Science in Society Archive

Letter to Nature Biotechnology: Systematic bias in favour of no adverse impacts from GM feed

I am writing in response to your article1 defending the ‘new format’ of your Feature2 on Ermakova’s findings of adverse health and reproductive impacts on rats fed genetically modified (GM) soya, and on the validity of the ‘scientific criticisms’ of her work.

You are still being unfair to Ermakova, especially in allowing the panel of critics in your original Feature2 - all well known for their writings and public appearances if not in declared financial interests to be strongly pro-GM - to have the last word.3

You asked for suggestions regarding the format you might use for Features of this kind. The real issue, however, is not the format but the journal’s policy on reviewing. When there is a debate or a controversy about an issue, reviewers must apply the same standards to papers on both sides, and that Chassy et al2,3 did not do. They explicitly wrote that Ermakova’s work should be judged by a more rigorous standard because it contradicts earlier work that showed no adverse effects from GM food. Yet the crucial earlier work4 that they repeatedly cited was indeed, not subjected to same rigorous standard they are demanding for Ermakova’s study; far from it. The same applies to other earlier research purportedly demonstrating that GM food is safe.

The specific GM food in question, Monsanto’s Roundup Ready soya (RR soya, event 40-3-2), has been commercially grown since 1996 if not before.  But contrary to the assertions of proponents such as Chassy et al,2,3 its market approval - as indeed the market approval of all genetically modified organisms (GMOs) - has been contested right from the start.5 At issue was a reductionist regulatory regime that allowed companies to present results of the most undiscerning tests to bolster the claim that the GMO is ‘substantially equivalent’ to, and hence as safe as, its conventional counterpart(s).

Feeding trials, typically conducted by the company seeking market approval,6 tended to focus on agronomic performance, not on safety, and were not of sufficient duration to assess any but the most acute, short-term effects.

Ermakova’s experiment is new in that it involves long-term transgeneration feeding through two generations (parent and their offspring), and in contrast with other investigations in which females started to be fed GM soya during their pregnancy, she began to feed the rats before and during mating, and continued during pregnancy and beyond.

The ‘multi-generational’ study by Brake and Evenson,4 which found no adverse effect, is widely seen as the most comparable to Ermakova’s study,2,3 though that’s far from being the case. Brake and Evenson used mice, not rats; more importantly, the experiment did not involve trans-generational feeding with RR soya. The misleading term ‘multi-generational’ merely refers to breeding the mice for three generations, and carrying out a separate feeding experiment for each generation. The study was otherwise also fatally flawed and we are surprised it got through peer review.

It claimed to have used a batch of RR soya harvested in a middle of a certain field in South Dakota, processed by a commercial company, and fed to mice of indeterminate age and body weight. The compositions of the RR and conventional soy-formulated diets were remarkably similar; so much so that 59 of the 78 components listed were identical to 2 or 3 significant figures, and the rest differed so slightly that they would have been within standard errors. Could it be that the researchers have been feeding both groups the same diet? No evidence was provided to indicate that the two diets were different; no PCR tests were performed to ascertain that one contained RR soya and the other conventional soya. If there is any doubt that Ermakova had fed RR soya to her rats, at least she has provided PCR data documenting that transgenic soya was fed to the experimental group and non-transgenic soya to the controls.7

The Brake and Evenson study, like many other ‘peer-reviewed’ published studies claiming no adverse health impacts, suffers from at least as many flaws as the largely unpublished study of Ermakova, if not more. This tells us that journals including Nature Biotechnology are applying different, mainly uncritical, criteria in favour of papers and commentaries claiming to find no adverse impacts.

More disturbing still, our regulators too, have been equally uncritical, if not systematically biased in accepting flawed data and unsubstantiated claims of safety in unpublished reports submitted by companies seeking market approval for a whole range of GM food and feed while ignoring and dismissing a long string of evidence preceding Ermakova’s findings suggesting that the GM technology may be inherently hazardous, as we have documented elsewhere.8

A recent case in point is Monsanto’s feeding study submitted in 2003 to the European Food Safety Authority (EFSA) for market approval of its GM maize MON 863 that claimed to find no adverse impacts. Independent review9 however, found serious flaws in the study at every stage, from experimental design, to data collection, reporting and analysis. Monsanto, supported by EFSA, kept the study from public scrutiny under a false claim of confidential business information until a German court order forced Monsanto to release the full report in 2005. Monsanto published the study in 2006,10 restating its position that the significant differences between transgenic and non-transgenic fed groups were “not biologically meaningful.” But a reanalysis of the data using the appropriate statistical tests and found signs of toxicity to the liver and kidney of rats fed the GM maize.11

Nature Biotechnology, like our regulators, have forgotten that the safety of RR soya is a matter of public health, and any new data suggesting it might not be safe should be taken most seriously, especially as one commentary puts it,12 “the prior research is inconclusive at best.” The ‘interview’ with Ermakova and the follow-up in which Chassy et al are given the last word indicates that, “Nature Biotechnology has for some time been unclear where the line that traditionally separates trade magazines from science journals lies.”

Mae-Wan Ho, Institute of Science in Society, UK

Article first published 07/01/08

  1. Marshall, A. Nat. Biotechnol. 25, 1359-1360 (2007).
  2. Marshall, A. Nat. Biotechnol. 25, 981-987 (2007).
  3. Chassy, B. et al. Nat. Biotechnol. 25, 1356-1358 (2007).
  4. Brake, D.G. & Evenson, D.P. Food Chem. Toxicol. 42, 29-36 (2004).
  5. Ho, M.W. & Steinbrecher, R. Journal of Nutritional and Environmental Interactions 2, 51-84 (1998).
  6. Hammond, B. et al. J. Nutr. 126, 717-727 (1996).
  7. Ermakova, I. Nat. Biotechnol. 25, 1351-1354 (2007).
  8. Ho, M.W., Cummins, J. & Saunders, P.T. Microbial Ecology in Health and Disease 19, 66-77 (2007).
  9. Preliminary report by Criigen on the “First public investigation of the crude data in Mon 863 toxicity tests on rats” 2005,
  10. Hammond, B. et al. Food Chem Toxicol 44. 147–160 (2006).
  11. Séralini. G.-E., Cellier. D., & Spiroux de Vendomois, J. Arch Environmental Contamination and Toxicology  published online 13 March 2007, Doi 10.1007/s00244-0149-5.
  12. The Bioscience Resource Project, Dec 4 2007, http://www=2Ebioscienceresource=2Eorg/commentaries/brc6=2Ephp

Got something to say about this page? Comment

Comment on this article

Comments may be published. All comments are moderated. Name and email details are required.

Email address:
Your comments:
Anti spam question:
How many legs on a cat?