Following an executive order from the Trump administration that promotes production of glyphosate, some Democrats have claimed that the herbicide causes cancer. The science, however, is nuanced. While there is some evidence linking glyphosate to cancers in lab animals or to the blood cancer non-Hodgkin lymphoma in agricultural workers, the findings have been inconsistent.
Regulatory agencies around the world, including the U.S. Environmental Protection Agency, have concluded glyphosate is unlikely to pose carcinogenic risks.
In a Feb. 18 executive order, President Donald Trump promoted production of glyphosate-based herbicides — originated in 1974 by Monsanto as the weedkiller Roundup — as necessary for national security. The move was widely viewed as counter to the Make America Healthy Again, or MAHA, movement, which generally opposes pesticides, and prominently glyphosate. Bayer, which acquired Monsanto in 2018, is the only company that makes glyphosate in the U.S., although there are also imported generic versions.
Health and Human Services Secretary Robert F. Kennedy Jr., the de facto MAHA leader, has long said that glyphosate causes cancer, although he defended the executive order.
Democrats quickly noted the contradiction — and proceeded to make claims of their own about glyphosate.
“This executive order is a slap in the face to the thousands of Americans who have gotten cancer from glyphosate,” Sen. Cory Booker, a Democrat from New Jersey, said in a Feb. 19 statement.
Democratic Sen. Ed Markey of Massachusetts, meanwhile, brought up glyphosate during the Feb. 25 confirmation hearing for the surgeon general nominee, stating that Trump is “siding with the chemical manufacturing company that is, in fact, causing the cancers.”
Even as he defended Trump’s action, Kennedy has continued to indicate that glyphosate is dangerous. In a Feb. 27 appearance on the “Joe Rogan Experience,” for example, he mentioned the link to NHL, the blood cancer found in some but not other studies of people who apply glyphosate.
Other Republicans, such as Rep. Nancy Mace of South Carolina, have also responded, although she did not make as strong of a claim about cancer, saying only that glyphosate “has been linked” to cancer.
“Glyphosate and other pesticides don’t belong on our food or in our children’s bodies,” she wrote in a March 8 post on X. “We are systematically poisoning ourselves.”
There is little to suggest glyphosate causes cancer in the trace amounts present in food. Some studies have identified associations between glyphosate exposure and cancer, either in humans who used the herbicide or in animals exposed in the lab. But the findings have been inconsistent, and researchers have come to differing conclusions about the overall evidence.
Results from a large National Institutes of Health study assessing exposure in agricultural workers, published in 2017, did not find an association between glyphosate and NHL or other cancers. This lack of a concrete connection has led many regulatory agencies to conclude glyphosate is unlikely to cause cancer.
At the same time, a widely cited 2015 report from the World Health Organization’s International Agency for Research on Cancer deemed glyphosate “probably carcinogenic to humans,” based on lab animal data and “limited” real-world evidence linking glyphosate to cancer in humans.
“The overall picture with glyphosate is messy,” David Eastmond, a professor emeritus at the University of California, Riverside, who studied genetic toxicology and chemical carcinogenesis, told us. He served on a 2016 committee of the WHO and the Food and Agriculture Organization of the United Nations that found human dietary glyphosate exposure was unlikely to cause cancer. “The human studies are messy, the animal studies are messy, the mechanistic studies are messy. And so within that messiness, you try and draw conclusions, and different people interpret that in different ways.”
Below, we will walk through the evidence about glyphosate that regulators and others have assessed, as well as more recent evidence being considered.
Widespread Exposure, But Little Agreement on Risks
Glyphosate-based herbicides are the most commonly used weedkillers in the world. As such, wide swaths of people come into at least some contact with them.
Monitoring by the Centers for Disease Control and Prevention has found that most people have some detectable glyphosate in their urine, although researchers from the agency have noted that this on its own “does not mean that glyphosate causes disease or adverse effects.” Glyphosate does not significantly build up in the body and is rapidly cleared.
Agricultural workers are likely to have the highest exposures to glyphosate. It can also be found in trace amounts in a variety of foods, particularly grains and legumes. People living near fields while they are being sprayed have been found to have elevated levels in their urine compared with those living farther away.
