Tuesday, December 13, 2005

Is wheat killing us?



Maybe you've heard someone say, "Well, in Asian countries they eat low fat diets and they don't have many heart attacks." This is mainly, mostly true. But interestingly, "The China Study", a recently published book by T. Colin Campbell, a leading researcher into diet and disease in China has one chapter about heart disease and it doesn't even MENTION China. How can that be? In a nutshell, Dr. Campbell is a vegan propagandist that will string together any number of flimsy correlations to support his cause, but not even he is bold enough to try to make it seem like there is a link between heart disease and animal food consumption - meat, milk and eggs - in China. Clearly, there is not.

I am fascinated by "The China Health Study". The whole thing was started in the early 1970's when the Chinese premier, Chou EnLai, was dying of cancer. He initiated a vast study correlating disease rates in China with different geographical areas, culminating in the publishing of a disease atlas of China. Later, the authors of The Study collected dietary and blood data from 65 mostly rural counties in China and correlated them with the data in the disease atlas. Then they printed the data in book form so that you can, for instance, look up myocardial infarction ( heart attack) rates, and see what foods are eaten in areas of China where people are prone to heart attacks. It is a truly epic study and I think it is far more interesting than most Western health studies because of certain factors in China:

1) The people in rural China eat what they grow, so the majority of people in the study have presumably eaten the same diet from birth to death. This contrasts with American diets which tend to change with the dietary breeze.

2) Due to wide variations in climate and therefore which crops can be grown, dietary staples differ by region. Rice is the staple in Southern China, wheat, corn and millet are staples in the middle latitudes and wheat and dairy predominate in the far North. There are pig-raising areas and sweet-potato raising areas. Therefore, we can see correlations of disease patterns among dramatically different diets.

3) There are dramatic differences in disease rates in different areas in China. The disease rates from region to region differ far more than they do in the US.

4) The authors published all of the tabulated data in a "monograph" titled, "Diet, Life-style and Mortality in China." Very few studies give you this type of unfettered access to the raw data. Usually, you only get the authors conclusions from their take on the raw data. You can probably find a copy and have a look yourself if you're REALLY interested.

The Meat of It

Unfortunately, we now have to talk about statistics for a minute. The data in the monograph is published with correlation scores between each pair of factors (like myocardial infarction and fat consumption). The scores range from -100 to 100. A score of 0 means that the two factors are completely unrelated, whereas a cc of 100 represents perfect correlation and a cc of -100 is perfect negative correlation. Come again? OK, so a cc of 100 between fish and myocardial infarction (MI) would mean that anyone who's ever tasted fish has died of a heart attack. That would be considered a "strong positive correlation". A cc of -100 would mean that no one who's tasted fish ever died of a heart attack. That's a "strong negative correlation". And a cc of 0 means that, from the perspective of MI, it made no difference whatsoever whether or not you've eaten fish (no correlation). Of course, in the real world the correlations tend to be a little more ambiguous.

OK, enough teasing. Here are the relevant correlations between dietary factors and MI:


Factors Correlated to Heart Attack Risk (MI)
Factor Correlation Score Significance
Meat Consumption -28 none
Dairy Consumption 6 none
Rice Consumption -58 Strong Negative Correlation
Wheat Consumption 67 Strong Positive Correlation
Other Grain Consumption 39 Moderate Positive Correlation
Green Vegetable Consumption 5 none
Blood Cholesterol 4 none

Whoah!! Did you catch that? I'll state it as plainly as I can:

The main dietary predictor of heart disease rates in China is the TYPE of grain you eat.

Rice eaters seem protected from heart attacks while wheat, corn and millet eaters are much more prone (the corn link is debatable). Meat, dairy and vegetable consumption play no obvious role. Blood cholesterol plays no role.

But we're not going to believe these statistics are we? Let's look at the raw data.

The first graph simply shows the raw correlation between eating grains other than rice and dying of a heart attack in China. There is a clear positive trend here. The counties with the nine highest rates of heart disease all eat more than 350 grams per day of grain other than rice. The majority of counties where more than 350 g/day of non-rice grain are eaten have a MI death rate of more than 7 per 100,000 per year whereas the majority of rice eating areas have heart disease rates of less than 4 per 100,000 per year.

