Of GM corn and rat tumors: Why peer reviewed doesn't mean "accurate"

UPDATE: After you read this story, make sure you check out the follow up piece. Editors at Embargo Watch have found evidence that The Sustainable Food Trust manipulated the media to prevent public criticism of this paper.

Yesterday, in an aside to a post criticizing an astroturf political campaign in California, Mark mentioned a new study that supposedly found GM corn causes tumors in rats. As Mark said in an update to that post, this study is severely flawed, but I wanted to follow-up on that with some discussion about why it's flawed.

After all, the study was peer-reviewed, right? Doesn't that mean we can trust it?

Here's the thing. Peer review is not perfect. It's not a panacea. It's simply the basic level of due diligence. By submitting work for peer review, a scientist has allowed people outside her own team to critique her work. And the journal might require some changes to the paper based on the critique — anything from edits for clarity to requesting that the scientist perform another experiment in a different way. If a paper hasn't gone through peer review, you should be more skeptical of it. Avoiding peer review means that the researcher decided to show the public her results before allowing those results to be critiqued by independent experts.

But, at the same time, just because something has gone through peer review doesn't mean it's been certified to be accurate. It just means that roughly three other experts have looked at the paper before publication. There's still a lot of room for things to go wrong. Peer review is like the bouncer at the door. The bouncer doesn't guarantee that every person in the bar would be a good person for you to date. Even if a paper gets through, you still have to think about it critically and evaluate it on its own merits. This recent paper on GM corn and rat tumors is an excellent example of that ...

Over at Discovery News, Emily Sohn has a great breakdown of everything that's wrong with the GM corn and rat tumors study. And there's a lot that's wrong with it. In fact, the laboratory that did this research — a French team led by Gilles-Eric Séralini — has been heavily criticized for the poor quality of their research into GM food on multiple occasions.

It's not Monsanto spokespeople doing the criticizing, either, Sohn writes.

One immediate problem, Newell-McGloughlin said, is that the line of rodents used in the study, known as Sprague-Dawley rats, are frequently used in cancer research because a large majority of them naturally develop tumors at a high rate, regardless of what they eat or how they're raised. What's more, the rats were allowed to eat an unlimited amount of food, which increases their chances of developing tumors. And two is a very old age for these rats, which could account for the large rate of cancer seen across all groups, including the controls.

The small size of the control group also raised red flags. Even experienced scientists in the field had trouble interpreting data in the study, as seen in comments collected by the UK's Science Media Center, but it appears that the study included just 10 or 20 control animals. That means there were at least nine times more test animals than control animals. If anything, studies of this kind usually include two or three times more controls than experimental animals.

The results don't make a lot of sense, either. No matter how much of either herbicide-laden or genetically modified maize the rats ate in proportion to their other food, rates of cancer and premature death remained the same. However, to be meaningful, toxicology studies like this should show a dose-dependent response, which means that if something is toxic, more of it should be more toxic.

Looking at the data, it appears that the study authors never tested their results to see if the numbers they turned up could have occurred by random chance, said David Tribe, a microbiologist at the University of Melbourne in Australia. And given the small numbers of animals used in the study, that's a real possibility.

Notably, the authors of the paper never responded to Sohn's request for an interview.

• Read Emily Sohn's report on the problems with the GM corn and rat tumors study at Discovery News

• Read the full study online

• Read an earlier BoingBoing post explaining how peer review works, and some of the flaws in the system. (Be sure to check out the extensive comments, which include further context from scientists.)

Image: Corn Maze, a Creative Commons Attribution (2.0) image from pinkmoose's photostream


  1. Maggie, thank you for these posts. This kind of readable, thorough science reporting is exactly what this world needs. 

  2. So is this going to turn out to be a case of follow the money? Egoboo? Or just “Science: ur doing it wrong”? What are the typical reasons for seriously flawed work?

    1. I think you kind of have to let go of the idea that flawed science always means “it’s a plot”. Researchers are people. Sometimes they simply have personal biases that lead them to not question themselves and their own work enough. Whether that’s a personal bias against GM food or a personal bias towards wanting to publish something groundbreaking or what, who knows. But it could really be as simple as that. 

      As for why flawed papers get past peer review and into journals, that’s also complicated. There’s a lot of different pressures (economic, personal, prestige) that might lead a journal to publish something that’s poorly done but will net them a lot of attention. 

