Authors: T. Colin Campbell,Thomas M. Campbell
THE CHINA STUDY
72
Because China is relatively homogenous genetically, it was clear that
these differences had to be explained by environmental causes. This
raised a number of critical questions:
• Why was cancer so high in some rural Chinese counties and not in
others?
• Why were these differences so incredibly large?
• Why was overall cancer, in the aggregate, less common in China
than in the U.5.?
The more Dr. Chen and I talked, the more we wished that we had
a snapshot in time of the dietary and environmental conditions in ru-
ral China. If only we could look into these people's lives, note what
they eat, how they live, what is in their blood and their urine and how
they die. If only we could construct a picture of their experience with
unprecedented clarity and detail so that we could study it for years to
come. If we could do that, we might be able to offer some answers to
our "why" questions.
Occasionally science, politics and financing come together in a way
that allows a truly extraordinary study to take place. This happened for
us, and we got the opportunity to do everything we wanted, and more.
We were able to create the most comprehensive snapshot of diet, life-
style and disease ever taken.
PULLING IT TOGETHER
We assembled a world-class scientific team. There was Dr. Chen, who
was the deputy director of the most significant government diet and
health research laboratory in all of China. We enlisted Dr. Junyao Li,
one of the authors of the Cancer Atlas Survey and a key scientist in Chi-
na's Academy of Medical Sciences in the Ministry of Health. The third
member was Richard Peto of Oxford University. Considered one of the
premier epidemiologists in the world, Peto has since been knighted and
has received several awards for cancer research. I rounded out the team
as the Project Director.
Everything was coming together. It was to be the first major research
project between China and the United States. We cleared the necessary
funding hurdles, weathering both CIA intrusiveness and Chinese gov-
e r n m e n t reticence. We were on our way.
We decided to make the study as comprehensive as possible. From
the Cancer Atlas, we had access to disease mortality rates on more than
73
LESSONS FROM CHINA
four dozen different kinds of disease, including individual cancers,
heart diseases and infectious diseases. 6 We gathered data on 367 vari-
ables and then compared each variable with every other variable. We
went into sixty-five counties across China and administered question-
naires and blood tests on 6,500 adults. We took urine samples, directly
measured everything families ate over a three-day period and analyzed
food samples from marketplaces around the country.
The sixty-five counties selected for the study were located in rural
to semi-rural parts of China. This was intentionally done because we
wanted to study people who mostly lived and ate food in the same area
for most of their lives. This was a successful strategy, as we were to learn
than an average of 90-94% of the adult subjects in each county still
lived in the same county where they were born.
When we were done we had more than 8,000 statistically significant
associations between lifestyle, diet and disease variables. We had a
study that was unmatched in terms of comprehensiveness, quality and
uniqueness. We had what the New York Times termed "the Grand Prix of
epidemiology." In short, we had created that revealing snapshot of time
that we had Originally envisioned.
This was the perfect opportunity to test the principles that we dis-
covered in the animal experiments. Were the findings in the lab going
to be consistent with the human experience in the real world? Were our
discoveries on aflatoxin-induced liver cancer in rats going to apply to
other types of cancer and other types of diseases in humans?
FOR MORE INFORMATION
We take great pride in the comprehensiveness and quality of the
China Study. To see why, read Appendix B on page 353. You'll find
a more complete discussion of the basic design and characteristics
of the study.
THE CHINESE DIETARY EXPERIENCE
Critical to the importance of the China Study was the nature of the diet
consumed in rural China. It was a rare opportunity to study health-re-
lated effects of a mostly plant-based diet.
In America, 15-16% of our total calories comes from protein and
upwards of 80% of this amount comes from animal-based foods. But in
74 THE CHINA STUDY
rural China only 9-10% of total calories comes from protein and only
10% of the protein comes from animal-based foods . This means that
there are major nutritional differences in the Chinese and American
diets, as shown in Chart 4.3.
CHART 4.3. CHINESE AND AMERICAN DIETARY INTAKES
Nutrient China United States
2641 1989
Calories (kcal/ dayV
14.5 34-38
Total fat (% of calories)
12
33
Dietary fiber (g/day)
64 91
Total protein (g/day)
0.8 10-11
Animal protein (% of
calories)
34 18
Total iron (mg/day)
The findings shown in Chart 4.3 are standardized for a body weight
of Sixty-five kilograms (143 pounds) . This is the standard way that
Chinese authorities record such information and it allows us to easily
compare different populations. (For an American adult male of seventy-
seven kilograms, calorie intake will be about 2,400 calories per day. For
an average rural Chinese adult male of seventy-seven kilograms, calorie
intake will be about 3,000 calories per day.)