In addition to being used on farms, glyphosate-based herbicides were historically sold for residential use, although beginning in 2023 Bayer has sold new products that include herbicides other than glyphosate, citing the need to “further reduce future litigation risk.”
Despite such litigation, it’s unclear what impact exposure to glyphosate-based herbicides — designed to interfere with a key pathway shared by plants and some microbes but not humans — has on people and at what level.
Glyphosate is not very acutely toxic. Scientists can test the acute toxicity of a chemical by feeding it to rodents and measuring the dose at which half of the animals have died. It takes more than 4,000 milligrams of glyphosate per kilogram of body weight to kill half of rats; this means glyphosate is less acutely toxic than table salt. However, for cancer, scientists are interested in long-term effects.
Some researchers say the evidence overall does indicate glyphosate can cause cancer. “Glyphosate and glyphosate-based herbicides (GBHs) harm human health and can cause cancer,” a group of 50 physicians, scientists and others — including the MAHA activist Kelly Ryerson — wrote in a March 27 statement. “The comprehensive evidence supports this conclusion, with the strongest epidemiological evidence linking exposure to increased risk of non-Hodgkin lymphoma, a cancer of the lymphatic system.” The statement followed a symposium on the health effects of glyphosate held at the University of Washington, which brought together academic and government researchers, consultants, lawyers, and representatives from nonprofit organizations.
Others have been less convinced, including, as we have said, regulators in a variety of regions and countries, including Canada, Japan and the European Union. Some epidemiologists and health communicators have pointed out that any cancer risks in rodents have generally been shown at doses higher than a person typically would be exposed to via their diet, while allowing that there may be concerns for people with more extreme exposures. And as we have said, a large, rigorous epidemiological study in humans did not show an association between glyphosate and cancer.
Adding complexity to this debate, there is a long history of concern over the influence Monsanto may have exerted over the scientific literature on its product’s safety. (Bayer acquired Monsanto in 2018.) In December, a journal retracted a 2000 review paper on glyphosate’s safety because a Monsanto employee had suggested in an internal email that it was ghostwritten.
A Dec. 4 statement from Bayer said that Monsanto’s role in the 2000 paper “did not rise to the level of authorship and was appropriately disclosed in the acknowledgments.” In a statement shared with us via email, a Bayer spokesperson emphasized the safety and extensive testing of the company’s glyphosate-based products: “The fact is that no health regulator anywhere in the world has ever found glyphosate to pose a threat to human health.”
Meanwhile, following the 2015 designation of glyphosate as “probably carcinogenic” by the WHO’s International Agency for Research on Cancer, people with NHL, working with lawyers including Kennedy, brought thousands of lawsuits against Bayer alleging harm from Roundup. (An aide for Booker, the senator from New Jersey, told us via email that the “estimate that thousands of Americans have gotten cancer from glyphosate is supported by the lawsuits brought by thousands of people in the United States who developed cancer after using glyphosate-based herbicides.”)
Bayer on Feb. 17 proposed a $7.25 billion settlement of current and future cases. Meanwhile, the Supreme Court will hear arguments this month over whether people can bring cases against Bayer under state law alleging failure to warn about harms on the labels for glyphosate-containing products. (The Trump administration filed a Dec. 1 amicus brief supporting Bayer’s position.) Advocacy groups have also challenged the EPA’s conclusions. The EPA is supposed to issue a revised decision by October.
“This year, EPA will undertake a comprehensive, transparent, and rigorous scientific review of glyphosate to evaluate its use and ensure decisions are fully aligned with the best available science as well as human health and environmental protections,” an EPA spokesperson told us via email.
The glyphosate litigation has brought in scientists to serve as expert witnesses for both sides.
“We all have biases to some degree, but some are influenced by external factors,” Eastmond said. He brought up stories about Monsanto’s ghostwriting, as well as the conflicts that can come from testifying as an expert witness. “If you’re working on one side or the other, you tend to study and focus research to support that point of view,” he said. He added that he is not aware of conflicts of interest on his part.
