Was Roundup approval based on "bad science" ?

A new peer-reviewed study shows that glyphosate, the main ingredient of Roundup, was approved on the basis of bad science that overlooked findings of birth defects in industry’s own studies, says Claire Robinson, co-author of the new study. She adds that those industry studies, currently kept secret, may contain evidence of cancer-causing effects that would lend support to Prof GE Seralini’s new research.

After Professor Séralini published his study revealing the toxic effects GM maize NK603 and Roundup herbicide, five of France’s former environment ministers called for the way in which pesticides and GMOs are evaluated for safety to be overhauled. [1] [2] Séralini himself has asked for this for years.

As the co-author of a new study [3] that questions the reliability of the EU approval of glyphosate, the main ingredient of Roundup, I support the ministers and Séralini in their call. Ours is the first peer-reviewed study to look in detail at how German government and EU authorities assessed the industry studies on glyphosate.

"Our findings are worrying."

We found that industry’s own studies dating back as far as the 1980s and 1990s, including many by Monsanto, showed that glyphosate caused birth defects in laboratory animals. The warning was there, even then, for those who cared to see it.

The task of reporting on the industry studies fell to Germany because it was the “rapporteur” member state for glyphosate, responsible for liaising between industry, the EU authorities and member states in the approval process.

But the German authorities minimized the findings of birth defects in their 1998 report to the EU authorities and member states. They used unscientific practices such as using irrelevant “historical control data” to make findings of birth defects disappear amid the resulting “data noise”. They even redefined a birth defect as a “developmental variation”.

As a result, the German authorities set – and the EU accepted in 2002 – a “safe” level for glyphosate exposure that is not safe at all.

Now, Séralini’s experiment shows that a level thousands of times lower than the EU “safe” level is, in fact, highly toxic, causing early death, increased tumours and organ damage in rats.

Germany remains the rapporteur for glyphosate and is currently re-assessing the chemical for the EU renewed approval scheduled for 2015. In October the German office for risk assessment, BfR, issued a statement dismissing Séralini’s study, citing long-term studies in rats and mice that it said showed “no … carcinogenic potential”. [4]

Glyphosate has been tested alone

There’s a problem with these tests, however, and BfR knows it. The tests, commissioned by industry and submitted in support of glyphosate’s approval, are on glyphosate alone. But the herbicide formulations as they are sold and used contain added ingredients that BfR admits “might affect the toxicity”. Studies by Séralini’s team and other scientists confirm that the formulations are more toxic than glyphosate alone. [5] [6] [7] [8] BfR notes “with interest” that Séralini’s study was the first long-term feeding study on a complete herbicide formulation.

It is absurd that BfR accepts industry studies on glyphosate alone as proof of the safety of Roundup when it is concerned about the safety of the added ingredients.

Nevertheless, BfR finds “deficiencies” in Séralini’s study design that it thinks justify consigning the study to the waste bin.

BfR’s reaction did not surprise me. Germany has defended the safety of glyphosate against a growing body of research by independent (non-industry) scientists – including Séralini – that show glyphosate herbicides can cause human cell death, neurological problems, birth defects, miscarriage, hormonal disruption, and certain types of cancer. Some effects are found at low, realistic doses.

Now Germany says that the industry tests on glyphosate show it doesn’t cause cancer. The problem is, they don’t show that at all.

Industry studies kept under "commercial confidentiality"

We cannot check the industry studies themselves, which are kept hidden under “commercial confidentiality” agreements between industry and regulators. Attempts by Pesticide Action Network Europe to force them into the open via the Dutch and German governments and the EU Commission have met with a “no”.

All that is in the public domain is the German authorities’ report on the industry studies, which we analyzed for our study in relation to birth defects. [9]

In fact, studies summarised in Germany’s report show that treatment with glyphosate, including at the lower doses tested, increased tumour incidence. But Germany dismissed the findings on the grounds that the tumours did not increase in a straight line as the dose increased (linear dose-response).

