The following originally appeared on The Upshot (copyright 2015, The New York Times Company)
Smoking tobacco causes cancer. So does eating salted fish, drinking alcohol, breathing polluted air and being exposed to the sun. All of these things are classified as cancer-causing by the World Health Organization.
This week, processed meat has been added to that list, meaning that the world’s attention has been focused on whether everyone should stop eating bacon, sausage or various charcuterie.
The short answer is no, you’re probably fine. As with many pronouncements about food, this one is being overhyped by the news media, and potentially over-interpreted by scientists.
I wrote about red meat here at The Upshot back in March, focusing mostly on the cardiovascular risks, rather than the cancer risks. But I still highlighted and discussed some key studies, including one that found that eating meat, especially processed meat, was associated with increased cancer and mortality in people age 50-65. As I said, it also found that the opposite was true in people over 65 years, but that gets mostly ignored.
Based on epidemiologic data like these, 700 studies on red meat and cancer and 400 more on processed meat, the International Agency for Research on Cancer felt comfortable making the declaration that processed meat causes cancer and that red meat probably causes cancer. The specific meta-analysis of cohort studies they cited for their relative risk point estimate of 1.18 was published in PLOS One in June 2011.
As Geoffrey Kabat pointed out, it’s worth noting that 25 years ago, the I.A.R.C. ruled that coffee was “possibly carcinogenic”. Despite the massive amount of evidence to the contrary that has been published since, they haven’t changed their minds. In fact, of the 985 agents the I.A.R.C. has classified, only one has been labeled “probably not carcinogenic to humans”.
I’ve written before about the dangers of making assumptions about causes from observational studies. My review of the literature and that of the experts at the I.A.R.C. use the same data. We just reach different conclusions. I also find it hard to ignore the fact that randomized controlled trials do exist. The Polyp Prevention Trial failed to show any effect of a low-fat, high-fiber, fruit-and-vegetable diet on tumor recurrence even after eight years of follow-up tests. The Women’s Health Initiative, which involved almost 50,000 women, also could not show that a change in diet reduced the risk of colorectal cancer after more than eight years.
The W.H.O.’s announcement makes trials like these much less likely to occur in the future, even though they haven’t shown meat to be harmful. After all, if the W.H.O. has declared that processed meat causes cancer, how can we ethically randomize people to eat it?
But let’s take the W.H.O. officials at their word that an actual link exists between processed meat and cancer. What’s missing from the discussion is any talk about the magnitude of that risk.
The I.A.R.C. delivers an opinion only on whether a link exists, not on how strong it is. That’s how tobacco smoke, an unequivocal and large risk, can be lumped in with alcohol, which probably has some benefits, and a rather small risk. The same goes for the sun, which surely can cause skin cancer, but which isn’t something anyone would tell you to avoid altogether.
The I.A.R.C. published that for each 50 grams of processed meat eaten daily, the risk of colon cancer goes up by 18 percent. That sounds scary. But that’s a relative risk increase. What we really need to know is the absolute risk increase. I went to the National Cancer Institute’s colorectal cancer risk assessment calculator, and plugged in all of my information. I had to say I’m 50, because it doesn’t have risks for people younger than that. It determined that 50-year-old me has a lifetime risk of 2.7 percent of getting colon cancer.
This means that, if I buy what the W.H.O. is saying, if I decided today to start eating an extra three pieces of bacon every day for the next 30 years, my risk of getting colon cancer might go from 2.7 percent to 3.2 percent. In other words, if 200 people like me made that decision, one extra person might get cancer. The other 199 would be unaffected.
That’s not nearly as scary as what many headlines would have you believe. Even with all that processed meat (which I am not going to eat), a 0.5 percent increase in the lifetime risk of something is still pretty small. Eating it occasionally, which is more likely, is not going to affect my lifetime risk measurably at all.
Let’s be clear. Rational people are willing to accept small risks of harm to obtain something they value. The example I always like to use is cars. The No. 1 killer of children in the United States is, by far, accidents. Every time we put a child in a car, we are exposing them to the thing most likely to kill them.
We don’t see headlines like “Cars Found to Kill Kids in Record Numbers!” or “Putting a Child in a Car Increases Their Risk of Death by 20 percent!” That’s because we have all recognized that while cars do increase the risk of a bad outcome, the gains from driving outweigh the potential and very small absolute risks of death.
The same is true of many things. I like Scotch. I like skiing. I like the occasional steak. All of these things may increase my absolute risk of death someday by some very tiny amount, but the daily happiness and satisfaction I gain from them outweigh those future, and most likely very small, risks.
What I said before about red meat still holds, as do my recommendations for healthy eating. If you’re consuming multiple portions of processed meat a day, then you may see some small benefit in the lifetime risk of cancer by cutting back. But if you’re like most people I know, enjoying bacon or prosciutto a couple of times a week, this news most likely doesn’t affect you at all.