    2. “What are the typical reasons for seriously flawed work?”

      Laziness? The desperation to court controversy to raise one’s credentials at their University?

    3. This might be a bit harsh, but it’s my experience that many bio/chem/med students simply don’t “get” maths/stats.  They may be otherwise very talented, but their lack of understanding in this area can mean they don’t appreciate the logic behind their analyses and are just applying textbook formulae where they seem appropriate.  All their good work is wasted.

      Ben Goldacre covered one particular stats technique used in neuroscience which was misunderstood and thereby rendered meaningless the results of large numbers of papers – http://www.guardian.co.uk/commentisfree/2011/sep/09/bad-science-research-error

      1. Completely agree. Stats is necessary, but not sufficient for good quality science. You can do good science without stats.

        1. You contradicted your second sentence with your third.  If stats are necessary but not sufficient to do good science, then you can’t do good science without them, but aren’t guaranteed to do good science by including them.  That’s how logical semantics work.  Instead you said they were necessary, but then immediately said they weren’t.  Which which sentence did you mean?

          Not to be a grammar nazi, but your sequentially contradicting statements undermine the logic of your message.

          1. Oops, yes, it was just a typo involving can instead of can’t in the third sentence. Probably I should qualify this by saying that appropriate use of stats is clearly important but often overlooked, and sometimes there just are no appropriate statistics for the situation. But most of the time studies have quantitative data which could be analyzed statistically, then the stats are necessary.

      2. Statistics and esp. probability can be extremely counter-intuitive. Even to respected mathematicians themselves. Remember the Marylin vos Savant piece on the Monty Hall problem a few years back? Famous mathematician Paul Erdös refused to believe the correct answer for some time. Along with about 1000 PhD’s.

    4. What are the typical reasons for seriously flawed work?

      Cops commit robberies.  Teachers screw their students.  Financial managers embezzle.  And in each case, some of their colleagues cover for them. The real question is why we assume that some professionals are always going to do everything ethically and competently when history has shown that some percentage has always gone off the rails.

    5.  Seriously flowed like not giving the Monsanto controlled media time to kill the story before it’s out? Label GMO foods and let it up to the customers. I live in Europe and most Europeans don’t want GMO food and yet the American government is threatening our governments to accept GMO crops. This is mafia stile and not free market. 90% of us Europeans DO NOT WANT GMO food and Americans don’t have even the right to know if their food is GMO or not. Nor is this the first study that reveals GMO’s are poisonous.

  3. Peer-reviewed or not, I still want to know if my food is a genetically modified organism. I’m not saying I won’t buy it (I ate GM corn all this year) but I am saying it’s my right to know. 

    YES ON 37. carighttoknow dot org

    1. This post isn’t about that, at all. I think it’s reasonable to want to know where your food comes from and what’s in it. I think it’s reasonable for Mark to critique astroturfing. This post is about the new study, though. 

        1. They already have labels; in this case, the label wasn’t deserved.

          I hope this incident leads to some replicating studies to amass a body of evidence one way or the other. It won’t be publicized, but it’s important information to collect.

          1. Well the label was deserved – I see no accusations of impropriety from the peer review process. I do see implied incompetence in the process but it was reviewed by their peers.

          2.  Stop messing with the narrative here.  The message is Roundup is safe and Monsanto is your friend and all research showing any other result is necessarily flawed and similar flaws in all research supporting Monsanto’s position are imaginary.  Stop thinking for yourself!

  4. I used to work for one of the largest publishers in the world and the way peer reviewers were chosen wasn’t particularly scientific.

    Usually reviewers were chosen based on their previous publications listed in PubMed, they had to have a previous publication based around the same topic within the last 6 months and be willing to respond to emails…

    Oh yes funny story we once had a author’s wife selected as a peer reviewer, because she decided to keep her maiden name.

    1. Also worth pointing out that peer reviewers are often “the people who happen to know the most about this and also happen to have the free time necessary to do peer review in addition to all of their regular work”. 

      Also, in some fields, the knowledge base is so specific that “peers” is an incredibly small group who might share personal biases towards or against one another. 

  5. It’s also become increasingly difficult for publishers to find people to do peer review. Everyone is busy. So the review that gets done may not be as thorough as everyone would like.

