What’s the truth behind the headlines? Can we eat bacon or not? Today the WHO issued a news release announcing ‘the consumption of red meat as probably carcinogenic to humans (Group 2A), based on limited evidence that the consumption of red meat causes [colorectal in the main] cancer in humans…’
They added, for good measure, processed meat is ‘carcinogenic to humans, based on sufficient evidence in humans that the consumption of processed meat causes colorectal cancer.’
The media reported on this at length, causing flurries and consternation. Of course, bad news sells, but why could the media not really read what the actual paper said and ask some proper questions? Even the UK’s flagship station Radio 4 fell prey to bad science on The Today Programme, The World at One, PM and the 6 o’clock news. They did ask ‘experts’ for their views, but the guests were patently not good scientists and trotted out the usual ‘we must eat less red meat’ maxims with no facts or data to back up what they were saying.
Here is the paper: Carcinogenicity of consumption of red and processed meat and here is the news release IARC Monographs evaluate consumption of red meat and processed meat from the WHO’s International Agency for Research on Cancer.
Let’s look at the truth behind the headlines. Before we do, let me say that if I see bacon with added sugar, I don’t buy it. Also, if I can find naturally cured bacon (ie, without nitrites and other nasties), I will choose that in preference; finding it is difficult, however, and this is where government policies can help make the changes.
UPDATE – 20 DEC 2018: Here’s a more recent report from the Guardian, and it’s well worth a read – https://www.theguardian.com/news/2018/mar/01/bacon-cancer-processed-meats-nitrates-nitrites-sausages. In the end we discover we can have nitrate- and nitrite-free bacon without being poisoned with botulism, but it also matters what else you eat besides.
But back to every-day bacon.
The data on bacon
The data for the review paper were taken from observational studies rather than meta-analyses of randomised controlled trials (the highest study standard). We are already at a disadvantage here; for example, stop now and try to recall what you have eaten over the last week (and in what quantities), and do the same for three weeks ago… and for last year. Unless you have been keeping a food diary, it’s hard to recall what you ate with any precision, isn’t it?
Of course, using a randomised controlled trial to look at how a certain amount of processed meat per day makes a statistically valid difference (or not) is really quite difficult, not least because it would be very expensive to conduct. What they do instead is take ‘baseline’ samples (blood profiles, weight, blood pressure, etc) from participants and follow them up for a period of time ranging from days and weeks to years. Some of the questionnaires are retrospective*.
The researchers then look to see if any patterns have emerged.
News agencies and the media read these papers and if there isn’t a proper pattern they take a few data and statements in isolation and make up their reports from the cherry-picked material.
Causation and association are often confused
Why, I don’t know; I think the media just do it to drum up headlines. (Either that or they have some very bad or stupid reporters.)
If you touch a hot stove top and get a burn, we all know the heat from the stove caused the burn. If you were wearing a green jumper at the time, the jumper is associated with the burn… but it did not cause it. The media, however, will find an expert who will warn against wearing green jumpers because the risk of burns goes up.
People who eat meat may get colorectal (bowel) cancer. We cannot say, given the trials undertaken, that red meat causes the cancer – only that there is an association. It could just as easily be the overconsumption of grain- and sugar-based foods the people consume… or something else entirely… or a number of factors working together.
Absolute or relative risk?
These are entirely different things, but often (deliberately or through ignorance) misread by the media. Relative risk tells you nothing about actual risk… as my husband often hears me shouting at the radio.
The ‘risk’ of something happening is its chance of taking place, and no guarantee that it will. Thus if, for example, you pick a card at random from a pack of playing cards, you have a one in four absolute ‘risk’ of picking out Hearts (ie, 25%) or a one in 52 absolute ‘risk’ of picking the King of Diamonds (just over 2%). Of course, it’s much more probable that you will pick a card that isn’t a Heart (three in four) or the King of Diamonds (51 in 52).
If I now tell you the ‘risk’ of picking any Heart is over ten times greater than picking the King of Diamonds, I have given you a relative risk; that is, I have compared two risks and come up with a relative risk – a comparison of risks – to tell you how much more likely you are to pick a Heart than a King of Diamonds.
(Want to know more about relative and absolute risk? Have a decko at this site, which has an animation showing there are 2,845 ways (yes!) to talk about a risk. As I write, the first example they show is about bacon sandwiches! You can click ‘absolute’ and ‘relative’ risk buttons to see the stats.)
Let’s look at one of the statements in the news release:
The experts concluded that each 50 gram (1.7oz) portion of processed meat eaten daily increases the risk of colorectal cancer by 18%.
Eek! 18% is a big scary, eh? We had better stop eating bacon now!
Or, because what they said is relative risk, instead of chucking the rest of the bacon in the bin, we could look at the absolute risk.
Cancer Research UK tells us the absolute risk rate of dying from colorectal cancer is 47 people in 100,000 people who are over the age of 50. (This cancer is much less likely in younger people.) That’s not 47 out of every 100 or 1,000 – that is 47 out of every 100,000 (one hundred thousand), or one person in 2,128.