In every category seen above, there are massive dietary differences
between the Chinese and American experiences: much higher overall
calorie intake, less fat, less protein, much less animal foods , more fiber
and much more iron are consumed in China. These dietary differences
are supremely important.
While the eating pattern in China is far different from that of the
United States, there is still a lot of variation within China. Experimental
variation (i.e., a range of values) is essential when we investigate diet
and health associations. Fortunately, in the China Study considerable
variation existed for most of the measured factors. There was exception-
al variation in disease rates (Chart 4.2) and more than adequate varia-
tion for clinical measurements and food intakes. For example, blood
cholesterol ranged-as county averages-from highest to lowest almost
twofold, blood beta-carotene about ninefold, blood lipids about three-
fold, fat intake about sixfold and fiber intake about fivefold. This was
75
LESSONS FROM CHINA
crucial, as we primarily were concerned with comparing each county in
China with every other county.
Ours was the first large study that investigated this particular range
of dietary experience and its health consequences. In effect, we are
comparing, within the Chinese range, diets rich in plant-based foods
to diets very rich in plant-based foods. In almost all other studies, all
of which are Western, scientists are comparing diets rich in animal-
b a s e d foods to diets very rich in animal-based foods. The difference
between rural Chinese diets and Western diets, and the ensuing dis-
ease patterns, is enormous. It was this distinction, as much as any
other, that made this study so important.
The media called the China Study a "landmark study." An article in
the Saturday Evening Post said the project "should shake up medical and
nutrition researchers everywhere."8 Some in the medical establishment
said another study like this could never be done. What I knew was that
our study offered an opportunity to investigate many of the most con-
t e n t i o u s ideas that I was forming about food and health.
Now, I want to show you what we learned from this study and how
twenty more years of research, thought and experience have changed
not only the way I think about the connection between nutrition and
health, but the way my family and I eat as well.
DISEASES OF POVERTY AND AFFLUENCE
It doesn't take a scientist to figure out that the possibility of death has
been holding pretty steady at lOO% for quite some time. There's only
one thing that we have to do in life, and that is to die. I have often met
people who use this fact to justify their ambivalence toward health in-
formation. But I take a different view. I have never pursued health hop-
ing for immortality. Good health is about being able to fully enjoy the
time we do have. It is about being as functional as possible throughout
our entire lives and avoiding crippling, painful and lengthy battles with
disease. There are many better ways to die, and to live.
Because the China Cancer Atlas had mortality rates for more than
four dozen different kinds of disease, we had a rare opportunity to study
the many ways that people die. We wondered: do certain diseases tend
to group together in certain areas of the country? For example, did
colon cancer occur in the same regions as diabetes? If this proved to
be the case, we could assume that diabetes and colon cancer (or other
diseases that grouped together) shared common causes. These causes
THE CHINA STUDY
76
could include a variety of possibilities, ranging from the geographic
and environmental to the biological. However, because all diseases are
biological processes (gone awry), we can assume that whatever "causes"
are observed, they will eventually operate through biological events.
When these diseases were cross-listed in a way that allowed every
disease rate to be compared with every other disease rate,9 two groups
of diseases emerged: those typically found in more economically de-
veloped areas (diseases of affluence) and those typically found in rural
agricultural areas (diseases of poverty) 10 (Chart 4.4).
CHART 4.4. DISEASE GROUPINGS OBSERVED IN RURAL CHINA
Diseases of Affluence Cancer (colon, lung, breast, leukemia,
childhood brain, stomach, liverL diabe-
(Nutritional Extravagance)
tes, coronary heart disease
Diseases of Poverty (Nutritional Pneumonia, intestinal obstruction,
peptic ulcer, digestive disease,
inadequacy and poor sanitation)
pulmonary tuberculosis, parasitic dis-
ease, rheumatic heart disease, meta-
bolic and endocrine disease other than
diabetes, diseases of pregnancy and
many others
Chart 4.4 shows that each disease, in either list, tends to associate
with diseases in its own list but not in the opposite list. A region in
rural China that has a high rate of pneumonia, for example, will not
have a high rate of breast cancer, but will have a high rate of a parasitic
disease. The disease that kills most Westerners, coronary heart disease,
is more common in areas where breast cancer also is more common.
Coronary heart disease, by the way, is relatively uncommon in many
developing societies of the world. This is not because people die at a
younger age, thus avoiding these Western diseases. These comparisons
are age-standardized rates, meaning that people of the same age are be-
i n g compared.
