When New Zealand regulators assess the long-term health risks associated with glyphosate, much of the toxicological evidence used to establish acceptable exposure levels relates to glyphosate itself—the declared active ingredient.
But glyphosate is not normally what farmers, councils and home gardeners spray on its own. They use formulated herbicide products containing surfactants and other ingredients intended to change how the herbicide behaves.
The EPA has identified 92 glyphosate-containing substances approved under New Zealand’s Hazardous Substances and New Organisms Act (HSNO). Those formulations are not necessarily toxicologically identical to glyphosate tested in isolation.
That raises an obvious question: how much does an assessment of glyphosate alone tell us about the products people actually use?
A 2018 study out of the University of Caen Normandy suggests that approach may be missing most of the picture.
What the Defarge Study Examined
Researchers Nicolas Defarge, Joël Spiroux de Vendômois and Gilles-Eric Séralini published their findings in Toxicology Reports (Vol. 5, pp. 156–163), comparing glyphosate on its own against 14 full glyphosate-based herbicide formulations, tested on both plants and cultured human embryonic kidney (HEK293) cells. They also ran mass spectrometry on 22 pesticide products — 11 of them glyphosate-based — looking for undeclared contaminants.
Séralini’s name will be familiar to anyone who follows glyphosate research; his 2012 rat feeding study remains one of the most contested papers in the field, retracted from one journal and later republished in another. That history doesn’t invalidate this study, but it’s worth naming upfront — readers evaluating the evidence deserve to know whose lab it came from.
What the Study Found About Glyphosate Formulations
Three findings stand out:
Glyphosate alone showed the lowest toxicity in the experiments. In the plant experiments, the formulated herbicides produced stronger herbicidal effects than glyphosate alone at comparable glyphosate concentrations. In separate experiments using cultured human kidney cells, the formulations also showed substantially greater cytotoxicity than glyphosate alone. Under the conditions used in these experiments, the commercial formulations were considerably more toxic than glyphosate alone. The authors concluded that co-formulants were important contributors to the observed toxicity.
One formulant family, POEA, stood out. POEA (polyethoxylated tallow amine, a family of surfactants used to help herbicide spread across and penetrate plant leaves) was identified by the researchers as a major contributor to the cytotoxicity they observed. In the kidney-cell assay, the particular POEA tested was approximately 3,450 times more cytotoxic than glyphosate when the researchers compared the concentrations required to produce a similar effect. That figure describes this laboratory experiment; it does not mean that every POEA-containing product is 3,450 times more dangerous to a person using it. In cultured human embryonic kidney (HEK293) cells, formulations containing POEA caused rapid cell death — with cells detaching, shrinking and breaking apart within about 90 minutes at concentrations where glyphosate alone produced little or no comparable effect.
Undeclared heavy metals turned up in the formulations. Mass spectrometry detected arsenic, chromium, cobalt, lead and nickel in many of the pesticide products analysed, including glyphosate-based herbicides. These metals were not listed as declared active ingredients. Of the 22 pesticide formulations examined, all but one contained detectable levels of one or more of these metals. The study did not establish where the metals originated, but their presence raises questions about impurities and the overall chemical composition of commercial pesticide formulations.
In a separate study published in 2016, some of the same researchers reported effects on aromatase activity—a hormone-related biological pathway—at concentrations below those producing overt cell death. Those effects were observed in cultured human placental cells and were attributed primarily to co-formulants rather than glyphosate alone.
Why This Matters for How Glyphosate Gets Regulated
This is the part that should concern anyone relying on a regulator’s safety assessment. Health-based reference values for glyphosate, including acceptable daily intake values, are generally derived from toxicological studies of glyphosate or technical-grade active material. They do not represent separate long-term assessments of every commercial formulation in which glyphosate is sold.
If co-formulants account for much of the observed toxicity, and if impurities such as heavy metals are neither routinely declared nor specifically considered as part of long-term toxicological assessments, then a safety case built solely around the active ingredient is answering the wrong question. It’s the regulatory equivalent of approving a recipe by testing one ingredient and ignoring the rest of the dish.
