Science in Society Archive

“Serious Deficiencies” in EFSA Glyphosate Re-assessment Signals EU Re-approval

In preparation for the EUs approval of glyphosate, EFSA has concluded it safe and unlikely carcinogenic; clashing with the IARC’s classification of “probable carcinogen” by selective dismissal of relevant data; EFSA’s conclusion should be rejected outright Dr Eva Sirinathsinghji

Sign the Independent Scientists Manifesto on Glyphosate here:

As the EU is set to decide on the re-approval of glyphosate, the European Food Safety Authority (EFSA) has ruled that the herbicide is “unlikely to be genotoxic (i.e. damaging to DNA) or to pose a carcinogenic threat to humans” [1]. The ruling is in contrast to that of the WHO earlier in the year, whose specialist oncology arm, the International Agency for Cancer Research (IARC), declared the top selling herbicide a “probable human carcinogen” [2] Glyphosate ‘Probably Carcinogenic to Humans’ Latest WHO Assessment, SiS66).

The opposing conclusions of the two assessments highlight the difference in approach between the two organisations. The IARC base its analysis solely on published peer-reviewed papers and publically available government documents, with the intention of maintaining a transparent assessment procedure by experts with no conflict of interest. The EFSA assessment on the other hand was performed in conjunction with industry and included unpublished industry data (see [3] Scandal of Glyphosate Re-assessment in Europe, SiS 63). As revealed in a leaked copy of a draft version of the assessment report that was analysed by toxicologist Peter Clausing [4], even significant, positive results of carcinogenicity were dismissed, and evidence misrepresented and wilfully misinterpreted in order to diminish existing evidence. Defending their assessment in response to EFSA’s decision Kate Guyton, senior toxicologist and member of the IARC panel, said they had conducted 'the most independent, rigorous and transparent evaluation”, based on “rigorous scientific criteria”. She reiterated the exclusion of unpublished data as well as industry-linked studies as “extremely important for independence and transparency” [5]. 

Not only is glyphosate now set to be reapproved for another 10 years, the levels of Acceptable Daily Intake (ADI) have been proposed to rise from 0.3 milligrams per kilo of body weight per day (mg/kg bw/d), to 0.5 mg/kg bw/d [6].  Recent reviews highlight toxicity at well below these levels. For example, oxidative stress - a potential precursor to cancer - had been documented at levels of 0.09 mg/kg bw/d, as has toxicity to the liver and kidney [7]  further questioning the validity of the regulatory assessments. Minute levels of only 0.169 ng/L to 0.169 mg/L have also been shown to induce proliferation of oestrogen-dependent breast cancer cells in vitro, and increase expression of oestrogen receptors by activating the genetic oestrogen response element (see [8] Glyphosate is Carcinogenic).

 As emphasised in [8], the IARC assessment, though comprehensive is not complete and could have been even more damning of glyphosate safety. Nonetheless, it made no difference in the end on the outcome of the EFSA assessment, which as explained below has little to do with science-based evidence.

Toxicologist declares “serious deficiencies” in EFSA assessment

The new version of the draft renewal assessment report (RAR) by EFSA, not published by EFSA, was leaked and analysed by Clausing for Pesticide Action Network, Germany, who found gaping inadequacies in its assessment of glyphosate’s carcinogenic and teratogenic toxicities.

The new draft was supposed to take account of the recent IARC decision. According to Clausing, the revised draft differs little from the initial version in cherry-picking and spinning evidence to support the safety of the chemical.

Burying the evidence of toxicity may well come back to haunt chief glyphosate producer Monsanto, now facing a growing number of personal injury lawsuits from cancer patients for knowingly misinforming the public and farmworkers about the dangers of the chemical.

As reported previously by ISIS (see [3]), the re-assessment was largely performed by a consortium of chemical companies that calls itself the ‘Glyphosate Task Force’ (GTF)  that includes Monsanto, Syngenta and Dow Agrosciences. Clausing highlights a “deliberate” dismissal of studies showing evidence of toxicity in the draft report, which he says [4]: “needs to be subjected to a complete and thorough re-analysis by a group of truly impartial experts before a decision can be made about the approval or non-approval of glyphosate in the European Union”.

The approval process

The renewal assessment report (RAR) is the basis for the legal decision for EU’s approval of glyphosate. It consists of experimental results provided by the applicant in the form of a dossier. In the case of glyphosate, the applicant is the Glyphosate Task Force (GTF). The European community legislation for marketing pesticides is included in Regulation [EC] No. 1107/2009, which states that each pesticide’s active ingredient alone must be re-assessed every 10 years, not the commercial formulations that are used in real life situations (and have been shown to be more toxic at levels 1 000 times that of glyphosate alone [9]). Notably, the IARC included studies on glyphosate formulations, analysing the public health risks in a more holistic manner that reflects the real life scenarios where people are exposed to the commercial formulations and not glyphosate alone.