Disease associations of this kind have been known for quite some time.
What the China Study added, however, was an unsurpassed amount of
data on death rates for many different diseases and a unique range of di-
etary experience. As expected, certain diseases do cluster together in the
same geographic areas, implying that they have shared causes.
These two disease groups have usually been referred to as diseases of
77
LESSONS FROM (HINA
affluence and diseases of poverty. As a developing population accumu-
lates wealth, people change their eating habits, lifestyles and sanitation
systems. As wealth accumulates, more and more people die from "rich"
diseases of affluence than "poor" diseases of poverty. Because these dis-
eases of affluence are so tightly linked to eating habits, diseases of afflu-
ence might be better named "diseases of nutritional extravagance." The
vast majority of people in the United States and other Western countries
die from diseases of affluence. For this reason, these diseases are often
referred to as "Western" diseases. Some rural counties had few diseases
of affluence while other counties had far more of these diseases. The
core question of the China Study was this: is it because of differences in
dietary habits?
STATISTICAL SIG NIF ICANCE
As I go through this chapter, I will indicate the statistical signifi-
cance of various observations. Roman numeral one (I) means 95+%
certainty; roman numeral two (II) means 99+% certainty; and ro-
m a n numeral three (III) means 99.9+% certainty. No roman numeral
means that the association is something less than 95% certainty.ll
These probabilities also can be described as the probability that an
observation is real. A 95% certainty means a 19 in 20 probability
that the observation is real; a 99% certainty means a 99 in 100 prob-
ability that the observation is real; and a 99.9% certainty means a
999 in 1,000 probability that the observation is real.
BLOOD CHOLESTEROL AND DISEASE
We compared the prevalence of Western diseases in each county with
diet and lifestyle variables and, to our surprise, we found that one of the
strongest predictors of Western diseases was blood cholesterol. l I I
IN YOUR FOOD-IN YOUR BLOOD
There are two main categories of cholesterol. Dietary cholesterol is
present in the food we eat. It is a component of food, much like sugar,
fat , protein, vitamins and minerals. This cholesterol is found only in
animal-based food and is the one we find on food labels. How much
dietary cholesterol you consume is not something your doctor can
know when he or she checks your cholesterol levels. The doctor can't
THE CHINA STUDY
78
measure dietary cholesterol any more than he or she can measure how
many hot dogs and chicken breasts you've been eating. Instead, the
doctor measures the amount of cholesterol present in your blood.
This second type of cholesterol, blood cholesterol, is made in the liver.
Blood cholesterol and dietary cholesterol, although chemically identi-
cal, do not represent the same thing. A similar situation occurs with
fat. Dietary fat is the stuff you eat: the grease on your French fries, for
example. Body fat, on the other hand, is the stuff made by your body
and is very different from the fat that you spread on your toast in the
morning (butter or margarine). Dietary fats and cholesterol don't nec-
essarily turn into body fat and blood cholesterol. The way the body
makes body fat and blood cholesterol is extremely complex, involv-
i n g hundreds of different chemical reactions and dozens of nutrients.
Because of this complexity, the health effects of eating dietary fat and
dietary cholesterol may be very different from the health effects of
having high blood cholesterol (what your doctor measures) or having
too much body fat.
As blood cholesterol levels in rural China rose in certain counties,
the incidence of "Western" diseases also increased. What made this so
surprising was that Chinese levels were far lower than we had expected.
The average level of blood cholesterol was only 127 mg/dL, which is
almost 100 points less than the American average (215 mg/dL)! 12 Some
counties had average levels as low as 94 mg/dL. For two groups of about
twenty-five women in the inner part of China, average blood cholesterol
was at the amazingly low level of 80 mg/dL.
If you know your own cholesterol levels, you'll appreciate how low
these values really are. In the U.s., our range is around 170-290 mg/dL.
Our low values are near the high values for rural China. Indeed, in the
U.s., there was a myth that there might be health problems if cholesterol
levels were below 150 mg/dL. If we followed that line of thinking, about
85% of the rural Chinese would appear to be in trouble. But the truth is
quite different. Lower blood cholesterol levels are linked to lower rates of
heart disease, cancer and other Western diseases, even at levels far below
those considered "safe" in the West.
At the outset of the China Study, no one could or would have pre-
dicted that there would be a relationship between cholesterol and
any of the disease rates. What a surprise we got! As blood cholesterol
levels decreased from 170 mg/dL to 90 mg/dL, cancers of the liver,1I
rectum/ colon, II male lung; female lung, breast, childhood leukemia,