1.2 The association between glyphosate / glyphosate-based herbicides, endocrine disruption and reproductive toxicity
Numerous single- and multi-generation studies have been performed with glyphosate in rats at daily doses of up to 1500 mg/kg. Despite thorough and systematic investigation of relevant parameters, they have yielded no evidence that glyphosate is toxic towards the male or female reproductive systems. No biologically significant effects occurred in a 13-week US National Toxicology Program (NTP) reproduction toxicity study in rats and mice at dietary doses of up to 5000 and 7500 mg/kg bw/d in the respective species. Furthermore, no effects indicative of endocrine disruption have been found in short-term repeat-dose, subchronic and chronic toxicity studies with glyphosate in laboratory animals, and glyphosate has negligible or weak effects on steroid hormone receptors and biosynthesis in vitro.
Little reliance can be placed on the study of Yousef et al (1995), which EOS claims to have demonstrated sperm damage in rabbits. When administered for six weeks, glyphosate may have caused fully or partially reversible decreases in ejaculate volume and the viability and activity of sperm, but the study used low numbers of animals and deficient experimental methods, was markedly affected by variation within the control group, and was poorly reported. It is even unclear what doses of glyphosate were administered.
Although EOS has described glyphosate as causing testicular cancer in rats, independent assessments of the relevant study (Lankas et al, 1981) by Australia, the WHO and the US EPA have concluded that the tumours were not treatment-related. Furthermore, neither testicular tumours nor other forms of cancer have developed in eight other carcinogenicity studies with glyphosate in mice or rats, respectively at doses of up to ca 5000 and 1200 mg/kg bw/d. Mink et al (2012) have reviewed the epidemiological literature (7 cohort studies and 14 case-control studies) to evaluate whether exposure to glyphosate is associated causally with cancer risk in humans. They found no consistent pattern of positive associations to indicate a causal relationship between total cancer (in adults or children) or any site-specific cancer and exposure to glyphosate. This provides strong evidence that glyphosate does not pose a carcinogenicity hazard to humans.
The APVMA anticipates that glyphosate’s potential to cause endocrine disruption will be clarified in the near future, as the active has been tested according to US EPA Series 890 Test Guidelines following its selection for Tier 1 screening under the EPA’s Endocrine Disruption Screening Program. As at June 2013, all data have been received by the EPA and are currently under review (see http://www.epa.gov/scipoly/index.html). So far three abstracts have been published, demonstrating a lack of potential to interact with oestrogen and androgen receptors in vitro, inhibit steroidogenesis in vitro, affect thyroid-mediated developmental endpoints in the amphibian metamorphosis assay, or cause endocrine disruption in the Hershberger and uterotrophic assays in rats (Levine et al 2012, Webb et al 2012, Saltmiras and Tobia 2012).
There is experimental evidence in support of EOS’s assertion that glyphosate-based herbicide formulations (GBHFs) cause reproductive toxicity in drakes and, in male rats, interfere with the maturation of the reproductive organs during puberty. In some studies (Oliviera et al 2007, Romano et al 2010) GBHFs were administered directly to the test animals, while other studies (Dallegrave et al 2007, Romano et al 2012) involved maternal exposure to GBHFs during pregnancy and/or lactation. However, the observed effects have been inconsistent, including increases and decreases in blood testosterone levels and sperm production, and delaying and hastening of the onset of puberty. Furthermore, most of the relevant studies are deficient in aspects of their design and reporting, have used novel, unvalidated test methods, and/or may have been subjected to interference by experimental artefacts. None of the studies have identified which component(s) of the test GBHFs caused the reported effects.
In vitro, some GBHFs have caused anti- androgenic and oestrogenic activity, changes in the expression of hormonally-regulated genes, inhibition of aromatase (an enzyme that converts testosterone to oestradiol), cell injury and death. However, many in vitro studies have used cancer cells or other novel test systems, and a 2009 Canadian PMRA assessment concluded that their findings are not representative of the exposure of live animals and humans (see http://www.hc-sc.gc.ca/cps-spc/pubs/pest/_fact-fiche/glyphosate/reconsideration-reexamen-eng.php). Furthermore, surfactants (including POEA) are a likely cause of cellular toxicity and interference with in vitro assays of hormonal regulation. Few studies have identified or controlled for the surfactants and other adjuvants present in test formulations, creating uncertainty as to which chemicals are causing the reported effects, and their mode of action.
Therefore, the APVMA believes it is premature to characterise GBHFs as endocrine disruptors.
1.3 Evidence for the genotoxicity and carcinogenicity of glyphosate / glyphosate-based herbicides
Only a small minority of the genotoxicity studies with glyphosate and GBHFs have yielded positive findings, some of which were inconsistent with negative results in other studies examining the same end-point. Interpretation of several published studies is hindered by methodological failings or inadequately-detailed reporting. Many instances of positive findings could also be explained by cytotoxicity, ie, generalised toxicity against the test cells, tissues or organs, as opposed to direct effects on genetic material. When the activity of glyphosate and GBHFs was compared under the same experimental conditions, the active constituent was usually inactive or much less active than the formulations. Where studies were performed with GBHFs without examining the individual ingredients, it is unknown whether the findings were caused by glyphosate, surfactants or other adjuvants, or depended on interaction between the various formulation components. A recent, comprehensive review of published and sponsored regulatory genotoxicity studies (Kier and Kirkland, 2013) has concluded that glyphosate and typical GBHFs do not appear to present significant genotoxic risk under normal conditions of human or environmental exposures. Studies of genetic injury within human populations have not yielded consistent evidence of a causal association between glyphosate exposure and genotoxicity. Therefore, weight and strength of evidence supports the view that glyphosate is not genotoxic.
Between them, the Australian DoHA, the US EPA, the EU and the JMPR have reviewed four dietary carcinogenicity studies with glyphosate in mice and six similar studies in rats, performed over dose ranges of 11 – ca 5000 and 4 – ca 1200 mg/kg bw/d in the respective species. Although the incidence of testicular tumours was increased in glyphosate-treated rats in one study (Lankas, 1981), the reviewing agencies agreed that by reference to HC data, the tumours were not related to treatment. Furthermore, tumours did not develop in the testis – or any other organs or tissues – in the remaining carcinogenicity studies.
A GBHF has been found to promote skin tumours when applied dermally to mice at 25 mg/kg bw/d (George et al, 2010), but carcinogenesis depended on prior treatment with a tumour initiator chemical, without which there was no development of cancer. The study did not demonstrate which component(s) of the product caused the promoting activity. The finding is of limited relevance to persons preparing GBHFs for use because the tumour promoting activity was relatively weak, and to achieve an equivalent level of exposure, operators would have to be exposed three times weekly for over a decade at doses unattainable while wearing the required protective clothing and equipment.
The Australian DoHA (2005) and the JMPR (2004b) have assessed nine epidemiological studies performed from 1999 onwards, including those cited by EOS as showing associations between glyphosate and blood system cancers. Some researchers have found increased odds of developing non-Hodgkin’s lymphoma, multiple myeloma or hairy cell leukaemia in persons who have used or been exposed to glyphosate. However, the evidence has been inconsistent both between and within studies, whose outcomes are potentially confounded by inaccurate exposure data and exposure to other pesticides and environmental agents. A recent review (Mink et al, 2012) of epidemiological studies relevant to cancer end-points considered seven cohort studies and 14 case—control studies; there was no consistent pattern of positive associations to indicate any causal relationship between total cancer (in adults or children) or any site-specific cancer and exposure to glyphosate.
Currently, the weight and strength of evidence does not support the conclusion that glyphosate causes cancer in either laboratory animals or humans.