Reassessing Roundup

Geoffrey Kabat’s illuminating essay “Who’s Afraid of Roundup?” (Issues, Fall 2019) should stir the reformational zeal of all who care about how scientific causal judgments are reached and implemented. As Kabat notes, the problems with the evaluations by the International Agency for Research on Cancer (IARC) are both substantive and procedural. The IARC’s disqualification of manufacturers’ consulting experts from its decisional process has silenced important voices, while permitting consulting experts of the plaintiffs’ bar, the lawsuit industry, to populate IARC working groups. The plaintiffs’ bar may be the largest rent-seeking group in the United States, and the IARC’s asymmetrical conflicts policy has allowed plaintiffs’ lawyers and other advocacy groups to have an undue influence on its evaluations.

Kabat’s illuminating essay should stir the reformational zeal of all who care about how scientific causal judgments are reached and implemented.

Kabat points to the confusing distinction between “hazard” and “risk,” which distinction is further obscured by extrapolations from high-exposure animal studies to humans with low exposures. Because of human defensive mechanisms, which must be overcome to induce cancer, the relevancy of high-exposure animal studies to lower human exposures may be lacking.

IARC evaluations are often undermined by undue methodological flexibility in the observational studies relied on by its working groups. Contrary to statistical best practices, many relied-upon epidemiologic studies analyze multiple agents, at various exposure levels, against dozens of outcomes. Similarly, the animal studies used in the glyphosate case, and in many others, lack prespecified endpoints, but recklessly declare “effects” whenever different rates yield p-values below 5%, regardless of how many statistical tests are conducted.

Although the IARC once enjoyed tremendous prestige and authority, its evaluations are now seen as suspect. Given its dubious methods and the rise of rigorous systematic review methodology, we should question the current role of IARC in our thinking about carcinogenesis.

Attorney, Schachtman Law
Former Lecturer in Law, Columbia Law School

Geoffrey Kabat has done an excellent job presenting IARC’s disregard for the scientific process and its arrogance toward the scientific community. The failures in IARC have undermined the public trust in regulatory science and provided ample kerosene for activist groups to pour on a chemophobic public. IARC, as Kabat rightly illustrated, has become a threat to the scientific method. He was, however, measured in his critique of the relationship that has been established between IARC and a group of scientists working for US tort law firms.

These American regulatory scientists frequently working with IARC have been implementing an alternative to the democratic risk assessment process they call “adversarial regulation.” They feel it is more effective to ban a substance or process by suing a company into compliance or bankruptcy via relentless lawsuits and activist campaigns. Unlike the risk assessment process, there is little need for evidence or scrutiny of data.

Kabat referred to the large number of lawsuits on Roundup as well as talc, both based solely on controversial evidence provided by IARC monographs. He indicated the important role IARC monographs play as (at times the only) evidence in these lawsuits. What needs more attention though is how the law firms suing these companies are combining forces with a large number of scientists who had worked on IARC monographs related to these substances (often, to IARC’s knowledge, acting as litigation consultants for these law firms during IARC Working Group meetings). In the case of benzene, IARC even held and produced a monograph solely on the basis that US tort law firms needed a direct link between non-Hodgkin’s lymphoma and benzene exposure. It appears that glyphosate was also rushed through to meet the tort lawyers’ objectives.

The Risk-Monger

Geoffrey Kabat properly takes the IARC to the woodshed for multiple offenses against science, reason, and the public good in its fallacious classification of glyphosate as a possible carcinogen. But he is far too kind and understates the case.

Regulatory agencies around the world have examined whether glyphosate is carcinogenic, and without exception concluded it is not. (IARC is not a regulatory agency.) The US Environmental Protection Agency’s position is representative of this consensus, and directly relevant to the consumer safety issues that are the entire and complete raison d’etre of the litigation over Roundup. The issue is not, in fact, a matter of scientific dispute, despite the claims of the plaintiff’s attorneys.

Kabat makes all this clear, and demonstrates that “what is at stake is society’s ability to rely on the best scientific evidence on questions that are entangled with competing interests and deeply held worldviews.”

