Wheelbarrows of suspected bad Science. I mean Surveys? This is such Click bait, but it needs to be exposed for the false-science it is.
Even moderate portions of red meat can cause cancer, study finds
Carolus_Holman
(Carolus Holman)
#1
New study says you will get cancer with one slice of bacon
New study says you will get cancer with one slice of bacon
Dread1840
(Chris)
#2
They didnāt even post a link to the study.
-
Healthy user bias
-
Food surveys
-
7th Day Adventist, big food, and pharma
CrackerJax
('Jackie P')
#3
Errrā¦each bottle of beer or small glass of wine raises bowel cancer risk by 8%!@Ć
I should of died in my teens! Iām 60 now.
Ridiculous!
FrankoBear
#4
Please post the rebuttal when it appears.
Junk epidemiology?
It looks like we will be revising absolute and relative risk reporting again.
Dread1840
(Chris)
#5
Instead of giving the guardian clicks, I saved you some: https://academic.oup.com/ije/advance-article/doi/10.1093/ije/dyz064/5470096
FrankoBear
#6
Yes, The Guardian report had the link to the study. Straight away we can see that the abstract is reporting relative risk very similar to previously debunked red meat and colon cancer epidemiological studies.
So much junk science with an agenda in the lead up to the 2020 version of the nutritional guidelines.
Azure
(*Tame Those Ghrelin Gremlins)
#7
I just do the opposite of SAD and nutrition guidelines. Itās made me healthier, thinner and given me more energy lol. I went Keto, eat good amounts of red meat, and eat way more salt then they recommend. Funny how my health has improved. Oh and I do drink a glass or two of Red wine now and then. Iām a real rebel.
PaulL
(You've tried everything else; why not try bacon?)
#8
Iām sure Georgia Ede will be whipping up a rebuttal soon. She already showed the flaws in the WHO report vilifying red meat, and Iām pretty sure the same will apply here, too.
ETA: To say nothing of Nina Teicholz.
djindy
#10
Anyone have some good, scientific reasons why this should be dismissed, other than that it doesnāt agree with our own biases? I think that would be more interesting and useful here.
So far, the only thing that jumps out as a bit odd to me is the comparison numbers:
Participants who reported consuming an average of 76 g/day of red and processed meat compared with 21 g/day had a 20% [95% confidence interval (CI): 4ā37] higher risk of colorectal cancer.
Well, for one, thatās not saying eating any red meat is a problem, as itās only looking at people that do eat red meat it seems. That said, those numbers are kinda odd and specific. Was this average an average for the group or an average for the individual? The former Iām not sure why you would do and what it would mean, unless colorectal cancer is contagious. The latter seems oddly specific to find a set of people for two groups, which both match a specific gram range of average red meat consumption per day, rather than a range.
Likewise, 20% increase risk from a ~362% increase consumption seems like anā¦ interesting correlation. I donāt know what to make of that, honestly. It could be meaningful, I suppose.
Looks like further down they break people into 4 or 5 groups though, but still at intervals I donāt really get. There is this that indicates how they got into the groups:
Participants are categorized according to their intake at recruitment.
Seems to me like it would have been better to categorize them after the data was received, unless they really were trying to fit those very specific g/day amounts for some reason.
Not sure about what this methodology really says though. If you look at Fig. 1, fish, dairy and cheese for some reason increase dramatically in incidents at slightly above the lowest group, but then drop down dramatically as consumption increased a bit.
This seems like probably the biggest confounded though:
Compared with those in the lowest category, participants in the highest category of reported total red-meat intake were slightly older, more likely to be smokers, had a higher BMI and body-fat percentage, had a higher alcohol intake and had lower intakes of fruit, vegetables and fibre. The same was true for processed-meat intake, with the exception of age, which was similar between the two categories.
I donāt want to dismiss this just because of all those factors of the group with more cancer also having more smokers, heavier, more alcohol, less veggies, etc, but these are a lot of confonders which include things already linked to poor health otherwise (and hey, maybe weāll find out that smoking doesnāt really cause cancer, it was more red meat the whole time? But this seems like a lot of unclear data).
So, right now, Iām not sure what exactly this indicates, nor what it means to people on a ketogenic diet whom we figure process meats differently anyway and donāt have the problem of a āmixedā diet. Maybe it does have impact? I donāt know.
OgreZed
#11
When I looked up a ā20% increaseā claim after the vegan propaganda piece What The Health, I found an observational study that showed an increase in mortality from 5 per 1000 to 6 per 1000. And that was on processed meats, not red meats.
Given what keto allows me to do in treating my obesity and my T2 diabetes, I consider that change to be less than trivial.
