Does the World’s Top Weed Killer Cause Cancer? Trump’s EPA Will Decide

Roundup has revolutionized farming. Now, human health and Bayer’s $66 billion deal for Monsanto depend on an honest appraisal of its safety.

Spraying a mix of Roundup and another product on soy fields in Wisconsin.

Photographer: Jesse Chehak for Bloomberg Businessweek

Every year, farmers spray, on average, almost a pound of the herbicide glyphosate on every acre of cropland in the U.S., and nearly half a pound on every acre of cropland worldwide. Glyphosate is the active ingredient in Roundup, a huge source of income for its manufacturer, Monsanto Co., and the foundation for its epochal foray into genetically modified organisms. If you know nothing else about GMOs and Monsanto, know this: The St. Louis-based company reengineered the DNA of corn, soybeans, and other crops for the primary purpose of making them resistant to Roundup.

Farmers spray the chemical on crops grown from Monsanto’s Roundup Ready seeds. The weeds die, harvests expand, and expensive, laborious tillage is no longer necessary. Large-scale agriculture is built on this model, and not only in the U.S., which is why Bayer AG, the German drug and chemical company, agreed in September to buy Monsanto for $66 billion, pending regulatory approvals. Other than government antitrust objections, about the only thing that could mess up the purchase would be for the U.S. Environmental Protection Agency to reverse its position on the active ingredient of Roundup, glyphosate.

Last December, the EPA convened a panel of outside scientists to peer-review the agency’s long-standing conclusion that glyphosate is unlikely to cause cancer. The peer reviewers, a mix of academics, federal scientists, and chemical industry consultants, gathered at an EPA conference center in Arlington, Va. From the agency’s point of view, this was something of a formality. Federal law requires an EPA health-effects review for every pesticide at least once every 15 years, and glyphosate has enjoyed a clean bill of health since 1991, when the agency cleared the way for Monsanto’s GMO breakout by classifying the herbicide as noncarcinogenic to humans.

Its use in global agriculture has soared almost fifteenfold since Monsanto introduced Roundup Ready seeds in 1996. As a result, traces of glyphosate have been detected in cookies, crackers, chips, breakfast cereals, and honey, and in human urine and breast milk. Monsanto says it’s nothing to worry about. “Glyphosate is about half as toxic as table salt and more than 10 times less toxic than caffeine,” the company says on one of its websites. More than 1,000 farmers and other agricultural workers stricken with non-Hodgkin lymphoma disagree. They’re suing Monsanto in state and federal courts across the country, claiming Roundup caused their cancer. Monsanto is vigorously contesting the claims. “There’s never been a more studied herbicide in the history of farming,” says Scott Partridge, Monsanto’s vice president for strategy. “In more than 700 studies, not one has associated cancer with the use of glyphosate. And in the more than 160 countries that have registered glyphosate for use, not a single government agency has found glyphosate is a carcinogen.”

December’s Scientific Advisory Panel meeting followed the typical script for a federal peer review, with some twists. Officials from the EPA’s Office of Pesticide Programs opened the public hearing by laying out 45 years of study data and describing why, in the agency’s view, they indicate that glyphosate is an unlikely human carcinogen at current exposure levels. The agency’s conclusion, elucidated in a 227-page issue paper released the previous September, reflects the “weight of evidence,” Jack Housenger, then chief of the pesticide office (he retired this year), told the panel. “Now it’s your turn to tell us what you think of our analysis,” he said, “and hopefully put the subject to bed so we can move on.”

Roundup waiting for dilution and application in Wisconsin.
Photographer: Jesse Chehak for Bloomberg Businessweek

Far from settling the matter, eight of the 15 experts expressed significant concerns about the EPA’s benign view of glyphosate, and three more expressed concerns about the data. Their skepticism also raised, again, questions about the independence of the Office of Pesticide Programs, which has the final say on permitting pesticides. The office relies on pesticide manufacturers for the data it uses in making health decisions—and got almost 30 percent of its operating budget from the industry last year.

