“To raise new questions, new possibilities, to regard old problems from a new angle, requires creative imagination and marks real advance in science.” Albert Einstein
World Health Organization (WHO) have stated that processed meats are on par with cigarettes as a “convincing” cause of cancer and that fresh red meat is a “probable” cause. These statements overstate the evidence and mislead the public about their cancer risk from eating meat. Driving the WHO conclusion was the epidemiological data. Epidemiology is a science that can establish associations but seldom cause and effect.
Epidemiological science’s biggest success was to link smoking to cancer as a “convincing” carcinogen in 1986. In that case, heavy smokers had a 9-to-25 times greater risk of contracting lung cancer than did non-smokers. This is a “relative risk” big enough to give researchers confidence that the association was a real. The decisions on meat were based on relative risks of 1.17 to 1.18, a tiny fraction of those for smoking! To keep things in perspective: for colon cancer, which was the focus of the WHO report, the absolute risk of contracting this cancer in one’s lifetime is less than 4.5%. An increased relative risk of 1.17 raises the absolute risk to no more than 5.3%! Is this a big enough risk to reduce a nutrient dense food such as red meat? I see this as a weak association.
The reason that weak associations are untrustworthy is that they could very well be due to bias but it has to do with genetics, lifestyle, and socioeconomic status. For example, vegetarians tend to be more alert to good health: they smoke less, exercise more, and have a higher socioeconomic status. By contrast, meat-eaters over the past 30-plus years are people who ignore their doctor’s orders and are likely to be engaging in other high risk behaviours (smoking, not exercising, drinking heavily) of which alone or in combination might explain the small relative risks associated with meat-eating. Bias against red meat is another factor easily observable in the scientific literature and the popular press.
Another important point to be made is that these foods based epidemiological studies rely heavily on a tool called a Food Frequency Questionnaire (FFQ). This is a tool that relies heavily on patient recall and is subject to large amount of bias. There have been several studies showing that FFQ are extremely unreliable and should not be used to make conclusive statements where only a small increase in relative risk is noted. Additionally, the weak associational data used to develop those WHO statements lump all types of red meat and red meat products together. I.e., grass-fed steak is not distinguished from a meat lovers pizza. Even if we were to give credit to epidemiological data, it’s very likely that the “high red meat” intakes were actually high in all types of processed foods associated with red meat intake (e.g., pizza, pies, sausage rolls, pastries, etc.).
Randomized clinical trials (RCTs) provide far more trustworthy evidence. It is therefore perplexing that this week’s WHO document does not even mention the relevant data: two large, multi-year RCTs, both funded by the National Institute of Health.
- The first was Polyp Prevention Trial, which tested a high-fibre, high-fruit-and-vegetable, low-fat diet on more than 2,079 people for four years. In this multi-centre trial, the intervention group significantly decreased red meat and processed meats, replacing them instead with chicken, yet researchers found no effect of this intervention, at the end of the trial or at the eight-year follow up, on the recurrence of colorectal cancer.
- The second was the Women’s Health Initiative, one of the largest randomized controlled trials ever conducted. The WHI tested a diet high in fruits and vegetables and low in fat, on nearly 49,000 women over 8 years. At year three, the only one for which the food data was published, the women on the low-fat diet reduced red meat by 20% compared to controls, a statistically significant amount, yet at the end of the trial, there was no effect on any of type of cancer, including colorectal cancer, ovarian cancer, endometrium cancer, or breast cancer. It’s possible that these trials didn’t last long enough to see cancer develop, but they were both designed as cancer trials.
These RCTs remain the most rigorous data to date, and neither support the hypothesis that red or processed meat causes cancer. To create and implement public policy requires objective data, at this stage I would argue this is lacking.