Shock! horror! unbalanced diets of the same stuff week after week can be bad for rats. Monotonous chow should be banned

Don't feed your daughter only soya beans every meal for 104 weeks Mrs Worthington.

Part of a long, long conversation about Seralini's rats at The Conversation. Go there for more back-story if you need it.

David,

I don’t accept anything without question. Seralini’s work, together with numerable other studies that question the safety of GM, DEMANDS further investigation. That is the ONLY reasonable scientific position to take and one that hundreds of scientists have taken. If that’s a bias towards proper scientific inquiry then yes my position is biased...

---

Lisa,
I completely agree that CRIIGEN report demands further investigation and we that we should welcome the appearance of further reports that fill any unanswered questions raised by CRIIGEN.

Fortunately, one such study has already appeared. It is Avis de l’Anses Saisine n° « 2012-SA-0227 » Maisons-Alfort, le 19 octobre 2012 AVIS  de l’Agence nationale de sécurité sanitaire de l’alimentation, de l’environnement et du travail relatif à l'analyse de l’étude de Séralini et al. (2012) “Long term toxicity of a ROUNDUP herbicide and a ROUNDUP-tolerant genetically modified maize”.

Among its findings it describes the results of the statistical tests that should have been supplied in the CRIIGEN publication (a big unanswered question).

I won't belabour ANSES findings: they are quite straight-forward- - none of the main claims of Seralini et al survive the standard appropriate stats tests.

This is not a surprise. The stats test results have been circulating in the scientific community for weeks. I did them myself to check what I read. I have not published them because they were the findings of others, and I was merely confirming other's work, but it is good that ANSES has now put them out to the public. (My translation of main statistical issues addressed by ANSES findings is here for those intimidated by French.)

But this new ANSES report provides some other fresh scientific findings:

• They point out that the previous 2008 long-term rat feeding study by a Japanese government institute found that feeding a rat high amounts of soyabean, compared to a diet with maize (GM or non GM) may be deleterious in terms of tumour incidence. [Dr Tribe suggests that] This finding may well relate to the content of phytoestrogen (plant female hormone mimics) naturally present in soy.

This interesting finding shows that certain conventional diets at the extremes can be harmful to health of rodents.  Are we to now to ban eating of all soyabeans till we know better the long term consequences of eating all soya products?


Inserted update from a reader on Google Plus

Who would have thunk unbalanced diets would harm you?
Hmmm. But then.... there's rabbit starvation! 

Rabbit starvation From Wikipedia, the free encyclopedia
Rabbit starvation, also referred to as protein poisoning or mal de caribou, is a form of acute malnutrition caused by excess consumption of any lean meat (e.g., rabbit) coupled with a lack of other sources of nutrients usually in combination with other stressors, such as severe cold or dry environment. Symptoms include diarrhea, headache, fatigue, low blood pressure and heart rate, and a vague discomfort and hunger that can only be satisfied by consumption of fat or carbohydrates....

Conversation quote continued:
Translation of relevant sections of Avis de l’Anses Saisine n° « 2012-SA-0227 » 
About Sakimoto et al 2008 (previous long term GM study)

3.3.1.2.2 Results
At the end of the study, the authors did not identify any difference related to GM soy on survival, weight rats or food consumption. Some hematological parameters (hemoglobin, hematocrit, mean corpuscular hemoglobin concentration) were significantly decreased in the GM group compared to non-GM group, but these changes are all less than 4% and have no biological significance. The only biochemical changes (transaminases, creatinine) and organ weights affect soybean treatments versus standard diet. The only significant differences in the appearance of neoplastic tumors are also observed in animals fed a diet soy versus standard diet. In conclusion, this study shows that GM soybeans administered for 2 years to F344 rats does not lead to significant changes compared to non-GM soy. In contrast, a diet containing 30% soy could pose potentially deleterious effects.

• The ANSES report reveals another section of errors in Seralini et al 2012 . CRIIGEN estimate glyphosate [herbicide] exposure risks that are quite unrealistic. This also is not a surprise -- discussion of these mistakes have been circulating in the scientific community prior to ANSES reporting them too.