The EPA paper had a whack-a-mole quality to it. Throughout, the authors included data sets suggesting that glyphosate could cause cancer, only to knock them down. On epidemiology studies, for example, they said farmers’ recollection of their own glyphosate exposure was biased and unreliable. On meta-analyses pooling human data from multiple studies to identify trends, the EPA assessors shaved decimal points from the results, which made it possible for them to shrug off data showing exposed farmers had an elevated risk of cancer.

Many of the reasons cited in the paper contradicted the agency’s own carcinogenicity guidelines, multiple panelists pointed out. “Every time there’s something positive there, you said there’s something wrong with the study,” Eric Johnson, an epidemiologist at the University of Arkansas for Medical Sciences, scolded EPA officials at the meeting. Lianne Sheppard, a biostatistician at the University of Washington at Seattle, says there was “consensus” on the panel that the agency hadn’t followed its own standards. “The available evidence did not fit with the conclusions drawn in the issue paper, particularly when put in the context of the guidelines,” she says.

The EPA’s report on the peer review, posted on March 16, raises obfuscation to a high bureaucratic art. While spelling out the panel’s criticisms, the report gives no indication which, or how many, reviewers felt strongly about which particular problems. Instead, it uses the phrase “some panel members” 76 times—as in “some panel members noted,” “some panel members emphasized,” “some panel members suggested.” The imprecision obscures that the majority of peer reviewers expressed doubts about the EPA’s methods or conclusions. Under the law, the agency must consider the panel’s input in its final evaluation of glyphosate, scheduled for completion later this year. By enshrining the reviewers’ comments in such vague terms, however, the EPA can more easily ignore them.

“I asked for a vote on the main issues to make our guidance clear, but this committee apparently never does that,” says Emanuela Taioli, an epidemiologist at Icahn School of Medicine at Mount Sinai in New York and one of the panelists who disagreed with the EPA’s interpretation of the evidence. “Not giving a tally leaves them more flexibility to interpret our advice.”

Monsanto, a company that has genetically altered vast swaths of cropland to vanquish farm pests, has lately been struggling to control an invasive species on its own turf: scientific doubt. It’s an exotic. Ever since President Bill Clinton awarded four Monsanto gene scientists the National Medal of Technology and Innovation in 1998, the company’s ag-biotech enterprise has been the closest thing America has to a Japanese-style strategic industry. From Clinton to Bush to Obama, successive administrations mobilized federal agencies and embassies around the world to promote GMOs, often against the vehement opposition of environmentalists and food purists, particularly in Europe. Dozens of diplomatic cables released by WikiLeaks show how U.S. missions received special funding from Congress to advocate for genetically engineered products and how Monsanto prevailed on American diplomats to lobby on its behalf when problems arose. In Argentina, for example, when the ministry of defense banned the use of glyphosate on its urban farmland in 2009, the U.S. embassy intervened, according to a diplomatic cable to Washington. “Post contacts within the Secretariat of Agriculture assure us that Argentina will continue to support biotechnology,” said the cable, signed by Thomas Kelly III, deputy chief of mission at the time.

The first breach in Monsanto’s fortress opened in 2015, when the International Agency for Research on Cancer labeled glyphosate a probable carcinogen. IARC, a France-based arm of the World Health Organization, has no regulatory power, but its carcinogenicity studies are widely cited in court cases and government health assessments worldwide. The agency’s assessment, based on published, peer-reviewed research, foreshadowed many of the concerns the EPA’s scientific advisory panel expressed last December. IARC acknowledged the studies were all flawed in different ways, but it concluded that their findings pointed toward cancer and couldn’t be dismissed. This led California on July 7 to list glyphosate as a known carcinogen, over Monsanto’s objection.

The question now falls to the Trump EPA and the courts. Led by Administrator Scott Pruitt, the former Oklahoma attorney general who sued the EPA more than a dozen times to stop environmental regulations, the agency has already canceled an Obama-era proposed ban on chlorpyrifos, a pesticide linked to cognitive damage in farmworkers and children. The chances that Pruitt will move against glyphosate, with all the attendant repercussions for industrial agriculture, appear slim.

