Support more videos like this at patreon.com/rebecca!
It’s been five whole months since I’ve talked about glyphosate and so I guess it’s time to talk about it again! I talk about glyphosate about as often as I remember to get my hair done. That reminds me, I need to get my hair done. Yikes.
Last time I talked about an overhyped study that claimed glyphosate was killing the bees. This time I’ll be talking about an overhyped study that claims glyphosate is killing the humans. Yeah, it got worse in five months, didn’t it?
As a reminder, glyphosate is the main ingredient in RoundUp, the world’s most popular herbicide, which is made by Monsanto. I’m sorry, I mean (evil voice) MONSANTO. I mean, Monsanto sold themselves to Bayer and now they’re just called Bayer which makes sense bc everyone associates Monsanto with death and Bayer with, like, aspirin so for the rest of this video I’ll call them Monsanto Bayer, because fuck their PR moves).
Before we dive in, let me just state my conflict of interest: I fucking hate Monsanto Bayer. I think they’re an evil corporation, which, yes, is redundant. Monsanto Bayer exists to make money and if they did find a chemical that made them a lot of money but cost a lot of human lives, they would 100% bring it to market. I know this because they already did it once — it was a chemical called Agent Orange and it injured or killed millions of people, including American soldiers, if you’re the sort of person who cares more about them then 400,000 Vietnamese.
But alas, RoundUp is no Agent Orange. It’s one of the most studied products on the planet at this point, because people keep insisting that it must cause cancer. Despite that, study after study has shown that that’s just not true. Until last month, when a new study was published claiming that glyphosate raises the risk of Non-hodkin Lymphoma (NHL) by 41%. That was the major finding, and that (of course) was the headline that ran in mainstream news, including the Guardian.
With any other chemical, this wouldn’t really make me bat an eye. Raising the risk of a fairly rare disease by 41% isn’t necessarily worth panicking over — about 1,000 of our 360 million Americans will die from NHL this year. No one wants to lose another 400 people a year, but your personal risk of developing this disease is still going to be ridiculously low.
But the fact that it’s a 41% increase that they say is due to an herbicide that has been repeatedly, exhaustively covered in previous research, and deemed safe time and time again, that made me pause and wonder where this study is coming from.
So this was a meta-analysis, which I cover a lot here as it’s a very useful form of research that collects a bunch of previous research into one giant study, combining all their results to see if any new patterns might emerge. But if you collect all the research on glyphosate, it’s pretty obvious at this point that the result is going to say it’s safe, because that’s what the vast majority of the literature says.
That’s where we run into problems with meta-analyses. You can’t just include all the research in one new analysis — you have to make sure the initial research was well done, and that each study you’re including used similar methods to obtain its data. And on the other hand, you can’t just pick and choose the studies that have the results you want to see.
I’ll give you a moment to guess what happened in this particular meta-analysis.
They included six studies in total: one of them is a “cohort” study and the other five are case-control studies. Case-control means that these studies look at people who already have a disease and compare them to people who don’t have it, and then they interview all those people about their lives and habits and they look for patterns, like hey, everyone who has this disease also eats shrimp, but no one in the non-disease group does. Maybe shrimp causes that disease. Those studies can be helpful but ultimately when studying a rare disease like NHL you don’t have a lot of data points, and you might end up with a lot of statistical noise. The other problem is that when you ask someone with a disease about their life, they’re more likely to remember odd things they think may have caused the disease, like that shrimp-eating contest they won. The people without the disease may have forgotten coming in second place in the very same shrimp-eating contest.
In the five studies the researchers chose for this meta-analysis, four found a faint correlation between NHL and glyphosate, and one found nothing.
Cohort studies are kind of the gold-standard, and the one they picked is a really good one in particular. Cohort means that the researchers took a large sample population and kept track of them over the course of many years to see who developed a disease. In this case, the Agricultural Health Study involved watching 45,000 people who came into contact with glyphosate and tracked them for 20 years. At the end the authors found “no association…between glyphosate and any solid tumors or lymphoid malignancies overall, including NHL and its subtypes.”
This was a huge study, with the specific purpose of finding any correlation between glyphosate and any of 20 different cancers, and they found nothing — the absolute worst thing they found was that after 20 years (but not 5, 10, or 15 years) there was a just barely statistically significant correlation between glyphosate and acute myeloid leukemia. They recommended someone follow up on that but it wasn’t a big deal.
So, guess which statistic this meta-analysis chose to include? Not the 5, 10, 15, or 20-year results for dozens of different cancers that showed absolutely no correlation. Not the 5, 10, or 15-year result for acute myeloid leukemia. Nope, they only included the just barely significant 20-year result for acute myeloid leukemia. That one had a relative risk (RR) of 1.12. Every other result was under 1, and had any of those been included in the meta-analysis, the overall RR average would not have been statistically significant, and then they never would have been published, let alone gotten this glowing write-up in the Guardian.
Does glyphosate cause cancer? No. It just doesn’t. This isn’t new data — it’s old, cherry-picked data that is tricking the general public into being scared of Monsanto. Don’t fall for it.