Why regulators conclude glyphosate safe while IARC, alone, claims it could cause cancer?


In March, the World Health Organization’s International Agency for Research on Cancer (IARC), issued a statement (also published in The Lancet) that re-classified glyphosate as “probably carcinogenic to humans”. It was a surprise to some in the scientific community because every major regulatory agency had determined that glyphosate, an herbicide often paired with genetically modified crops, was not carcinogenic. In fact, it’s lethal dosage measure is about that of common table salt. So why how did IARC reach such a classification? According to the IARC summary–the full report has not yet been released and may not be for months–because of:

Limited evidence of carcinogenicity in humans for non-Hodgkin lymphoma. The evidence in humans is from studies of exposures, mostly agricultural, in the USA, Canada, and Sweden published since 2001. In addition, there is convincing evidence that glyphosate also can cause cancer in laboratory animals.

Anti-GMO groups jumped on the report, led by anti-GMO crusader, NaturalNews.com, acting as point in a global campaign to ban glyphosate, which is better known as Monsanto’s Roundup (although it is widely sold in generic form).

In response to a recent International Agency for Research on Cancer report, which found that the Monsanto herbicide glyphosate “probably” causes cancer in humans, a cohort of international doctors is now petitioning the European Union Parliament, the EU Commission, and several other health and food safety authorities to take action by banning the use of this prolific chemical.

The website for Moms Against Monsanto declared that “Top Medical Journal, WHO Confirm: Monsanto’s Flagship Product Probably Causes Cancer.” Not to be outdone, Common Dreams declared “Glyphosate, Favored Chemical of Monsanto & Dow, Declared ‘Probable’ Source of Cancer for Humans.” Anti-GMO NGOs have been relentlessly flogging this report since.Even more scientifically oriented publications couldn’t resist overstepping in characterizing the review.

Scientific American declared “Widely used herbicide linked to cancer.” The science magazine had earlier published a post quoting a study by Gilles-Éric Séralini in an article entitled “Weed-Whacking Herbicide Proves Deadly to Human Cells.”

Some countries even took the IARC report to heart enough to suspend uses of glyphosate.

Comparing apples and Roundup

Does glyphosate pose a genuine danger to humans? The mainstream press has tried to separate the scare from the science, but it’s been a challenge because of the confusion of what IARC was evaluating and scientists in general assess the potential hazards and risks of chemicals. IARC does not evaluate actual human risks–a fact widely misunderstood by the public in general. Regulatory agencies do that. Rather, IARC looks at what is called “hazards”.

Note the focus of the IARC review. Quoting Nature’s summary:

The IARC review notes there is limited evidence for a link to cancer in humans. Although several studies have shown that people who work with the herbicide seem to be at increased risk of a cancer type called non-Hodgkin lymphoma, the report notes that a separate huge US study, the Agricultural Health Study, found no link to non-Hodgkin lymphomas. That study followed thousands of farmers and looked at whether they had increased risk of cancer.

But other evidence, including from animal studies, led the IARC to its ‘probably carcinogenic’ classification. Glyphosate has been linked to tumours in mice and rats — and there is also what the IARC classifies as ‘mechanistic evidence’, such as DNA damage to human cells from exposure to glyphosate.

Like glyphosate, almost anything can present a hazard, from the sun to chemicals to everyday foods like coffee depending upon exposure. British based Sense About Science just issued an explainer on the glyphosate controversy to help dispel the fog of confusion about what this review actually means. What is IARC, it asks?

The IARC is an agency of the World Health Organization (WHO) which aims to identify causes of cancer. It brings together groups of scientists to review scientific evidence in order to recognise chemicals, physical and biological agents, and lifestyle factors that can cause cancer in humans.

The IARC do not carry out a risk assessment but rather assess the potential of an agent to be carcinogenic. It does not take into consideration how much of or how commonly a risk it poses in the real world.

We’ve translated the IARC’s carcinogen list into something you can read hereWarning: You might be shocked…

Image made by Mabon ElisWhat’s been overlooked is that the classification that IARC assigned glyphosate—a “2A, Probably carcinogenic to humans”—is the same classification the organization gave to grapefruit juice, fruits (including apples), and working the night shift. At least glyphosate didn’t rate a “1, carcinogenic to humans,” so it’s not as dangerous as sunlight, sunlamps, oral contraceptives, Chinese style salted fish and alcoholic beverages, among a long list.

