Internet Myths about Red Meat

 "an odour assailed his nostrils, unlike any scent which he had before experienced." from "A Dissertation Upon Roast Pig" by Charles Lamb

Charles Lamb wrote that humorous fiction about the first time roast meat was accidentally tried. While the above quote describes the first accidentally roast pig, it could easily describe a grilled rib-eye steak. Grilled, bloody steaks are one of the great pleasures in life for many of us. 

There is enough evidence that saturated fat, and especially trans fats, are not heart-healthy. Cardiovascular disease is about 20 times the killer of elderly men compared to prostate cancer. Most of us take statins and exercise (I hope), so the occasional rib-eye won't clog our veins, as long as we eat it only occasionally. Of course, ask your cardiologist if you've already had evidence of obstruction.

There are many reasons to change diet, including cardiovascular health, and diabetes. Many men with prostate cancer use some form of androgen deprivation therapy (ADT) and are in danger of getting metabolic syndrome, especially if they are already overweight and sedentary. Men using ADT should monitor caloric intake and increase exercise, with their doctor's permission. Many patients like to change diets because it gives them a feeling of control in a situation where they have lost control of their own bodies. Others torture themselves (or are tortured by well-meaning loved ones) by depriving themselves (one more deprivation!) of their favorite foods. Eating and sharing meals with others is one of the pleasures of our lives, and is part of our social cohesion. But what is the evidence that diet plays a role in lengthening survival due to prostate cancer itself?

There are dietary fads. Current fads include plant-based/vegetarian/vegan diets, Mediterranean diet, ketogenic/low carb diet, alternate-day fasting, macrobiotic diets, microbiome diets, and so-called "cancer diets" that cut out sugars, animal fats, red meat, chicken, eggs, and dairy products, or that add foods with certain phytonutrients. It is extremely difficult to prove that diet has a causal effect in cancer progression. 

Associations between red meat intake and breast or colorectal cancers have been noted. Patients are highly motivated to seek out dietary changes they can do to regain control of their bodies.  Most patients search the Internet for clues but don't know enough about research methodology to distinguish gold from garbage. Many prostate cancer patients (and doctors, who should know better) have succumbed to the "conventional wisdom" that grilled red meat is bad for their cancer too. Is there any truth to that conventional wisdom? Here's the evidence.  

Only "Level 1" can prove that red meat consumption contributes to prostate cancer. (Follow these links to understand Levels of Evidence and GRADE). Higher levels of evidence, that are also high GRADE, replace lower levels of evidence. There is no Level 1 evidence (well-done randomized clinical trial), but there is some Level 2 evidence. A cohort study is Level 2 evidence only if it is prospective; i.e., the hypothesis to be investigated is formulated from the start. Longitudinal retrospective cohort studies are Level 3 evidence. See below for reasons why such studies almost always provide conflicting results.

Level 2 Evidence

The NutriNet Santé Cohort Study is a huge (61,476 adults over 35) prospective web-based cohort study in France. France has the highest rate of internet use of any large country, and it cuts across age and socio-economic classes. Participants fill out food questionnaires (3 random records of 24-hour food intake during a two-week period). They also provide information on physical activity, BMI, height, sex, age, lifestyle, health status, family history of cancer, and socio-economic conditions. They also may provide blood and urine samples and get a physical exam. Red meat can be beef, pork, veal, or lamb. Processed meat can be cured meats, sausage, bacon, ham, salami, spam, etc.

With a follow-up of 4 years, and 222 prostate cancers in the cohort, Diallo et al. reported:

• Red meat intake was not associated with prostate cancer risk
• Processed meat intake was not associated with prostate cancer risk
• There was no association for any Gleason score.

Also, Level 2 evidence is provided by the UK Biobank Mendelian Randomization Study. They found the genomic mutations associated with red and processed meat intake and found the correlations with incidence of prostate cancer in over 140,000 people.

• Red meat intake was not associated with prostate cancer risk
• Processed meat intake was not associated with prostate cancer risk

Allen et al. reported the results of The European Prospective Investigation into Cancer and Nutrition (EPIC). It was a very large cohort study with 142,251 men and 2,727 cases of prostate cancer (none at enrollment) and 8.7 years median follow-up. It randomly sampled the cohort to ask for 24-hour dietary intake. They found that:

• Red meat and processed meat intake was not associated with prostate cancer risk
• There was no association for advanced or localized PCa or by Gleason score

Level 3 Evidence

Cross et al. prospectively reported on red meat associations with prostate cancer in the PLCO screening trial. They found no association with any meat products, but a small positive association with very well-done meat.

Contradictory findings were reported in several observational cohort studies.

The MEAL Randomized Clinical Trial (RCT) - Level 1 Evidence

The MEAL RCT is the only Level 1 evidence we have on dietary causes of prostate cancer.

Parsons et al. randomized 478 low-risk men who were using active surveillance at 91 cancer centers to one of two groups:

• The intervention group received telephone prompts to eat 7 or more servings of vegetables per day, including 2 servings of each of cruciferous vegetables and tomato products.
• The control group only received written information

They all kept detailed food diaries, and plasma carotenoid assessment was used to prove compliance. After 2 years, the intervention group increased their consumption of vegetables and reduced consumption of red meat and animal fats. However, there was no difference in time to progression.

