Does glyphosate in food cause cancer?

The evidence says no at the exposures food actually contains. Residues in monitored food supplies run orders of magnitude below the FAO/WHO acceptable daily intake, and every regulatory risk assessment — EFSA, ECHA, US EPA, Health Canada, JMPR — has concluded dietary exposure is unlikely to pose a carcinogenic risk. The IARC 'probably carcinogenic' classification that headlines cite is a hazard judgment about the molecule's capability, largely at occupational and experimental doses, not a statement about trace residues. If glyphosate poses any human cancer risk, the candidates are the most heavily exposed professional users, not consumers.

Then why did juries award billions to cancer patients?

Because courts answer a different question with different rules. Civil juries decide liability on the balance of probabilities, hear individual sympathetic plaintiffs rather than population data, and were shown internal documents demonstrating genuine corporate misconduct — ghostwriting and lobbying that made the company an easy defendant to disbelieve. Several verdicts were reduced or overturned on appeal, and the science presented in court was contested throughout. Jury verdicts are evidence of how litigation works, and of how the manufacturer behaved, not of what glyphosate does at dietary doses — a distinction that matters in both directions.

Should farmers and gardeners worry?

Professional applicators with years of high-volume exposure are the one group for whom an NHL question genuinely remains open — case-control studies suggest elevated risk at the highest exposures while the large prospective Agricultural Health Study found none overall. The rational response is exposure minimisation regardless: gloves, long sleeves, avoiding spray drift, washing after use — measures justified for any pesticide. Home gardeners applying occasional spot treatments sit far below the exposure range where any signal has been reported; drenching oneself weekly in concentrate is a different activity from the label-directed use the safety data describe.

Is Roundup more toxic than glyphosate alone?

For some organisms and endpoints, yes — the surfactants that help the herbicide penetrate leaves, particularly the older POEA class, are more toxic to aquatic organisms and to cells in culture than glyphosate itself, which is why the EU banned POEA co-formulants and why 'glyphosate-based herbicides' is the accurate term in the toxicology literature. Regulatory assessment historically centred on the active ingredient, a legitimate criticism of the framework. For consumers the distinction changes little, since exposure to either is trace; for aquatic buffer zones and applicator protection it is the more relevant chemistry.

Why is glyphosate detected in urine and beer if it is safe?

Because modern analytical chemistry detects almost anything almost anywhere. Instruments measure parts per billion; finding a molecule at such concentrations demonstrates that agriculture uses it, not that it harms anyone. Urinary levels in general populations correspond to intakes roughly a thousand-fold below the acceptable daily intake, and the amounts reported in beer or cereal would require consuming physically impossible quantities daily to approach thresholds of concern. 'Detected' is an analytical statement; 'dangerous' is a toxicological one, and the distance between them is the entire science of dose.

Would banning glyphosate make food safer?

Probably not, and possibly the opposite in the short term. Weeds would still be controlled — by tillage, which costs soil carbon and erosion, or by older herbicides, several with worse toxicological profiles, a substitution already visible where resistance has degraded glyphosate's usefulness. The defensible case for restricting glyphosate is environmental and agronomic — biodiversity, resistance management, reducing total herbicide dependence through integrated weed management — rather than consumer safety. The EU's 2023 renewal to 2033, decided after member states split, reflects precisely that balance: no consumer-risk basis for a ban, unresolved environmental argument, political deadlock.

Home › Topics › Glyphosate

Glyphosate

GMJ News knowledge hub · last reviewed September 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

Glyphosate is the world’s most used herbicide and the centre of a decade-long scientific and legal war that turns on a single methodological distinction most coverage never explains: in 2015 IARC classified it as “probably carcinogenic to humans” (Group 2A) as a hazard judgment, while EFSA, the US EPA, ECHA and the joint FAO/WHO residue committee — assessing real-world risk at actual exposure levels — concluded dietary residues are unlikely to endanger consumers. The largest prospective study of exposed people, the US Agricultural Health Study of more than 50,000 pesticide applicators, found no overall increase in non-Hodgkin lymphoma. A genuine open question persists for the most heavily exposed occupational users; the claim that trace residues in food cause cancer is not supported by the evidence, however large the jury verdicts built on conflating the two (JMPR evaluations, FAO/WHO).

