Um, Actually, Eating Meat Isn’t Even Remotely As Bad As Smoking
A recent study proclaiming that eating meat could be as dangerous as cigarettes sent carnivores—and the media—into a tizzy. But a closer look suggests the science is as bad as the sensational journalism.
Outside's long reads email newsletter features our strongest writing, most ambitious reporting, and award-winning storytelling about the outdoors. Sign up today.
Unless you've been on an extended spring break, on an extremely remote beach, you've no doubt caught wind of the recent headline that spread faster than the smell of good BBQ: Eating meat and cheese is as deadly as smoking. The idea was broadcast in numerous media outlets, from NPR to Scientific American to the L.A. Times. Vegans rejoiced at the vindication of tofu (according to the study, protein from plant sources was “non-harmful”). In the carnivore camp, meat eaters shook their collective fists, while rational, health-conscious folks who take neither side simply wound up confused, and—rightfully—worried. What's the deal? Is a meat-and-cheese plate really as bad for you as a pack of Marlboro Reds?
The research in question has the insomnia-curing title, “Low Protein Intake Is Associated with a Major Reduction in IGF-1, Cancer, and Overall Mortality in the 65 and Younger but Not Older Population,” published in the March 2014 issue of Cell Metabolism. There were multiple pieces to the study, but I'm only looking at the human information here. This data focused on a bit of epidemiology, pulling information from the National Health and Nutrition Examination Survey III (NHANESIII). This is a long-running study in which participants were asked to fill out questionnaires about what they ate. The data was collected, then these people were tracked with regards to the occurrence of a wide variety of disease, including cancer.
In the Cell Metabolism report, the researchers claimed to have found a correlation between consuming animal protein from sources like meat and cheese, and increased cancer rates. This sounds quite serious, but one can find a number of alarming correlations that seem scary but beautifully illustrate how correlation should not be mistaken for causation. For example, one such report would have you believe there is a “clear” correlation between an increase in autism cases and… organic food sales.
The best that correlation can offer is a proposed mechanism, which may be investigated in rigorous settings like Randomized Controlled Trials (RCT's). It's not surprising that many media outlets irresponsibly jumped on this research and spun sensationalistic headlines—the media is pretty good at that. But more concerning to me, as a scientist and author, is that the researchers themselves presented this material as if it were gold-standard science. It. Is. Not.
Here's the problem: Researchers too-often rely on tools like NHANESIII, which is just a survey. People are asked to recall things like: What/how much did you eat? How often/much did you exercise? The funny thing is that people are terrible at recalling information like this. People forget. Sometimes they even “lie.” NHANESIII is actually the result of tweaking two earlier iterations of the same study in which it was clear the study participants were vastly underreporting food intake. Here is a quote from a critique of the NHANES studies:
“Across the 39-year history of the NHANES, EI data on the majority of respondents (67.3% of women and 58.7% of men) were not physiologically plausible… The ability [of the study] to estimate population trends in caloric intake and generate empirically supported public policy relevant to diet-health relationships from U.S. nutritional surveillance is extremely limited.”
So, in my firm opinion, the very dataset that this study is based on is, at best, deeply flawed. It is certainly not rigorous enough to make a sweeping, outrageous statement like animal proteins are as hazardous as smoking.
Oh, and here's another problem—a serious conflict of interest: Health researcher Zoe Harcombe recently conducted a great analysis of this same paper, revealing that several of the authors are involved in a plant-based protein company called L-Nutra. This might not seem like an important issue until one reads the paper and notes that the researchers go out of their way to say that plant-based protein is not a problem, but animal protein is. Hm.
I have my own biases in all this. Full disclosure: I’m a Paleo diet researcher, and it’s my position that animal protein is a vital part of the human diet. In my work, the science appears quite clear: Not only is meat intake not bad for you, it can, in fact, be quite healthy when managed correctly. Here is an example of an RCT that shows how the inclusion of red meat (along with some strength training) improved body composition and fitness, while also reducing inflammation. This study was conducted in a retirement-community setting, which means we were not relying on suspect food questionnaires. Researchers actually had great control over what food participants actually ate. This is the type of research that should make headlines, but “Old People Ate Meat, Lifted Weights, Got Jacked, and Reduced Inflammation” isn't nearly as catchy as “Meat: As Bad As Cigarettes.”
I’ve worked with many individuals who have lost weight and cleared up chronic problems, ranging from autoimmunity to diabetes. I have worked with a local municipality that implemented a Paleo-type diet among its Police and Fire Departments. A two-year pilot study indicated that the Paleo-based diet helped employees dramatically reduce the likelihood of diabetes, heart attack, and stroke (and saved the city a potential $22 million in health-care costs). In the next few years we will see some gold-standard RCT’s comparing the Paleo diet with other popular nutritional approaches.
All that said, keep in mind that the vast majority of our health and nutrition stories these days tend to come from epidemiological research that is at best correlation, not causation. Add to that the knowledge that these epidemiological studies rely on sources like NHANESIII, and you begin to realize that the data is suspect long before the headlines come out. No wonder they've come up with a saying about how research now tends to fall into three categories: Lies, Damn Lies, and Statistics.
Robb Wolf is the author of The Paleo Solution, and the founder of NorCal Strength and Condition. He lives in Reno, Nevada.