(I’m not saying bowel cancer isn’t serious, it is – and if you want to find out more about its incidence and mortality/recovery, it’s worth reading the whole page.)
OK, if there really is an 18% relative difference (and I’m not convinced here given the design of the study), what change would eating processed and red meats make to the absolute risk?
It would add eight people (well, 8.46 to be exact) per 100,000; that is, we would increase the risk from 47 out of every 100,000 to 55 out of every 100,000.
Put another way, the risk goes up from one person in 2,128 to one person in 1,800.
To put this into perspective here are some stats on the risk of dying from particular causes**.
Salami does not exist in isolation
We are not told what else the participants ate, nor indeed what their lifestyles were like. Someone who pops to the charcuterie to get some Parma ham and peppered salami for their Sunday lunch and follows it with a walk is not the same as someone eating cheap and nasty hot-dogs in a white bun and sugary ketchup whilst slumped in front of the TV
….and they are not the same as anyone in between these camps
…and someone who pops to the charcuterie to get some Parma ham and peppered salami for their Sunday lunch and follow it with a walk is not the same as their neighbour who does the same.
Why? One of the first things we were taught in science lessons at school was to only test for one variable at a time (we didn’t learn multivariate analysis techniques at the time). These studies do not do that. Confounding factors (other variables) include:
- Weight, BMI and waist measurements of the participants
- How much visceral fat people have
- How much processed foods altogether people eat
- What sort of processing the meat has had; for example, there is a difference between salt cured, sugar cured, nitrate cured, just dried, fermented or smoked
- How much sugar people have in their diet (including foods that turn quickly into sugars, such as bread, and including drinks high in fructose); sugars and the subsequent inflammation are more closely associated with cancers than meat is
- How much people use polunsaturated vegetable oils (such as sunflower) for cooking with; these oils transmogify easily into harmful tansfats
- How much unprocessed foods people eat
- The provenance of meats – what the animals have been fed will alter their own nutrient makeup; for example, grass fed, free range, organic beef used in burgers is not the same as that found in ‘4 burgers for 50p’ where the beef has come from corn fed animals kept in sheds under stressful conditions and injected with various antibiotics and hormones
- Genes – your genes mean you will have a different outcome even if you eat and do exactly the same things as your friend
- Gut microbes – a different biome (microbe population) means you will have a different outcome even if you eat and do exactly the same things as your friend
- The state of their gut – for example, do they have colitis or an ulcer, or ‘leaky gut’?
- Their general state of health – are they diabetic or issue-free, for example?
- What, if any, medications they are on
- How much exercise people do, and the type and intensity of exercise
- How much stress a person is under (or not)
- How many soft drinks people consume
- Their age(s)
- Pollution/environmental toxins they are subjected to
- … and much, much more
Meat is good
I’m not entirely sure why red meat was chucked in with processed meats anyway by the reporters/expert speakers. It supplies the nutrients we need, and you need to eat much less of it than non-meats to get your daily requirements.
What was actually said in the paper was:
Red meat contains high biological-value proteins and important micronutrients such as B vitamins, iron (both free iron and haem iron), and zinc. The fat content of red meat varies depending on animal species, age, sex, breed, and feed, and the cut of the meat.
and
Cooking improves the digestibility and palatability of meat, but can also produce known or suspected carcinogens, including heterocyclic aromatic amines (HAA) and PAH. High-temperature cooking by pan-frying, grilling, or barbecuing generally produces the highest amounts of these chemicals.
This does not say red meat per se is bad for you! See this bit:
a majority of the Working Group concluded that there is sufficient evidence in human beings for the carcinogenicity of the consumption of processed meat.
Chance, bias, and confounding could not be ruled out with the same degree of confidence for the data on red meat consumption, since no clear association was seen in several of the high quality studies and residual confounding from other diet and lifestyle risk is difficult to exclude.
and
The Working Group concluded that there is limited evidence in human beings for the carcinogenicity of the consumption of red meat.
I recommend you go for meat that has been reared naturally, making sure you cook it at lower tempertaures (eg, slow roast or in a slow cooker). In addition, consider keeping processed meats more as a treat and choose those that have been preserved naturally.
Two absurdities
What really makes me laugh is the thought of all our red meat-eating ancestors dropping dead from cancer. They didn’t.
What also makes me chuckle is the fact humans are red meat! That’s what our muscle is: are we indeed bad for ourselves?
*Here’s an example of a food frequency questionnaire: http://www.srl.cam.ac.uk/epic/epicffq/ffq.pdf; this is European Prospective Investigation into Cancer questionnaire.
**Here are two other sites looking at absolute death rates: http://www.dartmouth.edu/~ears5/handouts/Prob_dying6_25_99.html and http://www.riskcomm.com/visualaids/riskscale/datasources.php s