This is incorrect reasoning, based on outdated scientific concepts. Scientists have published papers since the 1990s showing that in the case of certain toxins, the toxic effect does not predictably increase in a straight line upwards as the dose increases (the “ski slope” pattern). Instead, some toxins can have a strong effect at a low dose and a weaker effect at a higher dose. Other weird and wonderful dose-response curves can take the shape of a “U” or even a mountain range. Such nonlinear dose-response curves are often found with substances that affect the hormonal system. [10] [11]

No linear dose effect

Séralini’s study confirms that Roundup’s toxic effects do not occur in a linear dose-response fashion. Robin Mesnage, a co-author of Séralini’s study, commented on Germany’s interpretation of the industry data: “I am very concerned about the carcinogenesis data of these tests. Many studies reported an increased incidence of tumours but Germany dismissed them because of the nonlinear dose-response.

“Toxic effects involving hormones are by nature nonlinear in relation to the dose. This was true of the mammary tumours we observed with Roundup. But regulation of pesticides still wrongly assumes that dose-response must be linear.”

So the question of whether Roundup causes cancer remains open. What is clear is that long-term tests must be carried out on all pesticides as they are sold and used, as well as on GMOs. Meanwhile, the industry tests and raw data on which pesticide and GMO approvals are based must be made public on the internet.

Author : Claire Robinson, Scientific journalist, Earth Open Source, november 22th 2012.

Claire Robinson is also the co-author of the book GMO Myths and Truths





[1] Le Nouvel Observateur with AFP (2012) OGM : Royal, Lepage et Voynet veulent revoir les autorisations [GM: Royal, Lepage and Voynet want authorizations reviewed]. 28 October.

[2] 20minutes.fr and AFP (2012). OGM: NKM et Chantal Jouanno se rallient à l’appel des trois anciennes ministres de l’Ecologie [Nathalie Kosciusko-Morizet and Chantal Jouanno rally to the call of three former environment ministers]. 20minutes.fr. 9 November.

[3] Antoniou M, Habib MEM, Howard CV, Jennings RC, Leifert C, Nodari RO, Robinson CJ, Fagan J (2012) Teratogenic Effects of Glyphosate-Based Herbicides: Divergence of Regulatory Decisions from Scientific Evidence. J Environ Anal Toxicol S4:006. doi:10.4172/2161-0525.S4-006 http://www.omicsonline.org/2161-0525/2161-0525-S4-006.php?aid=7453 Note: This study, published online on 7 November 2012, includes material previously published as: Antoniou, M., M. Habib, et al. (2011). Roundup and birth defects: Is the public being kept in the dark?, Earth Open Source. http://bit.ly/IP2FWH The new study has been updated with new research. This includes a re-evaluation of findings by Rull et al (2004, 2006) that found a modest association between exposure to Roundup and neural tube defects, a type of birth defect; and Argentine research showing an association between birth defects and pesticide exposure in an area of intensive cultivation of GM soy, confirming long-standing reports by residents and doctors.

[4] BfR (Germany) (2012) A study of the University of Caen neither constitutes a reason for a re-evaluation of genetically modified NK603 maize nor does it affect the renewal of the glyphosate approval. 1 October. http://bit.ly/Sz6lRI

[5] Dallegrave, E., F. D. G. Mantese, et al. (2002). Acute oral toxicity of glyphosate in Wistar rats. Online J Vet Res 1: 29–36

[6] Richard, S., S. Moslemi, et al. (2005). Differential effects of glyphosate and roundup on human placental cells and aromatase. Environ Health Perspect 113(6): 716-720.

[7] Mesnage, R., B. Bernay, et al. (2012). Ethoxylated adjuvants of glyphosate-based herbicides are active principles of human cell toxicity. Toxicology

[8] Koller, V. J., M. Furhacker, et al. (2012). Cytotoxic and DNA-damaging properties of glyphosate and Roundup in human-derived buccal epithelial cells. Arch Toxicol 86: 805–813.

[9] Rapporteur member state Germany (1998). Monograph on glyphosate, German Federal Agency for Consumer Protection and Food Safety (BVL). Vol 3-1 Glyphosat 04.

[10] Vandenberg, L. N., T. Colborn, et al. (2012). Hormones and endocrine-disrupting chemicals: Low-dose effects and nonmonotonic dose responses. Endocr Rev.

[11] Fagin, D. (2012). Toxicology: The learning curve. NATURE 490: 462-465.