A problem with any study of this magnitude is that there are so many variables. These variables are very, very tricky to tease apart, but I'll give it a go. In some areas of China, corn and millet are widely eaten. Since the link between wheat and heart disease is particularly strong, it would be interesting to see the link between rice and wheat consumption and heart disease in areas where corn and millet are not eaten. The following graph shows the relationship between wheat consumption and heart disease only in counties where they eat less than 100 grams per day of corn and millet.

Wow, no question about the correlation here!

I'm also trying to tease out the effect of fat consumption. Notice that there are three data series in the graphs. The blue diamonds represent areas where total fat consumption is less than 40 grams per day. The purple squares represent areas where fat consumption is more than 40 grams per day. If fat consumption plays a role in heart disease risk, you would expect to see the purple squares a little higher than the blue diamonds, on average. But as you can see, this is not the case, suggesting that it doesn't matter how much fat is eaten; in China, the more wheat you eat the greater your risk of heart disease, regardless of fat consumption.

The final data point, the yellow triangle, is Tuoli county, where 186 grams of fat are eaten per day, mostly from dairy and meat. If dietary fat plays a role in heart disease risk in China, you would expect that point to be elevated far above the others. It is not. Which brings me to my next topic.

The Forgotten Ones

There are populations in China in and around the steppe whose diets are more American than here in the good ole US of A. They eat wheat, dairy products and meat. And nothing else. You can read about some of them here, but I'll give you this quote as a prelude:

"The vegetables were, incidentally, an accommodation to the visiting lowlander. Self-respecting nomads don't eat plants; it's just not their place on the food chain."

I don't think the people studied in the China Health Study are actually nomadic, but they eat far more saturated fat than we do in America - most of it from dairy and meat. Yet, as you can see by the location of the yellow triangle in the graph, their rates of heart disease seem to be determined solely by how much wheat they eat.

Why do I refer to them as the forgotten ones? In this interview, Dr. campbell had this to say, "Firstly, the kind of diet that was consumed in China is very low in total fat. The range was between 6-24% of calories, for example, whereas in this country the range is more of the order of the high 20s-high 40s. " Actually, the fat content of the diet of the people of Tuoli, according to Dr. Campbell's study, was 45%. Has he forgotten them? My guess is that he simply prefers not to think about them. They don't support his views, you see.

Here's a table comparing three wheat eating areas and the Chinese averages. I have picked three areas with extreme diets to prove my point. The first is Longxian county, the area of China with the highest wheat consumption and below average meat and vegetable consumption, a very tedious diet by American standards. The next is Tuoli, where wheat is eaten along with large portions of meat and dairy but no vegetables. This area represents the stereotypical american diet. The final is Jingxing, an essentially vegan area where they eat a variety of grains other than rice and gorge themselves on green vegetables. Jingxing had by far the highest consumption of green vegetables of any county in this study. I suppose it represents the ideal vegan diet of low fat consumption with lots of vegetables and grains (except rice).

Three Chinese Counties
  Longxian Tuoli Jingxing Chinese Median
Rice Cons (g/day) 0 3.9 6.1 398
Wheat (g/day) 629 372 126 22.4
Other Grain (g/day) 75.1 6.8 342.1 1.1
Meat (g/day) 6.1 121 0 20.3
Dairy (g/day) 0 857 0 0
Green Veg (g/day) 36.3 0 691 76.8
MI/CHD 18.7 5.53 11.9 3.16
Stroke 61.2 34.3 42.3 25.2

As you can see, Tuoli county had the lowest levels of heart disease and stroke (which is also correlated with wheat consumption) of the three. The three counties followed the general trend:

In China, the more non-rice grain you eat, the higher your risk of heart disease (and stroke). Eating meat, dairy and vegetables don't affect your risk.


Correlation does not imply causation. What this study does NOT tell us is that wheat categorically causes heart disease. It tells us that there is something different about wheat eating people in China that predisposes them to heart disease, which may or not be the wheat itself. For instance, in the study the number of times per year than one ate green vegetables had a strong negative correlation with heart disease risk while the number of times per year that one ate carrots had a strong positive correlation correlation. Should we conclude from this that green vegetables prevent heart disease while carrots cause it? No. First off, the actual amounts of carrots or green vegetables consumed have no correlation with heart disease risk. These two parameters are actually telling us about the length of the growing season in an area. In the South, green vegetables can be grown for more of the year, while in the North carrots have to be grown since they can be kept in a root cellar during the longer winter. Rice growing is confined to the warmer South. So the reason that there is a strong correlation between how often you eat carrots and heart disease risk is that carrots are eaten in wheat growing areas. Wheat is the REAL correlation.

