Background: This series contains brief reports on each of the "600 studies" held on the GENERA database, published at Bifortified.org, in the order in which they were listed on 18 Sept 2013. The general purpose is to counter the extraordinary claims that the entries on this list demonstrate the human food safety of GM crops. See the Introduction to this series [link].
Great Big List of Studies: Entry 15 – Appenzeller 2008 [pdf link]
This is a ~93 day GM feeding study on Sprague Dawley rats, from age ~7-8 weeks to ~21 weeks. It was included in a dossier for the regulatory approval of DuPont Pioneers' herbicide tolerant GM soybean, "DP-356043-5" at FSANZ [link] and approved in 2010. The study was conducted by authors at Pioneer and Dupont, who under the "Conflict of interest statement" made the astonishing declaration:
"The authors declare that there are no conflicts of interest."
A study conducted in-house, by the owners of a commercial product, to advocate for the regulatory approval of their product, must by its nature expect questions about commercial bias and merit very close observation. This will be further explored.
This GM soybean contained two novel genes and other supporting genetic material to enable it to resist various types of herbicides. One of the plots of soybeans used in the feeding trial had been sprayed with three herbicides – a glyphosate from Syngenta and two herbicides from DuPont, chlorimuron and thifensulfuron.
Unsurprisingly for a study being used in a regulatory application of a GM crop the company authors concluded that "no biologically-relevant, adverse effects were observed in the rats" over a fairly broad range of measured variables. They reported a range of numbers that would support this statement, subject of course to questions of commercial bias. However their subsequent statement that study results indicate "356043 soybeans are as safe and nutritious as conventional non-GM soybeans" is a leap that cannot be supported by any numbers from the study.
The trial had been on healthy young rats. Judging by the weight charts the rats were still in a growth phase by the trial end. The GM material trialled had been the "dehulled/defatted toasted meals and toasted ground hulls" of soybeans within lab rat diets. The oil had been extracted from the beans; the meal and hull residues had been further processed and toasted. Such processing and actions outside the protocols can alter the GM components within the feeds. In the human experience soy can be consumed in a wide variety of styles after differing processing, by the full range of human ages, from soy received in-utero and in infant formula, through to soy derived artificial meats in age care homes. There was an essential merit to Séralini extending the GM feeding trials on Sprague Dawley rats through to the older ages where diet related chronic health impacts begin to be felt.
Further, in respect of the word 'nutritious' the claim can be immediately refuted if the diets were adjusted prior to the start of the trial to make them nutritionally equivalent as well as adequate, as indicated in the earlier companion chicken study [link – McNaughton et al 2007] and as is a frequence practice in GM feeding studies. No specific declaration was made in this study, but it was indicated. The argument for pre-trial nutritional adjustment goes that one should be testing the unintended effects of the genetic modification, not differences that may arise from nutritional variation in the feed source. However this argument fails where nutritional variation is a consequence of the GM event, or a consequence of using pesticides enabled by the GM event. In this particular case it has been established that glyphosate is a metal chelator. It can bind to micronutrients and limit their availability. If all other conditions were equal I would expect seed derived from a trial sprayed with glyphosate to show a reduced level of some micronutrients. I don't know what the other herbicides could do, nor the other impacts of the GM event. While the authors said that the mineral levels were tested and found to 'similar', they didn't declare there were no significant differences, nor did they provide the results, viz "data not shown".
Further GM feeding studies often contain full advice about the total feed components of the diet. But this was not provided in this study, and so it is impossible to determine whether there were ingedients in the diet that could have masked the effect of the trial. The only statements made about the composition of the diets concerned the inclusion of 20% w/w meal and 1.5% w/w hulls from the GM soy crop, processed as described above, and that the diets were nutritionally comparable to
- "PMI— Nutrition International, LLC Certified Rodent LabDiet— 5002 by Purina TestDiet (Richmond, IN)".
These two ingredients thus accounted for 21.5% of the diet - what else was included?
Certified Rodent Diet 5002 [link] contains the following ingredients: Ground corn, dehulled soybean meal, ground wheat, fish meal, wheat middlings, brewers dried yeast, can molasses, wheat germ, dried beet pulp, dehydrated alfalfa meal, ground oats, soybean oil, dried whey, ground soybean hulls, calcium carbonate, casein, salt, and many other nutrients.
If the major ingredient had been corn, is there a possibility that it could have been a GM corn? It's something I've noticed in many other feeding studies – corn or soy will be used as a major feed component and no statement will be made on whether it was a GM variety or not. I wonder about the masking effects. Is it possible a GM corn could have been the major component of all the diets? Further, could DuPont Pioneer's GM corn "DP-Ø9814Ø-6 (Event 98140)" be a major ingredient in all the diets? This corn is reported on the CERA website to contain two GM genes of the same general type as those in this GM soy. I wonder if the ELISA antibodies would be cross-reactive? The presence of the GM proteins were not tested for in the diets, both of the rats and of the chickens, which I found to be unusual. Indeed the authors had reported that the GM proteins were not detectable in the processed meal and hulls. I found this to be unusual, thinking the meals must have been subject to considerable heat treatment to denature proteins beyond the point of detection. GM proteins are often found in complex processed baked food matrices. Using a simple strip test I can still detect GM proteins in a dried GM canola plant that is some years old. Perhaps these GM proteins are more labile, and the antibodies used to detect them resultingly ineffective?. Of course I'm not saying that a GM corn was used as the major component of the diet, but it's something I always wonder about. A less than full reporting of an in-house study leaves these questions unanswered. This study reported that it complied with "Good Laboratory Practice" requirements. These are requirements that were put in place to try to prevent or limit lab fraud, though I don't know these requirements are specific in these aspects. It would be good to read the actual declarations which would be included in the dossiers. In the past FSANZ has approved GM crops despite the self-reported failure of GM companies to comply with Good Laboratory Practice on these types of feeding studies. I would like to read further into the comparable sub-chronic feeding study of the GM corn on rats, but it is behind a paywall at the moment.
I've actually gone into this study at some depth although only what two or three days can provide; I would like to trace and compare all of the companion studies, including the dossiers held at FSANZ. But for the purpose of describing studies I've declared an intention to be brief, although there is far more that could be said.
Perhaps in summary, the best such a study design could do is add one piece of specific information in support of a claim of equivalence, after the specific case of feeding toasted meals and hulls to young Sprague Dawley rats had been independently repeated in the scientific method.