Relevant Section translation

3.3.2.3 The study Séralini et al.[2012 CRIIGEN study]...
Given the purpose of this study referral has been analyzed in detail.
3.3.2.3.1 Protocol ...
The choice of doses
Regarding the administration of doses of GMOs in the diet, 11%, 22% and 33%, the experimental protocol is complete and adopted standard. These doses (11 and 33%) correspond to the doses normally used in regulatory studies subchronic 90-day rats.
However, the choice of levels of GT PLUS ROUNDUP administered in the drinking water of rats raises two questions. The first level of variation among the three tested doses ranging from 50 ng / L to 2.5 g / L is a multiplicative factor of the order of about fifty million. The gap between high-dose and low dose is too large to conclude that any dose effect relationship.
The second question concerns the significance of these three doses in terms of exposure or duconsommateur user glyphosate formulations.
Three concentration levels tested in this study were compared with exposure data available:
The first dose level corresponds to a level that the author refers to contamination of tap glyphosate: 50 ng / L. Regulatory standard in France is 100 ng / L in water up to drinking water. For water distributed over 43,741 analyzes looking glyphosate (period 2007-2009) only 95 (0.2%) can find measurable levels of glyphosate. These quantifiable results are found in a limited number of distribution stations (between 0.2 and 0.4% of the 21,864 stations tested). The concentration of 50 ng / L is a realistic value can potentially be observed but on a very limited number of French resorts. Moreover, these analytical data relate only to glyphosate and glyphosate association non-co-formulants. Co-expressing the ROUNDUP GT PLUS is not mobile in soil (Koc = 2500-9600, DT50 soil = 1-2 days). The probability to find the quantities tested in groundwater is negligible.
The second dose level tested is that the author describes contamination "equivalent" to the U.S. MRL for glyphosate in genetically modified feed for animal feed (400 mg / kg). At European level (source: website of DG Sanco), the MRL for glyphosate is 0.1 mg / kg for sweet corn (as fruiting vegetables) and 1 mg / kg for maize as cereal grain. Therefore, the level at which the European consumer is exposed is much lower than that tested in this protocol. In addition, as for drinking water, the co-formulating is not systemic, the consumer will be predominantly exposed to glyphosate and not mixing glyphosate + co-formulating.
The third dose level tested is that the author calls "half the prescribed dose applied glyphosate" (2.25 g / L). Taking into account the concentration of glyphosate in the preparation (ROUNDUP GT PLUS), the dose rate per hectare (data based e-phy of the Ministry of Agriculture) and the dilution recommended by the manufacturer, glyphosate concentration of "slurry" would be applied to about 7 g / L, the same order of magnitude as the dose tested in the publication. However, the user will be exposed primarily dermal and inhalation potentially preparing GT PLUS ROUNDUP diluted. Route of administration (oral) described in the study protocol is not the most suitable for risk assessment when applying the product.

____

Finally, interviews with Prof Seralini have been published where he has been questioned about his failure to do statistics (the big unanswered question on the CRIIGEN report).

A translation of key sections of the interview is as follows (about 8 minutes into the recording , in French):

QUOTE of G-E Seralini captured in video
Statistics are not the truth in biology, they are a linking element with respect to the rest, that is, rats, pathology, biochemistry, organs; and what we do is to compare for instance 12 with 4 and say: « it's three times more. »
You cannot compare whole numbers statistically, and here we have two whole numbers of tumours in the course of time.
And there is no statistical model allowing us to follow the rat mortality or tumour curve in those circumstances, so we cannot invent a statistical model from which to derive hypothesises such as the fact that it is linear or at first sight not linear in order to find something.
We therefore avoided those mathematical assumptions to, on the contrary, correlate all the arguments we had in order to affirm what we had.

These remarkable statements suggest that Professor Seralini is completely unaware of -- or more worryingly deliberately ignores-- standard medical science teaching about statistics and its application to the experiments he reports ( eg see here on how to do such tests ). I take them to show that anything he says about rats should be considered very cautiously and sceptically.

Update.
And, the Pundit adds upon further reflection, that M. Seralini should also watch Simon Singh on You-tube.