When IARC comes to a determination of what may cause cancer, it combs through existing literature (which does raise the risk of cherry-picking studies that satisfy your point of view). But it’s assessing the hazard of a chemical. A hazard assessment simply states that a certain chemical, environmental element or behavior is somehow related to cancer. It’ll then note whether something “is,” “is probable” or “is possible”, or “isn’t,” so far as we know.

What a hazard evaluation does not tell you is how likely you are to get cancer. That’s the domain of a risk assessment, which will use the same words–“is,” “probable” and “possible”–but in a different way. Here’s a very informative video explainer by Andrew Maynard, director of the Risk Innovation Lab at Arizona State University, that covers how IARC makes its hazard assessments:

Risks versus hazards

While a hazard just shows you that “somebody out there linked this to cancer,” a risk measures how likely you are going to come into contact with this hazard. So, in the case of apples and pears, IARC looked at the existence of amygdalin, or formaldehyde, both of which are considered class 1 carcinogens and occur naturally in apples. But apples contain 22 parts per million of formaldehyde, far below amounts necessary to cause cancer. In short, dose matters.

Likewise, a report trumpeted by Moms Across America, for example, claimed the existence of glyphosate in mother’s milk, but it was not actually a study and has been challenged by many scientists, most recently by researchers at Washington State University.

Most foods do contain certain chemicals that are associated with toxicity. But each chemical has a dosage curve showing how much ingestion is needed to cause harm, and most foods contain very low doses of these toxins. As the US Centers for Disease Control and Prevention states,

Just because we can detect levels of an environmental chemical in a person’s blood or urine does not necessarily mean that the chemical will cause effects of disease.

This explains why IARC, a WHO subsidiary, can issue reports on cancer hazards, while the World Health Organization itself declares that the IARC study does not indicate a need for more regulation of glyphosate. In fact, several other government agencies, including the German government and (so far) US Environmental Protection Agency, have issued statements on the doses of glyphosate that cause harm and the low cancer risk of the popular and targeted weed-killer:

  • The German Federal Institute for Risk Assessment re-examined data on glyphosate and declared that “the available data do not show carcinogenic or mutagenic properties of glyphosate nor that glyphosate is toxic to fertility, reproduction or embryonal/fetal development in laboratory animals.” The institute did find toxicity that originated from surfactants and other co-formulants used in the making of some glyphosate products.
  • The EPA and other US agencies have considered glyphosate’s cancer risk to be low (the EPA declared glyphosate as noncarcinogenic in 1991), but the EPA is currently reviewing the chemical for weed resistance as well as other properties.

Sense About Science and other groups maintain that even as a hazard evaluation, IARC badly botched its job.

Scientists have criticised the IARC glyphosate assessment for numerous reasons:

  • The selection of literature for reviewing was unbalanced and data has been ‘cherry picked’
  • No new scientific evidence was included in this evaluation
  • This classification is based mostly on animal studies and the report states that there is limited evidence of carcinogenicity in humans
  • It contradicts the conclusions of several national regulatory agencies around the world that have reviewed the large body of glyphosate research and deemed it a safe herbicide

…The dose makes the poison

It’s important to remember that any chemical, whether natural or synthetic can hurt us if we consume too much of it.  The dose is the crucial factor.

The responsibility of IARC and similar organizations doing hazard evaluations is to review the scientific literature (or sometimes conduct their own limited research) to classify the potential theoretical dangers that a chemical could pose–but not in real world usage. Then, it’s up to regulatory agencies using risk metrics to compare what’s known about toxic exposure levels of the hazard with actual exposure in humans or animals. It’s this assessment that informs us whether something might pose genuine health dangers, such as cancer.

So what does science tell us about glyphosate? It’s not carcinogenic and is safe as used. If you question that based on the IARC hazard study, perhaps you should consider advocating bans of grapefruits, apples or night shift work.

Andrew Porterfield is a writer, editor and communications consultant for academic institutions, companies and non-profits in the life sciences. He is based in Camarillo, California. Follow @AMPorterfield on Twitter.

  • Well said Andrew. It is really important to understand the difference between hazard and risk. Electricity can be extremely hazardous, but our risk is low because of the steps we take to prevent exposure to enough to hurt us. Same for chemicals – many are low in hazard (including glyphosate – it would take a lot to be hazardous), but we also minimize exposure. By doing so we can enjoy the benefits safely, as anyone is doing that is reading this on a device powered by electricity

  • There should probably be an “IARC, ” after the very first word in this article. A confusing start to a good article: particularly the pithy last paragraph.

