Split Within EPA on Glyphosate Carcinogenicity

The recent release in litigation of email correspondence between the EPA Office of Pesticide Programs (OPP) and the pesticide manufacturer Monsanto reveals a significant difference in the method of assessing the potential carcinogenicity of the widespread pesticide glyphosate, with the Pesticide Office using methods aligned with Monsanto’s approach and contrary to the Agency’s own Cancer Guidelines.

As the EPA faces a devastating 31 percent budget cut under President Trump, various programs are counting on support from environmental and public health groups to validate the important role played by the EPA. The pesticide office, on the other hand, has Crop Life America, the pesticide industry trade group, going to bat to protect its budget so that pesticide approvals won’t be delayed. 


GreenWire, March 7, 2017

The cache of documents released reveals a sharp division between EPA’s science program, the Office of Research and Development (ORD) and the pesticide office over its glyphosate cancer assessment. The pesticide office―and Monsanto―characterizes glyphosate as “not likely to be carcinogenic to humans.” Glyphosate is one of the most widely used pesticides in the world, with about 250 million pounds used annually on agriculture crops in the U.S., mainly―but not only―on corn and soybean crops in the Midwest, and an additional 25 million pounds used for non-agriculture purposes, such as parks, golf courses and residential lawns. It is a key ingredient in Round Up, Enlist Duo and other herbicides.

USGS Pesticide National Synthesis Project, Pesticide Use Maps, Glyphosate

The declaration of glyphosate’s safety in September 2016 by EPA’s pesticide office and prior to that by Monsanto is in stark contrast to the finding of the International Agency for Research on Cancer (IARC). IARC convened a meeting of 17 scientific experts from 11 countries and in March 2015 finalized its assessment concluding that glyphosate is “probably carcinogenic to humans.”  IARC has been aggressively targeted by Monsanto ever since. But an internal memo released in the litigation reveals  that EPA’s science program (which has been threatened with a 50% budget cut by the Trump Administration) was more in line with IARC’s science assessment and illustrates how,  in conducting its own analysis, the pesticide office failed to follow EPA Cancer Guidelines (ORD memo, Dec 14, 2015).

The difference between the two divergent conclusions is as follows:

The Problem

Monsanto and  EPA Office of Pesticide Programs (EPA-OPP)

IARC and EPA Office of Research and Development (EPA-ORD)

What the EPA Cancer Guidelines say

Monsanto and EPA-OPP disregarded evidence of harm in human population studies (epidemiology).

Used a “yes/no” approach for the epidemiological evidence, so not enough evidence for a “yes” meant a default conclusion of “no” cancer risk.

Concluded there was some (limited) epidemiologic evidence of an association between glyphosate and non-Hodgkin's lymphoma (NHL).

EPA includes gradations of causality, including ‘likely’ and ‘suggestive’ evidence for causality. IARC and EPA-ORD did it correctly.

Monsanto and EPA-OPP dismissed tumor evidence in animal studies by comparing them to tumor rates in animals from different labs over different years in the past.

Cherry-picked the control groups by comparing animals with tumors to historical control animals from other labs in past years resulting in the impression that the elevated cancer risk was within “historical control” levels, therefore not significant.

Compared animals with tumors to concurrent (within the same experiment) control animals, demonstrating that the glyphosate-treated animals had an elevated cancer risk, compared with control (untreated) animals.

The default position should be to not rely on historical control data except when concurrent controls yield clearly unreliable results. IARC and EPA-ORD did it correctly.

Monsanto and EPA-OPP dismissed tumor evidence in animal studies in the high dose treatment groups.

Disregarded cancer tumors at high doses, using the inappropriate excuses that the dose was too high, and the absence of a linear (monotonic) exposure-response relationship.

Included tumors in high dose treatment group. Concluded there is sufficient evidence of carcinogenicity in experimental animals.

Effects at high doses are presumed to be appropriate, unless evidence demonstrates they are the sole result of toxicity. The absence of an exposure-response relationship does not exclude a causal relationship. IARC and EPA-ORD did it correctly.

Monsanto and EPA-OPP cherry-picked the statistical method that didn’t find elevated cancer risk.

Focused on pairwise comparisons (which were generally not statistically significant).

Conducted pairwise and trend tests, which yielded several significant results, indicating elevated cancer risk.

Trend tests and pairwise comparison tests are both valid – significance in either test is sufficient. IARC and EPA-ORD did it correctly.

Monsanto and EPA-OPP dismissed evidence that glyphosate damages DNA (genotoxic).

Failed to identify evidence of genotoxicity (a mechanism of cancer), relying on a review article sponsored by Monsanto (Kier and Kirkland 2013).

Reviewed original studies. Concluded there was strong evidence that glyphosate is genotoxic (causes DNA and chromosomal damage in human cells).

“A lack of mechanistic data is not a reason to reject causality.” Even if Monsanto and EPA-OPP didn’t think that glyphosate was genotoxic, had it acknowledged the evidence from animal and human studies, it would have to find that glyphosate was associated with elevated cancer risk.

(See details in NRDC’s comments to the EPA Science Advisory Panel, November 2016)

In early 2017 noted cancer expert Dr. Christopher Portier used the Freedom of Information Act to obtain the raw data for the animal cancer studies for glyphosate conducted a careful reanalysis. He found that the data show eight instances where significant increases in tumor response following glyphosate exposure were identified, but had not been included in previous cancer assessments by either the European authorities or the U.S. EPA. Dr. Portier says that, "this suggests that the evaluations applied to the glyphosate data are scientifically flawed, and any decisions derived from these evaluations will fail to protect public health."

From years of submitting comments to this office, I believe that it’s not just glyphosate and Monsanto that has received favorable treatment. NRDC sued the EPA pesticide office―and won―after proving it had inappropriate contact with Syngenta, the maker of atrazine, during the period that atrazine was undergoing regulatory approval.

Pesticides are poisons. They are designed to kill things. Recognizing this danger decades ago, Congress passed the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), which gives the EPA pesticide office the authority to regulate, restrict, and even ban pesticides. Under FIFRA, all pesticides must be registered by the EPA before they can be legally sold or distributed in the United States and then undergo registration review every fifteen years. Under this same law EPA cannot renew the registration of a pesticide until it ensures that the pesticide’s use will not pose unreasonable adverse effects on the environment or human health.

The public will not be getting the protection it needs and deserves until the EPA Pesticide Office―under all Administrations―begins to independently and objectively evaluate the science, without a thumb on the scale to help industry, and, at a minimum begins to follow the Agency’s own scientific guidelines. 

About the Authors

Jennifer Sass

Senior Scientist, Health program

Join Us