Key messages

THE CENTRAL DISTINCTION: hazard is not risk

The entire glyphosate war turns on two agencies answering two different questions. IARC assesses hazard — can this agent cause cancer under any circumstances, at any dose? — and in 2015 answered 'probably' (Group 2A), citing limited human evidence for non-Hodgkin lymphoma, sufficient animal evidence and mechanistic data. EFSA, the US EPA, ECHA, Health Canada and the joint FAO/WHO residue committee assess risk — does it cause cancer at real-world exposures? — and concluded dietary residues are unlikely to pose a carcinogenic risk to consumers. Both conclusions can be simultaneously correct; most public argument consists of quoting one against the other as though they contradicted.

SETTLED: dietary residues are far below levels of concern

Residue monitoring across the EU and North America consistently finds glyphosate in food at levels orders of magnitude below the acceptable daily intake set by the FAO/WHO Joint Meeting on Pesticide Residues, and urinary biomonitoring of general populations confirms intakes a thousand-fold or more below regulatory thresholds. The recurring headlines about glyphosate 'found in' cereals, beer or urine reflect analytical sensitivity, not toxicological significance — detection limits in parts per billion will find almost anything almost anywhere.

GENUINELY OPEN: high occupational exposure and non-Hodgkin lymphoma

The legitimate residual question concerns the most heavily exposed users — professional applicators with years of high-volume use. Case-control studies and some meta-analyses of the highest-exposure groups report elevated NHL risk; the largest prospective study, the US Agricultural Health Study following over 50,000 licensed applicators, found no overall association, with a non-significant signal for one leukaemia subtype in the highest-exposure quartile. Recall bias in case-control designs and exposure misclassification in cohorts leave this genuinely unresolved. Protective equipment and exposure minimisation for professional users are justified on ordinary precaution regardless.

THE COURTROOM IS NOT A LABORATORY: the litigation

US juries have awarded billions of dollars to plaintiffs attributing NHL to the herbicide, and the manufacturer has paid settlements in the tens of billions — facts constantly cited as scientific evidence. Tort verdicts answer a legal question (liability under a preponderance standard, heavily influenced by internal-document narratives about corporate conduct) rather than a scientific one, and the corporate behaviour revealed — ghostwritten papers, attacks on IARC — genuinely damaged trust and deserves the criticism it received. One can hold simultaneously that the company behaved badly and that trace residues do not cause cancer; the evidence supports exactly that combination.

UNSUPPORTED: the broader charge sheet

Claims that glyphosate causes autism, coeliac disease, obesity or the general rise of chronic illness — often traced to correlation charts overlaying glyphosate sales on disease trends — are textbook ecological fallacy; the same curves fit organic food sales. Effects on the human gut microbiome at dietary exposure levels are not demonstrated: the shikimate pathway glyphosate inhibits is absent in human cells, and while gut bacteria possess it, dosing studies show no meaningful microbiome disruption at realistic intakes.

THE BETTER ARGUMENT NOBODY LITIGATES: agroecology, not oncology

The strongest criticisms of glyphosate are environmental and agronomic — herbicide-resistant weeds now driving higher and more complex herbicide use, effects on plant diversity and consequently on insect and bird populations, and the monoculture systems glyphosate-tolerant crops enable. These concerns are better supported than the human cancer claims and are the actual grounds on which reasonable pesticide policy is contested, including the EU's renewal debates. A public conversation fixated on residues in breakfast cereal has largely missed the questions where the evidence is real.

Key statistics

Group 2A

IARC hazard classification (2015): probably carcinogenic to humans — the same category as red meat and very hot beverages

IARC

5+ agencies

EFSA, ECHA, US EPA, Health Canada, JMPR — risk assessments concluding dietary exposure is unlikely to be carcinogenic

Regulatory assessments

54,000+

licensed pesticide applicators in the Agricultural Health Study — no overall NHL increase with glyphosate use

AHS/JNCI

1 mg/kg bw

FAO/WHO acceptable daily intake — general-population intakes run a thousand-fold or more below it

JMPR

2033

EU renewal horizon: glyphosate approval extended for 10 years in 2023 after member states failed to reach a qualified majority either way

European Commission

Tens of billions

US dollars in litigation settlements and verdicts — a legal outcome routinely miscited as scientific evidence

Court records

Glyphosate — where the disagreement actually lies

Source: Bars show strength of supporting evidence. High values indicate well-supported propositions; low values indicate claims tested and not supported.

