Monsanto and regulators have known glyphosate causes cancer for nearly four decades, according to new paper

Monsanto has known for almost four decades that glyphosate causes cancer, according to a new paper by researchers Anthony Samsel and Stephanie Seneff.

Samsel is the first independent researcher to examine Monsanto’s secret toxicology studies on glyphosate. He obtained the studies, which have been denied to other inquirers, via a request to his senator. With his co-researcher Dr Stephanie Seneff of MIT, he reviewed Monsanto’s data.

Samsel and Seneff concluded that “significant evidence of tumours was found during these investigations”.

However, they add, “to create doubt and obscure the statistical significance of inconvenient findings, which may have prevented product registration”, Monsanto introduced irrelevant historical control data from other experiments. This is data from the control animals in other unrelated experiments, which may have been carried out under widely differing conditions.

This practice had the effect of creating experimental noise which cancelled out the statistically significant findings of increased tumours in any one experiment.

In various cancer experiments, Samsel and Seneff found, Monsanto introduced spurious data from 3, 5, 7 and even 11 unrelated study controls to effectively eliminate results, as needed.

In some instances, Samsel and Seneff noted, the experiments’ internal controls showed zero incidence of tumours, while the results for the glyphosate-treated groups showed a statistically significant increase. However, through what Samsel and Seneff call the “dishonest magic of comparing the findings to data from unrelated historical controls”, these findings were “explained away as a mystery and deemed not to be related to administration of the glyphosate”.

In their paper, Samsel and Seneff add, “Using these deviations effectively neutralized the inconvenient results and thus allowed the product to be brought to market. Had they not engaged in this deception, glyphosate may never have been registered for use.”

The authors also note disagreement within the US EPA about the safety of glyphosate: “EPA documents show that unanimity of opinion for product registration was not reached. Not all members of the EPA glyphosate review committee approved the registration of glyphosate. There were those who dissented and signed ‘DO NOT CONCUR’.”

Bad experimental practice

Use of historical control data to dismiss statistically significant findings of harm in an experiment is bad practice, since experiments will vary in crucial factors like dietary components and contaminants, laboratory conditions, and different animal genetics. Even the year in which an experiment is carried out can affect the results.  Introducing control data from these unrelated experiments only serves to create experimental “noise” which masks the findings of harm in the experiment in question.

For these reasons, even the OECD, which designed protocols for industry experiments on pesticides, warns against the practice being used to dismiss statistically significant findings of harm in any given experiment.

Using such data in this way is, however, a common practice in regulatory processes for pesticides and other chemicals. Several years ago, another team of independent researchers, including GMWatch’s Claire Robinson, found the same practice was used by Monsanto and regulators in dismissing findings of birth defects in industry studies on glyphosate.

Effects do not always increase with dose

Another phenomenon that was likely relied upon by Monsanto and regulators in dismissing the carcinogenic effects of glyphosate in the studies reviewed by Samsel and Seneff was the fact that according to the data presented in the new paper, the number of tumours in the animals did not always increase in a straight “ski-slope” line with the dose. In some cases, the mid dose produced fewer tumours than the lowest dose, though both doses produced more tumours than controls (no glyphosate administered).

In the case of birth defects, this lack of a “linear dose-response” was relied upon by industry and regulators to dismiss increased malformations in animals fed glyphosate as not related to the chemical.

It was likely used in the same way to dismiss the tumours in glyphosate-treated animals.

However, this is a scientifically outdated idea. It is now well known that some substances are more toxic at lower than higher doses, especially when endocrine (hormone) disruption is involved. The regulatory system has not caught up with this idea, which has been known to independent scientists since the early 1990s.

How does glyphosate induce cancer?

Samsel and Seneff conclude their new paper by explaining biological mechanisms through which glyphosate could induce cancer. These include glyphosate’s ability to bind (chelate) manganese, reducing its bioavailability, which could contribute to oxidative damage to cells. This in turn could lead to cancer.

The authors conclude in their paper, “We believe that the available evidence warrants a reconsideration of the risk/benefit trade-off with respect to glyphosate usage to control weeds, and we advocate much stricter regulation of glyphosate.”


The new paper:
Glyphosate, pathways to modern diseases IV: cancer and related pathologies
Anthony Samsel and Stephanie Seneff
Journal of Biological Physics and Chemistry 15 (2015) 121–159
Received 5 August 2015; accepted 24 August 2015; published 30 Sept 2015
Available online:
http://bit.ly/20HnvgA

The paper can be accessed on the website of the journal – it's behind a paywall: http://www.colbas.org/jbpc/poap-art-4.htm