At present there are no internationally agreed mandatory guidelines on how GM crop animal feeding safety trials should be conducted. Although OECD guidelines for chemical safety testing are a reasonable staring point, there is an urgent need for internationally agreed, scientifically sound experimental designs specifically for evaluating the long-term safety of GM crops.

Séralini based his experimental design on previous studies conducted by industry in support of their applications for market authorization of GMOs. However, he extended the study design in duration, range of test feeding groups, and breadth of anatomical, cellular, and biochemical analyses. Therefore the criticisms levelled at Séralini’s study apply equally or to a greater degree to the studies conducted by industry, which were narrower in scope.

Critics of Séralini’s study are thus applying double standards in accepting the industry studies as sound but condemning Séralini’s study as flawed. Critics must apply the same set of standards by which they find Séralini’s study inadequate to the industry studies. The only conclusion that can logically result from such a comparison is to dismiss both Séralini’s study and the industry studies, or to accept that both have merit within the range of their respective analyses.

If it is agreed that both have some merit, then the following facts must be addressed.

Those who continue to argue that Séralini’s study is unacceptable because it used too few animals must address the fact that it used the same number of animals that were analyzed for blood and urine chemistry by Monsanto in the 90-day studies accepted as proof of safety for GMOs.

This same number of animals has been deemed valid by the OECD for some chronic toxicity protocols for tests performed by industry for regulatory purposes. For example, Seralini’s study used the same number of animals that must be analyzed for blood and clinical biochemistry according to OECD protocol 452, which can be extended for up to two years.

If Séralini’s findings are dismissed because the numbers of animals were too few, then this suggests that toxic effects found in comparable industry chronic toxicity studies are also routinely ignored and the product being tested is allowed to be marketed. If true, this would be a matter of great public concern.

Criticisms based on claims that Séralini’s study is a failed carcinogenicity study are spurious. It is not a carcinogenicity study, but a chronic toxicity study that happened to find tumours.

Even so, generalized claims that long-term studies must use large numbers of animals for findings of tumours to be considered valid are suspect. The larger numbers of animals required, for example, by OECD carcinogenicity protocols, are to avoid “false negatives” – false conclusions of safety. This was not an issue in Séralini’s study, which did not find safety but toxicity. If industry and regulators wish to argue that Séralini’s findings were a “false positive” – an incorrect conclusion of toxicity – they must produce independent toxicological data to prove it.

However we judge the strengths and weaknesses of Séralini’s study, certain conclusions are clear:

  • The short 90-day maximum feeding studies conducted by industry as part of their application for marketing approval of GM crops are inadequate to detect toxic effects arising from their long term (life-long) consumption. Even the French regulatory agencies ANSES and HCB, which criticized Séralini’s study, concluded that long-term studies are needed.
  • The GM industries’ and European Food Safety Authority’s routine dismissal of signs of toxicity observed in 90-day feeding tests as “not biologically meaningful” are false.
  • When evaluating the safety of pesticides, it is not enough to test only the active ingredient (e.g. glyphosate). The commercial formulation as sold and used (e.g. Roundup) must be tested in order to make evidence-based decisions.
  • Long-term tests (at least two years in rodents) should be conducted on all GM foods and pesticide formulations – both those already on the market and those in the pipeline.
  • GM crop feeding studies should be designed to distinguish whether any negative health outcomes are due to disruptive effects from the GM transformation process, the pesticide engineered into or used with the GM crop, or a combination of the two.
  • A minimum of three doses of the GM crop and/or pesticide should be administered to determine any dose-response effects.

Until such tests have been carried out, it is not possible to claim that a given GM crop is safe to eat on a long-term basis. In the interests of safety, Monsanto and other GM seed and pesticide developer companies must not have market access for inadequately tested products. Laws must be re-written to require mandatory long-term and multigenerational feeding trials on GM foods and whole pesticide formulations prior to commercialization, to be paid for by industry but conducted by independent scientists.