I am a scientist, not specifically in the field of nutrition, though I have touched on it as an aside of my main focus, and I know how to research and interpret (correctly!) scientific work. I have read The China Study and the China Project (a publication of the actual data from the study). Normally when reading bad science information I would just shrug it off and move on but thought that there would be people out there that would just take the authors word of the truth of the book, resulting in unnecessary dietary alterations and damage to health all with the aim of trying to get healthier. As such I have written my thoughts on the book - take from them what you will.
The China Study is an attempt by Campbell to promote veganism as a dietary lifestyle through scientific research. Unfortunately the scientific basis of the book if full off misinterpretations, omissions of conflicting data, and conclusions and statements based on unreferenced facts (possibly not facts?). I began reading the book with an open mind but from the outset it was clear that Campbell had one mantra - animal based food is bad, plant based food is good, and this is repeated over and over throughout the book.
Let's first look at Campbells own laboratory studies. In the presence of Aflatoxin, a carcinogen, rats fed a diet of 20% casein, a milk protein, develop cancer while those that are fed 5% casein do not. Okay, I am willing to accept that study on face value. How much casein causes cancer then? In a dose response study Campbell found that 10% casein doesn't contribute to cancer development, but above 10% does. Again, I am happy to accept that. A diet made up of 10% casein contributes to cancer development. How does that apply to humans? After describing a study about nitrosamines and how the dose wasn't relevant to the human population (page 45), Campbell has done the exact same thing with his Casein study. Casein is a milk protein. In 100ml of whole milk, the macro nutrient content is 5.2g of carbohydrate, 3.25g of fat and 3.2g of protein that equals 11.65g of nutrients, the rest of the 100ml mostly made up of water. Milk protein is 80% casein, 80% of 3.2g is 2.56, so out of that 11.65 total, 2.56 is casein which equals 22% of the total. Oh no! Milk will cause us to develop cancer! But don't worry, as long as we get the casein down to 10% we will be safe. How do we do that? Eat 13.95g of anything that is not casein. Pretty easy to do. So as long as we are not living of more than about 50% milk, then we are safe from cancer as a result of the casein in the milk. Do you know anybody that has that much milk? And that is ignoring the fact that casein extracted from milk for the purposes of his study is not exactly a healthy, natural source of protein purely as a result of the chemical extraction.
But hang on, what if other proteins contribute to the development of cancer? Campbell thought that so he investigated gluten and soy and found that neither of them had the same impact as casein. That clearly shows that not all proteins contribute to cancer, and having tested 2 plant proteins and 1 of the many animal proteins, we must therefore conclude that ALL animal proteins lead to cancer and ALL plant proteins do not. Does anybody else see a problem with this? All that we can conclude from these studies is a diet made up of above 10% casein, may contribute to the development of cancer and a diet below 10% casein does not contribute. That is all. Other proteins, both animal and plant, like gluten and soy, may behave differently and unless you have a milk fetish or you are downing large amounts of casein based protein powder (like the rats in the study) then the study is largely irrelevant to your diet or your health.
Before moving on I have one more observation; To test the impact of decreased protein from 20 to 5% they replaced some of the protein with carbohydrates to keep the calories the same. Commenting on the addition of carbohydrate he says "the extra starch and glucose in the low-protein diets could not have been responsible for the lower development of foci because these carbohydrates, when tested alone, actually increase foci development" (page 351). So carbohydrates, which come from plants, increase the development of foci? PLANTS CAUSE CANCER TOO?? Could this be something worth elaborating on or including in a conclusion? No, better not, lets keep that brief mention of carbohydrates causing cancer stuck away in an appendix in case anybody gets the wrong idea.
It is apparent from his casein studies that Campbell has come to the conclusion that "20% casein causes cancer, therefore all animal protein is bad". It is with this mindset that he then set out on the giant study of the China Project, a commendable effort that could have had many beneficial outcomes. Unfortunately, possibly as a result of his previous work, Campbell has gone in with blinders on, and all he can see is animal protein and the negative health outcomes associated with its' consumption. The project itself and the original publication arising from it produced a vast amount of data that provides some interesting insight into health and disease. However, what Campbell has shown in the China Study is but a fraction of the information to be gained from the project. It would require a whole new study (unbiased this time preferably) to go into all the beneficial knowledge we could gain, but I will touch on a few things here.
Campbells main conclusion in the China Study is that all animal protein contributes to disease and all plant protein prevents disease. In the original project, they performed a diet survey over 3 days, analyzing all the food consumed per person in that time. Guess how many of the measured mortality factors (about 50 of them), were associated with animal protein consumption measured from the diet survey. Zero. Zero. Zero. Okay, so Campbell can't have come to his conclusions from there. They also had study participants fill out a questionnaire that included one question on meat consumption. Guess how many mortality factors correlated with that? One type of cancer (naso-pharyngeal or something I think it was). An example of some of the many other inclusions in the questionnaire are canola oil and potatoes (not sweet potatoes) which both had a number of positive associations with the development of different types of cancer. Apparently that wasn't worth mentioning in the China Study. Speaking of oil, Campbell makes reference to %fat in the diet being a good indicator of animal protein consumption, despite the fact they clearly use enough canola oil (a vegetable fat) to measure in the study.
So a 3 day food consumption survey shows no association between animal protein and mortality and a questionnaire shows an association between meat and one of many cancers measured. From where can Campbell come to his evil animal protein conclusion then? They also took plasma samples and measured them for blood biomarkers of animal protein consumption. These biomarkers, listed in the references for chapter 4 #39 are "plasma copper, urea nitrogen, estradiol, prolactin, testosterone and, inversely, sex hormone binding globulin, each of which has been known to be associated with animal protein intake from previous studies". No mention of these previous studies of course. So the associations with most of those biomarkers and mortality rates are dubious, and the only biomarker statistically associated with cancer mortality is copper. Many places show food sources of copper and I went to [...] find the best sources of copper. The best? Calfs liver. The next 40 best? All from plants. 42 and 43 are shrimp and venison, the only other animal source in the list on the site. So for copper to be a biomarker of animal consumption then the participants in this study must be eating a lot of calf liver and avoiding a lot of vegetables. Sound realistic?
So from an association between blood biomarkers, the only real one being copper, and cancer mortality, Campbell has concluded that animal protein gives you cancer, despite the fact that the majority of dietary sources of copper are actually from plant sources. So that basically leaves Campbell with no actual evidence between animal consumption and mortality as a result of the original China project.
A final note. In his eating right section Campbell says supplements are bad (principle 2). Principle 3 then says "there are virtually no nutrients in animal-based foods that are not better provided by plants"(page 230), but over the page he says plants are not a good source of vitamin B12 and you probably should take a supplement. What? Then in the how to eat section on page 242 he says "the findings from the China Study indicate that the lower percentage of animal-based foods that are consumed, the greater the health benefits-even when the percentage declines from 10% to 0% of calories". As I've clearly shown, the China Study does not show this, and his own study with Casein proved that there was no benefit in eating less then 10% of your diet from Casein.
Clearly Campbell is a vegetarian, as he states in the book, and promoting vegetarianism is his main goal, which he tries to back up with scientific research that actually disagrees with him, but that he has interpreted in a way that makes it agree with him. Bad science, bad book and definitely bad recommendations as far as health. While I'm not saying go out and live on animal products alone, I don't think you should stop eating them, especially because they are tasty, but even if only for a natural source of vitamin B12.