Red Meat and Our Health: Making Sense of the Messy Literature
Seeking to answer if eating red meat causes or increases the risks of metabolic diseases and cancers.
Is red meat good or bad for health? Ask different experts, and you might get different answers. One reason for this is that nutrition research literature is messy, and questionable funding sources are a big culprit.
A 2016 research review identified 26 papers showing that soft drinks are not associated with diabetes or obesity. But all these 26 papers had ties to the food industry. In contrast, 34 papers found that soft drink promotes diabetes or obesity; only one of these papers is industry-funded.
Another study — published last month — analyzed studies in the top 10 nutrition-related journals and found, “Of articles with food industry involvement, 55.6% reported findings favorable to relevant food industry interests, compared to 9.7% of articles without food industry involvement.”
Such biased funding sources also apply to other food group research, including red meat, dairy, fats, plant products (e.g., cocoa, oil, avocado, oats, potatoes, etc.), and dietary supplements (vitamins, protein shake/bars, probiotics, etc.)
“Is it really true that food companies deliberately set out to manipulate research in their favor? Yes it is, and the practice continues,” stated a 2016 commentary in the highly reputable JAMA Internal Medicine journal.
While it’s generally accepted that a whole-foods, plant-based diet is healthy (although exceptions may exist for certain autoimmune diseases) and sugar overconsumption is harmful, the topic is less clear about red meat. So, this article seeks to answer if red meat is bad for health, carefully assessing the funding sources or conflict of interests of relevant studies.
Before going further, I should disclose my conflict of interest: I don’t eat processed meat, but I sometimes eat unprocessed red meat and would make some $ with this article. Yes, nutritional science isn't my field of study, but at least I hope that makes me more unbiased, taking no sides, to begin with.
Guidelines on red meat and diseases
Dietary guidelines —e.g., the 2015–2020 Dietary Guidelines for Americans, Public Health England Eatwell Guide, World Cancer Research Fund, and World Health Organization International Agency for Research on Cancer — often suggest lowering the intake of red meat, especially when processed, to reduce risks of cardiovascular diseases, obesity, type 2 diabetes, and certain cancers in adults.
Systematic review (and meta-analysis)
For example, a 2018 paper in the journal Nutrition Reviews analyzed dietary guidelines from over 30 countries. They found that “The majority of included guidelines were updates of previous guidelines, used existing scientific reports from authoritative bodies, and were based on other countries’ national dietary guidelines. Few guidelines were based on systematic reviews, and only 3 countries commissioned systematic reviews specifically for the guidelines.”
In systematic reviews, at least two researchers screen the literature for relevant studies. When there are discrepancies in study identification, a third researcher will make the final decision. All relevant studies — regardless of positive, negative, or neutral results— are then synthesized and analyzed with appropriate methods to derive a conclusion. Systematic reviews ensure no cherry-picking or biased selection of studies to argue for a point.
If statistical or quantitative analyses are used to pool numeric data from multiple studies, then the systematic review involves meta-analysis. If only qualitative analyses are used, then it’s just a systematic review.
“A growing body of evidence supports the notion that there are serious deficiencies in the methods used to develop dietary guidelines,” stated another 2019 paper in Annual Review of Nutrition.
“Such deficiencies include the failure to access or conduct comprehensive systematic reviews, a lack of systematic or rigorous evaluation of the quality of the evidence, a failure to acknowledge the limitations of the evidence base underlying recommendations, and insufficiently stringent management of conflicts of interest.” There are “hundreds of conflicting and misleading observational studies published every year,” they added.
Observations only tell us an association, but not causation or how that association is formed. For example, it’s observed that increased ice cream sales correlate with increased murder, but how?
After checking some of the cited studies in the dietary guidelines of WHO International Agency for Research on Cancer myself, they were indeed observational.
But observational studies are not useless. They have their value in suggesting possible causation. It’s only that results derived from observational studies shouldn’t be regarded as conclusive but warrant further RCTs, the gold-standard research design in determining cause and effect.
Unfortunately, over 70% of nutritional studies are observational. After all, it isn’t easy to prescribe diets to people for an extended period and on a large scale to conduct a successful RCT.