  • JoeFarmer

    My grandma used to shake the excess electricity out of an appliance cord when she unplugged it. Sewing machine, waffle maker, coffee pot, didn’t matter. She didn’t want to get shocked.

    • JoeFarmer, I can imagine why she did that. I used to help my 80+neighbor put up his strings of Christmas lights that were probably from the early 1960s. They were scary in terms of shock potential. We’ve come a long way on safety for both electricity and pesticides. Its a good thin

      • JoeFarmer

        You are 100% correct.

        I repair old tube-type guitar amps in my spare time. The old Fender, Ampeg and Marshall amps are a serious safety risk. No ground wire, the chassis ground is tied to the neutral. But the old plugs don’t have the big blade, so the “ground” on those old amps can be reversed by simply plugging the cord in the wrong way. Lots of failure modes that can send dangerous voltage up the guitar cord and into the player if there is a path to ground. I sprung for a wireless unit back in the early ’80s when I was in a band just for the shock protection.

  • Brian Sandle

    “The institute did find toxicity that originated from surfactants and
    other co-formulants used in the making of some glyphosate products.”

    What percent by volume of use? And are they giving glyphosate itself more access to cells? If so not much use in studying it alone.

    • Allan Felsot

      Surfactants all work on cells the same way; they disrupt membranes by interaction with the phospholipid bilayers. If it wasn’t for the use of surfactants in the early days of cell biology, organelles and other bitty cell constituents would not have been extractable and thus studied with any efficiency. When you use an in vitro cell culture in a tox test, and you add a chemical ingredient dissolved with surfactants, membranes will be severely disrupted, affecting the enzyme kinetics associated with membrane bound enzymes and other functional proteins. Indeed, even everyone’s friend Seralini showed this in his earliest studies, but only recently has he “discovered” the well known phenomenon of membrane disruption. Surfactants added to a formulation have the purpose of reducing the surface tension of the water carrier droplets, thereby enhancing coverage of mass active ingredient per unit area of plant surface. For glyphosate, the POEA surfactant means rapid crossing of the cuticle into the underlaying cells. Too much surfactant can disrupt cuticular wax deposits also, thereby causing phytotoxicity. However, although the physical mechanism of surfactant action is common, the potency of various surfactants differs (bear in mind that commercial surfactants are mixtures of differing chain length polymer, not one substance). To get to an answer to your question about volume of use, a typical spray formulation when it is diluted in water will often have under one percent of surfactant. Here’s a quick example: professional strength Roundup has about 15% POEA surfactant in it. If you dilute a liter of it in a spray tank that holds 30 gallons of water (that’s a very small tank), you’ll end up with less than 0.2% surfactant. More importantly, because surfactant molecules are very large compared to glyphosate molecules, they will have vastly different physicochemical properties. Thus, on any surface, the various types of molecules will behave independently, making the presence of the surfactant irrelevant once the water has dried. In other words, if any Roundup entered your body (which it won’t if you wash after use or exposure owing to extremely poor penetrability through the stratum cornem), the surfactant molecules will not distribute in lockstep with the glyphosate molecules owing to their differences in physicochemical properties. The bottom line is that in vitro experiments with surfactant make nicely wrapped up biochemistry experiments but they are irrelevant to the environmental chemodynamic principles affecting molecules in the real world of organisms.

      • hyperzombie

        Excellent comment, well once I put it through the scientific jargon to simple folk translator. Keep up the good work.

      • Rickinreallife

        Allan Felsot said: ” More importantly, because surfactant molecules are very large compared to glyphosate molecules, they will have vastly different physicochemical properties. Thus, on any surface, the various types of molecules will behave independently, making the presence of the surfactant irrelevant once the water has dried. In other words, if any Roundup entered your body (which it won’t if you wash after use or exposure owing to extremely poor penetrability through the stratum cornem), the surfactant molecules will not distribute in lockstep with the glyphosate molecules owing to their differences in physicochemical properties. ”

        That is one question I had. If I understand correctly, Seralini essentially put some surficant in a petri dish with some cells and found, voila!, toxicity. I suspect that is not a surprising finding, and wouldn’t be surprised if you got the same results if you exposed cells directly to common dishsoap, (e.g. the soap that is part of the mixture of soap, Epsom salts and vinegar that is popularly touted as a safe alternative to roundup). But, I have wondered whether the surficant in Roundup remains in solution and accompanies the glyphosate through the plant. Are you saying that the surficant likely has a different environmental fate, largely remaining on leaf and other plant surfaces?

        Anyway, thank you for the clarity, and for respecting people’s intelligence.

