Monsanto have been found liable for the terminal cancer of a former groundskeeper exposed to glyphosate and other chemicals in their Roundup herbicidal product.

On Friday, a California jury hit Monsanto with $289 million in damages in a lawsuit brought by a patient suffering from terminal cancer, accepting the plaintiff's claims that his disease was caused by the company's popular herbicide, Roundup. The suit neatly sidestepped the complicated epidemiology of the active ingredient in the herbicide—glyphosate—and instead made the claim that the cancer was the result of glyphosate's interactions with other chemicals in Roundup—a claim for which there is even less evidence.

However the scientific evidence to support the above verdict seems questionable at best. The WHO has listed glyphosate as a "probable" carcinogen, but European regulators (who, as a rule, tend to be a lot stricter) consider it to not be carcinogenic.

The degree of exposure can be an issue with glyphosate. High levels of exposure in animal testing has hinted that the chemical could cause cancer, and some small epidemiological studies found a link between cancers and extensive exposure during agricultural work. That was enough for the World Health Organization to label the idea that glyphosate caused cancer as "probable."

But there have been questions raised about the significance of the animal studies even as the WHO report was being prepared. And a meta-analysis of epidemiological studies found no consistent association of glyphosate with cancer. European safety regulators have come to an opposite conclusion to that of the WHO, determining that glyphosate is not a carcinogen.

The above article is very dismissive of the claims regarding glyphosate, but the popular press seem far more convinced, as did the jury in the described case. The problem is that the press is rarely if ever a reliable source for scientific stories, especially not ones that have an emotional component such as fear of cancer.

  1. Is glyphosate, on its own, a carcinogen?
  2. Is glyphosate, when combined with other chemicals in Roundup, a carcinogen?

Related Questions:

  • 2
    I do not fully understand this, so not an answer, but apparently the WHO makes a difference between "health hazards" and "health risks", where classifying something as a hazard is the first step to assess if there is actually a risk. As far as I can tell the WHO considers Glyphosate a hazard (i.e. something that warrants a closer look), but not a risk. Try if you can make more sense from this than I : who.int/foodsafety/faq/en. – Eike Pierstorff Aug 12 '18 at 13:55
  • 4
    Pure water is a poison and everything else is worse. – blacksmith37 Aug 12 '18 at 14:25
  • 1
    @blacksmith37 I'm well aware that the dose makes the poison, but given the trial verdict it seems sensible to ask if there's anything about glyphosate that makes it especially bad to the point where selling it would constitute actionable negligence. – GordonM Aug 12 '18 at 15:02
  • There is an ambiguity in the core claim here because the carcinogen might not be the active ingredient and it isn't clear which is the claim. The formulation sold in Europe uses different surfactants to the US version. Moreover, plenty of other firms now sell glyphosate-containing weedkillers so it isn't just Roundup (the Monsanto brand) that matters. I think the title should reflect this. – matt_black Aug 12 '18 at 18:52
  • @matt_black The question was originally about glyphosate either on its own or in combination with some other chemical. The question got edited. – GordonM Aug 12 '18 at 19:47

There is no consistent evidence showing a relationship between glyphosate and cancer

There are a lot of studies on the relationship between glyphosate use and cancer. Some are small and poor; others are large and more reliable. It is inevitable and likely that some of these studies will report a connection between the weedkiller and cancer just through random chance alone (if you randomly sample a population, some samples will have higher in-sample rates of cancer than others through chance alone.) If those random results are selectively quoted they will provide food for lawsuits that cannot be justified in the context of the overall weight of evidence. So it is important to look at the weight of evidence from all studies and from well-conducted studies.

Here are the conclusions of two of those studies. The first, a meta-analysis of many other studies was conducted in 2012 by Mink, et. al. in Regulatory Toxicology and Pharmacology. The abstract reports (my emphasis):

Glyphosate is widely considered by regulatory authorities and scientific bodies to have no carcinogenic potential, based primarily on results of carcinogenicity studies of rats and mice. To examine potential cancer risks in humans, we reviewed the epidemiologic literature to evaluate whether exposure to glyphosate is associated causally with cancer risk in humans. We also reviewed relevant methodological and biomonitoring studies of glyphosate. Seven cohort studies and fourteen case-control studies examined the association between glyphosate and one or more cancer outcomes. Our review found no consistent pattern of positive associations indicating a causal relationship between total cancer (in adults or children) or any site-specific cancer and exposure to glyphosate.