The dossier contains a volume called Toxicology and Metabolism. These include studies published in the scientific literature over the last 10 years as well as unpublished industry studies based on the OECD guidelines for good laboratory practice (GLP). These industry studies are confidential and only their summaries are made public in the dossier. If a study from the published scientific literature is omitted from consideration for re-approval, the applicant must give a reason for the omission. As summarised below however, many studies were omitted without explanation, or reasons for dismissal are arbitrary and appear dependent on the results of the study. The applicant then selects a country, the Rapporteur Member state (RMS), in this case Germany, to assess the application. Overall responsibility for the assessment lies with the Bundesinstitut für Verbraucherschutz (BVL, Federal Office for Consumer Protection and Food Safety), but the Bundesinstitut für Risikobewertung (BfR, Federal Institute for Risk Assessment) is  actually responsible for assessing the human safety aspects.

When the application was first submitted, the rapporteur state had opened it up for public comment in December 2013, and the finalised report released in 2014, which the BfR claims to have all the comments included. According to available information, there were two further drafts re-submitted to the RMS and the European Food Safety Authority (EFSA) (January and March 2015), the earlier being available only to scientists close to industry. It is the March version that Clausing analysed.

Disappearing evidence of carcinogenicity and reproductive toxicity

Some of the highlighted inconsistencies include the selective dismissal of studies during the literature search on carcinogenicity studies and mechanistic studies of cancer including genotoxicity – DNA damage that can lead to cancers. EU legislation relating to genotoxicity tests follows a two-tiered approach starting with in vitro tests; and if these fail to detect toxicity, in vivo tests are performed. Under EU legislation for assessing cancer risk of pesticides, genotoxicity is one of the cancer mechanisms studied. This is important because specific forms of genotoxicity are valid biomarkers for cancers, such as increased micronuclei, whole segments of chromosome that have become separated from the nucleus (indicative of DNA damage). Positive tests for genotoxicity therefore generally preclude the approval of a pesticide.

The RAR included 48 studies on genotoxicity from the literature search, 14 of which have negative results, while 31 show genotoxic effects. Three were described as inconclusive by the GTF. This means that the GTF explained away 31 studies to come to the conclusion that glyphosate is not genotoxic, despite the EFSA guidance for the literature review stating that “relevance criteria should not be too restrictive”. Many of the disqualified studies were not even discussed; especially those that followed established international protocols that validate them for regulatory assessments. One such study by Koller et al. (2012) received no negative comments [10]. It followed established protocols and included both glyphosate and glyphosate formulations, a positive control and also showed genotoxicity at levels several times lower than concentrations sufficient to induce cytotoxicity. Furthermore, many studies were misrepresented in the summary table with for example, the table only displaying the highest doses used, as well as only some of the supplementary experiments while omitting the important tests such as micronuclei assays that had been performed in the studies. Also omitted were in vivo tests performed on blood from people in regions sprayed with glyphosate (see [7]), which clearly show that glyphosate in the air is absorbed into the body and is toxic. The most recent study from 2015, which was also not included in the IARC report assessed DNA damage in children living in a high agrochemical use zone in Argentina. Researchers found that children living within 500 metres of sprayed fields had 66 % more cells with micronuclei than those living 3 000 metres away; 40 % of these children also had persistent illnesses [11].

The BfR, according to the German Ministry of Agriculture, claimed to have made a “detailed, quality-assured examination” of all the studies, independent of the GTF’s summary. The IARC in comparison included an extra 21 publications in its carcinogenicity assessment absent from the RAR. Clausing rightly questions what the conclusion would have been if the additional 21 studies were included in the analysis, and how the BfR could allow these glaring omissions when they claimed to have conducted independent literature searches.

Oxidative stress, like genotoxicity is another key mechanism for cancer development. Studies reporting increased oxidative stress following glyphosate exposure were included in the IARC’s assessment. In contrast, the RAR failed to mention it at all. When pressed by Testbiotech and PAN Germany, the BfR admitted that oxidative stress is “without doubt” a possible mechanism of carcinogenicity. So how can they have omitted it unintentionally? This is an important point considering that the EU legislation for testing carcinogenicity of pesticides requires long-term 2-year, large-scale animal studies, very few of which are conducted by scientists outside of industry due to the high costs of such experiments. Mechanistic evidence therefore, should be given careful consideration, and not dismissed or ignored. Compiling evidence of carcinogenicity from these different fields, as the IARC did, would make the evidence too compelling from industry’s point of view. As reviewed in [12] A Roundup of Roundup® Reveals Converging Pattern of Toxicity From Farm to Clinic to Laboratory Studies, Special ISIS report, the consistent pattern of toxicity across scientific fields, clinical studies and farmer experiences provides the strongest evidence of glyphosate’s harmful effects.

