Total no. foetuses with cardiovascular malformations
Total no. foetuses with skeletal malformations
Foetal loss (%)
Stauffer Chemical Co (1983b)^^
Coles and Doleman (1996)
Kimmel et al (2013)
Embryofetal deaths (Mean no./litter)
Postimplantation loss (%)
Total no. foetuses with cardiovascular malformations
Kimmel et al (2013)
Embryofetal deaths (Mean no./litter)
Postimplantation loss (%)
Foetuses with cardiovascular malformations (%)
Foetuses with skeletal malformations (%)
Foetuses with skeletal variations (%)
Statistical significance vs concurrent control group: *p < 0.05 **p<0.01
CC = Concurrent control group mean HC = Historical control group range NR = None reported
^Glyphosate administered in the diet; otherwise, gavage dosing ^^Glyphosate trimesium
Brooker et al (1991b) [Reviewing Agency: BVL] The BVL assessed this study for the JMPR (2004b) and EU (1998) reviews of glyphosate. The evaluation for the EU was less detailed, but both assessments established the same NOELs / NOAELs and reached the same conclusions as to the biological significance of foetal mortality and heart malformations, based on HC data. In female rabbits orally gavaged from GD 7 – 19 at 0, 50, 150 or 450 mg/kg bw/d, there were dose-related increases in the incidence of soft / liquid faeces and inappetence and decreases in food consumption and bodyweight gain. The NOAEL for maternotoxicity was set at 50 mg/kg bw/d. Late embryonic deaths were increased significantly at 450 mg/kg, but not at the mid and low doses. At and above 50 mg/kg bw/d, total embryonic deaths and post-implantation losses were significantly higher than in the concurrent controls. Although no explicit rationale was given the BVL did not attribute embryo mortality at 50 and 150 mg/kg to treatment, possibly because the incidence of total (early + late) embryonic death was not dose-related and lay within the HC range from 21 studies performed over 1989 – 1990. The proportion of malformed foetuses was slightly increased at 150 and 450 mg/kg, due to increased incidences of interventricular septal defect and other cardiac abnormalities. However, the BVL did not consider the cardiac abnormalities to be treatment-related, as their incidences lay within the HC range in 13 studies performed in 1989. The NOAEL for developmental toxicity was set at 150 mg/kg bw/d, based on the increased incidences of late embryonic death and postimplantation loss at 450 mg/kg bw/d.
Comment: Postimplantation losses at 50 and 450 mg/kg exceeded the HC range by 2.0 and 3.5%, respectively.
Bhide and Patil (1989)[Reviewing Agency: BVL] When rabbits were gavaged with glyphosate at 0, 125, 250 or 500 mg/kg bw/d between GD 6 and 18, abortion occurred in 2/15 does from the high dose group, which also displayed depression in food consumption and bodyweight gain. A maternal NOEL of 250 mg/kg bw/d was set. Fetotoxicity, skeletal variations and visceral malformations were noted at 500 mg/kg, seen as decreased foetal viability, increased foetal non-viability and increased incidences of unilateral 14th rib, ventricular septal defect, absent kidney and absent postcaval lung lobe. A NOEL of 250 mg/kg bw/d was established for developmental toxicity. No reference was made to historical control data.
Comment: EOS’s disagreement with the BVL evaluation focuses on increases in the incidences of ventricular septal defect, absent postcaval lobe and absent kidney at 125 and 250 mg/kg, even though the increases are small compared with those seen at 500 mg/kg. EOS also contends that the increase in rudimentary 14th rib at 250 mg/kg was treatment-related.
Moxon (1996b) [Reviewing Agency: BVL] This particular assessment was performed only for the JMPR review. In female rabbits orally gavaged from GD 8 – 20 at 0, 100, 175 or 300 mg/kg bw/d, the NOAEL for maternotoxicity was 100 mg/kg bw/d based on clinical signs (diarrhoea and reduced faecal output) and reduced food consumption and bodyweight gain at and above 175 mg/kg bw/d. At 300 mg/kg, mean foetal bodyweight was depressed by ca 8%, there were significant increases in the incidence of partially or un-ossified vertebrae and sternebrae (see Table), and slight increases in manus and pes scores11. The proportion of foetuses with minor skeletal defects was statistically significantly increased at the low and high doses but not at 175 mg/kg bw/d, which the BVL assigned as the NOAEL for developmental toxicity based [probably] on reduced foetal bodyweight at 300 mg/kg.
