In June 2011, Earth Open Source (EOS) published an article titled “Roundup and Birth Defects: Is the public being kept in the dark?” in which the organisation made a number of claims about the safety of the herbicide glyphosate and products containing it. These were said to include:
Developmental malformations affecting the skull, face, brain and spinal cord in frog and chicken embryos at concentrations lower than used in agricultural and garden spraying;
Endocrine disruption, reproductive toxicity and a range of developmental malformations in humans and experimental animals;
Damage to DNA and genetic material in laboratory animals, humans a variety of in vitro test systems;
Cancer of the testis in rats, skin cancer in mice, and blood system cancers in humans; and
Neurotoxicity, including the development of Parkinson’s disease in humans.
EOS was also highly critical of the European Union’s review of glyphosate (EU, 2002 and 1998); challenged the design, conduct and scientific independence of industry-funded toxicology studies; and questioned some of the scientific principles normally applied to the assessment of hazard and risk from chemicals.
Given the widespread use of glyphosate in Australia for weed control in agricultural, home garden and other settings, the APVMA has investigated1 the claims made in the EOS article and created this web-based publication to facilitate communication of its findings with the public and other stakeholders. The APVMA has:
evaluated the key published studies cited in the EOS article together with some newer related publications and archived toxicology studies;
examined the EU review of glyphosate and compared its findings with those of similar reviews prepared by the Australian DoHA, the US EPA (1993) and the Joint FAO/WHO Meeting on Pesticide Residues (JMPR, 2004a,b);
assessed the scientific merit of the EOS arguments and the research upon which they are based; and
considered whether there are implications for the registration of products containing glyphosate in Australia.
The APVMA’s findings are summarised in this section of the publication. More detailed evaluations and scientific discussions of each main issue are presented in Sections 2 to 5 and Appendices 1–5.
1.1 The association between glyphosate / glyphosate-based herbicides and developmental malformations
As stated by EOS, Paganelli et al (2010) have shown that glyphosate and a glyphosate-based herbicide formulation (GBHF) cause malformations including microphthalmia and microcephaly (abnormally small eyes and head) in toad and chicken embryos. However, the routes of administration (incubation with, or injection into toad embryos, and injection into chicken eggs) are not relevant to humans and other mammals, whose foetuses can only become exposed to chemicals if they are absorbed by their mother and transferred across the placenta from her blood circulation.
In 1996 the APVMA reviewed glyphosate products because of evidence of toxicity to amphibians when applied in and around aquatic areas. Toxicity was attributed to polyethoxylated tallow amine (POEA) surfactants in some glyphosate products. The APVMA consequently strengthened label warnings and restricted the use of glyphosate products around waterways and water bodies to reduce the risk of aquatic contamination (see http://www.apvma.gov.au/products/review/completed/glyphosate_history.php), until less toxic formulations could be developed and registered. Today, over a third of registered glyphosate products contain low toxicity surfactants, and can be used in or around waterways (see http://www.apvma.gov.au/news_media/community/2010-13_glyphosate_au.php). Nevertheless, the APVMA will refer Paganelli’s findings to the Department of Sustainability, Environment, Water, Population and Communities (DSEWPaC) for consideration.
Eight developmental toxicity studies with glyphosate in rats and seven in rabbits have been reviewed by pesticide regulatory agencies and scientific organisations including the APVMA, the US EPA, the EU and the JMPR. These and additional studies have also been evaluated by Kimmel et al (2013). The reviews have concluded that at high oral doses, glyphosate causes toxicity to mother rats and foetuses but is not a teratogen (ie, does not cause foetal malformations). The APVMA is satisfied that the German BVL has not misused historical control (HC) data in its evaluations, despite the EOS claim to this effect.
The lowest NOEL for maternal and foetal toxicity in rats was 300 mg/kg bw/d (1000- times the Australian ADI for glyphosate) and the lowest LOEL in foetuses was 1000 mg/kg bw/d. In rabbits, visceral abnormalities including heart dilation and intraventricular septal defect were reported in six of nine developmental toxicity studies. By the most conservative interpretation, these effects were confined to a doses of 450 and 500 mg/kg bw/d. The lowest NOEL for foetal toxicity in rabbits was 100 mg/kg bw/d, or 333-times higher than the Australian ADI. The margins between women’s dietary exposure to glyphosate and the NOELs in laboratory animals are even higher; following a dietary survey of pregnant Australian women and analysis of composite food samples they provided, McQueen et al (2012) estimated that maternal dietary exposure to glyphosate is 0.001 mg/kg bw/d. This dose is 0.33% of the ADI, and is also only 5% of the National Estimated Dietary Intake (NEDI) of 0.02 mg/kg bw/d.
Dallegrave et al (2003) have reported skeletal abnormalities in foetal rats whose mothers were treated at 500–1000 mg/kg bw/d during gestation with an herbicide containing 36% glyphosate and 18% POEA surfactant. However, the study has been criticised for reporting deficiencies and anomalies, and its results may have been affected by non-standard methods used to fix and prepare foetuses for skeletal examination (Williams et al, 2012). The APVMA notes that POEA is not a developmental toxin and has a NOAEL of 300 mg/kg bw/d in foetal rats (Holson, 1990).
The APVMA has investigated EOS’ claims that agricultural use of glyphosate is causing adverse reproductive outcomes in exposed human populations. However, the published body of epidemiological research has produced inconsistent, equivocal or weak evidence of reproductive harm. In particular, most epidemiology studies rely on self-reported exposure information, do not measure exposure, and cannot demonstrate causal associations between glyphosate and reproductive harm. Many studies are also affected by confounding variables including exposure to other possible risk factors and the use of, or potential exposure to, other chemicals.
Furthermore, evidence suggests that exposure of glyphosate product users to glyphosate contained in herbicide products is relatively low, possibly due the low dermal absorption rate of glyphosate, which the EU (2002) has estimated to be less than 3%. In a urinary biomonitoring study of American farming families, the maximum absorbed doses from a single mixing / loading / application event were 0.004 and 0.00004 mg/kg bw in the farmers and spouses, respectively (Acquavella et al, 2004 and JMPR, 2004b). These values represent 1.3 and 0.013% of the Australian ADI for glyphosate.
Therefore, the APVMA is satisfied that glyphosate does not pose a risk of developmental toxicity through public or occupational exposure, despite EOS’s claim to this effect.