China Study: Fact or Fiction?

This is entirely stolen from a friend of mine who wrote an email to a private group. I’m publishing it here, without his consent, knowledge, and without identifying him, because I think it’s helpful for other people to learn from and he probably would like that. If he tells me I’ll remove it or add his information.

I think the biggest lesson for people to learn here isn’t about nutrition. The biggest lesson is learning to be skeptical of claims, even if red flags are subtle. People have bias, their own agendas, and their own ethics, and they may not match up with what you expect.

I apologize in advance for the length of the email. Since [my wife] has been into diet and nutrition for most of her adult life, I have also learned a fair amount about the subject over the last 10 years (mostly from her). Without going into why, for the past 10 years [my wife] and I have chosen to eat close to a 100% whole grain, organic diet at home. We avoid almost all processed foods, and avoid all partially hydrogenated fats. We do eat meat (though [my wife] was once a vegetarian for 11 years), though not much red meat and almost always organic, and we once bought ¼ of a grass fed organic cow and kept it in our freezer for a year. Our general philosophy is that it is what you eat 90% of the time that matters, so we don’t stress too much about what we eat away from home.

Below are my thoughts and experience reading The China Study. I am always up for a chat in person or over the phone if anyone wants to talk more.

— [Me]

My initial reaction as I started to read The China Study was that the guy writing it really has his stuff together, is very well spoken/written, and seems to have a good grasp of science and statistics. It was very drawn into the book and was excited to learn the results of a many decades study in China that covered millions of people.

As I progressed into the book, my reaction started to change and red flags started going up in my mind as I read. Most were small red flags, a few were large, but the number of them started getting me worried. I will not list all of the small red flags in this email, but I will mention the largest one for me. I was most concerned about the fact that this book is called The China Study, yet the book lacks any formal presentation at all of any data from the actual “china study.” If the data is so conclusive, then why does the book not contain any of it? Since there were only 65 counties studied in the “china study,” it would only take 1 page of the book to show the amount of protein consumed in each county and the corresponding mortality and cancer rates for each county. Shouldn’t this, at a minimum, be in an appendix of the book? It made me wonder whether or not the data actually backed up the authors claims.

Being a natural skeptic, I decided to spend some time doing some research online to see what I could find. I ended up doing about a day of research; 2/3 of it was online and 1/3 of it was in the UCSF Medical Library. I went to the UCSF Medical Library because I discovered that they have the actual book (written in English and Chinese) that contains the statistical results of the original study that was performed in China.

Unfortunately, my research has led me to the conclusion that the author of The China Study is being manipulative with data, presents the results of other people’s studies out of context, lies through omission in many instances, and cannot present the data results from the actual “china study” in his book because most of the statistical results disprove the protein / animal protein premise of his book. I will go over some of what I found in this email, but am also more than happy to talk with anyone over the phone or in person if you would like more detail:

The China Study (the real study, not the book)

Let me start out by saying that the authors claim that the “china study” is essentially the grand-daddy of all nutritional studies appears to be a massive overstatement. The author leads the reader to believe that the “china study” followed millions of people over a many decade long period to get its results. The reality is the complete opposite of this. The study only gathered blood, diet and survey information from 6,500 people, which certainly does not make it massive in size or scope. More importantly, the study had inherent flaws in its organization and structure that most of its data might actually be irrelevant anyway:

MORTALITY & DIET DATA: The study collected from blood samples, diet information and surveys from people in 65 counties in China and compared it to the mortality data availably from each of those counties. This all seems well and good until you learn that the blood samples, diet information and surveys were all collected in 1983 and the mortality data used for the analysis was collected from 1973–75. Obviously, the mortality data from 1973–75 is primarily a reflection of the diet and lifestyle of the 1950s and 1960s, not of the 1980s. The largest known famine in the world took place in China from 1959–61. I would think that this event alone would have a significant effect on mortality data from 1973–75.

DATA SIZE AND COLLECTION METHOD: Although the mortality data is collected for almost the entire population during the collection period (1973–75), only 100 people from each county were used for the blood, diet and survey data collection. There are multiple problems with this. To start, this number is so small that it is not statistically likely that data from 100 people is a true representation of that county. The next problem is that the study then took the information from each set of 100 people and aggregated them together to get 1 set of data to work off of for each county. This is a major problem because it means that the entire study now has only 65 data points to go off of to generate all of its results. 65 data points is not enough data points to get results with much statistical certainty across the number of mortality end-points they were studying.

Now, if we ignore these two major problems with the study, you would assume that the data would strongly backup the author’s premise that diets high in meat, protein and dairy would result in higher mortality rates. But the data actually tends to prove the opposite. For example, the data for all mortality causes (which is obviously the most important) shows that the following items in people’s diet had a statistically significant (95% or greater) correlation with reduced mortality, listed from higher to lower correlation:

  • Soy Sauce
  • Eggs
  • Phosphorus
  • Total protein

None of the diet items studied had a statistically significant correlation with increased mortality. On a side note, although the correlation was not statistically significant, milk and meat consumption both correlated with reduced mortality.

