Study questioned world-wide - Here's why
About the only thing that can really be said about this French study is "RATS GET TUMORS AND DIE".
The combination of biased, inept and unqualified (one researcher is a homeopath!) researchers along with Mike Adams and Joe Mercola loudly touting this study makes it the epitome of ideologically motivated pseudoscience, unsupported conclusions and mindless sensationalist promotion of gross misinformation.
Both the scientific research community and the scientific press has taken particular notice of this study - NOT because of the controversial focus or conclusions but because there are enough methodological and analytical flaws in the study itself to make it absolutely laughable. Certainly no future research by this group will be taken seriously. It could be a prime example for "How not to perform an animal study" in a graduate school course. Professor David Spiegelhalter, Winton Professor of the Public Understanding Of Risk, University of Cambridge, said:
"In my opinion, the methods, stats and reporting of results are all well below the standard I would expect in a rigorous study – to be honest I am surprised it was accepted for publication."
Just some of the more grievous elements of the study:
- 70% of females died early - Really? Read the causes of death in the paper... The age of death with the females was manually manipulated! Almost all the females were selectively euthanized while only a minority of the males were. Any euthanized female that would have been left to die naturally, like most of the males, could have reached the age where they would not have been counted as "death by GMO".
- Lack of blinding. With any well planned and executed animal research project like this, the research staff should NOT know which group is receiving which diet. Knowing these details lends itself to observational bias like selecting particular rats for euthanization early and thus skewing the death rates.
- No dose response evaluated. This would have been a critical measure for demonstrating the hypothesized toxic effects of different levels of GMO corn and Roundup fed to the rodents.
- Control group is way too small. A significant asymmetry exists with 180 test rats and only 20 controls. With 180 rats being subjected to various GMO diets and exposures to Roundup herbicide, there should have been at least 180 control rats instead of only 20. A larger number of controls would have been a much more valid comparison.
- Tumor rate in control group is suspiciously low. The particular rodents used in the study, the albino Sprague-Dawley variety from Harlan Labs, have a documented very high frequency for spontaneous tumor development. Using this particular rat variety with such a high natural incidence of tumor growth makes it difficult to detect a significant increase in the number tumors when the background rate is already so high. This is well known among researchers as a 1979 study reported 72% tumor incidence over time for females and 86% for males, regardless of diet. This alone should have precluded them from use in a study like this. The very small number of controls used introduced a wide variability of tumor genesis where as a larger control group would regress towards the mean and likely mirror the 1979 study.
- Large number of groups created with such small numbers of subjects in each group makes meaningful analysis next to impossible. Because of differences between males and females, there were really 2 control groups and 18 experimental groups, all with only 10 rats each, a very small sampling. Observed variations within small groups can be HUGE. The researchers further confused the study by attempting to measure a very large number of parameters within each SMALL group at multiple time windows. Attempting to analyze this many variables introduces a virtual guarantee that results "positive" to a biased researcher's liking will certainly be extracted at some point. The paper states: “All data cannot be shown in one report and the most relevant are described here” which makes one wonder exactly what they left out. Having too much data for publication is not unusual but most researchers make the extra data available at an on line location. What are they hiding? Was the data "cherry picked"? Transparency in science is a critical element of any valid research.
- Voodoo statistical analysis. The statisticians involved did not utilize standard analytical mythologies for a study of this type. Instead of using "Analysis of Variation" or even a Cox Analysis, they used multiple regression models. Perhaps analysis of variation failed to show them the results they wanted? This hit or miss method of applying multiple non-standard statistical tests until you get the result you like spawns false positives and can manifest biased conclusions - Maybe that was the objective?
- No controls on amount of food consumed or testing for fungal contaminants in food. Both will increase tumors in this particular variety of rat.
- What did the control group rats eat? Harlan, the company who raised and provided the rats for this study, was contacted by Tim Worstall, who blogs at Forbes.com. Harlan staff told him that they do not exclude rat feeds that contain GMO corn. So quite likely, the control rats did not have a totally non-GMO diet so they are not an appropriate control. If they had already been exposed to GMO feed, why were there so few tumors in the control group?
- Polycarbonate plastic cages used to house rats. Polycarbonate plastic leaches PBA which is oncogenic in rodents, a confounding factor not taken into account.
- The use of Roundup is puzzling. It's not DDT! Roundup is about of the most innocuous herbicide that can be purchased - that's why it's still sold while so many others have been removed from the market due to toxicity. The surfactants in Roundup are more toxic than the active glyphosphate ingredient! The same surfactants are found in many household products used daily - even ecologically friendly products.
Mark my words, I don't expect that any other independent researcher using appropriate study methodologies will ever be able replicate Séralini's work. Replication is the gold standard in scientific research. Come back when you fully understand the workings of a valid animal research project and have some quality research to quote!
BTW, the very same journal that published Séralini's paper recently also published a review of the long-term health impact of gmo plant diets by another French group. This was a review of 24 other GMO related animal feeding studies and they found no effect at all! Why has this review been ignored by Mike Adams and Joe Mercola?
"Assessment of the health impact of GM plant diets in long-term and multigenerational animal feeding trials: A literature review"
Disclaimer - I don't work for Monsanto and I don't know anybody who does!