If farmers lose the freedom to use glyphosate, which has been shown to be safer than table salt, baking soda, coffee, chocolate, and beer, they will be compelled to revert to obsolete weed control measures that are less efficient and less sustainable, increasing agriculture’s greenhouse gas emissions at a time we least need it. If this comes to pass, IARC will have been instrumental in perpetrating grave harm against the common good. They are deserving not of polite correction but frank and forthright condemnation. It’s long past time for funders to withdraw support from IARC and for the United Nations Food and Agriculture Organization to clean house.

Senior Fellow
Information Technology and Innovation Foundation

The apparent acceptance by so many in the scientific community of the classification by the IARC Monographs Program of glyphosate as a “probable carcinogen” is evidence of the failure of the vaunted self-correction mechanism of science. It is scandalous that despite the absence of credible evidence that glyphosate is associated with cancer risk from animal, human, or laboratory studies, the IARC classification still stands almost five years after it was announced. It is incumbent upon the Monographs Program either to refute the demonstration of serious scientific flaws in the Monograph 112 Working Group deliberations on glyphosate, or to retract its erroneous classification of glyphosate. My published critiques of the IARC glyphosate deliberations accepted that IARC was evaluating hazard not risk, that the IARC criteria for identifying a carcinogenic hazard were appropriate, and that the body of studies relied on by IARC was sufficient to come to a valid conclusion regarding the potential carcinogenicity of glyphosate. The IARC criteria do not support the conclusion that glyphosate is a probable carcinogen based on an honest, rigorous, and complete synthesis of all of the evidence in studies relied upon by IARC.

In his article, Geoffrey Kabat notes procedural irregularities in addition to serious misrepresentations of scientific evidence in the IARC glyphosate report. The failure of the Monographs Program to follow strictly its published protocol is disturbing. In March 2014 an advisory group chaired by Christopher Portier added glyphosate to the list of agents given “medium priority” for evaluation in future Monographs. A call for experts for Monograph 112 in July 2014 included an agenda that mentioned only organophosphate insecticides. In an October 2014 IARC announcement of upcoming meetings, glyphosate had been added to the agenda for Monograph 112. This was after the call for experts period had closed. The Monographs Program should disclose when the addition of glyphosate to the Monograph 112 agenda was first publicly acknowledged, and explain the rush to evaluate a medium priority agent so quickly.

It is past time for the IARC Monographs Program to be held accountable for its erroneous glyphosate classification. The absence of such accountability puts the long-term scientific value of the Monographs Program at risk.

Retired; National Cancer Institute (28 years) and International Epidemiology Institute (14 years)

(Dr. Tarone participated as an unpaid fact witness in two California trials involving Monsanto, the maker of Roundup, in January 2020)

Geoffrey Kabat has misrepresented the difference between hazard assessment (HA) and risk assessment (RA). This is important to the Roundup story because the IARC, which conducts HA, has classified glyphosate, the active ingredient in Roundup, a “probable carcinogen,” while several national and international regulatory agencies cited by Kabat conduct RA, and their conclusions support the safety of the herbicide. The evaluations conducted by IARC and by the regulatory agencies are designed to answer different questions, and so it is not surprising that they reach different answers.

In the Preamble to its Monograph series in which agent evaluations are published, IARC states: “A cancer ‘hazard’ is an agent capable of causing cancer, while a cancer ‘risk’ is an estimate of the probability that cancer will occur given some level of exposure to a cancer hazard.” HA tries to answer a fundamental question about whether a substance can cause cancer, explicitly separating that question from a second important question: how large is the risk of cancer from a substance in a specific exposure scenario? RA, which addresses this second question, is conducted somewhat differently by different agencies, but fundamentally requires two critical additional pieces of information that are not needed for HA: first, a judgment about how large a risk is “acceptable” or “unacceptable,” and second, an assumption about a specific exposure scenario.

There are uncertainties inherent in both HA and RA, but Kabat’s characterization of the difference is not helpful to readers wishing to understand the debates surrounding the safety of glyphosate. He attempts to trivialize the HA process by saying that “IARC considers any scientific evidence of possible carcinogenicity, no matter how difficult to interpret or how irrelevant to actual human exposure.” Setting aside the pejorative tone (according to Kabat we should use only evidence that is easy to interpret?), one can read between the lines a correct distinction—that IARC does not consider human exposure scenarios because these are not relevant to the fundamental question of whether a substance has the capacity to cause cancer.

