Response to a feeding study in goats by Tudisco et al

• kid weights were significantly higher in the control group at day 30 and at slaughter
• kids in the controls groups were significantly taller
• colostrum from the treatment groups was significantly lower in fat and protein content although this difference disappeared after day 15
• colostrum and kid serum IgG concentration were significantly lower in treatment groups
• transgenic DNA was detected in colostrum from goats fed GM soybean meal in their diet

FSANZ response 

• Insufficient information is provided about the diet fed to the goats.  Soybean varieties can vary considerably with regard to nutritional quality, digestibility and anti-nutritional factors. Variability can also occur within a single variety due to agronomic and climatic variation, processing methodology, and storage conditions and duration. Tudisco et al do not provide any information on the variety of the conventional soybean or how and where it was produced, nor is the data provided on chemical composition of the diets extensive enough to determine if they were truly equivalent (e.g. the level of anti-nutrients is not provided).

For example, Gu et al (2010) found significant differences in body weight gain and feed conversion ratios for rats whose diets differed only in the variety of soybean meal (all non-GM varieties). The study indicated that anti-nutritional factors, such as trypsin inhibitor and lectin, varied as much as 2 to 3-fold among the varieties.

• The level of protein in the colostrum (first milk) of GM fed goats was only around half that of the conventional soy diet fed goats. This is a remarkable finding given there were no signs of any maternal toxicity prior to delivery and the kids were of normal weight and size. Reduction in milk quality is typically associated with poor maternal condition.
• Several large farm animal feeding studies exist with the same GM soy (MON40-3-2) in which this claimed effect is not observed (e.g. pigs -  see Świątkiewicz et al (2013)). Tudisco et al do not comment on the lack of concordance of their results with other similar studies.
• The fact that the different treatment and control groups of goats were housed separately prior to kidding may be a confounding factor. The authors do not provide any information in the paper to indicate the housing conditions were the same for each group. Differences in housing conditions between groups could affect the lactation of the dams. 
• Although the dams in the study were selected as homogeneous in parity and milk production during the previous lactation, the authors provide no evidence they were homogeneous for colostrum composition in the previous lactation. By chance, the treatment groups might have had more dams producing colostrum that is lower in immunoglobulins. 
• The authors provide data on various body dimensions of the kids (height, chest dimensions, rump length and width etc.) but there is no information about the conformation of their dams. Differences in conformation between the kids might have been inherited from the dams. Again, there is no evidence that the dams were homogeneous for size and conformation. The body dimensions of the kids should have been corrected statistically for the size of their dams. 
• A number of organs from the kids were weighed but not their brains. This is an important omission because central nervous tissue mass is very highly conserved. It is therefore usual to weigh brain and to calculate both organ-to-bodyweight and organ-to-brain weight ratios. The only data provided in the paper are the group mean organ weights. The parenchymatous organ weights are fairly consistent across groups, which suggest the differences between groups are attributable to the musculoskeletal system. This again raises the question about the extent to which the size and conformation of the kids was due to genetics (i.e. inherited from their dams).
• While the smaller size of the kids may be attributed to a lower protein and fat content in the colostrum/milk it could also equally be attributable to other factors (different genetics, husbandry conditions). The authors failed to address these other possibilities in the discussion of their results.
• It is also of note that no dose response was observed in either the protein levels in the milk, or the slaughter weights of the kids. This suggests the reported effects are more likely to do with animal husbandry, rather than the amount of GM soy fed.
• The authors do not report the volume of milk consumed by the kids following parturition and whether this was similar between groups. Nor do the authors report the group body weight gains, as would normally be expected for a study such as this. It is well known, in lambs at least but probably also for kids, that separation from mothers soon after birth can affect milk consumption.
• No evidence is provided to show that the colostrum and milk were so low in protein/fat that it would have inhibited growth. That is, the lower protein and fat levels observed in milk for the first 15 days of lactation may still have been more than adequate to ensure the kids met their full genetic potential for growth.
• In relation to the detection of transgenic DNA in some of the milk samples, FSANZ notes that soybean DNA was detected in the majority of samples from both the control and GM-fed dams. It is therefore not surprising, particularly given the sensitivity of the analytical technique used, that some of the transgenic DNA would also be detected. There are no safety concerns about the presence of transgenic DNA or other DNA fragments in milk.

FSANZ conclusion