But if you want to show that wheat is really the cause of elevated heart disease risk and not just a marker for something else, you have to look a little deeper. And in China, there are a whole bunch of differences between wheat and rice eating peoples. Wheat eating people are taller and heavier. They tend to have higher levels of blood triglycerides, associated with increased risk of heart disease in the US, and sex hormone-binding globulin(shbg), which the authors of the study claim is a sensitive indicator of insulin resistance. Cholesterol levels were the same in both populations. Wheat eaters had lower levels of monounsaturated fats (like in olive oil) in their blood cell membranes and higher levels of omega 6 fats (like in soybean oil).

So a reasonable possibility, based on the triglyceride and shbg data is that eating wheat and certain other grains leads to insulin resistance, which is associated with increased heart disease risk. But the really interesting thing is why that should be. Why does eating wheat, but not rice, seemingly lead to insulin resistance in China? I don't know. That, my friends, is the million dollar question that we should be asking but aren't.

Where is T. Colin Campbell on this? He doesn't even MENTION the fact in his book! It's not like he didn't notice. His name is on the paper about it from 1996. The only reasonable explanation is that Dr. Campbell is more interested in pushing his agenda than on doing good science.

East and West

Might the difference between heart disease levels in Asia and Western countries be explained by the fact that Western peoples eat more wheat and Asian peoples eat more rice? Maybe partially. It's important to note that even in wheat eating areas of China, heart disease levels were still below those of Western countries.

If the causal link does indeed turn out to be insulin resistance, then there is one other major dietary difference that ought to be noted. In China, virtually no sugar is eaten, so there is nowhere in this study where the carbohydrate intake is similar to the US. In the US, we eat about the same amount of white wheat flour and sugar (or high fructose corn syrup) and not much else as far as carbs. In China they eat rice or wheat or maybe rice and millet or corn and wheat but never wheat and sugar. What if they did?


In China, the main predictor of heart disease rates in a given population is how much wheat flour (and other grains except rice) that population eats. The consumption of vegetables or animal products doesn't play an obvious role in heart disease rates. Tuoli county, where they eat far more saturated fat than in the US, had far less heart attack deaths than the US and no more heart attack deaths than you would suspect based on the amount of wheat they consume compared to their Chinese Colleagues.

A Few Notes

A couple more notes. Sweet potatoes, widely eaten in China, had a neutral effect on heart disease risk.

Apparently, in China, they eat mainly white flour and white rice, so this probably isn't a whole grain versus processed grain issue, although you could argue that the wheat, being ground into flour, is more processed than the whole white rice.

The study didn't publish any figures on physical activity, but it is clear that the Chinese get more physical activity and are less overweight than their Western counterparts. This is interesting in that it may help explain the overall differences between Asia and Europe, but I don't think it affects the relationship between wheat or meat and heart disease. I have no reason to believe that rice eaters got more physical activity than wheat eaters or that meat eaters had more physical activities than non meat eaters.

Some notes about comparing CHD rates in China and Europe/US. The statistics in this paper are done in a much different manner than are the CHD statistics available from the British Hearty Foundation, so the numbers in this article and "The Axis of Paradox" are not comparable. Not being a statistician, I'm not really sure how these CHD rates compare to the ones in Europe, but I'll trust Dr. Campbell on this issue that the areas with the highest CHD levels in China are still lower than European levels. One probable reason for that is that the Chinese life expectancy is significantly shorter than the European life expectancy and the leading causes of death in this study were infectious diseases such as tuberculosis. But you have to die of something, so presumably if infectious diseases in China were cured, life expectancy rates would increase and more people would die of heart attacks.

The Chinese in this study died much more frequently of strokes than heart attacks, which is common in Asian countries. In Asia it is considered honorable to die of a stroke but dishonorable to die of a heart attack, so the bias of the doctors when the cause of death is in doubt in to call it a stroke rather than a heart attack. Presumably the bias of Western doctors is the opposite. I have no idea if there is a real difference where Asians have more strokes and Westerners have more heart attacks (I presume there is), but it's interesting food for thought.