The considerations are much different, however, for U.S. District Court Judge Vince Chhabria in San Francisco. The judge is presiding over multidistrict litigation composed of 310 plaintiff lawsuits against Monsanto filed by cancer victims around the country. (It will likely consolidate hundreds more suits.) Chhabria has told both sides that the question of whether Roundup can cause cancer will turn on the scientific evidence presented at trial, not on what agencies such as IARC and EPA say. In this instance, the difference between Roundup and glyphosate is crucial. The EPA focuses on the latter. The plaintiffs in the court case claim that Roundup contains ingredients that exacerbate the effects of glyphosate alone.

Chhabria has allowed the plaintiffs wide latitude to collect evidence on Monsanto’s health-effects research over the years, which the plaintiffs hope will show the company manipulated the data. In March he unsealed dozens of Monsanto’s confidential documents for the public to see. The records show internal deliberations on how to present the science on glyphosate’s health impacts and manage a global public-relations campaign to assure consumers and regulators that Roundup is safe.

Containers of Roundup that are kept outside the cockpit of the sprayer.
Photographer: Jesse Chehak for Bloomberg Businessweek

Monsanto documents show the company commissioned scientists to publish papers rebutting IARC. Reminiscent of tobacco companies, it also funneled money to front groups, according to a plaintiffs’ court filing. The groups, with names such as Genetic Literacy Project and American Council on Science and Health, published articles praising the EPA and attacking IARC, which they called on Congress to defund. The plaintiffs claim that Monsanto established a program called Let Nothing Go, through which it made payments to people with no apparent industry ties to post online comments defending Monsanto, its chemicals, and GMOs in news articles and Facebook posts. “That’s simply false,” says Monsanto’s Partridge. “We don’t need to do that type of stuff.”

In Washington, where Monsanto has spent almost $60 million on lobbying since 2008, the company cultivates allies on both sides of the aisle and in the relevant federal agencies. It deployed five lobbyists in 2015 to trash IARC’s findings at the White House, Congress, and the agencies. Monsanto employees are regular visitors to the EPA’s Office of Pesticide Programs, according to logs obtained through the Freedom of Information Act.

Relations were warm, even jocular. “So Jess called me out of the blue this morning,” wrote Monsanto’s lead EPA liaison, Dan Jenkins, to William Heydens, the company’s chief of regulatory research, in an April 2015 email released in the court case. Jess was Jess Rowland, a senior official in the EPA’s pesticide office who was chairing the agency’s cancer assessment of glyphosate at the time. Heydens had emailed Jenkins the day before, asking his colleague to reach out to the EPA and find out “what area they see as most problematic (e.g., human epidemiology vs. animal bioassays vs. genotoxicity), or just ask if there is anything that would help them defend the situation?”

Rowland was all set, Jenkins reported back to Heydens. “We have enough to sustain our conclusions,” Rowland told Jenkins on the phone, according to Jenkins’s email. “I am the chair of the [cancer review],” he added, “and my folks are running this process for glyphosate.”

On the same call, Jenkins wrote, Rowland said he was working to control a separate glyphosate assessment by another federal unit, the Agency for Toxic Substances and Disease Registry (ATSDR), a division of the Centers for Disease Control and Prevention. “If I can kill this, I should get a medal,” Jenkins quoted Rowland as telling him.

“Wow!” Heydens wrote back to Jenkins. “That’s very encouraging. Thanks for the news update.”

Does the World’s Top Weed Killer Cause Cancer?

In April 2016, as Rowland was preparing to retire after 26 years at the EPA, his assessment that glyphosate is unlikely to cause cancer leaked online—just in time for Monsanto’s lawyers to cite it at an important court hearing in San Francisco. The EPA quickly characterized the report as “preliminary” and “not final,” but Monsanto’s lawyers still told Judge Chhabria, “the scientists have spoken.” As for the ATSDR review, another internal Monsanto document reported that the CDC division had “agreed, for now, to take direction from EPA.” This May, EPA Inspector General Arthur Elkins Jr. confirmed he’s investigating if EPA employees colluded with Monsanto. Through his lawyer, Rowland declined to comment for this story. Monsanto’s relationships with Rowland and other EPA employees were professional and proper, Partridge says.