    1. You mean people with Jobs are busy. Couldn’t we get some unemployed people to do the review? /kidding.
      But seriously how does funding fit into this issue of peer review. People who are getting funded to do the study vs. people who are getting funded to NOT do other work that might be critical.

      BTW, Maggie was on Virtually Speaking last night talking about the whole peer review process and explaining how some people don’t know what the even means. I know the words,  but I don’t know the process and I especially don’t know the politics of funding stuff and funding the process. How do various companies avoid the process of peer review in order to go right to the press with “exciting news” that might be found wanting later?
      Do listen to Maggie on Virtually Speaking, it was really interesting, http://www.blogtalkradio.com/virtually-speaking-science/2012/09/19/maggie-koerth-baker-tom-levenson

      If I was in PR for a GM corn company my job would be to attack anything in any fashion that got in the way of selling GM corn, including casting doubt on any kind of research, even peer reviewed, that made it look bad..  And if I found research that made it seem safe I would push that info equally hard trumping the whole “peer reviewed!” process. And, from the PR guys point of view, if I found out on Friday that Peer Review was good for me I would say it’s great. If I found it was bad I would make sure everyone knew that Peer Review was not perfect.

      This is why the right wing attacks science, because sometimes it doesn’t fit their ideology. They have no problem being inconsistent from one day to the next.  This is part of Bob
      Altemeyer’s the Right Wing Authoritarian Mind set.  http://home.cc.umanitoba.ca/~altemey/

      We would hope that they would see the inconsistency from one statement to another, but Altemeyer’s work shows that they never ‘merge the files” and that they have no problem saying one minute, “I hate science and evolution isn’t real” and the next saying, “I want my anti-biotics, MRIs, iPhones and microwave ovens”

  6. I find it reassuring that you would openly critique a study that is “on your side” so to speak. I would imagine that most of Boing Boing’s authors and readers are in favor of transparency, so I think it takes integrity to critique a study that would aid in the effort to promote transparency.

    1. this++
      Maggie, I really love your objectivist, contrarian posts.It’s great to have someone challenging our preconceptions and helping buck our own cognitive dissonance.

  7. Of course, the real test is repeatability. A single study is suggestive, but hardly dispositive. Show me multiple studies by independent groups, and I’ll think you’re on to something.

    1.  Essentially, by the flawed methodology, the only real result of this study is that mice bred to rapidly grow tumors tend to rapidly grow tumors.  The control is so small that the margin of error includes the whole population of the control group.  Using a control of 10-20 individuals when they have a roughly 86% chance of developing tumors means that there’s a decent chance all of them will get tumors anyway.

      It reminds me of reading the literature that came with a prescription benzoyl peroxide gel I once had.  The study’s abstract claimed it didn’t raise cancer rates in mice.  But when I looked through the 4pt fine print on 2 8″x11.5″ sheets, I saw that the experimental group was mice with topical benzoyl peroxide and the control were mice injected with TAR!

      1.  Accusations without providing details mean nothing. 100 rats were fed with GMO corn and another 100 rats were fed with non GMO corn in the control group. The main critique has been that those members are to small. I think 100 + 100  rats is sufficient when 80% of the GMO rats get tumors.

  8.  What about writing about the bias introduced by research funding? In order to secure funding from bodies who have a vested interest, scientists are constantly at least incentivised if not obliged to formulate hypotheses and methods so as to arrive at conclusions favourable to the agendas of the funders. Whilst this study may sufffer from exactly this problem, the history of GMO research is a sorry story of exactly this problem with industry funded research. Monsanto et al prefer to spend huge amounts of money lobbying governments on how exhaustive research on long-term health outcomes are not necessary rather than on the research itself. That may be why, remarkably, this is the first ever animal study to be pursued over this time period. The question ought to be, why didn’t Monsanto want to know about this enough to sponsor two year studies (with a large enough sample) before now? Why weren’t the FDA insisting on them? Michael Taylor?
    This is why basic necessary research on cross-pollination spread, or the effects of genetic modification on gut bacteria are hopelessly inadequate after all these years.
    Without harping on about the specific issue of GMOs, it is a typical case of scientific objectivity being instrumentalised by corporate interests. In the fields where this is at it’s most perniscious, most scientist will tell you that it is a much more important factor in the deficiencies of research than the weakness of the peer review process.