In the analysis of the animal studies, misinterpretations and twisting of results continues. Of the four regulatory animal studies discussed by the RAR, malignant lymphoma was increased upon glyphosate exposure. How were these findings explained away? First, the significant increase in lymphoma incidence was seen in a study using Swiss albino mice, which have a higher propensity to develop tumours, so the RAR concluded that this strain of mice was not suitable. This was also a criticism levelled at Dr Séralini’s group when they published increased tumours in Sprague Dawley rats exposed to GM foods and pesticides 2 years ago, drawing a concerted attack from pro-GM proponents (see [13] Retracting Séralini Study Violates Science and Ethics, SiS 61), despite the study being a toxicological and not carcinogenic study for which Sprague-Dawley rats are standardly used. The other three studies were performed in CD-1 mice and showed dose-dependent non-significant increases, though when using alternative statistical methods explicitly recommended by the OECD (Cochran-Armitage-Trend Test) they become significant. The latest RAR draft confirmed the significance of these studies when applying the Cochran-Armitage statistical methods, and admit that it could be due to real treatment effects with “evidence of carcinogenicity conclusive”, but then goes on to conclude that a lack of consistency across studies makes the evidence “insufficient” for classification.  Further, tumour incidence in the Swiss albino mice was outside the range of historical controls, above the mean incidence of the historical control, and also the incidence was dose-dependent. If they had fallen within the range of historical controls, no doubt industry would have used this to dismiss the findings. Even with results that are not significant, this does not rule out biological significance, just as we see in the reverse situation where industry dismisses statistically significant results as being not biologically relevant (as has already been the case with glyphosate where evidence of teratogenicity was dismissed) [14]. When it came to analysis of epidemiological studies included in the IARC report, the RAR dismissed them, claiming they “see a particular problem with the co-formulants of glyphosate-based formulations” due to the higher toxicity of formulations over the active ingredient alone.

It comes as no surprise then that the BfR concluded [2] “there is no in vivo genotoxicity and mutagenicity potential of glyphosate or its formulations expected under normal exposure scenarios”. In contrast, the IARC concluded that “there is strong evidence that glyphosate causes genotoxicity”.

Thus, while there are strict guidelines for regulatory study protocols, industry has been granted complete freedom in how the studies are interpreted for risk assessment. 

The regulatory process in the EU is clearly exposed in the glyphosate re-approval process as toothless, corrupt and unscientific (see [3]).  Both industry and our regulatory bodies are fully aware of the dangers of glyphosate herbicides.

A similar practice of dismissing and ignoring evidence was repeated with regard to the reproductive toxicity of glyphosate.

Lawsuits piling up as public react to Monsanto’s lies

It has become clear that Monsanto have known about the carcinogenic properties of glyphosate for many years (see [2] Glyphosate is Carcinogenic), and now that the toxicity of the herbicide is finally being recognised by the WHO, the public are beginning to challenge the company. Recent months have seen the filing of class action lawsuits for falsifying safety claims in the labelling of the herbicide (see [15] Fallout from WHO Classification of Glyphosate as Probable Carcinogen, SiS 67) as well as ‘mass tort’ personal injury lawsuits. Mass tort lawsuits let one attorney or several attorneys represent several injured parties at once. In the US, a Colorado-based law firm is putting together cases for 50 individuals, while in Delaware three lawsuits were filed by three firms representing three plaintiffs in recent weeks. In the case of 58-year old Enrique Rubio who lives in Colorado, the plaintiffs claim that he developed cancer after a decade of spraying Roundup on vegetable and fruit crops, unaware of the cancer risks [16]. Another plaintiff, Joselin Barrera who is only 24 years old and a child of migrant farm workers claims that her non-Hodgkin lymphoma was related to glyphosate exposure. A third plaintiff is a horticultural worker who contracted leukaemia [17].

Upcoming review on exposure levels by WHO also compromised by conflicts of interest

The WHO will also soon re-evaluate the findings of its various scientific bodies that differ in their interpretation of glyphosate safety. The JMPR (Joint Meeting of Pesticide Residues) are administered by the WHO as well as the UN Food and Agriculture Organisation (FAO) and meet regularly to review pesticide residue safety and set maximum exposure levels and recommended daily intake of chemicals. They last evaluated glyphosate in 2011, claiming it to be safe. Due to the divergence of opinion between the JMPR and the IARC, the WHO created an ad-hoc expert taskforce to explain the divergence of analyses, recommending a complete re-evaluation by the JMPR. The JMPR also have clear conflicts of interest, with 3 of the 8 members reportedly having financial and professional ties to industry [18], while one is the head of the BfR. The taskforce concluded that differences stem from the IARC including peer-reviewed literature as well as studies using commercial formulations in its analysis, calling for the JMPR to review internal guidelines for criteria for data inclusion/exclusion. In October 2015 the JMPR published a call-out for data for a meeting on glyphosate in May 2016. Calls for data are advertised on their website [19].