Bayer and Wall Street are betting none of this matters. Monsanto has built the kind of virtuous circle that management experts and business school professors rave about. More sales of Roundup Ready seeds beget more use of Roundup; more herbicide use drives up demand for Monsanto’s GMO seeds. The global chemical dependency could be too big to kick.

It’s hard to quantify what a shift to glyphosate-free farming would look like. A study by chemical industry consultants in the U.K., where about a third of the nation’s wheat fields are sprayed with the herbicide, estimated that yields of the grain would fall 12 percent if glyphosate were banned. A study last year by Andrew Kniss, an associate professor of plant sciences at the University of Wyoming, showed the yields from organic farms were roughly two-thirds of those from conventional farms for corn, wheat, soybeans, and barley, and less than half for grapes, tomatoes, bell peppers, and onions.

Kniss did another study that found planting genetically modified sugar beets saved farmers $200 an acre, equal to about 15 percent of their revenue, compared with planting non-GMO seeds. With no glyphosate, farmers would have to resort to using more-toxic chemicals for weed control, Kniss says, or revert to grueling tillage by hand. “Getting rid of glyphosate would have a major impact on farmers and their bottom lines,” he says. “It’s not like there’s a risk-free scenario here.”

Robert Fraley grew up on a farm in Hoopeston, Ill., in the 1950s and 1960s, and one of his earliest memories is of the fields turning black each November. He recalls tractors churning up a foot of dark topsoil to keep the weeds from taking over. His dad cranked up the tractor every morning by 4 a.m. and plowed until Fraley got home from school and took over until midnight. In summers the boy walked the rows of beanstalks with his friends and cousins, pulling the weeds by hand. Fraley, now 64, helped develop Roundup Ready seeds and is Monsanto’s chief technology officer. “Any kid my age who spent time on the farm, the first thing you realize is the greatest challenge is weeds and bugs,” he says. “We freed America’s farm youth.”

By 1980, when Fraley joined Monsanto as a 27-year-old with a doctorate in microbiology and biochemistry, scientists were experimenting with recombinant DNA in yeast and animal cells, but no one had introduced a new gene in a plant. Fraley’s team, working with a germ called agrobacterium, which normally causes blight in plants, isolated the part of the germ that binds to plant cells and can inject its DNA into plants, and eliminated the blight-producing sequence. They worked with petunia plants to fine-tune a gene insertion process, using the flower’s color genetics to map the plant’s chromosomes. It took more than a decade to develop commercially useful traits such as resistance to insects and glyphosate. The breakthrough finally made glyphosate, which was invented in 1970, a significant commercial power.

Even as environmentalists vilify Monsanto for its link to GMOs, its mantra internally remains stewardship and sustainability. “We get to enjoy more of our forest and wetlands and prairies because we’ve increased yields on the land we’re already farming,” Fraley says. The alternative to genetic engineering and the accompanying chemicals, he says, is plowing up an additional 30 million or 40 million acres of land to feed a hungry planet.

That probably overstates the trade-off. Land spared from cultivation is seldom set aside for conservation. And there are alternatives to Roundup Ready farming other than going organic. “It’s such a Monsanto-based perspective to say food prices will spike if we use less glyphosate,” says Claire Kremen, a conservation biologist at the University of California at Berkeley. “There are other methods, besides organics, that are just as productive as conventional farming and don’t rely on toxic chemicals that endanger lives and harm the environment.” A more realistic alternative, she says, is finding middle ground between the industrial and organic farming models. Researchers at Iowa State University, for example, have shown that rotating diverse crops in three- and four-year cycles and controlling weeds with limited herbicide spraying produce similar yields and profits to conventional farming—with only 1 percent of the water toxicity. And such alternatives are becoming necessary anyway, because weeds are developing resistance to glyphosate at an accelerating rate.

