The China Study

Before writing about China, Colin Campbell spent years studying childhood malnutrition in the Philippines, in the 1960s and 1970s. The question was why so many Filipino children developed liver cancer, a problem that science at the time linked to aflatoxin, a toxin produced by a fungus that contaminates peanuts and corn stored in damp places.

The odd thing is that, looking at the cases up close, Campbell noticed that the children who developed cancer were not always the poorest ones. Often they were from slightly better-off families, who ate more animal protein. That went against what he had learned in college, where animal protein was almost a synonym for healthy growth.

That discomfort, a researcher watching his own training collide with the data he was collecting himself, is the real starting point of the book, and it came from years of accumulated doubt, not from a sudden change of mind.

Back in the lab at Cornell, Campbell exposed rats to aflatoxin and then split the animals into groups on diets that varied only in the amount of casein, the main protein in cow's milk: one group got 20% casein, a proportion similar to consumption in a wealthy country, and another got just 5%.

The result repeated itself experiment after experiment: rats on 20% casein developed far more precancerous liver lesions than those on 5%, even though they received the same dose of the cancer-causing agent. Campbell could even switch the growth of those lesions on and off just by adjusting the casein midway through the experiment.

Plant proteins, like those from wheat and soy, did not reproduce the same effect even in high amounts. To him, that suggested the problem was not protein in general but casein in that specific model, a narrower finding than the book's reputation tends to suggest.

Those results led Campbell to look for a larger-scale test, outside the animal lab. In the 1980s, he joined Chinese researchers and the University of Oxford to build the China-Cornell-Oxford Project: a survey of diet, lifestyle, and causes of death across dozens of rural Chinese counties, where access to animal food was limited compared to the United States.

The scale was huge for the time: dozens of counties, thousands of families examined, and hundreds of variables cross-checked, from blood cholesterol to rice consumption, producing thousands of statistical correlations between diet and disease.

The team took advantage of a rare feature of rural China in that period: within a given region, people ate practically the same thing their whole lives, which made it easier to compare a county with a more plant-based diet to another that already included more meat and dairy.