PMRA’s Proposal to increase glyphosate in legumes is not consistent with science or the law

Here’s an update to the glyphosate MRL story, the story of Health Canada’s Pest Management Regulatory Agency’s (PMRA) proposal to permit more glyphosate in foods.  Safe Food Matters Inc. (SFM) and Prevent Cancer Now (PCN) received information on the “confidential test data” underlying PMRL 2021-10 (the Proposal). The data was received from PMRA on January 28, 2022, more than eight months after the date it was requested, May 12, 2021. 

The groups report that the scientific and legal basis for the Proposal is ill-founded. This article provides a summary of the submission of SFM and PCN on the Proposal, which was provided to PMRA on April 13, 2022 (further to an extension granted to the two organizations).  The full submission can be downloaded here.  

The comments make various points. First, PMRA did not have jurisdiction to increase the glyphosate maximum residue limits (MRLs), and even if it did, it used inappropriate data.  The field trial data and reports PMRA used were either not applicable to the Canadian context or not scientifically justified. The MRLs were “overestimated”.

Plus PMRA did not conduct the evaluation as required. Risk assessment requires examination of hazards (toxicology) and exposure to the pesticide.  On the hazard side, PMRA relied on old science from 2015, used outdated assumptions (like glyphosate doesn’t affect the microbiome), and ignored more recent toxicological and epidemiological evidence showing genetic effects, developmental toxicity, and multi-generational effects, among others. It did not conduct a cancer risk assessment, and dismissed a required safety factor.

On the exposure side, PMRA did not account for the fact, shown in the field trial studies, that residues in beans/seeds actually increase, rather than decrease, with longer preharvest intervals between spraying and harvest.  In  the dietary exposure assessment, PMRA used outdated consumption data from the US, even though more recent Canadian data was available, and applied an unjustified data analysis technique.  It did not conduct an occupational or residential exposure assessment. Examination of the sensitivities of subpopulations (like children and vegetarians) to increased levels of glyphosate did not occur.  PMRA did not update its drinking water assessment even though available evidence showed significant increases in glyphosate levels in surface water.

Finally, PMRA did not assess the risks arising from exposure to the entire glyphosate pesticide product. Science shows the product can contain ingredients that are more toxic than just the ingredient glyphosate alone.

Please note that the comments and below summary relate to PMRL2021-10, not to Discussion Document DIS2021-01 regarding the “Targeted Review” of the Pest Control Products Act (PCPA) that closes May 20, 2022, although these comments could inform the Targeted Review question on setting MRLs. 

Summary of SFM/ PCN Comments

  • No Jurisdiction. PMRA does not have the jurisdiction to increase domestic MRLs pursuant to the Proposal.  The PCPA says MRLs can be specified when PMRA is making a decision on registration of a pesticide (s. 9) or adding a new use to a previous registration (s. 10). However the Proposal was neither – it was just a response to a request from the registrant Bayer to “align MRLs with those proposed” by the FAO and WHO.
  • Inflated Estimates. The Proposed MRLs accommodate inflated estimates of glyphosate residues in crops. MRLs are supposed to be  based on actual results of residues found in field trials where glyphosate is sprayed according to Canadian label directions. But PMRA uses the OECD Calculator to include statistically possible extremes, which inflates or “overestimates” the actual results by statistical extrapolation  – by a factor of 11 for dry beans, 4.4 for lentils and 5 for dry peas, respectively. 

  • Decline Studies show Increases. PMRA and other regulators assume that after preharvest spraying of crops, residues in the bean/seed/grain portion of the plant (the seed) will decline over one or more weeks.  Most of the “decline” field trials showed, unexpectedly, that glyphosate residues didn’t  decline but INCREASED in the seed over the time frame between preharvest spraying and harvest. The seed residues at the time of harvesting the plant were higher when spraying occurred 21 days before harvest (21 day “preharvest interval” or PHI), than the residues at a 14 day PHI and higher again than at a 7 day PHI.  

  • Data Points Not Supportable. The Proposed MRLs and the field trial studies on which they were based came from the 2019 Extra Joint FAO/WHO Meeting on Pesticide Residues (the JMPR). As mentioned, registrants want Canadian MRLs to align with those proposed by JMPR.  The data points used by JMPR either aren’t relevant for Canada or aren’t scientifically defensible:
    • In setting the Proposed MRLs, the JMPR selected residues at only a 7 day PHI. Canada’s PHI is 7-14 days, so the Proposed MRLs do not account for higher residues associated with a 14 day PHI that can result in Canada. 
    • In setting the Proposed MRLs, the JMPR field trials used a rate of preharvest spray application that is 2 to 3 times higher than the rate allowed in Canada. 
    • Reliance on the field trials requires adherence to the label direction of spraying when the seed moisture content is “30% or less”.  The seed moisture for the lentil and pea field trials were not measured or recorded, which means the trials are not verifiable and cannot be relied upon. 
    • The field trial studies and the JMPR report averaged results that were widely different, and thereby obscured and discounted high reside values. This calls into question the reliability of the very few data points, and is not scientifically justified.
    • The seed residue results in the field trials were “not corrected for moisture content” which means there were lower residues on a parts per million (ppm) basis in the freshly harvested legumes at the time of sampling than in the ultimate products, the “dry” beans, peas and lentils. This means the glyphosate residues in the dried commodities are underestimated, although the extent is not known because of lack of data. 
  • Crop Groupings Miss Risks.  The Proposal set MRLs for “each commodity in the listed crop groupings” in accordance with its Residue Chemistry Crop Groups, and in theory the crops are grouped  based on “botanical”  and “growth habit” criteria. A key characteristic is whether a plant is “determinate” (matures with a single flush of flowers and seed production) or “indeterminate” (continues growing, flowering and producing seeds until killed by frost). Growing plants transport glyphosate to the seed, thereby increasing the residues in the food (the seeds people eat). At least the “beans” grouping of the Proposal includes both determinate and indeterminate verities. Thus the grouping is not supportable. The risk is heightened because seeds of indeterminate crops can accumulate higher residue, and this factor was neither acknowledged or addressed. Furthermore, the crop groupings of PMRA in the Proposal have since been changed (but don’t fix the concern).  

  • Dietary Exposure Assessment Flawed. PMRA did not conduct a valid health risk assessment for the MRL increases. In its Evaluation Report it only updated the dietary exposure assessment (DEA). However, the DEA was flawed:
    • It looked at the diet of Americans in 2005-2010, not Canadians, even though PMRA was aware that Canadian data was available from 2015 and met all methodology requirements. PMRA was also aware that the diets of Americans and Canadians differ – Canadians  eat more fresh vegetables and fruit and less processed food than Americans. 
    • PMRA conducted a “deterministic”, not a “probabilistic” analysis, which was not justified by its own policies and statements. It reported on the results of the DEA at the “95th percentile”, rather than at the 99th or 99.9th percentile, because it did “not want to unreasonably overprotect” certain people. At the 99th and 99.9th percentile, the results of PMRA showed exceedances and risks of concern!

  • Cancer Risk Assessment Required but Not Done. In its DEA, PMRA did not assess the risk for cancer, but said it was relying on the dietary exposure assessment published for the 2017 re-registration of glyphosate in PRVD2015-01 and RVD2017-01 and related monographs. Concerns are the following:
    • Policy of PMRA requires examination of both hazard and exposure in risk assessment.  PMRA did not set a hazard endpoint or conduct an exposure assessment for cancer.
    • At the time of the RVD2017-01 decision, PMRA had identified cancer hazard concerns relating to oncogenicity, cytotoxicity and immunotoxicity, and was aware of IARC’s hazard determination that glyphosate was probably carcinogenic to humans, but unjustifiably dismissed these findings.
    • Epidemiological studies merit considerable weight, and must be presumed to have been conducted in the context of adherence to label directions. PMRA dismissed them.  
    • New evidence of cancer since the time of RVD20017-01justifies a cancer risk assessment:
      • Animal toxicology studies  show that glyphosate causes hemangiosarcomas, kidney tumors, malignant lymphomas, kidney adenomas, liver adenomas skin keratoacanthomas, skin basal cell tumors, adrenal cortical carcinomas and hepatocellular adenomas and skin keratocanthomas in mice and rats.
      • The 2020 ATSDR Toxicological Profile on glyphosate of the US Department of Health found positive associations for lymphohematopoeitic cancers.  (As an aside, it also contains an error introduced following the 2019 draft).

  • Toxicological Evidence Requires Endpoint Examination. New evidence showing the hazards of glyphosate requires a lowering of the thresholds (endpoints) for glyphosate toxicity. Strong evidence of genotoxicity, colorectal cancer, immunotoxicity, prenatal developmental toxicity and others are found in the literature. Significant lines of evidence not previously reviewed by PMRA include: 
    • Glyphosate affects the human microbiome. PMRA assumed this was not the case in RVD2017-01. The effects relate to neurological impacts, endocrine disruption, inflammatory bower disease and colorectal cancer, among others. 
    • Animal studies confirm effects in utero.
    • Glyphosate causes multi-generational reproductive harms. 

  • Indirect Effects of Glyphosate on Dietary Risk. Glyphosate also affects the soil microbiome and is associated with higher levels of mycotoxins in crops. It also mobilizes cadmium, which is high in Canadian potash and some prairie soils. Cadmium in Canadian commodities has exceeded European tolerances. These risks warrant assessment.

  • Reduction of Safety Factor without Scientific Rationale. PMRA reduced the 10-fold safety factor the PCPA prescribes without adequate scientific justification.  The application of the factor would substantially adjust the acute reference dose (ARfD) and acceptable daily intake (ADI) endpoints for the DEA. This would have resulted in exceedances of thresholds of concern in the DEA. 

  • Incomplete Health Risk Assessment. A valid health risk assessment requires:
    • examination of new risks arising from exposure, which includes the glyphosate-associated risks of fungal diseases and cadmium contamination in food. PMRA did not assess the risks.
    • an occupational exposure assessment, but PMRA did not conduct one.
    • PMRA to consider “available” information on various factors, but PMRA did not consider such information or update its assessments:
      • regarding residential exposure; 
      • regarding the drinking water component of the aggregate risk assessment, even though the detection of glyphosate in waters has “increased significantly” over the years; 
      • regarding the sensitivities of subpopulations, even though the Environmental Working Group published data showing high levels of glyphosate in children’s cereals. With respect to the diets of vegetarians, certain minorities, and those who wish to eat a healthy diet, the DeTox Project shows high levels in plant based diets and healthier foods because of preharvest spraying, and Moms Across America has shown high levels of contamination in Canadian wheat. 

  • Entire Product Not Assessed.The PCPA defines “Pest Control Product” to mean either the whole product consisting of an active ingredient, formulants and contaminants, or just the active ingredient. PMRA in PRVD2015-01, 2017-01 and the Proposal assessed only the active ingredient, glyphosate. This is problematic because:
    • People are exposed to the whole product, not just the active ingredient, so unless the assessment looks at “real world” exposures of the whole product, PMRA cannot be reasonably certain that risks of concern won’t arise from exposure.
    • It is well established that formulants and contaminants can be toxic in their own right. 

Conclusion: PMRA’s Proposal Not Sound  

In conclusion, the Proposal of PMRA to increase glyphosate MRLs does not have a sound basis in law or in science. The Proposal relies on science and approaches provided by the JMPR, which are not scientifically sound or appropriate for Canada. It did not contain a full health risk assessment, or include current scientific literature showing risks of glyphosate. The dietary exposure assessment referenced in the Proposal was not applicable to Canada, outdated, and used flawed analysis.

It is clear that PMRA did not use a “scientifically based approach” for Proposal PMRL2021-10. As a result, there is no assurance that the increase in glyphosate levels in food set out in the Proposal will not harm the health of Canadians. 

Write a comment