Suresh (1993a)[Reviewing Agency: BVL] Rabbits were gavaged with glyphosate at 0, 20, 100 or 500 mg/kg bw/d over GD 6 – 18. The 500 mg/kg dose caused inappetence, clinical signs, a possible depression in bodyweight gain and the death of 8/16 does. A further 4/16 does died at 100 mg/kg without displaying signs, but the BVL attributed their mortality to treatment and set the maternal NOEL at 20 mg/kg bw/d. Abortion did not occur at any dose but one doe displayed complete resorption at 500 mg/kg. At caesarean section on GD 28 there were 20 / 133, 13 / 78, 12 / 77 and 6 / 28 pregnant does / foetuses in the respective groups.
There was no treatment-related effect on external or skeletal malformations. A slight, dose-related upwards trend in the incidence of extra 13 rib was evident in the treated groups, attaining statistical significance (p<0.05) at 500 mg/kg only. There were also eight foetuses with major visceral malformations at 500 mg/kg (significant, but p value unstated), compared with four in the control group and six at 20 and 100 mg/kg. Of these foetuses, four, four and five at 20, 100 and 500 mg/kg had dilated heart, compared with none in the control group. The percentage incidence was significant vs control (p<0.05) at all doses; see Table. In contrast to the study author, who interpreted the lowest dose (20 mg/kg bw/d) as an effect level, the BVL reviewer assigned a NOEL of 100 mg/kg bw/d based on the increased incidence of 13thth rib and heart dilation at 500 mg/kg.
The BVL’s rationale for the choice of NOEL was as follows:
The absolute number of foetuses with dilated heart was small.
The number of affected litters (3/13, 2/12 and 2/6 at 20, 100 and 500 mg/kg) was also low.
The numbers of affected foetuses or litters did not differ markedly between the treated groups.
The study author provided no information about the severity of heart dilation, and the consequences of such a finding in a foetus were “equivocal”.
There was no evidence of other and much more common visceral anomalies.
Therefore, it was “rather unlikely” that the isolated finding of heart dilation was indeed related to treatment, but nevertheless
Based on the [foetal incidence data], a treatment-related effect could not be completely excluded, at least at 500 mg/kg.
The BVL did not identify the other major visceral malformations found in four, two, two and three foetuses at 0, 20, 100 and 500 mg/kg.
No reference was made to HC data; hence, it is unclear whether the control group was unrepresentative of the background rates of cardiac abnormalities at the study laboratory.
Heart dilation was classified both as a malformation and a major visceral anomaly (final paragraph of p 109 and Table B.18.104.22.168.1-1, Annex B-5). Combined with the lack of information as to the severity of the finding, this creates ambiguity as to the functional significance to the developing foetus.
Tasker et al (1980b)[Reviewing Agencies: BVL, US EPA and Australian DoHA] According to the BVL evaluation for the EU, rabbits gavaged with glyphosate at 0, 75, 175 or 350 mg/kg bw/d over GD 6 – 27 displayed clinical signs and potentially treatment-related maternal mortality at and above 175 mg/kg. The NOEL for maternotoxicity was therefore set at 75 mg/kg bw/d. There were no effects on foetal survival, growth or development, and so the foetal NOEL was set at 350 mg/kg bw/d.
The US EPA (1993) assessment differed in setting a NOAEL of 175 mg/kg bw/d for maternotoxicity. However, the EPA agreed that there was no developmental toxicity at any dose tested.
The DoHA (1985) set a NOEL for maternotoxicity at 175 mg/kg bw/d, based on diarrhoea, soft stools, nasal discharge and the death of 10/16 rabbits at 350 mg/kg. In common with the BVL and EPA, no treatment-related effects were considered to have occurred on foetal survival, growth, sex ratio or development. This assessment has been corroborated independently by Williams et al (2012).