This trial proves that increasing vegetable intake has no effect on even the "lightest" prostate cancer. It also shows there is no "dose response" due to meat consumption. A "dose response" means that the more one uses some intervention, the greater the effect will be. We see this frequently with drugs - the higher the dose, the greater the response, up to a certain point; and the lower the dose, above some minimum, the lower the response. Some argue that meat-eating has no dose response but is binary -- eliminating all non-vegetable protein sources can make a difference. 

A typical dose-response curve:

Other Vegetarian Trials

It is worth mentioning Dean Ornish's  Prostate Cancer Lifestyle Trial (PCLT)  randomized 93 men on active surveillance to either:

• A vegan diet with complex carbs and very low fat. They also exercised, took a variety of vitamins and food supplements, managed stress, and met in groups.
• The other group just had usual care.

After 2 years of follow-up, fewer men in the treatment group had prostate cancer treatment. There was no difference in PSA. How much of the difference is due to exercise, may be speculated upon, and is the argument against unwarranted claims, such as this one. Early results of the ERASE trial suggest that exercise can make a difference. Cardiovascular parameters improved more in the intervention group.

A pilot trial among 36 recurrent men randomized them to either:

• 11 weeks of mostly plant-based foods and oily fish, with no or less animal proteins. They also practiced mindfulness.
• usual care

After 3 months, there was no difference in the rate of PSA change. 

There have been no other randomized studies in men with prostate cancer. There have been a couple of prospective trials without a control group to compare results.

Why Only Level 1 Evidence is Important to Patients

Level 2 or 3 studies are only hypothesis-generating for other researchers to conduct Level 1 trials. They should never be used by patients to make life-changing decisions. Observational studies suffer from "selection bias." This means that the patients who got the treatment (here, red meat) were in some ways different from the patients used as the control group (here, vegetarians). Researchers use statistical techniques like "propensity score matching" in the hope of correcting for this bias. But a new data analysis proves why they don't work. 

Wang et al. looked at 15 published studies on the topic of red meat and mortality, and at a huge dataset, The National Health and Nutrition Examination Survey 2007-2014 (NHANES), a longitudinal observational study (level 3 evidence) of over 10,000 persons. 

Their review of 15 published studies analyzed in 70 unique ways found:

• Red meat intake was associated as originally analyzed with anywhere from a 37% reduced risk of death to a 131% increased risk of death.
• They were able to re-analyze that data 10 quadrillion different ways (each way used a different set of variables like age, health status, etc.).

They randomly chose 1200 different ways that all seemed appropriate and applied them to the NHANES dataset.

• Analysis of the 1200 methods for analyzing the data found:
• There was no statistically significant association between red meat intake and death. With 95% confidence, it ranged from a 49% reduced risk of death to a 75% increased risk of death.
• In total, 64% of analyses showed a reduced risk of death, while 36% showed an increased risk. Only 4% of analyses had statistically significant results. Of those, 83% showed reduced risk of death, and 17% showed increased risk of death.
• Most analyses resulted in very little association (±10%)

Conclusion: with studies that are low-level evidence, the method chosen for the analysis changes the conclusion. Similar studies of analytic choice have similar results (see this link and this one).

Other Protein Sources

There is no Level 1 evidence for any protein source. Lower level evidence yields conflicting associations for every protein source. Below is just one example of each (note: negative associatiation means consuming more was beneficial, positive association means consuming more was deleterious).

Dairy: There are observational studies that have shown positive, negative, and no correlation with prostate cancer outcomes.

Chicken: Observational studies show no association, a positive association, or a negative association with prostate cancer progression.

Eggs: Observational studies show no association, or a positive association, with prostate cancer progression.

Fish: Observational studies show no association, a negative association, or a positive association with prostate cancer outcomes.

Vegan/Vegetarian: The complete protein source for vegans is high in carbohydrates (e.g., soy, beans and pasta). Observational studies with diets high in soy show no association, negative association, or a positive association. A small randomized trial designed to see if there was any effect of a diet low in carbs on PSADT. It was stopped early because of no effect.

Humans evolved to be omnivores. Our gut bacteria co-evolved and thrive on a varied diet. Plants are necessary for good gut motility.

ADVICE

There is no usable evidence that cutting back on red meat is beneficial. Given the lack of convincing evidence, it is a good idea to:

• Vary one's protein sources (unless cardiologist dictates otherwise)
• Don't deprive yourself! You have cancer -- treat yourself well, at least occasionally.
• If you are on ADT, your metabolism is slower, so consume fewer calories and exercise more.
• Eat plenty of vegetables, especially cruciferous and highly-colored vegetables.
• Avoid vitamins, minerals and supplements unless you are deficient. Get your micronutrients from food. Your body will take what it needs and discard the rest. Don't try to outsmart your body -- you will lose. It has millions of years of evolution on its side.