Glossary of key terms

Hazard versus risk assessment
Regulatory scienceA hazard assessment asks whether an agent can cause harm under some exposure condition; a risk assessment asks whether it does so at exposures people actually experience. IARC runs a pure hazard programme — which is why its Group 1 contains both plutonium and processed meat — while food-safety regulators run risk programmes incorporating dose. Glyphosate is the canonical teaching case: Group 2A as hazard, negligible dietary risk in every major regulatory assessment. Journalists and juries repeatedly treat the IARC classification as a finding about breakfast cereal; it is a finding about the biological capability of the molecule, mostly at doses no consumer approaches.

The Agricultural Health Study
EvidenceA prospective cohort of more than 50,000 licensed pesticide applicators in Iowa and North Carolina followed since the 1990s — prospective design eliminating the recall bias that plagues case-control studies of pesticide exposure. Its 2018 glyphosate analysis found no association between glyphosate use and NHL overall or its major subtypes, with a non-significant elevation of acute myeloid leukaemia in the highest exposure quartile flagged for follow-up. It is the largest and methodologically strongest human study available; its main limitations are exposure self-report and the healthy-worker structure of any occupational cohort. IARC's 2015 review predated its key publication, which partly explains the divergence between IARC and later assessments.

Formulations, surfactants and the POEA question
ToxicologyCommercial herbicide is not glyphosate alone but a formulation with surfactants that help the active ingredient penetrate leaves — historically polyethoxylated tallow amine (POEA), demonstrably more toxic to aquatic life and to cells in vitro than glyphosate itself. Regulatory carcinogenicity assessment has focused on the active ingredient, while real-world exposure is to formulations; this gap is one of the more legitimate criticisms of the assessment framework and one reason the EU banned POEA co-formulants in 2016. It is a genuine open question in the science and simultaneously a modest one for consumers, whose exposure to either is minimal.

The shikimate pathway argument
MechanismGlyphosate kills plants by inhibiting EPSP synthase in the shikimate pathway, which synthesises aromatic amino acids. Human cells lack this pathway entirely — the basis of glyphosate's favourable mammalian toxicity profile — but many gut bacteria possess it, generating the hypothesis that dietary glyphosate disrupts the human microbiome. Quantitative work has not supported the hypothesis at realistic doses: most gut bacteria obtain aromatic amino acids from food rather than synthesis, concentrations reaching the colon from dietary residues are far below inhibitory levels, and controlled animal studies at exaggerated doses show minimal community shifts. The mechanism is real botany and, at human dietary exposure, unsupported medicine.

Resistance and the herbicide treadmill
AgronomyMore than fifty weed species have evolved glyphosate resistance since glyphosate-tolerant crops made the herbicide ubiquitous — a predictable evolutionary outcome of applying one selection pressure across hundreds of millions of hectares. The agronomic consequence is the treadmill: rising doses, tank mixes with older and often more toxic herbicides such as dicamba and 2,4-D, and new tolerant-crop stacks. Whatever one concludes about human health, the resistance economics steadily erode glyphosate's original value proposition of replacing more hazardous chemistry — the strongest practical argument for integrated weed management rather than chemical dependence.

The IARC-industry war and ghostwriting
History/IntegrityLitigation discovery revealed that the manufacturer had ghostwritten or heavily shaped ostensibly independent review articles defending glyphosate, and had organised campaigns to discredit IARC after 2015 — conduct that drew formal criticism and permanently damaged the company's credibility as a scientific actor. The episode matters beyond glyphosate: it handed every future critic a documented example of corporate science manipulation, and it obliges honest communicators to separate two things the discovery documents fused in public perception — the company's misconduct, which is proven, and the claim that residues cause cancer, which remains unsupported by the exposure science regardless of who behaved badly.