        • SageThinker

          Yes, indeed. This is an excellent comment. The surfactants do not reach the food product in any notable amounts, by my readings on glyphosate and glyphosate-based-formulations. I am concerned about glyphosate in the food supply, but the experiments that show toxicity of adjuvants and surfactants on cells and then make the leap to assume that it’s in our food are being disingenuous.

          • Rickinreallife

            I am appreciative of your willingness to be objective in evaluating this highly speculative claim of Seralini’s. I think Seralini’s surficant study and his latest gmo’s in rat chow studies are actually an own goal setback for the anti-gmo movement. Mr. Felsot comments infrequently, but i find the quality of his posts highly intelligent and highly informative. His posts are reasoned, grounded, thoughtful and courteous, in contrast with the inane and inflammatory comments that marks so much of this discussion, and yes contributed by advocates on both sides.

            Because of your posts, I am keeping an open mind regarding glyphosate and human gut gut biome interactions. I responded to hackety man some time ago that i do think gut biome and human health is an under appreciated and underresearched field of study. I will be truthful, i am unconvinced with your at posts that glyphosate residues in diets are causing all kinds of havoc, but you have at least been thoughtful and backed up your claims with evidence and rational arguments. The fact that you are willing to acknowledge that Seralini’s surficant claims are highly speculative indicates you are interested in truth, not just what is politically advantageous at the moment, increases my respect for your perspective.

          • Rickinreallife

            Actually, I don’t like the term anti -gmo. I prefer the term GMO skeptic communirty.

          • How would you define a GMO skeptic? Labeling groups have ALL said their goal is to eliminate all genetic engineering in agriculture, so they are not skeptics. Anti-GMO groups can’t suddenly be reclassified as skeptics. I think the only people who are really GMO skeptics are biotechnology scientists, who realize the strengths and limitations of the technology and are the first ones to shout ‘smoke’ when issues arrive. Frankly, I’ve not yet encountered a GMO skeptic community as you fashion it to me. There are very few GMO critics who are not in drag on this issue (such as all the pro labelling types, whose real agenda has come out in less guarded talks).

          • Rickinreallife

            Jon, if you have followed my posts on the topic of labelling, you will see that I am very opposed to to the rationale for gmo labelling. When I use the term gmo skeptic community, I am referring to the laymen common Joe’s that are inundated with scientific sounding scare crap, and adopt a position of uncertainty and a natural tendency to distrust corporations like Monsantvo. I very much agree there there is an element of professional, political opportunists that puposely distort genetic engineering, that I would agree are classified as anti-gmo who are not the least interested in truth, just what is politically and propoganda expedient at the moment. But there are what I consider to be responsible critics, which I would not include e.g. gm watch, food and water watch, and God no gmo free usa, (gmo free usa is the most intelligence insulting, cheap, dialogue polluting idiots inflicted on us to date). Rob Walberg is for example one I would consider a responsible critic.

          • Jon don’t you agree: How can you determine the hazard/safety risk of glyphosate if you don’t test for it in food/feed commodities?
            The US government (EPA, FDA, USDA) does not test for the world’s most widely used herbicide, glyphosate even though it has established Maximum Residue Levels (MRLs) for glyphosate.

          • Glyphosate has been the most evaluated pesticide in the history of agriculture. It’s been evaluated by state of the art risk evaluations by the world’s top scientists and regulatory agencies. Unanimous finding, even by IARC: At level found in food, it’s harmless. Latest two organizations to test it are those in Europe and Canada. Here is the one from Europe: http://www.efsa.europa.eu/en/press/news/151112

          • The EPA, USDA, FDA have never tested for glyphosate in crops/feed food commodities using the Maximum Residue Levels (MRLs) established by the EPA, except for soy in 2011.
            The USGS national drinking water tests show glyphosate in USDW.
            I don’t live in Europe or Canada.
            Why isn’t the US government testing glyphosate using EPA MRLs
            and MCLs along with other “safe” herbicides they test for?
            The IARC World Health Organizations finding glyphosate as a probable carcinogen has not been denied by the EPA, USDA, FDA, CDC, NIH, or NCI.
            If it is so safe why hasn’t the US government responded to the IARC finding Glyphosate and its products as a probable carcinogen?
            And if you read the EFSA response it says glyphosate products are a probable carcinogen.

          • JP

            Because everything is a conspiracy to kill you. Duh. PARANOIA4LYFE

          • Guest

            x10E6 for Stryc9nine

          • Stryc9nine

            “Currently, no regulatory authority requests mandatory chronic animal feeding studies to be performed for edible GMOs and formulated pesticides. This fact is at the origin of most of the controversies. Only studies consisting of 90-day rat feeding trials have been conducted by manufacturers for GMOs.”

