Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

Points to Consider: Claims About Pesticide Toxicity are Based on Discredited Methods


Having failed to convince the public that biotech improved seeds present novel or unexamined risks, professional opponents are now working overtime to tar GMOs with the same brush they’ve used successfully to denigrate the use of synthetic chemistry in agriculture.  Papers that look scientific to the casual observer are frequently cited in these attempts. We take a closer look at one particularly good example of a bad example.

Original PaperRobin MesnageNicolas DefargeJoël Spiroux de VendômoisGilles-Eric Séralini.  2013.  Major Pesticides Are More Toxic to Human Cells Than Their Declared Active Principles. BioMed Research International (Impact Factor: 2.88). 12/2013; 2014(Article ID 179691).  DOI:10.1155/2014/179691

Published Analysis:

ScienceInsider:  “Pesticide Study Sparks Backlash

ravingscientist: “Séralini Has Done it Again!

In the Pipeline: “Pesticide Toxicity?

European Crop Protection Association (ECPA) Statement: “Séralini Study Fails to Meet Basic Scientific Standards

Primary Claims of the Original Paper:

  • Pesticide formulations as sold and used are up to 1000 times more toxic than the isolated substance that is tested and evaluated for safety
  • Roundup the most toxic of herbicides and insecticides tested
  • Flawed safety evaluations put public health at risk


  • The experimental conditions utterly fail to represent the exposure pathway under real-world uses of the compounds studied, and ignore that cells in humans are organized into organs and tissues, and protected from direct exposure by skin.
  • The experiment exposed cells in culture directly to surfactants. Surfactants are well known and widely used precisely for their ability to dissolve lipids, one of the major constituents of cell membranes.  Beginning biology students would expect to see the kind of negative effects on cells in culture exposed to lipid dissolving agents as reported.
  • The testing methodology was directly to expose cells in tissue culture under starvation conditions to unrepresentative high levels of several compounds.
  • As usual, poor experimental design omits the proper use of controls (e.g., standard procedure would call for a negative control treatment with 0.5% DMSO; exposure of non-starved cell lines).
  • Similar results would be expected from exposure to baby shampoo or hand soap.
  • As one blogger noted, “…juiced broccoli straight from the organic farmer’s market might well have similar effects under these conditions.”
  • Another noted  “The endpoints observed are so general that we could probably find the same kind of toxicity with lemon juice or grapefruit extract… It’s not new or shocking. It is what I would have expected at the level he is giving this to the cells.”
  • The Cairo-based publisher is a “pay-to-play” journal well known for lax peer review standards, and thus not highly regarded by the scientific community.
  • One of the journal editors was so unhappy with the poor quality of the study and flawed peer review that he immediately resigned his post.

Analysis of Specific Claims Made by the Author:

  • “Ethoxylated adjuvants found in glyphosate-based herbicides were up to 10.000 times more toxic than the so-called active AP glyphosate, and are better candidates for secondary side effects.”
    • Contrary to the implications of the paper, adjuvants and inert ingredients are subject to prior review and approval by EPA.
    • The exposure pathway used fails utterly to reflect the conditions of real-world use, and renders the results and claims meaningless.
    • According to ECPA,  “Pesticide products, including the active substance and the pesticide formulations, are closely regulated in the European Union. No pesticide product can be authorised for sale in the EU unless data are provided and accepted by regulatory authorities showing that it does not pose an unacceptable risk to human health for the designated use. The assertion by the authors that industry or regulatory authorities assume pesticide co-formulants (called “inerts” in the publication) are non-toxic is incorrect.”
  • “These adjuvants also have serious consequences to the health of humans and rats in acute exposures.”
    • The claimed validation of the methodology used in this experiment is to a paper by some of the same authors, who lack toxicology competence) and to which access is obscured behind a paywall. The claimed validation is without merit.
  • “Upon the introduction of herbicide tolerant genetically modified organisms (GMOs), designed to tolerate Roundup and to accumulate unusual levels of its residues, Roundup quickly became the major pesticide in the world and a major food or feed contaminant.”
    • False claim: None of the crops improved through biotechnology on the market or in development was “designed…to accumulate unusual levels of [Roundup residues]”. They have been improved to tolerate exposure to the compound;
    • Seralini deliberately uses the more evocative and negative term “pesticide” rather than the more accurate “herbicide.”
    • The assertion that Roundup is “a major food or feed contaminant” is untrue, and ignores Roundup’s favorable safety profile  and superiority to alternatives.
    • The authors ignore the fact that Roundup’s mode of action targets metabolic pathways found in plants but absent in humans/mammals, which leaves them immune to its intended effects.
    • The suggestion of hazard from Roundup is contradicted by a wealth of peer reviewed scientific literature encompassing extensive animal testing, and over 40 years history of safe use to manage weeds in agriculture worldwide.
  • “We used the embryonic (HEK293), placental (JEG3), and young adult hepatic (HepG2) human cell lines because they are well characterized and validated as useful models to test toxicities of pesticides [19-21], corresponding to what is observed on fresh tissue or primary cells [22-24].”
    • While in vitro toxicology studies may have value in some circumstances they do not reflect real-world conditions for herbicide or pesticide use and exposure. They are therefore not useful, and are not used to evaluate real-world risks to humans.
    • In view of the wildly unrealistic and unrepresentative exposure pathway, the choice of cell line is immaterial.
  • “All formulations were cytotoxic and far more toxic than their AP, except for isoproturon and its formulated pesticide Matin which were both not soluble over 100 ppm; Matin does not have any declared adjuvant (Table 1)… In fact, 8 formulations out of 9 were clearly on average several hundred times more toxic than their AP, ranging from 2-3 times more toxic for pirimicarb or prochloraz, to 1056 times more toxic for tebuconazole. Results were similar for all cell types.”
    • A finding of cytotoxicity from direct exposure of cell membranes to lipid dissolving compounds is completely unsurprising.
    • The unrealistic exposure pathway deprives the reported results of any practical value.
    • From ECPA: “All of the pesticides studied in the publication have already been fully evaluated with in vivo toxicity studies. These studies confirm the absence of significantly increased toxicity of the pesticide formulation when compared with the toxicity of the active substance contained in those formulations. This in vivo data is publicly available and remains the most robust and relevant data upon which to assess to potential toxicity of pesticides. In addition, co-formulants, including “adjuvant” co-formulants, are specifically mentioned in the EU regulatory requirements for pesticide product approval and are part of the overall assessment of the potential risk of the use of a pesticide product.”
  • “ This is the first time that all these formulated pesticides have been tested on human cells well below agricultural dilutions. The three different cell types reacted very similarly and the toxicities were observed on several biomarkers; this confirmed our results.”
    • As noted above, the findings reported are entirely unsurprising, and the unrealistic exposure pathway renders them devoid of practical value.
  • “Adjuvants in pesticides are generally declared as inerts, and for this reason they are not tested in long-term regulatory experiments. It is thus very surprising that they amplify up to 1000 times the toxicity of their AP in 100% of the cases where they are indicated to be present by the manufacturer (Table 1).”
    • All adjuvants and “inerts” are subject to toxicity assays and conditions of use, per 40CFR (Part 180) “TOLERANCES AND EXEMPTIONS FOR PESTICIDE CHEMICAL RESIDUES IN FOOD”.
  • “The definition of adjuvants as “inerts” is thus nonsense; even if the US Environmental Protection Agency has recently changed the appellation for “other ingredients”, pesticide adjuvants should be considered as the first toxic “active” compounds.”
    • Contrary to the implications of this paper, all “inert ingredients” and adjuvants in pesticides require prior approval by EPA (see 40CFR (Part 180)).
  • “In the scientific literature, in contrast with regulatory beliefs, some harmful effects of the adjuvants present in this study are reported.
    • These are evaluated by EPA and factored into the permissible conditions of use stipulated under pesticide registration.
  • “It is commonly believed that Roundup is among the safest pesticides. This idea is spread by manufacturers, mostly in the reviews they promote [40, 41], which are often cited in toxicological evaluations of glyphosate-based herbicides. However, Roundup was found in this experiment to be 125 times more toxic than glyphosate. Moreover, despite its reputation, Roundup was by far the most toxic among the herbicides and insecticides tested. This inconsistency between scientific fact and industrial claim may be attributed to huge economic interests, which have been found to falsify health risk assessments and delay health policy decisions [42].”
    • The safety of Roundup is well documented and widely recognized and not challenged by anything reported or claimed in this paper;
    • The unrepresentative, unrealistic and non-validated experimental conditions render the authors’ conclusions meaningless.
  • “In conclusion, our results challenge the relevance of the ADI, because it is calculated today from the toxicity of the AP alone in vivo. An “adjuvant factor” of at least a reduction by 100 can be applied to the present calculation of the ADI if this is confirmed by other studies in vivo. As an example, the present ADI for glyphosate is 0.3 ppm, for glyphosate-based herbicides it would be 3 ppb or less. However, this will never replace the direct study of the commercial formulation with its adjuvants in regulatory tests. Anyway, an exposure to a single formulated pesticide must be considered as co-exposure to an active principle and the adjuvants. In addition, the study of combinatorial effects of several APs together may be very secondary if the toxicity of the combinations of each AP with its adjuvants is neglected or unknown. Even if all these factors were known and taken into account in the regulatory process, this would not exclude an endocrine-disrupting effect below the toxicity threshold. The chronic tests of pesticides may not reflect relevant environmental exposures if only one ingredient is tested alone.”
    • The artificial and unrealistic exposure pathway and the lack of appropriate controls, make the entire study worthless. None of the reported results are surprising, nor do they have any practical value.

Print Friendly, PDF & Email