He also fails to point out that along with those additional assumptions needed for RA come many additional opportunities for error. Because of its complexity and the number of required assumptions, it can be difficult for policy-makers or the public to understand how dependent the results may be on biases and limitations in the data. About 20 years ago, a now-famous experiment was conducted in which four different RA teams were provided with the same set of data with which to evaluate the cancer risk to children wearing pajamas impregnated with a flame-retardant chemical known as tris. Their calculations of the numbers of additional lifetime kidney cancers per million exposed children differed by three orders of magnitude—from 7 to 17,000 additional cancers, or perhaps you might say, from reassuring to not so much.

I think it is useful for the public to know that a group of experts has evaluated all the evidence and decided that glyphosate is probably a carcinogen—a statement that acknowledges the uncertainties in the evidence.

Kabat provides a long list of regulatory agencies that have declared glyphosate to be safe. I went to the public website of the first one on his list, Health Canada. Here’s what it says: “products containing glyphosate do not present unacceptable risks to human health or the environment when used according to the revised product label directions.” I do not doubt the integrity and rigor of Health Canada’s review, but I find these words only modestly reassuring: What is unacceptable risk? And what happens when the pesticide is not used as directed? IARC’s hazard assessment does not depend on assumptions about acceptable risk and about how the pesticide is used or misused. These are the kinds of uncertainties and easily hidden assumptions that probably underlay the thousandfold differences among the RA calculations in the tris experiment mentioned above.

There will always be substantial uncertainties about the safety of newly invented chemicals such as glyphosate. And there is urgent need to know whether to allow their widespread use. Under these conditions, I think it is useful for the public to know that a group of experts has evaluated all the evidence and decided that glyphosate is probably a carcinogen—a statement that acknowledges the uncertainties in the evidence. It doesn’t say what to do. But that’s someone else’s job.

Professor and Director
Lowell Center for Sustainable Production
University of Massachusetts Lowell

The glyphosate affair and subsequent rulings exemplify two critical issues. One is how science is generated and communicated to the public and how that may be co-opted by single-issue advocates. The other is the role of corporate power in manipulating scientific processes, institutions, and public opinion when profits are at risk. Geoffrey Kabat focuses on the former while ignoring the latter. Both are of importance to science and society.

The glyphosate court rulings showed how Monsanto, as a company that perceived the IARC report as a threat to its profits, reacted in ways that were “anti-scientific,” much as the tobacco industry did before it. These tactics included trying to engineer journal retractions from unhelpful studies, ghostwriting scientific and journalistic articles, and coordinating efforts to criticize not just the IARC determination on glyphosate, but IARC as an institution and its role more generally.

Third-party groups joined in with these tactics. These efforts ironically included social media and meme-ification efforts. For example, a group called the Campaign for Accuracy in Public Health Research, which arose from the American Chemistry Council, also then funded by Monsanto, posted graphics on social media with statements such as “IARC’s cancer assessments are nothing but smoke and mirrors.”

Kabat is right to point out that hazard determinations should not become unscientifically weaponized on social media. However, the use of such tactics by commercial interests underline how important it is to also ensure the integrity and independence of those institutions that assess toxicity. What’s concerning is that court proceedings also revealed evidence that Monsanto funded the American Council of Science and Health, where Kabat serves as an adviser, with a view to help in discrediting IARC. There are evidence-based and commonsense reasons for disclosing relationships of this nature.

In making the moral and empiric case for science, it is more important than ever that we view risk clear-mindedly. To focus solely on single-issue advocates as barriers to progress, while ignoring the challenge that corporate conflicts of interest pose to scientific evidence, processes, and institutions, is a risk we cannot afford to take.

2019–2020 Harkness Fellow in Health Care Policy and Practice
School of Public Health
Boston University

Cite this Article

“Reassessing Roundup.” Issues in Science and Technology 36, no. 2 (Winter 2020).

Vol. XXXVI, No. 2, Winter 2020