11/03/2015 12:54 EST | Updated 11/03/2016 05:12 EDT

What The Headlines Didn't Tell You About That Bacon And Cancer Study

They put processed meat into Group 1 -- "exposures known to be carcinogenic to humans." But categorization caused misunderstandings. The report simply put processed meat in the same category (Group 1) as cigarette smoking, but did not claim it was as dangerous as smoking.

Tannis Toohey via Getty Images
TORONTO, ON- OCTOBER 26: World Health Organization says bacon, sausage and other processed meats cause cancer. WHO says bacon, sausage and other processed meats cause cancer. at the in Toronto. October 26, 2015. Tannis Toohey/Toronto Star (Tannis Toohey/Toronto Star via Getty Images)

You've heard about the WHO's announcement that eating processed meat is associated with colorectal cancer. As an epidemiologist and educator, I needed to weigh in.

How Was The Study Done?

The report is a systematic review, thus its authors did not perform any new experiments. Rather, they reviewed existing studies, assessed them for strengths and biases, then synthesized an overall estimate of risk. The hierarchy of quality of evidence places randomized controlled (or clinical) trials (RCTs) near the top. We call RCTs the "gold standard" of evidence, though a systematic review comprised of good RCTs is oft considered better than a single RCT.

The authors: "The greatest weight was given to prospective cohort studies done in the general population. High quality population-based case-control studies provided additional evidence." Cohort and case-control designs are not RCTs, and are therefore lower-quality designs with poorer "internal validity." This means that while associations can be measured using these designs, causality is more difficult.

In a cohort design, researchers would identify some people who eat processed meat and some who do not, then wait to see which group manifests a greater proportional incidence of cancer.

In a case-control design, they would identify some people with cancer and some people without, then investigate to determine how much meat both groups had consumed in the past.

Neither scenario can control for other potentially causal factors, such as a family predisposition toward cancer. There is a concept called "confounding" wherein a variable either masks a real relationship or creates the impression of an illusory relationship. Is it possible that people who consume a lot of processed meat are also people who engage in other unhealthy activities, like smoking, drinking, or inactivity? Tobacco, alcohol and exercise would then be confounders, creating a false impression of an association between meat and cancer.

Knowledge of causality is often not required for public health action. If meat is a strong risk factor for cancer, then reducing consumption will quite possibly result in lowered cancer risk, regardless of the causal pathway.

The authors assessed 800 studies, finding that 18 cohort and 14 case-control studies were of sufficient quality to be examined. Their assessment of the risk posed by meat was placed into the appropriate risk category, along with other carcinogenic exposures.

They put processed meat into Group 1 -- "exposures known to be carcinogenic to humans" -- and red meat into Group 2A, "exposures probably carcinogenic to humans."

But categorization caused misunderstandings. The Guardian's misleading headline was, "Processed Meats Rank Alongside Smoking as Cancer Causes." The report simply put processed meat in the same category (Group 1) as cigarette smoking, but did not claim it was as dangerous as smoking.

Statistical vs Clinical Significance

They also conducted something called a meta-analysis, collecting individual risks from the reviewed studies to compute a grand statistical estimate of overall risk. This is the number that has been reported in the media: from data extracted from 10 cohort studies, the overall increase in risk of colorectal cancer was 17 per cent from red meat and 18 per cent from processed meat.

These numbers were statistically significant, which is not the same as clinical significance. Statistical significance means that if the study were repeated an infinite number of times, there is a vanishingly small probability that the measured difference in risk would be zero.

They also reported something called a confidence interval: if the study were repeated an infinite number of times, 95 per cent of the time the increased risk posed by red meat would be somewhere between five per cent and 31 per cent, with the most likely estimate being 17 per cent. Similarly, the increased risk posed by processed meat is between 10 per cent and 28 per cent, most likely 18 per cent.

Thus, those who eat processed meat have an 18 per cent higher likelihood of developing colorectal cancer than those who do not eat it.

Relative vs Absolute Risk

The biggest misunderstanding concerns the nature of relative risk. If I say the incidence of malaria in a large city tripled since last year, you might be horrified. But looking more closely, you see that the incidence of malaria did in fact increase -- from one case to three cases.

The proportion of new cases in a given group is called an absolute risk. The ratio of two absolute risks is a relative risk, or risk ratio.

Which is more important, the relative or absolute risk? That depends on your clinical or policy question.

An 18 per cent increase in risk from a non-meat eating to a meat-eating state is a relative risk, representing a risk ratio of 1.18. The question then is, how much baseline risk is there, initially? It's been argued that pretty much everything is carcinogenic, meaning that there exists a risk for colorectal cancer even for vegans. But if that baseline risk is small, then an 18 per cent increase really is not appreciable.

In fact, it's already been reported that eating these meats increases lifetime risk for colon cancer from a baseline of five to six per cent. Is a one per cent increase a big deal? Again, that's debatable.

According to the WHO, there are about 8 million cancer annual deaths globally. About 700,000 of those (less than nine per cent) are colorectal cancers. In the USA and Canada, about 36 per cent of new diagnoses of colorectal cancer will be fatal, compared to a global fatality rate of 51 per cent.

To keep things in perspective, in the U.S. more people will be killed by unintentional poisoning or influenza. But among cancers, the colorectal type is usually the 10th or 11th most proportionately fatal.

What To Do?

While avoiding processed meat reduces risk, there is another option that offers even greater protection. Screening for the early stages of colon cancer is enormously beneficial. If it's identified early enough, the five-year survival rate is 90 per cent (compared to 11 to 12 per cent at much later stages of diagnosis).

So, you can probably continue to have processed meat occasionally without exposing yourself to an appreciable risk of cancer, as long as you are conscientious about getting screened. That's my take on this, as a non-clinician numbers wonk.

Of course, there are other reasons to eschew the meat -- environmental, economic, or ethical -- but that's an entirely different conversation!


Photo gallery Processed meats that 'are carcinogenic', according to WHO See Gallery