            You don’t see the problem here? And what is your science background?

          • Therese Olsen

            Please share with us what the “real” agenda is.

          • SageThinker

            I appreciate this comment more than you can know.

          • Rickinreallife

            I will not form any judgement from Seralini’s website or anything quoting GMO Watch. I would prefer to read the original source document for myself, the actual court decision, free of the rhetoric and interpretation of the site you linked to, and of course, the usual six degrees of separation indictment of everybody.

            So Seralini was successful in a defamation suit that the story suggests hinged on falsification of documents or at least of attribution. That in no way exonerates his research. This court decision was not decided on the question of whether Seralini’s research arrived at well founded conclusions. All the critiques of his methodology, presentation and interpretation of results remain valid. Also, “independent” research is not synonymous with “objective” research. I have a hard time believing Seralini by any contortion of the meaning of the word is truly “objective.”

      • Peter Olins

        The largest source of surfactants in our intestines, by far, is the bile acids produced by the liver, which help us absorb fats. If Seralini had been honestly investigating this phenomenon, he might have considered use bile acids as a negative control.

  • Allan Felsot

    It’s critical to read the epidemiology papers that will go into the final IARC determination. What you will find is very shoddy interpretation of the statistical estimates of the odds ratios by which conclusions are made that one variable is correlated with another. A careful read of many of these papers (and this statement is generally true for a lot of chemical epidemiology) reveals the odds ratios to be close to one or even below one if the 95% confidence intervals are taken into account. To add insult to the injury of ignoring one’s own statistical estimates of the confidence intervals bounding the calculated odds ratios is that the papers nearly all the time have absolutely no real exposure information. Surrogate information is used, often based on exposures that may have happened many years ago, and furthermore, based on memory of users and in extreme cases, relative surrogates when the cases are already dead. Thus, while risk is the most important outcome to think about, I would argue strongly that there is not even evidence of hazard present based on the human epidemiology. If one then examines animal studies or biomarkers from human blood and applies a weight of evidence analysis based on hypothesis testing, the conclusions of IARC on glyphosate are even more spurious. I dislike ad hominem arguments, but at the risk of going there, it is important to understand that each subject of classification (whether it be whole food, extract of natural product, Gingko biloba [I didn’t forget that was a possible human carcinogen], or pesticide is determined by empaneling putatively objective scientists that are not tied to the product. When you investigate the panel for glyphosate, you’ll find several have published chemical epidemiology papers. Indeed, at least one was involved in the 2,4-D chemical epidemiology papers from Canada. If you read that reviewers papers, you’ll find all my aforementioned comments are applicable. I’m not impugning anyone’s character here, but your focus will not be skeptical analysis of data if you are in the business of publishing chemical epidemiology papers that inevitably conclude, yes there is an association with so and so.

    • Peter Olins

      Apparent associations found in epidemiology can never demonstrate causation: they’re just starting points for further research. The IARC panelists are well aware of this, but this point seems to have been lost in their final assessment.

  • Andrew Porterfield, misses the point. How can you determine the hazard/safety risk of glyphosate if you don’t test for it in food/feed commodities?
    The US government (EPA, FDA, USDA) does not test for the world’s most widely used herbicide, glyphosate even though it has established Maximum Residue Levels (MRLs) for glyphosate.

  • SageThinker

    Here’s the actual conclusion of the IARC monograph, for reference. Note that they conclude that it does cause cancer in lab animals. On humans there is not the same data available of course.

  • Peter Olins

    The IARC has reviewed 985 chemicals: only one was “probably not carcinogenic”.

    One simple interpretation is that almost any chemical can cause cancer in lab animals if the dose is high enough. Animal models can be useful, especially if there is an established mechanism of carcinogenesis, but I seriously question our over-interpretation of the predictability of these tests.


    • 360_Viewer

      Let’s remember that the reason IARC decides to evaluate a material is because there is some evidence of a causal relationship. Considering an estimated 50 million chemicals have been found or made, it’s not like they are classifying everything as carcinogenic, as the statistic you quoted infers.

  • AndyMatts

    Good points, but incomplete. Science is finally starting to not just focus on the active ingredients in pesticides and herbicides, but also the so-called “inert” ingredients, and the conclusions are not good.

  • guruurug

    Everyone will die of some disease, vegans too. You were born with cancer, deal with it crybaby.