updated August 2012                                                  Cliquez ici pour la version française

A recent publication by Aziz Aris and Samuel Leblanc in the journal Reproductive Toxicology (Maternal and fetal exposure to pesticides associated to Genetically Modified Foods in Eastern Townships of Quebec, Canada) claims to have detected traces of:

herbicides (used on herbicide tolerant ‘genetically modified’ plant varieties) or their major metabolite,

and the insecticidal protein Cry1Ab (produced by certain varieties -called Bt- resistant to insect pests)

in the blood of Canadian women, pregnant or not pregnant, and in umbilical cords.

This site will publish any credible information about the validity of these claims and this article will be updated periodically.

The critical views posted on this site are in agreement with the comments of Food Standard Australia New Zealand (FSANZ) and of Prof. A. de Weck (immunologist). Read also a publication by Utz Mueller & Janet Gorst (FSANZ) reminding that the compounds allegedly detected can originate from other sources than GMOs (Bt proteins are also used on organic products for example). 

Read also the information provided by A. Blacker et al. (Bayer CropScience) which suggests an enormous error in detecting glufosinate and a metabolite. Read also the answers by D.A. Goldstein et al. (Monsanto).

A publication lacking credibility

Only claims of Aris and Leblanc on Cry1Ab are discussed here for the time being.

The Cry1Ab protein is produced by some Bt cotton and corn (e.g. MON810). Aris and Leblanc claim they detected this protein in 93% of pregnant women and 69% of non-pregnant women tested and believe that this is linked to the consumption of foods derived from Bt varieties, which in Canada must mean corn rather than cottonseed oil. 

Surprisingly, the authors do not consider that the origin of Cry1Ab could be food from organic farming (which sprays Cry1Ab, or bacteria producing it, on fruit or vegetable crops) or from its use in gardening (CryA1b is part of available "natural insecticide" formulations).

If we examine the possibility of a Bt corn food origin for Cry1Ab, since these proteins do not bioaccumulate, it is necessary to consider recent consumption.

First question: do 93% of pregnant women in Canada actually consume corn almost daily?

Second question: are the values in blood reported by Aris and Leblanc consistent with the levels present in Bt corn kernels? The answer is no. Here is why:

The authors reported average values of 0.19 nanograms per milliliter (ng / ml) of blood from pregnant women. Knowing that, in corn MON810 for example, levels of Cry1Ab in the grain are between 190 and 390 ng / g fresh weight, assuming that 1% will pass into the blood (which is on the high side taking into account losses during corn storage, cooking, gastric digestion and the intestinal barrier), this would require a woman of 60 kg to consume 120 g of corn (for the mean blood value of 0,19 ng / ml, assuming a plasma volume of 2.5 liters) and about 1.5 kg (for the maximum reported blood values of 2.28 ng / ml), which seems unrealistic ... And even more if one takes into account all extracellular fluids (10 liters, which would imply an average consumption of 490 g of corn and 5.8 kg in order to reach the maximum value in blood).

Third question (which follows logically the above-mentioned findings): is the Cry1Ab detection method used by Aris and Leblanc reliable?

Note first that the test used, marketed by Agdia, is claimed to detect the protein Cry1Ab from 1 ng / ml (read the introduction to this article). While Aris and Leblanc claim to have detected average concentrations lower than the detection limit, e.g. 0.04 ng / ml in umbilical cords!

One can cite the publication by Lutz et al. (J. Agric. Food Chem. 2005, 53 (5) :1453-6) showing that the ELISA test used by Aris and Leblanc is not sufficient to guarantee the identity of positive signals (« to avoid misinterpretation, samples tested positive for Cry1Ab protein by ELISA should be reassessed by another technique »).

Note that Aris and Leblanc did not discuss this issue, nor the results of Chowdhury et al. (J. Animal Sci., 2003, 81:2546-2551) which indicate that these ELISAs do not work for blood (from pigs)...

Moreover, they do not cite the publication by Paul et al. (Analytica Chimica Acta 2008, 607: 106-113) that discusses the validity of the tests available on the market...

(Provisional) answers to the questions that arise: in the absence of the validation of the detection of Cry1Ab, it is likely that the authors, incorrectly, conclude that any signal was indicative of the presence of the Cry1Ab protein, whereas they most likely correspond to false positives.

A possible validation, which surprisingly is lacking in the work of Aris and Leblanc, is the electrophoretic separation of plasma proteins and immunodetection of the protein Cry1Ab ('Western blot', a common laboratory technique).

It therefore appears that this publication, in its present state, is of unsufficient quality to be convincing. It has not undergone a proper review process according to the standards of a scientific journal, which would have required the validation of the results and their discussion in relation to available literature.