To conclude

EFSA have, perhaps surprisingly, admit to the potential genotoxic properties of glyphosate formulations [20], going against the general assumption that adjuvants are “inert” and can thus be kept as industry commercial secrets.  So, while declaring glyphosate safe and aiming to approve its use for another 10 years, EFSA are also admitting that the analysis by the IARC cannot be thus far disputed. Furthermore, adjuvants are added to all pesticide formulations, implicating a universal toxicity of chemical pesticides.

The US Environmental Protection Agency may be acknowledging the toxic nature of glyphosate herbicides to some extent, as it has just withdrawn approval for the currently in-use Enlist Duo herbicide from Dow Agrosciences that contains both glyphosate and 2,4-D. The decision came following new information provided by Dow to the EPA of its toxicity to non-target plants being more serious than previously understood [21].

In short, EFSA’s conclusion should be rejected outright as it results from a corrupt assessment process in which relevant data are arbitrarily dismissed in favour of confidential unpublished data from industry.

Article first published 30/11/15


  1. Addendum: Renewal Assessment Report. Glyphosate Addendum I to RAR, 31 August 2015. Clausing P.
  2. Ho MW and Swanson N. Glyphosate ‘Probably Carcinogenic to Humans’ Latest WHO Assessment,Science in Society 66, 16-18
  3. Swanson N and Ho MW. Scandal of glyphosate reassessment in Europe. Science in Society 63, 8-9, 2014
  4. The Glyphosate Renewal Assessment Report. An Analysis of Gaps and Deficiencies.
  5. IARC expert defends glyphosate cancer verdict., accessed 24th November 2016
  6. EFSA advises higher safety limits on glyphosate weedkiller., accessed 24th November 2015.
  7. Mesnage R, Defarge N, Spiroux de Vendômois J. Séralini GE. Potential toxic effects of glyphosate and its commercial formulations below regulatory limits. Food and Chemical Toxicology 2015, 84, 133e153
  8. Ho MW & Saunders P. Glyphosate is Carcinogenic. Chapter 8, Banishing Glyphosate, Special ISIS report 2015
  9. Mesnage, R., Defarge, N., Spiroux de Vendômois, J., Séralini, G.E. Major pesticides are more toxic to human cells than their declared active principles. BioMed Research International Volume 2014, Article ID 179691, 1-8   
  10. Koller VJ, Fürhacker M, Nersesyan A, Mišík M, Eisenbauer M, Knasmueller S. Cytotoxic and DNA-damaging properties of glyphosate and Roundup in human-derived buccal epithelial cells. Archives of Toxicology 2012, 86, 805-13. doi: 10.1007/s00204-012-0804-8
  11. Bernardi N, Gentile N, Mañas F, Méndez Á, Gorla N, Aiassa D. Assessment of the level of damage to the genetic material of children exposed to pesticides in the province of Córdoba. Archivos Argent Pediatría 2015, 113, 126-31
  12. Sirinathsinghji E. A Roundup of Roundup® Reveals Converging Pattern of Toxicity From Farm to Clinic to Laboratory Studies.Science in Society 65, 26-31, 2015.
  13. Ho MW and Saunders PT. Retracting Séralini study violates science & ethics. Science in Society 61, 20-21, 2014.
  14. Antoniou M, Habib M, Howard CV, Jennings RC, Leifert C, Nodari RO, Robinson C, Fagan J. Roundup and birth defects: Is the public being kept in the dark? Earth Open Source, 2011.
  15. Sirinathsinghji E. Fallout from WHO Classification of Glyphosate as Probable Carcinogen, Science in Society 67, 2015.
  16. Monsanto in Another Huge Lawsuit for ‘Lying About Roundup-Cancer Link’., accessed 23rd October 2015
  17. U.S. lawsuits build against Monsanto over alleged Roundup cancer link., accessed 23rd October 2015
  18. Groups Urge WHO to Set Safety Standards for Herbicide Glyphosate it Classified as Carcinogenic., accessed 27th October 2015.
  19. Joint FAO/WHO Meeting on Pesticide Residues (JMPR)., accessed 15th November 2015
  20. EFSA explains the carcinogenicity assessment of Glyphosate., accessed 15th November 2015.
  21. EPA moves to withdraw approval of controversial weed killer., accessed 25th November 2016

Got something to say about this page? Comment

Comment on this article

Comments may be published. All comments are moderated. Name and email details are required.

Email address:
Your comments:
Anti spam question:
How many legs on a cat?