Nonetheless, a lot of farmers are deeply committed to glyphosate. “The cancer issue doesn’t concern me at all,” says Paul Jeschke, 64, who farms 4,000 acres of corn and soybeans with his brother-in-law and nephew in Mazon, Ill. Before he started using Roundup in the 1980s, Jeschke says, his topsoil, after constant tillage, would wash away in the rain. The quack grass would get so bad some farmers would have to put up a fence around their fields and turn the pigs loose, losing a year’s harvest. “Every one of us takes risks every day when we take our car on the road or get on an airplane or dump table salt into something we’re cooking,” Jeschke says. “Looking at the cost-benefit ratio, I’m extremely reluctant to give up glyphosate.”

The first cancer concerns came from within the EPA’s Office of Pesticide Programs in 1984. This was despite the traditionally close relationships between the agency and the companies it regulates. Pesticide makers, called registrants, pay the office to review their compounds for registration. In 2016 they provided $47 million, or 28 percent, of OPP’s budget. The OPP is also the only EPA branch that does its own health assessments; the agency’s National Center for Environmental Assessment is in charge of evaluations for the other EPA branches. The OPP’s studies are based, by law, on data provided by pesticide registrants themselves. This has fueled intra-agency criticism that some of its cancer reviews are too lax.

Soy fields treated with Roundup.
Photographer: Jesse Chehak for Bloomberg Businessweek

Still, its scientists did come down hard on glyphosate early on. In February 1984, EPA toxicologists sounded the alarm internally when a Monsanto-sponsored study showed that 4 of 100 mice that were fed large amounts of glyphosate developed rare kidney tumors called tubular adenomas, compared with zero of 98 mice given little or no glyphosate. Monsanto insisted glyphosate wasn’t the cause, invoking an argument that would become a pillar of its—and the EPA’s—defense of the chemical for the next 30 years. The mice with the tumors weren’t the problem, the company said. The healthy mice were. The control group ought to have had more tumors.

An OPP statistician was having none of it. “Our viewpoint is one of protecting the public health when we see suspicious data. It is not our job to protect registrants from false positives,” wrote Herbert Lacayo, who analyzed Monsanto’s complaints in a memo in February 1985. A week later, OPP’s toxicology branch cited the mouse tumors in a “consensus report” naming glyphosate a possible human carcinogen.

Monsanto was apoplectic. The Roundup Ready world that was coming to life inside Fraley’s gene-splicing lab depended on the copious use of glyphosate. To change the EPA’s mind, the company solicited 10 outside experts to examine the mice in the control group. One of the scientists, who’d been a Monsanto consultant for many years, reported to the company that he’d found cellular changes in the kidney of one mouse. Monsanto argued it was another adenoma. If so, it would render the four tumors in the exposed mice statistically insignificant. The other experts then supported the finding of the single scientist.

The EPA’s toxicology branch recut new sections of all the mouse kidneys for another look. Its scientists reconfirmed there was no control adenoma. Nevertheless, in February 1986, the EPA’s scientific advisory panel overruled the agency scientists, asserting that the “vast majority” of pathologists who’d looked at the control kidney in question saw a tumor. If the reviewers knew those experts were brought on by Monsanto, they didn’t care. The panel reclassified glyphosate as a chemical of uncertain carcinogenicity. Five years later, another EPA panel looked at the same data and a new rat study and reclassified glyphosate again, this time as noncarcinogenic to humans.

By 1999, with Roundup Ready soybean, cotton, and corn seeds already changing global farm markets, Monsanto was facing questions about how glyphosate affected animal genes. Emails unsealed in March by Judge Chhabria show that the company hired James Parry, a prominent genetic toxicologist at Swansea University in Wales, to publicly advocate that the chemical wasn’t genotoxic—that is, it didn’t have a destructive effect on DNA and RNA. But after reviewing studies that Monsanto provided, Parry reached the opposite conclusion.

He wrote a report for Monsanto that said glyphosate appeared to damage genes through a biochemical process called oxidative stress—the same cancer-causing mechanism IARC identified 16 years later. He recommended Monsanto do a series of studies to find out. If glyphosate was confirmed to be genotoxic, Parry said, the company should analyze the lymphatic cells of exposed humans to check for chromosome damage.

