The results in [Séralini’s] article add new knowledge to our scientific literature. These include the surprisingly similar physiological response of the rats to all three treatments (Roundup, Roundup Ready maize, Roundup Ready maize + Roundup) relative to the control. Hormonal imbalances were associated with Roundup in previous studies. But this new study suggests such imbalance may lead to greater problems with tumours in older female rats. Furthermore, the low threshold of response to Roundup in drinking water was biologically very significant, and demands further research. It is also significant that negative effects were found testing commercial Roundup formulations, which underlines the importance of testing real products rather than so-called active ingredients.

In contrast to the many (often excessive) critiques that can be read on the net these days about the study, my review of the paper suggests that it is:

1)    based on careful, state-of-the-art research technique;

2)    does not reflect any misplaced biases but rather follows a literature-based line of reasoning;

3)    clearly shows statistically significant and biologically meaningful differences between the control group and treatments with Monsanto products.

It seems to me that the most cogent critique is that insufficient numbers of rats were used for the test, especially as results involved tumours. However, the numbers of rats utilized (10 per sex per treatment) were in line with numbers suggested by OECD non-mandatory [chronic and subchronic toxicity] regulations. It is probable that the authors held the number of rats to the minimum in line with the high expense involved in such experiments. It also seems probable that the authors did not expect tumours to be an outcome of the study.

Monsanto’s feeding trials using glyphosate tolerant corn (Hammond et al. 2004) utilized twice as many rats per treatment, but most of the clinical work (clinical pathology, haematology, serum chemistry, urine chemistry) was based on analyzing only ten rats per sex per treatment as in the Séralini study. Thus logically, the same critique leveled at the Séralini study should also be leveled at the Hammond study. On the other hand, the Hammond study could be criticized for not being as clean a design as the Séralini study as it utilized lots of unnecessary controls, thereby potentially weakening or cluttering statistical analysis.

Irrespective of all this, the overall biological results of the Séralini study go in a consistent direction that is surprising, shocking, and stimulating. The social implications seem to be that:

1)    we may have been short-sighted or misled in believing short-term feeding trials gave assurance of effect;

2)    more long-term feeding trial research is needed by independent researchers on the subject with more rats per treatment;

3)    GM crops and Roundup may not be as safe as many have wanted to believe.

Walter A. Goldstein, research director, Mandaamin Institute, USA