Anon (1981) [Reviewing Agency: BVL] In this study, which the EU classified as “supplementary” due to serious reporting deficiencies, glyphosate was administered in the diet to rabbits over GD 6 – 19 at calculated actual doses of 0, 10.5, 50.7 and 255 mg/kg bw/d. There was no evidence of maternal toxicity, but foetal losses were markedly enhanced at the mid and high doses (incidences were 0.9, 0.8, 6.1 and 7.0% in the respective groups). Foetal bodyweight was not affected and no malformations were noted. The BVL assigned a NOEL of 10.5 mg/kg bw/d for fetotoxicity.
Comment: The evaluator remarked that it was unclear why “...an increase in intrauterine mortality would be elicited in a feeding study at doses far below those at which foetal effects were observed in the gavage studies. Thus, it is very doubtful whether this finding was actually related to glyphosate administration. Against the background of the data obtained in more valid, GLP-like studies, it can be concluded that the NOEL for developmental toxicity in rabbits is much higher.” Presumably, the BVL reasoned that foetal exposure to glyphosate after maternal dietary dosing at 50.7 and 255 mg/kg would have been lower than attained at doses up to 350 mg/kg in the gavage studies.
Stauffer Chemical Company (1983b) [Reviewing Agency: Australian DoHA] When pregnant rabbits were gavaged with glyphosate trimesium at 0, 10, 40 or 100 mg/kg bw/d from GD 7 to 19, maternal mortality and abortion occurred at 100 mg/kg bw/d and clinical signs were observed at 40 mg/kg and above. Significant decreases in maternal bodyweight gain and food consumption were noted throughout the dosing period at 100 mg/kg, while there was depression in bodyweight during the first seven days of dosing at 40 mg/kg. The maternal NOEL was 10 mg/kg bw/d. There were no effects on foetal survival, bodyweight gain or development at any dose.
Kimmel et al (2013) [Reviewer: Scitox Assessment Services] These authors assessed seven proprietary developmental studies with glyphosate in rabbits. Five studies (Moxon, 1995b; Brooker et al 1991b, Tasker et al, 1980b; Suresh, 1993a; Bhide and Patil, 1989) had been reviewed previously by the BVL, US EPA and / or Australian DoHA (see above).
Kimmel et al corroborated the BVL assessment of Brooker et al (1991b), describing cardiovascular malformations including intraventricular septal defect, retroesophageal right subclavian artery, dilated or narrowed aorta or pulmonary artery, and disproportionally sized ventricles, seen either alone or in combination.
In the study of Moxon (1996b), Kimmel et al noted three foetuses (one each in the control, 100 and 300 mg/kg groups) had “heart defects involving effects on septation”, together with statistically significant increases in the incidences of minor skeletal malformations at 100 and 300 mg/kg and skeletal variations at 300 mg/kg only. The NOAELs for maternal and developmental toxicity were set at 100 and 175 mg/kg bw/d, respectively, the same doses assigned by the BVL.
Kimmel et al confirmed that there were no cardiovascular malformations or treatment-related skeletal malformations in Tasker et al (1980b), and in common with the BVL assigned a NOAEL of 75 mg/kg bw/d for maternal toxicity. Kimmel et al set a developmental NOAEL of >175 mg/kg bw/d because they considered that too few foetuses were available for adequate morphological assessment of the 300 mg/kg group.
With respect to Suresh (1993a), Kimmel et al corroborated the BVL’s reporting of maternal mortality and clinical signs but set a maternotoxicity NOAEL of 100 mg/kg bw/d. They also confirmed the BVL’s stated incidences of cardiac dilation among foetuses, while adding that Suresh reported (but did not define) “seal-shaped” heart in one control foetus and one 100 mg/kg foetus, the latter also displaying cardiomegaly. Kimmel et al also clarified that two visceral malformations (single cases of liver haematoma and absent gall bladder) seen at 500 mg/kg were unrelated to the cardiovascular system. Given that only 28 foetuses were available for examination at 500 mg/kg, Kimmel et al established the developmental NOAEL at 100 mg/kg bw/d. They commented that the observation of dilated hearts (which was unique to this study) may have been due to overly stringent inspection compared to criteria used by other laboratories.