Absolute changes in percentage can be meaningless. For example, if I said something increased the risk 10-fold, would you consider it important? For all you know, it changed the odds from 1 per billion to 10 per billion?
atomicspacebunny
(Bunny)
#13
People who eat carrots 4 times a week get colon cancerā¦
People who eat eat fish 4 times a week get colon cancerā¦
People who throw frisbees at 12:00 noon on Saturday get colon cancer?
People who drive blue cars on Sunday get colon cancer?
People mixing cancers with random associations get colon cancer?
LeCheffre
(Banting & Yudkin & Atkins & Eadeses & Cordain & Taubes & Volek & Naiman & Bikman )
#14
Relative risk is a place to hide many sins.
As Zed notes, the actual risk increase is from 5 to 6 out of 1000. Thatās a 1000 Number Needed to Treat to see a difference in outcomes.
If we want to take it apart further, the dataset they have used is the UK Biobank, which collects genetic, physical and clinical information, from 500K volunteers across the UK, and is generally concerned with genome sequencing and predictive medicine. I donāt see much discussion of the method of diet collection, but the link shared by Chris above will explain:
Men and women aged 40ā69 years at recruitment (2006ā10) reported their diet on a short food-frequency questionnaire ( n = 475 581). Dietary intakes were re-measured in a large sub-sample ( n = 175 402) who completed an online 24-hour dietary assessment during follow-up.
Aside from the inherent biases in a short food frequency questionnaire (for instanceā¦ they looked at Red & Processed, Red, Processed, Poultry, Fish, Dairy Milk, Cheese, Fruit, Veg, Fiber, Booze, Tea, Coffeeā¦ whereās the bread, the sugar, the sauces), the recall of folks is questionable. Of course, they backed it up with a 24 hour dietary assessment, where folks punched their food into a website. Of course, maybe it was someoneās birthday or anniversary, or cheat dayā¦ or maybe it was meatless monday for someoneā¦
Iām just gonna leave this right here:
Although painful to admit, it is possible that epidemiologists have been deluded in their acceptance of food frequency questionnaires (FFQ) as the standard tool for dietary assessment in large studies of diet and cancer. The substantial limitations of FFQs have been known for some time (1) and published studies based on FFQ-derived data have long included in their discussion sections a litany of weaknesses due to suboptimal dietary assessment. However, few of us expected the astonishingly poor measurement characteristics of FFQs when compared with doubly labeled water (a gold standard for energy intake; ref. 2), nor had we expected to learn that diet and cancer associations detected when dietary assessment is based on dietary biomarkers (e.g., ref. 3) or food records (4) are undetectable when based on FFQs. We are facing a crisis: hundreds of millions of dollars and many scientistsā careers have been invested in studies using only FFQs to measure diet, but it is possible that these studies have not been, and will not be, able to answer many if not most questions about diet and cancer risk.
Yeah, torture that data and find whatever you like.
Arimathean
(Cranford Coulter)
#15
Exactly. I saw a study 40 years ago that showed that eating green beans caused cancer.
Iām sure it was done by some doctoral student as revenge against his mother who made him eat his vegetables as a kid.
VirginiaEdie
(Edith)
#16
The other thing I would like to know is: why are processed meats always lumped together with red meat? That is not a scientific way to test if red meat causes any kind of health issue. Fresh red meat is very different from processed meat.
PaulL
(You've tried everything else; why not try bacon?)
#17
The first thing is that a 20% increase in risk is too low to draw any inferences from. The likelihood of confounders is just too high to make any categorical statements. In the epidemiological studies that were used to show that smoking causes lung cancer, the risk for smokers was 16-30 times (1600 to 3000%) the risk of lung cancer for non-smokers, and the size of the risk matched the number of cigarettes smoked. That is one of the only cases in the whole medical literature, where it could safely be claimed that the correlation indicates causality.
Dividing a population into quartiles or quintiles makes sense only when there is a fairly uniform distribution of behavior or effect in the population. For this purpose, the larger the population, the better.
For example, if you wanted to sort everyone into quintiles by age, you would first list everyone by their age, then you would start at one end and count. When you reached 20% of the population, you would start assigning to the next quintile, and so forth. If the ages and the quintiles break more or less together, then great, but when you have a lot of people at the same levelāsay a lot of 42-year-olds, in this caseāsome of those 42-year-olds would have to be randomly assigned into one quintile, and some into the other. If the distribution were skewed enough, you might end up with a median age of 40 in one quintile and a median age of 44 in the other, but most of the people in both quintiles could still be 42-year-olds. In a case where almost everyone eats eggs once a day, and very few people eat eggs more than once, dividing the population into quintiles on egg consumption is statistical nonsense.
ETA: In the physical sciences, if the p-value is above 0.001, itās hardly worth reading the abstract.
Rclause
(Scott)
#18
This is where I suggest people pay attention to family history and recommended guidelines on getting a colonoscopy done. I was glad and lucky I did.