Another possible explanation for varying conclusions between IARC and pesticide regulators is that the groups had different procedures and were assessing different questions. IARC was assessing whether glyphosate is a hazard — i.e., whether it has the theoretical ability to cause harm. Some other groups were assessing glyphosate’s risk, or how likely glyphosate is to be causing harm under certain circumstances, such as under typical exposures.
For example, the 2016 committee from the WHO and U.N. Food and Agriculture Organization that assessed glyphosate was tasked with determining whether dietary exposures from very low levels of pesticide residues came with cancer risk, which is different from the question of whether some very high level of exposure could cancer. Regulators also tend to assess risk under realistic levels of exposure.
However, a look at different groups’ and scientists’ arguments also reveals more fundamental disagreements on how to interpret the science, and multiple situations where evaluating carcinogenicity is not cut-and-dried.
Inconsistent Evidence in Humans
The available studies in humans come to differing conclusions about whether glyphosate is associated with cancer in people who apply the herbicide. Meanwhile, there isn’t evidence in humans that low-level exposures in food are associated with cancer. It is challenging to study whether glyphosate causes cancer in humans both because cancer takes many years to develop and because it is tricky to assess how much of the herbicide people have been exposed to over a stretch of time.
At the time that IARC assessed the human evidence of glyphosate’s carcinogenicity as “limited,” there were half a dozen studies assessing glyphosate and NHL in humans, Laura Beane Freeman, an epidemiologist at the National Cancer Institute, explained during a March 25 presentation at the Seattle Glyphosate Symposium. “Most, but not all, of the studies had some evidence of an association with non-Hodgkin lymphoma overall,” she said. “And I’m using that term loosely. It doesn’t necessarily mean statistical significance, it just means some evidence of a positive association.”
The studies that initially raised concerns were case-control studies. This type of study identifies people who developed a type of cancer in a population, as well as controls from the same population who did not have cancer, and then assesses their exposure in retrospect. The studies relied on asking participants or their family members about past glyphosate exposure.
In a review of the evidence published in 2017, the EPA pointed out that not all of the studies took into account whether people were exposed to other pesticides, which could have had their own health effects, and that many studies had small sample sizes. “In epidemiological studies, there was no evidence of an association between glyphosate exposure and numerous cancer outcomes; however, due to conflicting results and various limitations identified in studies investigating NHL, a conclusion regarding the association between glyphosate exposure and risk of NHL cannot be determined based on the available data,” the agency review concluded.
The Agricultural Health Study is a prospective cohort study that enrolled licensed pesticide applicators and has followed them for many years. An advantage of this sort of forward-looking study is that people’s estimates of how much pesticide they used cannot be biased by knowing whether they later went on to develop cancer, unlike in studies that ask people with cancer to look back at their past exposures. In addition, it is easier for this sort of study to look at a greater variety of cancer types.
A 2005 analysis of the study did not find an association between glyphosate and cancer. A 2018 updated analysis of the more than 54,000 participants also found no association between glyphosate use and any cancer type. (For acute myeloid leukemia, there was a numerically higher number of cases in farmers with the highest exposures, but the result was not statistically significant.)
For some, the negative results in the AHS are convincing, particularly given the fact that glyphosate use has increased since it came to market in the 1970s but NHL has slightly fallen overall since its peak in 2007. “The strongest study to date in my understanding is the Agricultural Health Study,” Eastmond said. “They just didn’t see any evidence” for cancer, with the exception of the possible increase in AML.
“That long-term study of agricultural workers, with a relatively well-defined exposure, over now approaching 20 years, shows no evidence of a risk of cancer,” Alan Boobis, an emeritus professor of toxicology at Imperial College London, told us. Boobis led the FAO/WHO committee that evaluated glyphosate in 2016.
Other researchers have been reluctant to interpret the AHS as vindicating glyphosate. “Even though the Agricultural Health Study was largely negative, there are other studies that were strongly positive,” Dr. Philip Landrigan, a pediatrician and public health physician at Boston College who signed the Seattle Glyphosate Symposium statement, told us.