A large single study has been ongoing for some time and, while it doesn't appear to have been published yet has had its results examined by several experts on risk. Its preliminary conclusions are reported in The Scientist (my emphasis again):

The new study, which was seen by Reuters, draws on long-term data collected through the Agricultural Health Study. This has monitored the health of nearly 90,000 people in Iowa and North Carolina from 1993 to 2010, including farmers licensed to apply pesticides to their crops, and their spouses. The researchers tell Reuters that among more than 54,000 pesticide applications taken into account in the study, 83 percent contained glyphosate. Yet they found no significant increase in cancers among those exposed to the chemical.

While the study did report:

According to Reuters, the study did note that there was “some evidence of increased risk of acute myeloid leukemia (AML) among the highest exposed group,” but reports that the correlation was “not statistically significant.”

This is not however strong evidence that there is a real link as this quote from the original Reuters report on the study suggests:

David Spiegelhalter, a professor of the Public Understanding of Risk at Britain’s Cambridge University who has no link to the research, said Thursday’s findings were from a “large and careful study” and showed “no significant relationship between glyphosate use and any cancer”.

He added that the possible association with AML “is no more than one would expect by chance”.

So it looks as though the legal case and many of the more bizarre regulatory judgements are based on selective quotation of underpowered studies. The good evidence and the overall weight of evidence do not show any meaningful connection between the use of glyphosate-containing products and cancer.

  • 2
    I feel its important to add that like most studies on the subject: 1. real world validity is very limited when studying the pure substance in a lab 2. epidemiological studies might suffer from (reversed) p-hacking (IRCA finding higher incidence for cancer being attacked for the stats) and most importantly 3.: Mink et al. (as were way too many others) were sponsored by Monsanto (read the full text instead of just the abstract. They are almost transparent about it.) Combine that with other problems of science, like pub-bias and we come to "we do not know dreck, evidence needed either way." – LangLangC Aug 14 '18 at 13:43
  • 2
    Also the constant reminder: "when used as directed on the label" is a nice canary over all. That substance, its formulations and scale or habitual application are certainly not harmless to human health overall and its ecological effects are quite certainly problematic. Meaning: focussing on just one definition of "causes cancer" might be interesting in itself, perhaps necessary for a Sekptic's Q, but it is probably distorting the essential, full picture of it. – LangLangC Aug 14 '18 at 13:48
  • @LangLangC Studies being sponsored by Monsanto isn't itself a problem: perhaps more careful analysis of methodology and data is merited but certainly not dismissal. Compare studies on drug safety and efficacy: we wouldn't have many drugs if we relied on people other than their owners to test them. Moreover, Mink et. al. wasn't original research, but a meta analysis: if they selectively quoted studies it would be obvious and you would be able to find a refutation of their results. – matt_black Aug 14 '18 at 14:26
  • @LangLangC And the other study in the answer was not a lab study but a study of real world use. "not harmless to human health overall" is the question and this study answers with a negative: if you have evidence to the contrary, show it. Equally with "ecological effects are certainly problematic". it is a weedkiller, of course it has ecological effects: that is its job. The question is is is worse than other ways of killing weeds or not? Again, if you think otherwise, citation needed. – matt_black Aug 14 '18 at 14:30
  • The difference between "dismissal" and "caveat because of conflicting interests": the way Mink is presented here just lends it a bit too much credibility: They selected the studies to include, designed the tests (stats) to evaluate them etc. – I crititcised mainly how the whole subject is handled in tendency in gov and academia; that seems to be insufficient for an A in itself, as I do not have "evidence for the contrary", just further evidence for (already in yours) lack of conclusive evidence. (Sidenote: do we need to kill weeds or just manage them to maintain yields?) – LangLangC Aug 14 '18 at 16:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .