Regulators are ignoring the legal obligation to use independent science in setting safety levels for pesticides including glyphosate, writes Claire Robinson

In 2009 Europe passed a new pesticides regulation mandating that studies from the peer-reviewed open literature be included in the risk assessment. These independent studies had to be considered alongside the dossier of unpublished, commercially confidential studies commissioned by the companies that want to commercialise the pesticide.

It was a major victory for public health and the environment. Independent scientists regularly find risks where industry scientists do not – probably because of the more sensitive and up-to-date methods of detection used.

But industry and regulators have shown themselves to be perfectly capable of ruining the provision at the implementation stage.

In 2011 the European Food Safety Authority (EFSA) produced a guideline making industry's life easy by saying "Klimisch" criteria could be used in judging whether a study should be included in the risk assessment.

These are criteria proposed by three employees of the chemical giant BASF. They judge studies that comply with OECD/Good Laboratory Practice protocols (generally industry-sponsored studies) to be highly reliable – and independent academic studies to be less reliable or completely unreliable. In its guideline, EFSA gave industry the perfect excuse to judge academic studies as irrelevant for risk assessment. In fact, industry doesn't even have to read them. "Not OECD/GLP" is enough of a reason to confine a study to the waste bin.

In a new report, the NGO Pesticide Action Network Europe analysed seven pesticide risk assessments to see whether independent studies were finally being taken into consideration. PAN Europe discovered that only 23% (99) of 434 important toxicity studies from the peer-reviewed literature were identified by industry in these seven risk assessments. Furthermore, not one of those 99 studies was seen as relevant and reliable enough to be used for decision-making, generally because they were not performed according to OECD/GLP protocols.

In other words, more than 300 academic studies with adverse effects were not identified by industry and 100% were judged irrelevant. Low-dose studies, which can show endocrine-disruptive effects of pesticides, were also not considered relevant.  

For the seven pesticides in PAN Europe's sample, several independent studies were found to falsify the alleged safe doses set in the risk assessment. Data from these independent studies showed that the alleged safe doses were between 2 and 1500 times too high.

The European member states overseeing the dossiers allowed industry to disqualify the independent studies without giving scientific reasons. Member states mostly did not even ensure that industry performed the search for independent studies mandated by the 2009 regulation. According to PAN Europe, "Spain even claimed, without rationale, that there is no such mandate."

One honourable exception, Sweden, did call for some independent studies to be considered in one risk assessment.


For glyphosate, only 52% of the independent studies found by PAN Europe were also identified by the industry applicants, and only 31% of the studies found by PAN Europe were discussed. The alleged safe level is falsified by independent studies. Often the industry applicants implied that toxic effects found for the complete herbicide formulations, such as Roundup, were due to the surfactant, an ingredient added to herbicides to enhance its toxicity to weeds. However, they never did any actual tests to determine if that is so.

Democratic decision to improve pesticide assessments stifled

PAN Europe concluded that they were "very concerned" at how a democratic EU decision to bring more and better science into pesticide risk assessments was stifled by the EU institutions and member states.