Overall, current dietary guidelines to limit red meat intake are not perfect due to the lack of systematic reviews and reliance on observational evidence.
The recent controversial recommendation
In November 2019, the Annals of Internal Medicine published a paper, titled “Unprocessed Red Meat and Processed Meat Consumption: Dietary Guideline Recommendations From the Nutritional Recommendations (NutriRECS) Consortium,” that made us rethink the health impacts of red meat.
Herein, researchers from multiple countries conducted a systematic review of the entire literature on the health outcomes of red meat (processed and unprocessed). They used the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) methodology to perform quality control of studies, so only studies of highest grade were synthesized and analyzed.
Reduced red meat intake was defined as eating 3 fewer servings per week. For reference, the average American eats about 4.5 servings of red meat per week. So, the research question was if there are “potential reduced [health] risks associated with a decrease of 3 servings per week of both unprocessed red meat and processed meat”?
Conflict of interests
This paper is not funded, and nearly all researchers declared no conflict of interest. Two co-authors declared that they had received funding from non-industrial sources, such as academia or governments that mainly advocate plant-based eating. So, no industrial funding was present.
However, after publication, the lead author was criticized for not disclosing a funding source he received from Texas A&M AgriLife Research to research saturated and polyunsaturated fats. However, AgriLife argued that it’s a state agency affiliated with academia, so its grants count as public sources, not industrial. Plus, the AgriLife funding was not for the red meat paper.
Still, at least all the remaining 17 authors had no such controversial conflict of interest. Thus, this systematic review is perhaps the most unbiased paper about red meat we have presently. With that said, let’s move on to the actual content of the paper.
Systematic review results 1: Randomized controlled trials (RCTs)
After screening over 13,000 papers, they identified 12 RCTs (mostly on unprocessed red meat). Only three trials received funding from the meat and livestock industry. Two trials received no funding, and the remaining seven trials from governmental or foundation sources.
However, after grading each trial's quality, only 1 trial — funded by the National Institutes of Health (NIH)— was deemed of sufficient quality. Yes, they omitted 11 RCTs due to low sample size, reporting bias, lack of blinding, early termination, etc. As the authors explained, “Eleven studies proved at high risk of bias overall, primarily because of lack of blinding and substantial missing participant outcome data.”
It’s a drastic but perhaps necessary measure to filter out flawed studies with serious limitations, especially when the nutritional literature is already messy.
Results did not find that decreasing red meat intake led to fewer deaths from diseases. Although there’s a trend towards fewer deaths, the results were not statistically significant and were of high uncertainty levels:
Note that the 95% confidence interval (CI) is wide for all rows. If the 95% CI crosses 1.00 or 0 cases, it means that the result is not statistically significant, which means no effect.
Taking the first row as an example, it means that reducing red meat intake could lead to 12 fewer to 7 more cases of all-cause mortality per 1000 persons about 95% of the time. In other words, unprocessed red meat neither increases nor decreases all-cause mortality about 95% of the time.
(Even at the 5% outlier, red meat might either hugely increase or decrease all-cause mortality, so the direction of effects is still uncertain.)
Therefore, unprocessed or processed red meat intake does not affect all-cause mortality, cardiovascular mortality, fatal stroke, breast cancer mortality, or total cancer mortality. “Our results highlight the uncertainty regarding causal relationships between red meat consumption and major cardiometabolic and cancer outcomes,” the authors concluded.
Systematic review results 2: Observational/cohort studies
Following a systematic screening of the literature, the researchers identified 62 relevant observational studies. After quality control, only 23 studies — totaling 1.4 million participants — were analyzed further. Since there’s enough numeric info from at least two studies, meta-analyses were done.
Results showed that decreasing unprocessed and processed red meat intake led to fewer deaths from diseases. However, while statistically significant, this effect is of small clinical significance and observational — rendering the evidence of “low- to very low-certainty evidence,” stated the authors.
To understand this table, again, understanding 95% CI is important. If the 95% CI stays crosses 1.00, there’re no effects. If the 95% CI stays below 1.00, there’s reduced risk. If the 95% CI exceeds 1.00, there’s increased risk.