At the symposium, he called a 2019 meta-analysis the “most noteworthy” of the newer studies in humans. (Meta-analyses also attempt to make sense of the data overall by combining results from multiple studies.) A spokesperson for Mace, the representative from South Carolina, had highlighted this study when asked about the data behind her concerns about glyphosate.
The study found that groups reporting the highest level of glyphosate-based herbicide exposure had a 41% higher rate of NHL than those who did not report use.
“I actually do think the scientific evidence is really strong” implicating glyphosate and glyphosate-based herbicides as carcinogens, Luoping Zhang, the first author of the study and an adjunct professor emerita of toxicology at the University of California, Berkeley, told us. Zhang was on a 2016 EPA panel that reviewed glyphosate and was one of the signers of the Seattle Glyphosate Symposium statement. She has been an expert witness for plaintiffs in glyphosate lawsuits.
However, a 2020 EPA review of Zhang’s meta-analysis questioned whether the researchers had a good rationale for zeroing in on the highest-exposure groups. The review emphasized that the updated AHS study — which it called “the largest study and of the highest quality” — found no sign of an increasing risk of NHL in people exposed to higher levels of glyphosate.
Zhang defended her team’s choice to look at high-exposure groups as common sense. “If you are thinking exposure to chemical A can cause cancer, everybody would believe the more you expose, the higher level you expose,” the higher the chance of cancer, she told us.
Divergent Readings of Rodent Studies
Scientists often look at data in rodents to better understand whether a chemical is likely to be harmful to humans, as it is possible to expose the mice and rats to precise quantities of the substance and assess its effects over a relatively short period of time. Again, groups diverged in their evaluation of the data on glyphosate, with IARC finding “sufficient” evidence in animals that it could cause cancer and regulators viewing the rodent cancer data more skeptically.
An important factor is that different groups reviewing glyphosate did not rely on exactly the same data, Eastmond said. IARC only considers data that the public has access to. Regulatory agencies consider proprietary data submitted from companies, and the FAO/WHO group also gained access to this data.
Some scientists have contended that IARC did not properly account for the many statistical comparisons in the rodent data. With more comparisons, it becomes more likely that there will be statistically significant results by chance alone. “That’s part of the reason people can interpret things quite differently,” Eastmond said.
In coming to its conclusion on glyphosate’s carcinogenicity, IARC cited an increased rate of a rare form of kidney cancer in a type of male lab mouse exposed to glyphosate and increased cancer of the blood vessels in exposed male mice, as well as increases in some benign kidney tumors.
Other groups interpreted the rodent data differently. “Based on the weight-of-evidence evaluations, the agency has concluded that none of the tumors evaluated in individual rat and mouse carcinogenicity studies are treatment-related,” for various reasons, the EPA concluded in its review. The agency did not find a significant increase in kidney tumors in mice, after a reanalysis found an additional tumor in the control mice that previously had not been seen. The EPA’s review also noted that some mice in the study received atypically high doses of glyphosate.
The European Chemicals Agency, or ECHA, similarly concluded in 2022 that the mouse data “did not demonstrate convincing evidence of glyphosate induced” tumors. The group did find some increased rare kidney tumors in male mice exposed to very high levels of glyphosate but called the relevance to humans “low” due to the high dose.
The FAO/WHO group that Eastmond and Boobis were a part of, meanwhile, “concluded that glyphosate is not carcinogenic in rats but could not exclude the possibility that it is carcinogenic in mice at very high doses,” according to the 2016 report released on its conclusions. However, the group — which was only tasked with assessing the effects of pesticides in food — concluded that “those effects were seen at such high doses that we did not think it was relevant for the decisions we were making about pesticide residues in the diet,” Eastmond said.
Some people with concerns about glyphosate cite a June 2025 study in rats as evidence that the herbicide can be carcinogenic at lower doses. (The aide for Booker, the senator from New Jersey, cited this study, among other sources suggesting glyphosate is carcinogenic.) The study found elevated rates of various cancers in rats exposed to glyphosate or glyphosate-based herbicides beginning in utero and through their lives. This included an increase in early-life leukemia, which is rare in the type of rats studied. The researchers used doses of glyphosate pegged to European regulatory limits for daily exposure.