Kimmel et al also reviewed the study by Bhide and Patil (1989), but concluded its data were unsuitable for setting NOELs because of reporting deficiencies and inappropriate experimental methods. Nevertheless, their assessment of embryofetal mortality and malformations was consistent with the BVL’s.
Two other studies in rabbits (Hojo, 1995; Coles and Doleman, 1996) have not been included in any available agency review. Hojo administered glyphosate by oral gavage at 0, 10, 100 or 300 mg/kg bw/d over GD 7 – 19 and observed hypoplasia of the pulmonary artery and ventricular septal defect in one foetus at 100 mg/kg, but no other cardiac abnormalities. No skeletal variations or malformations were ascribed to treatment. Based on clinical signs (soft / liquid faeces) at 300 mg/kg, a NOAEL of 100 mg/kg bw/d was assigned for maternal toxicity. The developmental NOAEL was >300 mg/kg bw/d.
Coles and Doleman gave oral gavage doses of 0, 50, 200 or 400 mg glyphosate/kg bw/d to pregnant rabbits from GD 7 to 19. Based on clinical signs (soft, liquid, mucoid faeces) and decreased bodyweight gain, a NOAEL for maternal toxicity was set at 200 mg/kg bw/d. Embryofetal deaths and post-implantation losses were increased at 200 and 400 mg/kg, but statistical significance was attained at 200 mg/kg only. At 400 mg/kg, the increase was due to one doe with nine late foetal deaths, which Kimmel et al considered to be of questionable biological significance. At 200 mg/kg, a heart and great vessel defect occurred in an acephalic (headless) foetus. However, there were no other cardiovascular malformations and no treatment-related skeletal malformations or variations. A NOAEL of >400 mg/kg bw/d was assigned for developmental toxicity.
After Kimmel et al aggregated the data for each dose level (excluding those from Bhide and Patil, 1989), the incidences of septum-related defects were 1/770 in controls and 6/1939 among glyphosate-exposed foetuses (i.e. 0.13 and 3.1%). Four of the six cases in treated groups occurred at the maternally toxic dose of 450 mg/kg.
Septal defects were not observed among 747 foetuses whose mothers received 175, 200, 300, 350 or 400 mg/kg bw/d.
Cardiomegaly was seen in one foetus at 100 mg/kg (i.e. in 1/374 foetuses or 0.27% incidence), while one case of dilated ventricles occurred at 0, 100 and 500 mg/kg (i.e. 1/770, 1/374 and 1/28 fetuses in the respective groups, = 0.13, 0.27 and 3.6% incidences). Dilated heart was reported in 4/78 (5.1%), 4/374 (1.1%) and 5/28 (17.9%) foetuses at 20, 100 and 500 mg/kg. None of the 954 foetuses whose mothers received glyphosate at 150 – 450 mg/kg bw/d displayed cardiac or ventricular enlargement or dilation. The aggregated data suggest that even if they are not a reporting artefact, the cases at 20 and 100 mg/kg bw/d were not treatment-related.
Kimmel et al concluded that “there was no increase in cardiovascular malformations at doses that were not overtly toxic to the pregnant rabbits (i.e. generally at doses over 150 mg/kg [bw]/d”).
Comment: Inclusion of data from Bhide and Patil (1989) in the aggregated dataset would make negligible difference to Kimmel et al’s analysis of the incidences of cardiac / ventricular enlargement or dilation, since these findings were reported only by Suresh (1993a). It would add single cases of septal defects at 125 and 250 mg/kg and a further two cases at 500 mg/kg bw/d, making a total of ten affected foetuses from treated mothers12 (one each at 100, 125, 150 and 250 mg/kg, with four at 450 and two at 500 mg/kg). In the APVMA’s opinion, this pattern is most consistent with septal defects having a relationship to treatment at 450 and 500 mg/kg, but not at <250 mg/kg bw/d.