For example, if the relative risk is 0.85 with a 95% CI of 0.80–0.87, it means there’s a 15% reduced risk. But if the relative risk is 0.85 with a 95% CI of 0.80–1.1, it means there’re no statistically significant effects since the outcome ranged from 20% reduced risk to 10% increased risk about 95% of the time.
Therefore, based on the table above, it can be observed that lowering intake of unprocessed red meat:
- ↓ risk of cardiovascular mortality by 10% (translating to 4 fewer cases/1000 persons compared to cases in the control group).
- ↓ risk of stroke incidence by 6% (1 fewer case/1000 persons).
- ↓ risk of myocardial infarction by 7% (3 fewer cases/1000 persons).
- ↓ risk of type 2 diabetes by 10% (6 fewer cases/1000 persons).
- No effect on all-cause mortality, cardiovascular disease, and fatal stroke incidence.
The same concept applies to the next table below on processed red meat, wherein limiting its intake:
- ↓ risk of all-cause mortality by 8% (translating to 9 fewer cases/1000 persons compared to cases in the control group).
- ↓ risk of cardiovascular mortality by 10% (4 fewer cases/1000 persons).
- ↓ risk of stroke incidence by 6% (1 fewer cases/1000 persons).
- ↓ risk of myocardial infarction by 6% (2 fewer cases/1000 persons).
- ↓ risk of type 2 diabetes by 22% (12 fewer cases/1000 persons).
- No effects on risks of cardiovascular disease and fatal stroke incidences.
The same concepts apply to cancer. Due to the extensive tabulated data involving many cancer types in the paper, I will just list the main findings of their meta-analysis of 118 observational studies — totaling 6.1 million participants — of satisfactory quality.
Reducing the intake of unprocessed red meat:
- ↓ risk of overall cancer mortality by 7% (translating to 7 fewer deaths/1000 persons compared to cases in the control group).
- No effects on risks of overall cancer incidence or prostate cancer mortality.
- No effects on risks of prostate, esophageal, gastric, colorectal, pancreatic, or breast cancer incidences.
Reducing the intake of processed red meat:
- ↓ risk of overall cancer mortality by 8% (translating to 8 fewer deaths/1000 persons compared to cases in the control group).
- ↓ risk of prostate cancer mortality by 23% (1 fewer case/1000 persons; yes, just one, apparently).
- ↓ risk of esophageal cancer incidence by 30% (2 fewer cases/1000 persons).
- ↓ risk of colorectal cancer incidence by 7% (1 fewer case/1000 persons).
- ↓ risk of breast cancer incidence by 10% (5 fewer cases/1000 persons).
- No effects on risks of gastric, colorectal, and pancreatic cancer mortality.
- No effects on risks of overall cancer incidence.
- No effects on risks of oral, gastric, small intestinal, hepatic, pancreatic, ovarian, endometrial, or prostate cancer incidences.
Based on quality evaluations, the authors chose not to emphasize much on dose-dependent analyses — i.e., the degree of risk with increasing red meat intake. “In our assessment of the certainty of evidence, we did not rate up for dose-response, given the potential for residual confounding,” they explained.
For reference, they did find that too much processed red meat intake further increased the risk of type 2 diabetes. But it lowered the risk of breast cancer incidence instead, despite that 3 fewer servings per week was calculated to decrease breast cancer incidence by 10%, as detailed in one of the bulleted points. (Yes, that’s weird; after all, it’s observational data). For other diseases, the dose-response analyses did not make much of a difference compared to results from eating 3 fewer servings of red meat per week.
To sum up, while existing observational studies of adequate quality found reduced risks of some diseases following reduced red meat intake, the effect sizes were tiny, which is only detectable when looking at huge populations. For example, in one of the bulleted points above, reducing the intake of processed red meat decreases eight cases of all-cause mortality per 1000 persons, so the chances of that applying to one person are 0.8%.
Plus, the observational nature of the evidence renders it less clinically meaningful. There might be uncontrolled factors confounding the results, the authors pointed out, such as socioeconomic status or geographical region. Most crucially, causation can only be drawn from RCTs, not observational studies. And no high-quality RCTs have confirmed that red meat is tied to human diseases.