“What that says to me is that the levels that people are being exposed to today in food … those levels have risk,” Landrigan said, adding that the study establishes that glyphosate causes cancer. “The risk to any one person may be relatively low, but when millions of people are exposed … there are always going to be some people who eat more contaminated food than others, and there are always going to be some people in the population who are biologically more sensitive than others … so across a population if you expose a whole population to a chemical that has the power to cause cancer, then you’re going to push up the risk across the population.”
However, some scientists have criticized the study as using unusual statistical and other methods, while noting that its conclusions contrast with those of other rat studies.
In a July 2025 review, for example, scientists from the German Federal Institute for Risk Assessment — the group that led the most recent European Union safety review of glyphosate – wrote that “due to its design, the study is only very limited in its comparability with the many long-term studies on glyphosate that are already available” and “does not refute their findings.” The German review said that prior studies using far higher exposures had not gotten similar results.
The “unusual” study design of the new rat study “doesn’t in itself invalidate the study, but it means that it needs to be open to scrutiny,” Boobis said. “They have been very reluctant to let outsiders access to the raw data, the pathology slides, etc., to do independent evaluation.” He also called the way the study counted the tumors and compared the groups of rats “extremely unconventional.”
Sifting Through the Mechanistic Data
The third line of evidence scientists use to evaluate whether a chemical is carcinogenic is whether there is a mechanistic explanation for how it causes cancer. Again, groups have come to divergent conclusions about whether glyphosate leads to cancer-related changes.
IARC found “strong” mechanistic evidence that glyphosate causes cancer, citing evidence that it damages DNA, called genotoxicity. The group also found evidence of oxidative stress, a more indirect measure of possible carcinogenicity. Cells are considered to be under oxidative stress when they fall behind on dealing with reactive oxygen-containing molecules. In the long-term, this can lead to cancer.
In contrast, the EPA review concluded that the available data showed that glyphosate does not cause DNA mutations when consumed by mouth. The FAO/WHO group also did not find genotoxic effects from glyphosate in mammals exposed orally, and the European ECHA evaluation also concluded glyphosate did not cause mutations.
Eastmond, who helped lead the FAO/WHO group’s efforts to weigh the mechanistic evidence, said that people may come to different conclusions about genotoxicity in part because there are so many studies on the topic, with widely varying quality, and because IARC only considered published studies while others had data from the manufacturer. “We focused on what we thought were the most relevant for human risk by the oral route of exposure,” he said. “When we did that, we thought the evidence was clearly pretty overwhelmingly negative for genotoxicity.”
More recently, National Cancer Institute researchers have also taken urine samples from agricultural workers in the AHS and found some signs of increased oxidative stress in urine that had more glyphosate in it.
However, Boobis and Eastmond noted that many substances cause oxidative stress, and that this does not always lead to cancer.
A different recent NCI study found that among agricultural workers in the AHS study, higher self-reported exposure to glyphosate over time was associated with certain chromosomal changes, although the authors said their results would need to be replicated.
Another question is whether there is a difference between exposure to glyphosate on its own versus glyphosate-based herbicides, which contain other ingredients which are in some cases proprietary. Some recent mechanistic studies have suggested that glyphosate is unlikely to cause cancer-related changes in cells but raise the possibility that glyphosate-based herbicides, which also include other ingredients, may lead to these changes.
Eastmond acknowledged that despite the large amount of data on glyphosate, there are still potential gaps. He noted that the original court case was brought by a person who was exposed “extensively” via the skin, where most studies are of oral exposure. “You could argue maybe there’s a difference,” he said. He added that he tells people to take precautions while applying pesticides but doesn’t in most cases “worry too much about everything I eat and drink.”
Editor’s note: FactCheck.org does not accept advertising. We rely on grants and individual donations from people like you. Please consider a donation. Credit card donations may be made through our “Donate” page. If you prefer to give by check, send to: FactCheck.org, Annenberg Public Policy Center, P.O. Box 58100, Philadelphia, PA 19102.