Questions for New Zealand’s EPA
The EPA cannot say it was unaware of the concern. Its own May 2022 report acknowledged submissions arguing that POEA surfactants may be more toxic than glyphosate itself. It then made a specific commitment:
“We will therefore review the available information on POEAs to determine if regulatory action is required.”
Four years later, the obvious question is not whether the issue has ever been raised. It is what became of that review.
Did it commence? What evidence did the EPA examine? Did it consider the Defarge study and the wider research comparing formulated herbicides with glyphosate alone? Was any conclusion reached about products containing POEA in New Zealand?
And if the review has not been completed, why not?
The Defarge study does not prove that every glyphosate formulation poses an unacceptable risk. It does, however, expose the weakness in speaking about “glyphosate safety” as though glyphosate and the products sold under dozens of different formulations were necessarily the same toxicological question.
They are not.
Study reference:
Defarge, N., Spiroux de Vendômois, J., & Séralini, G.E. (2018).
Toxicity of formulants and heavy metals in glyphosate-based herbicides and other pesticides.
Toxicology Reports, 5, 156–163.
Further reading
The Defarge study is not an isolated piece of research. It sits within a broader scientific and regulatory discussion about whether testing glyphosate alone provides a complete picture of the products people actually use. If you’d like to explore that discussion further, these sources are worth reading.
Ignoring Adjuvant Toxicity Falsifies the Safety Profile of Commercial Pesticides
Mesnage & Antoniou, Frontiers in Public Health (2018).
A broader review making the same regulatory point as the Defarge study: pesticide risk assessments worldwide test the “active principle” alone while the co-formulants — never subject to an acceptable daily intake — go unassessed, despite documented toxicity in their own right.
Ethoxylated Adjuvants of Glyphosate-Based Herbicides Are Active Principles of Human Cell Toxicity
Mesnage, Bernay & Séralini, Toxicology (2013).
An earlier, more targeted lab study isolating POEA (POE-15) as a toxicity driver in nine commercial glyphosate formulations, tested across liver, kidney and placental human cell lines — effectively a precursor to the 2018 study’s findings.
Major Pesticides Are More Toxic to Human Cells Than Their Declared Active Principles
Mesnage, Defarge, Spiroux de Vendômois & Séralini, BioMed Research International (2014).
Extends the same comparison (formulation vs. declared active ingredient) across nine pesticides beyond glyphosate, including insecticides and fungicides — useful for showing this isn’t a glyphosate-only problem.
Carcinogenicity of Glyphosate: Why Is New Zealand’s EPA Lost in the Weeds?
New Zealand Medical Journal (2018).
A direct critique of the NZEPA’s own glyphosate carcinogenicity assessment, written for a domestic audience — good companion piece for grounding this in local regulatory failure rather than only overseas evidence.
Glyphosate in Aotearoa New Zealand – Call for Information Summary Report
New Zealand Environmental Protection Authority (May 2022).
Summarises the EPA’s 2021 Call for Information and includes the agency’s commitment to review POEA surfactants after submissions raised concerns about their toxicity.
APP204718 — Grounds for Reassessment: EPA Assessment Report
NZ Environmental Protection Authority.
The primary source confirming the 92 glyphosate formulations approved under the HSNO Act, and the EPA’s own reasoning on whether grounds for reassessment exist — worth citing directly since it’s the regulator’s own words.
PSGR Submission to the NZEPA Glyphosate Call for Information
Physicians and Scientists for Global Responsibility (2021).
A formal NZ submission that cites the Defarge study directly and raises the tank-mixing and formulant issue in a New Zealand regulatory context — shows this argument has already been put to the EPA once.
Together, these studies and reports illustrate why the distinction between glyphosate and glyphosate-based herbicide formulations continues to be debated. They also help explain why questions about New Zealand’s regulatory approach remain both relevant and worth asking.
Image Source & Attribution
The feature image on this page was created using AI-assisted image generation from an original concept developed by No More Glyphosate NZ and refined for publication in Canva.