Given the collective evidence, the authors recommended that people continue eating processed or unprocessed red meat. But these are “weak recommendations, low-certainty evidence,” the authors stated. “Our weak recommendation that people continue their current meat consumption highlights both the uncertainty associated with possible harmful effects and the very small magnitude of effect, even if the best estimates represent true causation, which we believe to be implausible.”
Indeed, as mentioned, observational studies shouldn’t be regarded as conclusive. “Observational studies will continue to be limited by challenges of accurate measurement of diet, the precise and accurate measurement of known confounders, and the likelihood of residual confounding after adjusted analyses,” the authors wrote.
The authors and journal editors then recruited 14 panel members from diverse health, medicine, and nutritional backgrounds from 7 countries with no conflict of interest to review and vote on the recommendations. Only 3 members opposed the recommendations, and the remaining 11 agreed.
The authors also checked nine other meta-analyses of observational studies published previously. While these meta-analyses found greater health benefits of reducing red meat intake, their statistical analyses and study design were flawed. The authors explained, “Some of the meta-analyses relied on extreme exposure categories (48, 49, 50)…and many did not consider absolute effects or certainty of evidence on an outcome-by-outcome basis (15, 16, 49, 50, 51, 52, 53, 54).” Hence, the current one is arguably the most up-to-date and rigorous systematic review we can rely on.
Finally, no research is flawless, and the authors pointed out the limitations of their systematic reviews and meta-analyses.
- No matter how rigorous the quality control of studies was performed, most evidence was still observational.
- Although the included studies distinguished between low and high intake of processed or unprocessed red meat, other food or diet patterns differ a lot between studies.
- There was not enough high-quality evidence to separate their analyses based on cooking methods, such as boiling vs. grilling or factory- vs. farm-raised meat.
- Animal welfare and environmental concerns were beyond the scope of their investigation.
A 2018 article in the Healthline reviewed several meta-analyses and RCTs on this topic too, although quality control of research papers was probably not conducted. In brief, they concluded that unprocessed red meat, especially when not overly grilled, is not harmful and may even benefit our health and support muscle growth. After all, red meat contains essential nutrients, such as bioavailable vitamins B6 and B12, iron, zinc, and certain amino acids.
Obviously, some experts disagree with this red meat paper —mainly drawing on years of research that consistently finds an association between red meat and diseases. Those ‘years of research’ are observational studies, however, which shaped the present dietary guidelines.
Overall, this systematic review’s conclusion still stands among the scientific and medical communities. If not, the paper would have been retracted by now, not to mention that Annals of Internal Medicine is a world-ranking medical journal.
An article in the New York Times made an important point about this red meat paper. “It is important to recognize that this group reviewed the evidence and found the same risk from red and processed meat as have other experts,” Marjorie McCullough, a senior scientific director of epidemiology research at the American Cancer Society, stated. “So they’re not saying meat is less risky; they’re saying the risk that everyone agrees on is acceptable for individuals.”
Since observational nutritional research has many inherent flaws, its results must not be regarded as conclusive. So, “People should be aware of the uncertainty and make their decisions based on that awareness,” advised David Allison, dean and provost professor at the Indiana University School of Public Health-Bloomington.
The red meat paper authors believed that it’s ‘implausible’ that red meat is a causative factor of diseases. It’s implausible because true causation would reflect a higher degree of change, even in observational studies. For example, it’s observed in human populations that habitual cigarette smoking increases lung cancer risk by 1500–3000% and up to 4000%. Thus, we can confidently say that smoking causes and increases the risk of lung cancer.
Perhaps we should focus on other health aspects in disease prevention, such as exercising and keeping a healthy body fat %. Having excess fat around organs is also a great risk factor for all sorts of diseases. Paying attention to these factors, plus avoiding smoking cigarettes or drinking alcohol, would most likely do much more good than stressing over red meat.
Thanks to Joe Duncan for informing me of the conflicting findings between red meat and cancer, which probed me to write this article.