy jars of cheap sugary sauces. We also know that they eat plenty of good quality vegetables, cheese, meat and fish.

Maybe its sugar that has been the main culprit all along, hurting our bodies and robbing our minerals. Maybe white flour is benign, not giving us any nutrition (just calories) but also not robbing us of it either.

Author Gabriel Land said...

I think white flour is benign pretty much. Gluten is easily broken down. Surely it is much easier to digest than animal proteins are. Especially if it has been leavened or boiled. I am eating pastries a lot lately and feel as healthy as I have ever been, and certainly much more healthy than when I was trying to avoid carbs.

Tiffany said...

Excellent comments. I don't believe that wheat is the boogie man itself, but possibly the changes in wheat over the past 50 years especially. Some research that I have found might offer some clues.

First of all, Murphy et al published in Euphytica (2008)163:381-390 "Relationship between yield and mineral nutrient concentrations in historical and modern spring wheat cultivars" that in regards to calcium, copper, iron, magnesium, manganese, phosphorus,selenium and zinc, that "while grain yield has increased over time, the concentrations of all minerals except calcium have decreased". Murphy then goes on to compare how many slices of bread would be required to meet RDAs of each mineral historically versus currently. Not surprisingly, the amount of slices required today is significantly higher.

Now combine this with Fan et all who submitted "Evidence of decreasing mineral density in wheat grain over the last 160 years" (Journal of Trace Elements in Medicine and Biology 22 (2008) 315-324) and stated that with the "Green Revolution" of the 60s came semi-dwarf cultivars of wheat that were high yielding, yet were 20-30% lower in essential minerals zinc, iron, copper and magnesium. In addition, as phytate levels did not decrease, the ratio of minerals to phytate actually went down meaning that the minerals in the wheat were probably lower in bioavailabilty in the modern cultivars than in the historical cultivar. (phytate binds to minerals in the gut rendering them unusable by our body). Possibly the new cultivars with their increased yields and decreased minerals, has left current strains of wheat not able to provide the minerals that historically they used to. Thus, obesity and a number of other deficiency issues could be the result.

Just my two cents worth. . .I am not a scientist so possibly I am not reading things correctly. Would recommend going to the sources I quoted for further clarification.

Anonymous said...

"You can mathematically determine if the association between wheat and CHD is simply due to the fact that wheat eaters live further North [...] may be partly responsible for the purported benefits of the Mediterranean diet" How does this theory account for the fact that for example Italians have a higher consumption of wheat-based pasta than anyone else on the European peninsula? All while still maintaining better cardiac profile than western Europe.

MikeBelanger said...

Forgive me if I'm wrong, I don't see exercise anywhere in the statistics. As far as cardiovascular disease is concerned, doesn't exercise ( burning the carbs ) help as well?

As far as statistics from long ago, most people in the pre-industrialized farmed in an un-mechanized way. Farming probably supplied enough exercise to make wheat a detail in overall health.

Bringing in the French - don't most of them bike? I can't find a statistic, but I've heard average French citizen spends way less time in a car, and more time walking and biking than the average American.

Unknown said...

Quite contrary to what I know...I always thought rice make the body gain weight and is better to have wheat products and even made a shift of diet to brown rice.Nice to know of latest research developments and analysis.You blog rocks.
Diet Plans

Dave said...

Denise's critique of The China Study is based on her interpretation of raw unadjusted univariate correlations. She makes a serious attempt to adjust for confounding variables but she doesn't take them all into account. However she also fails to take into account several other factors that are even more important than confounding variables. These are:

1) The range of values for saturated fat and animal protein is too limited to detect what their effect on heart disease and cancer would be in the US where on average we eat about 10 times as much animal protein and about 3 times as much saturated fat.

Here's an example to illustrate the point. If every county has 0% smokers the correlation with lung cancer and smoking would be zero of course. But if every county has 100% smokers the correlation with lung cancer and smoking would still be zero. Even though everybody knows that smoking causes lung cancer this would be impossible to detect with statistical analysis unless you compare groups that have a different percentage of smokers. And the greater the difference the easier it would be to see the effect.

The same is true for the kind of animal protein and saturated fat most Americans eat. So if you compare 10 counties in China with 10 states from the U.S. (using US averages with equal up and down variation) the correlation between animal protein and heart disease is over 0.900 and the same is true for saturated fat. For all cancers the r value is over .500 for animal protein and over .700 for saturated fat.

2) The fact that nutritional variables have an optimal range and a U-shaped curve for benefits. This means that if the average value for a variable in rural China is on a different point of the U-shaped curve than the average value for that same variable in the US, then in the US we could get the exact opposite correlation or no correlation at all.

Here's an example to illustrate the point. In China, when comparing counties eating almost no animal protein (who after adjusting for PDCAAS would be getting less than the RDA of protein), to those eating a few additional grams, we see a reduction in heart disease for the people getting more protein. But using data from page 74 of The China Study, that's only 3.6% of total calories for a person weighing 143 pounds and eating 2671 calories a day.

In Dr. Campbell's experiments with rats, as long as the casein was less than 10% of total calories the correlation between casein and cancer was close to zero. But once that threshold was crossed, the more casein the more cancer.

That result was paralleled in the recent study we have been discussing. Compared to the controls the group eating 12% animal protein had a hazard ratio for cancer of 0.96. But the group eating 18% protein had a hazard ratio of 1.28.

So depending on the amount of animal protein you can get a strong positive, strong negative or close to zero correlation.

3) Heart disease and cancer have several independent risk factors that have been established by professional researchers and if animal protein or saturated fat are greatly reduced in a population their r value will also be greatly reduced and the r value for the others risk factors will be greatly magnified and they would become confounders that would have to adjusted for.

Dr. Campbell also made the same mistake of trying to justify his conclusions with correlations that are only comparing counties in China with low levels of animal food. Denise correctly points out that Dr. Campbell's effort to do this contains many errors. If this was the only evidence he relied on, he would totally fail to prove his point. But in the book, Dr. Campbell relied on other evidence which included comparing the US with rural China, other epidemiological studies, plausible biological mechanisms as well as clinical trials.

Stephan Guyenet said...

Hi David,

You said "Denise's critique of The China Study is based on her interpretation of raw unadjusted univariate correlations. She makes a serious attempt to adjust for confounding variables but she doesn't take them all into account."

I don't understand what you mean here. Denise adjusted for variables that made a significant contribution to the associations, which is all you can do given the data set.

I agree with your point that associations in the China study may not reflect associations in other cultures, due to dramatically different nutrient intakes. I suspect Denise would agree with that as well. I don't recall her trying to prove that the China study associations apply to other cultures.

Your example of comparing 10 counties in China to 10 counties in the US does not in any way support the hypothesis that animal protein or fat cause heart disease. You could find the same association with TV, boxer briefs, English speaking, or any other variable you wanted to choose that is common in the US but not in China. As a whole, the best individual-level prospective studies we have to date indicate very clearly that there is no association between SFA consumption and heart disease in Western populations with 2-3 fold different SFA intake levels, when other dietary factors are adjusted for.

You said "in the book, Dr. Campbell relied on other evidence which included comparing the US with rural China, other epidemiological studies, plausible biological mechanisms as well as clinical trials."

In my opinion, Campbell has selectively cited and stretched the evidence to make it fit his hypothesis, and the data as a whole do not support his conclusions. For example, see this post on Campbell's claim that animal protein causes cancer: