The ever-changing nutrition recommendations we see in the news can be mystifying.
One day, the story is "coffee is good for us," and then it's bad for us. Reputable doctors say pregnant women can drink red wine, and then another says hold up. This spring, over the course of a month, two large-scale and contradictory studies came out: One claimed eating eggs can contribute to heart attacks, while the other says eggs ward them off.
The confusion continued Oct. 1, when the Annals of Internal Medicine released new guidelines suggesting that adults can continue their unlimited consumption of red meat and unprocessed meat — because the evidence saying you shouldn't is weak.
The guidelines caused a stir because they oppose the long-held advice — from places like the American Heart Association and American Cancer Society — that red meat and processed meat increase a person's risk for cardiovascular disease and cancer, namely colorectal tumors. In 2015, the World Health Organization declared definitively that processed meats cause colorectal cancer and that red meat most likely causes the condition as well.
So, what should you believe? Is there a right and wrong?
Here's the thing: The new advisory does not — because it cannot — explain how much meat is safe to eat, according to the authors and other researchers who spoke with the PBS NewsHour. The new guidelines also do not claim consuming more red meat and processed meat is now O.K., nor does the review claim that doing so comes without danger to a person's health.
Moreover, the main finding — the assertion that weak evidence has propped up the warnings about red and processed meat — has been found by past medical reviews.
The latest study "really did not come up with anything new," said Dr. David Seres, director of medical nutrition at Columbia University's Institute of Human Nutrition, who wasn't involved with the study. "The difference [here] is the willingness of the authors to be really quite devout in their adherence to making recommendations based on the quality and the strength of the findings."
How can a medical trend be both well-established across dozens of studies, but also based on weak evidence? Perplexing, I know.
Here's why all of those things can be true, plus what it means about the meat in your diet.
Are 4 burgers better for you than 7 burgers?
This new study started with a very specific question: Does subtracting three servings of red meat or processed meat from a diet make a difference in people's health?
Bradley Johnston, an epidemiologist at Dalhousie University in Canada, who co-led the endeavour behind the new meat recommendations, said the research teams considered a three-serving reduction as realistic and attainable.
That's because the average person in North America or Europe eats about two to four servings of meat per week. One serving of red meat is about the size of a deck of cards or a small, 120-gram hamburger, while a serving of processed meat takes up about half as much space.
As we mentioned earlier, the teams didn't actually set out to answer this question anew. They looked up thousands of studies other organizations had already done on the topic through a process called GRADE.
Aside from this three-serving criteria, they also excluded a study from their review if it: didn't precisely report the quantity of meat consumed by the subjects; if it involved pregnant women; and if the subjects were asked to recall dietary patterns from prior periods of their lives. Except for one subset that compared vegetarians versus non-vegetarians, all of the analyses only looked at meat-eaters.
After weeding, the team had about 130 studies left.
Some fell in the category of being randomized trials — a higher quality-standard for medical research when it comes to accounting for the biases intentionally or unwittingly introduced by researchers when picking subjects. The rest were observational studies, wherein past scientists had classified or assigned people as low-meat or high-meat dieters and then monitored health outcomes.
The team found that higher red meat consumption is indeed associated with increased risk of chronic disease and mortality, which mirrors what had been found by others in the past. A 2015 meta-analysis reported that less red and processed meat reduced overall mortality, including through cardiovascular disease and cancer, by 11 to 28 percent, which falls in the ballpark of the new guidance.
The same similarities applied to the occurence of colorectal cancer cases, which the World Health Organization has said reduced by 17 or 18 percent for every serving of red or processed meat cut from a person's diet, respectively. The meta-analysis behind the new Annals of Internal Medicine guidance found a similar drop in colorectal cancer rates after less meat consumption.
But this risk reduction may have nothing to do with red meat, Johnston said, because the studies that showed any effect on health were all rated as "low" or "very low" quality by the GRADErs. Almost all of those studies were observational and not randomized trials; the latter is the better tool for showing cause and effect.
"We did find a very small risk reduction in cancer, diabetes and heart disease," Johnston said. "But it's based on low to very low quality evidence, which means we're uncertain if it exists at all."
If the risk is low and comes with a lot of uncertainty, then his team posits that most people do not need to worry about cutting a few burgers out of their diets. Hence, the new guidance.
A passing GRADE?
The Annals — a journal run by the American College of Physicians, the second-largest physician group in the United States — issued these recommendations based on review of more than 130 studies. The new guidelines were ratified by a panel of scientists and regular members of the community — though not unanimously. Of 14 panelists, three unnamed members dissented.
Everyone involved with the research had signed a disclosure form that they did not have any conflicts of interest to report during the past three years, which Johnston said is the standard set by the International Committee of Medical Journal Editors. But on Friday, The New York Times reported that Johnston had received funding from the International Life Sciences Institute — an industry trade group backed by agribusiness, food and pharmaceutical companies as recently as 2016 — for a similar study on sugar.
Johnston said in an email to the NewsHour that "it is tenuous at best" to suggest that his earlier work on sugar had any influence on how his team made the new meat recommendations. "We have no relationship with the meat industry," he added.
Seres said that Johnston "was shortsighted and perhaps naive by not disclosing, even if by the journal's rules he might not be required to."
"He was opening himself up to justifiable criticism and that his credibility would tainted," Seres added. "But this does nothing to discredit the findings."