  9. Often these investigators are portrayed as isolated, idealists fighting billion dollar corporations. Nowadays environmentalist organizations have fairly deep pockets, even though it`s nothing compared to behemoths like Monsanto. But they certainly have their revenue streams, that they`re busily protecting, and their agendas, that they pursue with vigor. 

    1. Isn’t that the standard global warming denialist view?   Those wealthy, greedy greens, faking science to keep the grants and donors happy!

      1. Yeah! Instead of acknowledging that people on both sides of the argument have the capacity for dishonesty, and that we need to be skeptical of any study like this until transparency is achieved, let’s accuse him of being a global warming denier!

  10. Glad to see this criticism, the study looked suspicious to me but not knowing about this kind of lab science, I couldn’t put my finger on the main problem with it. It looks to me that by far the biggest flaw is what is known (since the 1970’s!) about the increased cancer incidence in the Sprague-Dawley rat strain:
    The next biggest problem is that they have no plausible biological hypothesis for why GM+pesticide should be causing cancer. But lots of stats-based studies are like this in the ‘soft’ sciences. Data crunching or stats alone cannot answer this so one should always be suspicious of general scientific claims without plausible underlying mechanisms; so the fact that there is no null hypothesis testing is problematic, but the fulminating criticisms on this point are overblown, e.g.:
    “You never, ever design an experiment to look for specific outcomes,” … “If you’re a real scientist, based on observations you find, you put forward hypotheses about how you are getting these types of outcomes. They do none of this.” – Martina Newell-McGloughlin, a plant biotechnologist at the University of California, Davis.
    Reliance on that kind of reasoning is what drives nutty conclusions based on the dumb use of hypothesis testing and ‘p-values’ as a substitute for good observational science. I very much doubt that re-analyzing the data from this flawed study with good statistical techniques would magically transform it into a reliable study, for instance.

  11. I do not want to defend Monsanto, glyophosphates, unlabeled GMOs, or uncontrolled release of GMOs into the wild.  Nor do I want to jump on the bandwagon of condemning research that serves as a clear indicator that these things need far more study than we are giving them – if you think this study was bad, check out the studies that supposedly prove Monsanto’s products are safe.

    But a worse problem seems to be people’s willingness to overuse chemical weapons and to overlook industrial and municipal overuse.

    Pesticides and herbicides aren’t really necessary and don’t work in the long run; pest organisms evolve past them rapidly.  They primarily function as a way to reduce staff on the corporate farms.  During the World Wars, that was a reasonable goal, but nowadays it just means the corporate boardroom class makes more money and the unskilled poor have fewer jobs available.

    1. I think the probability that GM foods will cause health problems in consumers is low. Ecological devastation or a faminocalypse seem much more likely.

    1. If it was labeled, you could vote in the marketplace.  If the claims made by GMO supporters are true (“we’d all starve without GMOs!” for example) labeling would prove them correct.

      I think brussels sprouts are a bad idea, and I’m glad they are labeled so nobody can foist them off on me as baby cabbages.

      1.  Nope! Europe for the most part has no GMO foods and is doing pretty well. Why is Monsanto fighting GMO labeling of foods? If they are “so good” more people would buy GMO foods? I think not.

  12. 10 or 20 control rats??

    If each rat in this high-cancer-susceptibility strain has a 50% chance of getting cancer, then you’ve got a 6% chance of 6 or fewer of these 20 controls randomly having cancer, and a 6% chance of 14 or more of them getting cancer — either of those are likely enough and would cause you to completely swap your conclusions.

    Not to mention that, looking at the original paper, they are measuring a whole host of different things (lymphocytes, liver weight, urine phosphorus, urine potassium, white blood cell count, and tons more…), and they are randomly finding different groups have different statistically significant variations. “Oh, females who ate 33% GM corn had lighter livers. Males who ate 11% GM corn had higher white blood cell counts, etc. etc.”

    Any student of statistics knows that if you look at more and more variables, the chance of getting a statistically significant result on any random one increases. There was even an XKCD on it.

    1. Yes, it would seem wiser to have a control population at least as high as the total of all the non-control populations, which probably should have been higher as well.

  13. Oh my goodness, it appears that Martina Newell-McGloughlin, who was linked above and is critical of the motivations of the Roundup Rat Study, might have few suspect motivations herself!   She’s a loud cheerleader for misinforming the public, in fact.

