104: The China Study Revisited - Science vs. Storytelling with Dr. Ben Bikman

Bikman argues that the fear of animal protein largely stems from the China Study's flawed interpretation of isolated casein. In contrast, real-world whole foods deliver a package where protein and fat work together. He cites epidemiological data showing animal protein linked to lower mortality, and specifically points to the InCHIANTI study. He also highlights that dairy fats like CLA and butyrate have anti-cancer mechanisms, so whole dairy is not merely neutral but potentially protective. The advice is to stop fearing eggs, beef, and full-fat dairy, viewing them as valuable components of a healthy diet.

Bikman contrasts the short-lived mTOR flick from leucine with the sustained hold from insulin, citing the 2011 Gran and Cameron-Smith study. He applies this to practical diet: the real driver of cancer-related mTOR overactivation is likely the insulin response to refined carbohydrates, not eating steak or eggs. The recommendation is therefore to minimize foods that spike insulin—those from bags and boxes with barcodes—and instead get carbohydrates from fibrous, slow-digesting sources like leafy greens and berries to maintain low, stable insulin levels.

Bikman emphasizes that the same China Study data could have been used to argue that wheat flour causes heart disease. He points to the omitted whey research and the Tuoli County outlier as examples of inconvenient facts left out of the narrative. His broader lesson is that nutritional epidemiology is rife with confounding, and a single study or book built on correlations should be treated as hypothesis generation, not proof. He urges listeners to question bold claims and look for independent, rigorous reanalysis.

Bikman explains that the China Study's human evidence is entirely correlational, and correlations are easily misinterpreted—illustrated by the 'ice cream drowning fallacy'. He credits Denise Minger's rigorous reanalysis of the raw data, which revealed inconsistencies: total cancer mortality was more strongly correlated with plant protein intake than with animal protein, and in some counties higher meat consumption correlated with lower mortality. He highlights Tuoli County, where residents ate more than double the US average of animal protein and high fat yet had low rates of both cancer and heart disease—a glaring outlier the book downplays. Wheat flour intake showed a stronger correlation with heart disease than animal protein did, yet the book singles out animal protein. Bikman argues that, based on the same data, one could just as easily write a version blaming wheat flour. Overall, the China Project should be seen as a hypothesis generator, not causal proof.

Bikman details three major problems. First, the studies used purified casein, not whole dairy (milk, cheese, yogurt); real dairy contains fats, whey protein, and vitamins that alter metabolism and cancer outcomes. He likens it to evaluating sugar by giving only isolated fructose. Second, the same research group published a 1989 paper showing that whey protein led to significantly fewer precancerous liver lesions than casein when both were accompanied by aflatoxin—a nuance the book omitted. Third, when casein was tested without aflatoxin, tumors did not develop spontaneously, indicating casein is not a primary carcinogen. He also notes that not all animal proteins behave like casein: fish protein reduced mammary tumors in rodents, and high beef intakes did not increase colon cancer risk. Painting all animal protein as carcinogenic based on this narrow, artificial model is nutritionally shortsighted.

Bikman discusses the Gran and Cameron-Smith 2011 study comparing leucine and insulin on mTOR activation in muscle cells. Leucine caused a modest 1.3-fold increase in mTOR phosphorylation at 30 minutes, which faded by 3 hours and was absent at 24 hours. In contrast, insulin produced a 2-fold activation at 3 hours and maintained a significant 1.5-fold activation even after 24 hours—the maximum measurement point, so the effect likely lasted longer. If chronic mTOR activation truly promotes cancer, then insulin is the more meaningful threat. He ties this to a 2022 Nurses' Health Study paper showing that diets with high insulinemic potential (rich in refined carbs and sugary drinks) linked to a 15% higher breast cancer risk, particularly ER-negative tumors. Thus, hyperinsulinemia from refined carbohydrates, not dietary protein, is the primary mTOR driver.

Bikman explains that CLA, found in butter and cheese, inhibits cancer cell growth by inducing apoptosis and disrupting tumor pathways. A 2015 meta-analysis in PLoS ONE associated higher CLA intake from dairy with lower breast cancer risk in human observational studies. Butyrate, a short-chain fatty acid produced from dairy fat, promotes colon cell differentiation and apoptosis, with human studies (e.g., a 2016 Nature Reviews paper) linking higher dairy fat intake to lower colorectal cancer risk. He argues that isolating casein as done in the China Study strips away these synergistic fats and vitamins; low-fat dairy and purified proteins miss the protective collective benefits. This aligns with keeping protein and fat together in whole foods like whole milk, cheese, and yogurt to maximize metabolic benefits while potentially reducing cancer risk.

Bikman cites the InCHIANTI study, a 20-year cohort of over 2,000 older Italians, which found animal protein intake associated with lower all-cause and cardiovascular mortality, supporting longevity and muscle health. A 2018 study in the European Journal of Nutrition found no link between animal protein and cancer, with fish and poultry showing protective effects. A 2016 study found no consistent link with red meat when controlling for lifestyle factors. A 2022 study across 175 populations linked higher total meat intake with greater life expectancy. Bikman's point: if weak correlational data carry weight, then the growing body of correlations in favor of animal protein must be considered, and they contradict the China Study's selective narrative.