Johnston's industry ties aside, there are other possibilities for potential bias in these recommendations. Dr. Frank Hu, an epidemiologist and chair of the nutrition department at the Harvard T.H. Chan School of Public Health in Boston, said GRADE can inherently introduce bias of its own because the system excludes so many studies and because the ratings can be heavily subjective.
For example, the team initially found 13,000 studies involving randomized trials about red meat, heart health and cancer outcomes. But they were left with 24 studies after whittling out the ones that didn't fit their criteria.
Granted, the trials in those 24 studies covered a lot of test subjects — 54,000. But a huge portion — 90 percent — came from only one trial, and it was primarily studying the effects of lower fat consumption, rather than meat.
"It's not a meat reduction trial because there are some very small changes in the meat intake between the intervention and control group," Hu said. "But they misuse the data from the Woman's Health Initiative to say that meat reduction has no effect on cardiovascular disease, cancer or mortality."
Hu said other findings defy the value of reproducibility, tossing more weight behind single studies that fit the GRADE criteria, versus consistent findings spread across a body of research. In one case, the evidence from one observational study on ovarian cancer was rated higher than the evidence from 18 studies with consistent findings on overall cancer mortality.
"That really doesn't make any sense, right? The most important criteria in science is reproducibility and replication," Hu said. "If the same procedure were used to validate secondhand smoking, for example, the evidence would be rated very low or low quality."
Notably, secondhand smoke, the smoke inhaled from tobacco smoked by other people, creates about as much as risk for cancer and heart disease as red and processed meat — and the underlying studies around secondhand smoke carry just as much uncertainty.
But no rational person who looks at the public health data around the effectiveness of smoke-free zones would argue that people should continue exposing themselves to secondhand smoke, Hu said.
So why do the same for red meat and processed meat?
What the study is really saying about eating meat
To recap, the Annals of Internal Medicine meta-analyses suggest that eating meat has a small effect on health. As we know, consuming red meat and processed meat can increase one's chances for cancer and heart disease, and one serving of those food items once per day remains about as risky as exposing yourself to secondhand smoke. It's a trend that's been reported for years, even as some researchers raise doubts about how much we can prove it.
But when scaled across a global population of 7 billion people, a small effect could mean millions of deaths. While the average American eats four and a half servings of red meat, a third of the country intakes more than seven servings in the same timeframe, Hu said. Another 10 percent consume 14 servings per week. In the United States alone, Hu estimates a moderate reduction in red meat consumption could prevent 200,000 deaths per year.
John Ioannidis, chair of disease prevention at the Stanford University School of Medicine, who wasn't involved in the study, agreed with Hu's characterization of this danger, but thinks the new guidelines still deserve our attention. They show that time and money are being wasted on studying diet prescriptions with weak methods, he said.
"People have done a very large number of nutrition studies that, just by their very design, they're unlikely to give us an answer that will be concrete," Ioannidis said.
That's because studying nutrition and other aspects of lifestyle has always been extremely difficult.
A scientist may suspect a food, beverage or other consumable might be bad for people, but how do you test that? Randomized trials work well for sussing out drug effectiveness because a doctor can administer a medication in an unbiased way and immediately gauge the health consequences. But it would be unethical to conduct a long randomized trial where you give some people high amounts of meat for decades just so you can wait to see if they develop cancer.
On the flip side, observational studies rely on subjects being honest about their food consumption habits or adhering to prescribed diets they may not like — which, as you know if you've ever been embarrassed about eating more or less of something than you should, doesn't always happen.
"The public in general have this deeply ingrained attitude that components of diet are deeply responsible for our health," said Alfred Neugut, a medical oncologist and cancer epidemiologist at Columbia University who wasn't involved with the new guidelines. "Yet despite decades of intensive research huge investments in research, the truth is that most of the research has been really pretty unimpressive in terms of showing a strong relationship."
Neugut, Ioannidis and Seres said researchers should focus on improving randomized trials for nutrition studies — perhaps by investing funds in finding better ways to pick truly random subjects.
But the meat guidelines do attempt something new by trying to account for people's individual circumstances. One of their reviews centered on studies that had previously tried to assess when people would be willing to give up meat. Johnston's team found most people are unwilling to change even in the face of undesirable health effects.
That forms the crux of these recommendations. It's not easy to make people change their habits, especially when it comes to poor nutrition. So if the risk of getting sick is low and most people won't change anyway, then why tell everyone to quit or cut back on meat?
"The system that's been in play for nutrition guidelines has been more patriarchal," Johnston said. "People can make their own decisions and they should be aware that the quality of evidence is low and that the risk reduction is very small."
Ioannidis said people should consider other personal factors — like how the meat is prepared, their family history of chronic disease, their access to exercise, stress exposures or their sleep patterns — while also deciding whether or not to cut meat, because chronic disease tends to result from a convergence of unhealthy behaviors. He said the environmental ramifications of meat production, which the new reviews did not address, also cannot be ignored.
Over the years, researchers have bombarded the general public with messages that they need to cut back on certain foods, even though they know that many people won't respond, Ioannidis and Seres said. In order to actually get people to change or adhere to nutritional guidelines — and thus save lots of lives — they both say that resources would be better spent on behavioral studies that might help make such diets more appetizing.