For over a decade, a controversial study published in the prestigious journal Cell has cast a shadow over protein consumption, with sensational headlines claiming that “high protein diets are as bad for you as smoking.”
However, new research utilizing more robust methodology and a larger dataset has effectively debunked these alarming claims, providing much-needed clarity on protein’s role in human health.
The original 2013 Cell study made provocative assertions about protein increasing mortality risk, particularly cancer mortality, by elevating IGF-1 levels. The researchers claimed that high protein intake dramatically increased the risk of death, especially in people over 50. These findings sparked widespread fear about protein consumption and influenced dietary recommendations for years.
However, leading protein researchers, including Dr. Stuart Phillips and Dr. Don Layman, immediately identified serious methodological flaws in the original study. The researchers had inexplicably eliminated nearly half of the available data from the NHANES dataset, which contained mortality information on over 12,000 people. Even more concerning, their overall analysis actually showed no effect of protein on mortality, but this finding was buried in supplementary tables while the alarming sub-analyses took center stage.
The original study’s definitions of protein intake were also problematic. Their “low protein” group consumed under 10% of calories from protein—equivalent to less than 42 grams daily, which represents frank protein deficiency. Most troubling was their use of hazard ratio analysis with drastically uneven group sizes, creating a situation where small sample sizes in the low-protein group could severely skew results. As experts noted, when dealing with tiny sample sizes, random chance can create misleading mortality percentages.
The new research addresses these shortcomings comprehensively. Using the complete NHANES dataset of over 15,000 people, researchers employed appropriate statistical methods including the multivariate Markov chain Monte Carlo method to account for the limitations of 24-hour dietary recall data. Instead of arbitrary protein categories, they created tertiles based on usual protein intake distributions, ensuring more balanced group sizes for proper hazard ratio analysis.
The results were striking: total animal and plant proteins showed no association with cancer mortality, cardiovascular disease mortality, or overall mortality. Surprisingly, animal protein showed a slightly protective effect against cancer mortality. IGF-1 levels, previously blamed for increased cancer risk, showed no association with any mortality measures examined.
This study examined usual protein intakes ranging from approximately 60 to just under 100 grams daily, representing normal consumption patterns rather than extremely high protein diets. The researchers found no dose-response relationship between protein intake and mortality risk across this range.
Importantly, the protective effect sometimes attributed to plant protein over animal protein wasn’t observed in this analysis. This challenges the assumption that plant protein is inherently superior, suggesting that benefits previously attributed to plant protein may actually come from accompanying nutrients like fiber and other beneficial plant compounds.
While this research doesn’t extend to extremely high protein intakes, it provides strong evidence that normal protein consumption—both animal and plant-based—poses no mortality risk. The study’s robust methodology and large sample size offer a more reliable foundation for understanding protein’s health effects than the flawed research that previously dominated headlines.
