After wrapping up our mini-series on the 2017 Bruning–Browning review, one thread kept tugging at us — the part of the story that sits underneath all the conclusions:
The evidence itself.
Not the findings, or the politics, or the court cases, but the foundation on which the Environmental Protection Authority (EPA) relied when it said glyphosate was safe.
So before we moved on, we decided to explore that foundation properly — the evidence that shaped the EPA’s view, the evidence that never made it into the assessment, and what the Bruning–Browning review had already documented about those gaps.
Because once you look closely at the science that should have been considered, it becomes much easier to understand how the EPA arrived at a conclusion so out of step with modern toxicology — and why those gaps still matter today.
Why Evidence Selection Matters More Than We Think
Most people assume a risk assessment is a neutral summary of all the available science.
But in reality, a regulator’s conclusion is shaped long before the final sentence is written.
It depends on:
- which studies count
- which don’t
- how evidence is weighted
- what is considered “reliable”
- and what is dismissed for not fitting the template
Two piles of studies can produce two different outcomes — not because the chemical changed, but because the selection criteria did.
The 2017 Bruning–Browning review made this painfully clear:
New Zealand’s glyphosate assessment didn’t fail because of one big mistake.
It failed because the evidence pipeline itself was too narrow to capture the science that mattered most.
How the EPA’s Evidence Pipeline Actually Works
To understand why so much evidence went missing, we need to understand the system that processes it.
Here’s the basic structure:
- Companies submit their own studies.
These are usually unpublished and remain confidential. - Independent researchers have no structured path to contribute equivalent data.
The EPA cannot commission its own long-term toxicology. - Regulators rely on industry summaries, not raw data.
Independent scrutiny becomes impossible. - Independent peer-reviewed studies are often excluded for technical reasons, not scientific ones — for example, for not following Good Laboratory Practice (GLP) paperwork rules.
This creates a predictable pattern:
industry studies dominate, and independent evidence is repeatedly filtered out.
The 2017 review described this clearly, noting the EPA:
“…appears to give exclusive consideration and weight to industry-paid and industry-supported studies and reviews as well as arguably out-dated and industry-developed guidelines.”
That’s not balance.
That’s structural bias.
What the 2017 Review Found in the Footnotes
Much of the power of the Bruning–Browning review lies in what it uncovered inside the details — the study lists, citations, tables, and omitted material.
Among the clearest findings:
- Industry-funded animal studies were treated as the most reliable evidence, even when unavailable for public review.
- Independent studies showing harm were downgraded or excluded, often without clear reasoning.
- Entire categories of modern toxicology were missing, despite being published before the EPA’s review.
- No scrutiny was applied to overseas regulators’ evidence selection, even when those regulators were later found to have relied on ghostwritten or unpublished industry material.
The authors summarised the problem simply:
“…fails to address twenty-first century scientific understanding of the factors that pre-dispose to risks of cancer development – ignoring new data from toxicology and cancer biology.”
And perhaps the most important line of all — one that should have triggered immediate reassessment:
“…the NZ EPA report considers just ‘glyphosate chemistry and not glyphosate-based formulations that are used in “the real world” and that are obviously of the essence’.”
This was not a small oversight.
It went to the very centre of real-world exposure.
The “Invisible Studies” — Evidence That Existed Before 2016 But Never Entered the Assessment
Many of the concerns people now associate with “new science” were already well established before the EPA finalised its 2016 glyphosate review.
These included:
Endocrine Disruption
Studies from the early 2000s onwards showed glyphosate and Roundup formulations could disrupt hormone pathways at low doses.
This science existed. It was peer-reviewed. It was relevant.
It simply wasn’t integrated.
Oxidative Stress
A substantial body of work—stretching back to the mid-2000s—linked Roundup formulations to oxidative stress in liver, kidney, and neural tissues.
Formulation Toxicity
Research had already shown that surfactants like POEA were significantly more toxic than glyphosate alone.
Some countries were restricting them before 2016.
New Zealand did not examine them at all.
Reproductive & Developmental Evidence
There were published studies documenting impacts on sperm, embryos, placental cells, and reproductive function.
Gut Microbiome Effects
Early but credible research was emerging pre-2016, yet this category was absent from the assessment.
Real-World Exposure & Biomonitoring
Worker exposure studies and ecological surveillance existed internationally.
None were incorporated.
Here’s the key insight:
The evidence didn’t go missing because it didn’t exist.
It went missing because the system wasn’t designed to include it.
The EPA used an older regulatory framework that prioritised GLP industry studies and did not require (or know how to integrate) modern independent toxicology.
That structural blind spot shaped the outcome.
Why Industry Studies Get More Weight — Even When Independent Science Is Stronger
In regulatory toxicology, a study’s format, not its scientific relevance, often determines its influence.
- Industry studies follow GLP rules → high regulatory value
- Independent studies often do not → lower regulatory value
But GLP is an administrative standard.
It ensures strict record-keeping — not real-world biological relevance.
Many independent studies:
- use modern endpoints
- identify early signals of harm
- replicate environmental exposures
- explore mechanisms like oxidative stress or endocrine disruption
Yet they are routinely discounted because they don’t follow industry-style paperwork.
The 2017 review called this out directly — and with good reason.
When format outweighs findings, the system becomes less about risk assessment and more about compliance.
The Transparency Gap — The Public Still Cannot See the Full Evidence
Perhaps the most striking problem is one that hasn’t changed in eight years:
Most industry studies remain confidential and cannot be accessed by the public.
This means:
- independent scientists cannot evaluate them
- NGOs cannot challenge them
- courts may face restrictions
- the public is asked to trust conclusions built on unseen data
When transparency is missing, trust cannot grow.
This gap was obvious in 2017.
It is still obvious in 2025.
Why Missing Evidence Changes the Entire Conclusion
When whole categories of modern science never enter the assessment pipeline, the outcome becomes structurally predetermined.
If studies showing harm are excluded…
and studies showing safety are given exclusive weight…
then the conclusion becomes a reflection of the filter, not the full body of evidence.
This is not a technical flaw.
It is a design flaw.
One that the 2017 review exposed clearly — and that New Zealand has still not addressed.
What This Means for New Zealand Today
As the EPA defends its glyphosate decisions in court again in 2025, the core question isn’t simply:
“Did the EPA get the science right?”
It’s:
“Did the EPA consider the science that matters most?”
Because if important evidence was never examined in 2016…
and the EPA still relies on that same foundation in 2025…
then the reassessment, by definition, inherits the same weaknesses.
This isn’t about blame.
It’s about updating a system built for a different era of toxicology.
Where This Leaves Us — A Case for Independent Evidence Review
If we want a credible glyphosate assessment in New Zealand, the solution isn’t small tweaks — it’s structural reform.
That means:
- independent toxicology capability, funded publicly
- formulation-level assessments
- transparent evidence rules
- reduced reliance on confidential industry submissions
- modern scientific endpoints (endocrine, oxidative stress, microbiome)
- real-world exposure testing
- rules that reflect how New Zealand actually uses these products
Because when evidence never makes it into the room, the conclusion was never built to protect us.
And we deserve better than that.
Mini-series: Revisiting the Glyphosate Review
This article is part of a 3-part mini-series on glyphosate regulation, industry influence, and public health oversight in New Zealand.
Part 1: The Review They Ignored — how the EPA dismissed a damning 2017 assessment of its glyphosate review process
Part 2: The Coziness Problem — how industry language and influence shaped our glyphosate assessments
Part 3: History Repeating — why the EPA’s 2025 court defence sounds a lot like 2017 all over again
Addendum: Missing Science — a standalone analysis examining what was included, excluded, and overlooked in the EPA’s evidence base (you are here)
Resources & References
Before stepping away from this addendum, it’s worth pausing on the sources themselves — because the story they tell is bigger than any single article.
When you look at the evidence that was available before 2016, the gaps in New Zealand’s assessment stop looking like oversights and start looking like the product of a system that was never designed to see the full picture.
The references below aren’t just citations; they map out the science the EPA did consider, the science it didn’t, and the science the Bruning–Browning review warned was being filtered out long before the public ever saw a conclusion.
Public Health Concern: Why Did the NZ EPA Ignore the World Authority on Cancer? (2017) [PDF]
Steffan Browning and Jodie Bruning’s independent assessment of the EPA’s glyphosate review. This document highlighted major methodological problems, gaps in the evidence considered, and systemic issues in how New Zealand regulators rely on overseas assessments.
EPA New Zealand — Review of Hazard Information for Glyphosate (2016) [PDF]
The EPA’s official response to IARC’s “probable carcinogen” classification. The review relied heavily on industry-funded studies and did not assess the toxicity of glyphosate-based formulations.
IARC Monograph 112 (2015)
The WHO-linked International Agency for Research on Cancer classified glyphosate as “probably carcinogenic to humans” after reviewing only publicly available, peer-reviewed evidence.
The Monsanto Papers – Carey Gillam (2021)
An investigative book based on internal Monsanto documents revealing how the company manipulated public messaging, influenced regulators, ghostwrote academic papers, and undermined independent scientists.
Whitewash – Carey Gillam (2017)
Award-winning investigative journalism detailing two decades of hidden science, regulatory failures, and industry influence surrounding glyphosate.
Court of Appeal to Hear Glyphosate Case – NoMoreGlyphosate.nz (2025)
Our coverage of the Environmental Law Initiative’s appeal challenging the EPA’s glyphosate assessment and its reliance on outdated and incomplete evidence.
Key Scientific Studies Referenced in This Article
These studies were not part of the EPA’s assessment, nor were their findings reflected anywhere in the 2016 review.
Formulation Toxicity & Surfactants
Ethoxylated adjuvants of glyphosate-based herbicides are active principles of human cell toxicity.
Mesnage, R., Bernay, B., & Séralini, G.-E. (2013)
This landmark study showed that the surfactants used in many glyphosate-based weedkillers — particularly POEA — are far more toxic to human cells than glyphosate itself. It’s one of the clearest examples of why assessing the “active ingredient” alone misses the real-world risks of full formulations.
Facts and Fallacies in the Debate on Glyphosate Toxicity
Mesnage & Antoniou, (2017)
Explains why glyphosate alone is not representative of real-world exposure and how surfactants dramatically increase toxicity.
Endocrine Disruption (Available before 2016)
Cell Cycle Disruption by Roundup
Marc et al., (2002)
Early evidence showing that Roundup formulations can disrupt cell-cycle regulation — specifically CDK1/Cyclin B activation — at concentrations far lower than those used in agriculture.
Glyphosate induces human breast cancer cells to proliferate
Thongprakaisang et al., (2013)
Identified estrogenic activity and endocrine disruption mechanisms.
Oxidative Stress & Mitochondrial Dysfunction
Comparative effects of the glyphosate formulation Roundup and glyphosate on mitochondrial oxidative phosphorylation
Peixoto, F. (2005)
Demonstrated mitochondrial toxicity of Roundup formulations — effects not seen with glyphosate alone.
Oxidative stress responses of rats exposed to Roundup and its active ingredient glyphosate
El-Shenawy, N. S. (2009)
Showed oxidative stress and liver damage in mammalian models.
Reproductive & Developmental Toxicity
Developmental toxicity of glyphosate-based herbicide in mammals
Dallegrave et al., (2007)
Found skeletal and visceral malformations in offspring following maternal exposure.
Long-term exposure to glyphosate-based herbicide affects testicular function in rats
Cassault-Meyer et al., (2014)
Provided evidence of hormone disruption and reproductive toxicity.
Microbiome Disruption
Effect of glyphosate on gut bacteria
Shehata et al., (2013)
Showed selective suppression of beneficial gut microbes, particularly Lactobacillus and Bifidobacterium.
Genotoxicity / DNA Damage
Biomonitoring of genotoxic risk in agricultural workers from five Colombian regions: Association to occupational exposure to glyphosate
Bolognesi & Carrasquilla, (2009)
Examined occupational glyphosate exposure and found markers of DNA damage in workers — supporting the genotoxicity concern.
This is not an exhaustive list of the available science — not even close.
It’s simply a representative snapshot of the kinds of studies that existed before or around 2016, across multiple areas of emerging concern. Many more papers were published both before and after the EPA’s review, and the scientific picture has only grown stronger since.
What this selection shows is that the problem was never a lack of evidence — it was a lack of recognised evidence. When whole categories of science aren’t required inputs in a regulatory assessment, they effectively disappear from view, no matter how robust or relevant they are.
Image Source & Attribution
The feature image on this page was generated using AI (ChatGPT/DALL·E) and adapted for the web using Canva.