So where is the data that backs up the author’s claim about protein? It is not in the data generated from any of the blood samples or the diet studies; these all show results that tend to disprove the author’s claims. There is only one strong data point in all of the data that backs the author’s thesis. This shows up in the questionnaire section in which there was a statistically significant correlation such that people who reported that they had a relative that had cancer tended to have a diet higher in protein. The author takes this correlation, ignores all the other data, and extrapolates from there. On a humorous note, the questionnaire section also showed a statistically significant correlation between smoking homemade cigarettes and reduced cancer mortality.

Dr. Lester Morrison’s Study

The author uses the results of Dr. Lester Morrison’s 1950’s study of heart patients as positive confirmation of his thesis. The author states that Dr. Morrison, a famous heart surgeon, took 100 patients that had heart attacks and had 50 of them continue on their normal diet and had 50 of them change their diet to significantly reduce their cholesterol. The result was that the 50 that changed their diet lived much longer and had significantly reduced reoccurrences of heart attacks.

What the author does not tell you is that Dr. Morrison achieved this by putting his patients on a high-protein, low fat diet. His patients ate 2-ounces of lean meat (ie. lamb) at lunch and dinner and drank 13-oz of low-fat milk each day. He also had them take large amounts of supplements, like calcium, phosphorus, vitamin c, etc. So while, 4oz of meat a day is probably less than the American average, it is hardly a low-meat diet. And when you factor in the large daily glass of milk, the diet certainly was not low in animal protein. It is also ironic that the author argues strongly against supplements in his book, and all the patients in the good group took large amounts of supplements.

Is the author of The China Study being deceptive by leaving out these details of Dr. Morrison’s experiment? I think so.

The Framington Study

The author also uses results of the Framington Study to prove his case. The Framington Study followed thousands of people in a small MA town over many decades. This study showed a strong connection between cholesterol and heart disease. However, the Framington study also found that there was no discernable correlation between diet and cholesterol levels among the people it was studying.

Casein, Rats & Cancer

The author continually refers to a study performed in China in which rats given high levels of Casein were shown to develop cancer and those given low levels were not. I found it very odd that the author continually used this single study as the primary rationale for linking cancer to animal protein. If the connection were so strong, I thought that surely there would have been many follow up studies by other researchers that he could reference as well.

Well, it turns out that there is also another protein that is in milk called whey. And, it happens to be that when you feed rats whey, it significantly inhibits cancer growth. My guess is that most of the follow up studies to the original casein experiment probably looked at multiple forms of protein. The author probably did not want to mention any of these because he would have also had to mention that there was an animal protein that reduced cancer. Obviously, when we drink milk or eat cheese, we are getting both casein and whey.

I read the results from a study that took a group of rats and fed 1/3 of them a diet high in casein protein, 1/3 a diet high in soy protein, and 1/3 a diet high in whey protein. It then gave them all chemically induced mammary cancer. They did this experiment with two different groups of rats. In one group the casein-fed rats developed the most tumors, in another group the soy-fed rats developed the most tumors, and in both groups the whey-fed rats developed noticeably less tumors. Overall, their conclusions were that diets rich in soy reduced the incidence mammary cancer development by 20%, but that diets rich in whey were more than twice as effective as the soy diets at reducing both tumor incidence and multiplicity.

So does dairy cause cancer or prevent cancer? It seems to depend on which protein you study. Regardless, it is obviously not as cut and dry as the author presents it to be, and it is worrisome to me that he just completely omits any discussion about whey protein and its anti-cancerous tendencies while bashing all animal proteins as cancer promoters. To me this is another example of deception by omission.

My Overall Conclusion

My general belief is that what the author of The China Study should be saying is the following:

Lowering your cholesterol is the best way to lower your chances of getting heart disease, which is one of the top killers in the US. A primarily vegetarian diet, combined with exercise, is an almost sure fire way to lower your cholesterol and improve your health. Here is how you can do it…

Unfortunately, the author instead seems to have cherry-picked data, omitted key information, and taken some results out of context in an attempt to prove that a vegan diet is the ONLY healthy diet alternative. This is where I strongly disagree with the author, his data, and his reasoning. I have no doubt that a primarily vegetarian diet (and a vegan diet) is very a very healthy diet, however, I am equally certain that it is just one of many possibly “healthy” diets available and certainly not the ONLY one.

Side Note to Vegetarians

When doing my little bit of research, I came across two studies that were done in Asia that showed that heart disease tended to increase as meat consumption increased. The studies also found that stroke incidence increased as meat consumption moved towards zero. So, a little meat (could be fish) at minimum in a diet appears not to be a bad thing.