Monsanto scientists deliberated for weeks about their consultant’s unwelcome advice. The company was in a “genotox hole,” wrote senior toxicologist Donna Farmer in a September 1999 email. “I am concerned about leaving Parry out there with this as the final project/his final impressions.”

“Maybe you should invite Parry to St. Louis to get him more familiarized with the complete database,” suggested another Monsanto toxicologist.

In an email Monsanto must surely regret, Heydens, the regulatory research chief, wrote that changing Parry’s mind would be expensive and probably not worth it. “Let’s step back and look at what we are really trying to achieve here,” Heydens wrote to Farmer and two others. “We want to find/develop someone who is comfortable with the genetox profile of glyphosate/Roundup and who can be influential with regulators and Scientific Outreach operations when genetox issues arise. My read is that Parry is not currently such a person, and it would take quite some time and $$$/studies to get him there. We simply aren’t going to do the studies Parry suggests.”

Parry’s report was never submitted to the EPA. (He died in 2010.) The episode points to an ongoing concern at Monsanto, which was concisely stated by Heydens in a later email: “Data generated by academics has always been a major concern for us in the defense of our products.” Partridge says Heydens sometimes made “an unfortunate choice of words,” but that subsequent studies by Monsanto resolved Parry’s concerns.

The transcript of the EPA’s scientific advisory panel meeting runs 1,300 pages. Reading the document is the only way to know that four of the six reviewers charged with evaluating the crucial epidemiological data lambasted the EPA. (The four critics are all biomedical researchers at major universities; the two who supported the EPA’s evaluation are private consultants.) The agency disregarded all but one meta-analysis of the epidemiological data, because agency evaluators said the results weren’t statistically valid. When several of the panelists reran the pooled data, they found the EPA was plain wrong. Not only were the meta-analyses statistically significant, but they also showed farmers exposed to glyphosate had an elevated risk ratio for non-Hodgkin lymphoma of 1.27 to 1.5, meaning they were at 27 percent to 50 percent higher risk than control groups.

“For a human epidemiologic study, an association of 1.2 or 1.3 is very meaningful and impactful,” says Mount Sinai’s Taioli. At the meeting, she pointed out that millions of American women no longer take estrogen after menopause, because studies found that it increased the risk of breast cancer by about 22 percent. Sheppard, the biostatistician with the University of Washington, said at the meeting that the EPA’s evidence assessment was “highly imbalanced” and that the agency downplayed statistical findings in favor of other criteria. “The agency’s conclusion is seriously flawed and needs to be strongly revised,” she said.

Several panelists asserted that while glyphosate probably doesn’t initiate cancer by causing gene mutations, it appears to promote malignancies by spurring tumor growth. Such a carcinogen is more dangerous to humans than to rodents, because people live much longer and thus accumulate more lesions susceptible to glyphosate’s catalytic effect, panelist Barbara Parsons, a molecular toxicologist with the U.S. Food and Drug Administration, said at the meeting. She warned that mixing a tumor promoter such as glyphosate in formulations with other chemicals that have “any genotoxic potential would be a significant public-health concern.”

That argument is the crux of the plaintiffs’ case in the consolidated federal suits. Their lawyers say they have evidence that Monsanto knew for years that some of the nonactive ingredients in Roundup are carcinogenic, and thus the danger of those chemicals is compounded when they’re combined with glyphosate. They say that the OPP, by focusing its concerns on the active ingredient instead of on the formulated product, has let Monsanto off the hook. Partridge denies any Roundup ingredients are genotoxic and says potentially carcinogenic impurities in the product are strictly controlled.

The doubt invading Monsanto’s prize product is as strong as it’s ever been, even as Roundup has become instrumental in industrial agriculture. Farmers and consumers have reaped huge savings from productivity gains made possible by taming the scourge of weeds. Improbable as it seems, suppose the EPA now moves glyphosate from the category of unlikely carcinogen to suggestive or even likely. That would trigger extensive cost-benefit analyses. Then the questions get really difficult.