    Labels are “very costly, are not going to be informative, and there’s absolutely no basis in science for this,” said Martina Newell-McGloughlin, director of life and health science research initiatives at UC Davis

    She says there’s “no basis in science” for labeling genetically modified organisms in the marketplace.  It’s pretty hard to take anything else she says seriously.

    1.  There isn’t any basis in science; they haven’t been shown to be harmful and people have been looking.  It would be like labeling cell phones with “Warning!  This device produces electromagnetic radiation!”

      And if you think that would be a good plan, then I KNOW you’re a troll.

      1. You know how I know you’re a troll?  Because you jumped to the conclusion that Ito Kagehisa is a troll simply on the basis of having an opinion you disagree with.

        Your analogy is flawed as well.  As far as I know, there are no human beings who ingest and digest cell phones.  Beyond that, to my knowledge there is no way of making a cell phone that doesn’t emit electromagnetic radiation, while it is quite easy to make an ear of corn without inserting viral DNA into its germ-line cells.

        Finally, your argument is more generally flawed: the risk of harm is not the only reason to label a product.  Organic foods have not been found to be significantly less harmful or more nutritious than factory-farmed foods but people still prefer them for a variety of reasons.  Even if you disagree with those reasons surely you must concede that the existence of organic certifications gives the consumer more choices.  Similarly, labeling GM foods, even if they’re not harmful, puts the choice in the hands of the consumer rather than the producer.  Again, you don’t have to agree that GM foods are harmful to believe that more choice for consumers is a good thing.

        If GM foods are so good for us, why do the folks making them feel the need to hide whether or not a particular food is genetically modified? Especially when it comes to food — stuff that will get digested and become parts of our own bodies — I think consumers deserve all the information they need to make their own choices.

      2. The labeling Newell-Mcloughlin opposes does not exist in some metaphorical Ayn Rand marketplace where the invisible hand writes; and, having writ, moves on to make us all taller and better looking.

        She’s opposed to labeling food to inform buyers in a real, literal, physical marketplace.  This is the most fundamental principle of the fair market – that people should know what they are buying.  It’s not some superannuated science argument in your head – it’s allowing people to choose what to ingest.  It’s preventing sellers from misrepresenting their products.

        Ever since Adam Smith we’ve known that free markets aren’t free if the customer isn’t fully informed.  If you are against accurate product labels, you stand against a free and fair marketplace and you are in favor of denying freedom of choice.  Buyers simply can’t choose freely when sellers are allowed to misrepresent products.

        The argument against fair labeling has never changed, from La Guardia’s time to the present. It goes like this: “Other people are so stupid my wise and noble team would lose money if we weren’t allowed to lie to them for their own good. My team’s income is so important that we can’t let people exercise their stupidity. So we need to be able to mislead them by hiding the truth about what we are selling”.

    2. Surely a climate scientist who has been vocal in support for a carbon emissions law can also offer an opinion about a crappy global warming paper? How is this different?

      1. Newell-Gloughlin’s actually traveled the world pushing the idea that genetically modified foodstuffs are safe and should not be subject to increased regulation.  But your point is still correct – the fact that she’s not an unbiased or reliable source of information does not mean all her criticisms are necessarily invalid, it just means you should examine her points more closely, and assume she’s putting the worst possible spin on everything.

  14. Labels are not really informative and can actually be more confusing, for example:
    Organic: Grown in a wonderful garden, more healthy, better flavor, no nasty chemicals – $ounds Great!

    Natural: Doesn’t mean a dang thing.

    Genetically Modified Organisms: Scientists mutated this plant so Monsanto can make tons of money and bankrupt the independent farmer.. Oh, and it’s a Mutant! I’ve seen movies. I don’t want my cornflakes to spawn and burst out of my chest.

    1. Yes, we have to protect people from sources of knowledge and information.  As you’ve demonstrated with your examples, education is totally bad.

  15. Scientific research is essential in investigating these questions. The motivation behind asking the questions isn’t the important part, many questions are are asked from extremely passionate or concerned people, take the examples of leaded gasoline or CFCs. It’s when that passion clouds the results is when it gets tricky, that’s why there is peer review and reproducability as the core of the scientific method.

    Don’t think this study is some profound or breakthrough hypothesis, GMOs have been around for years and studied carefully from all sides. 
    The companies creating them have tons of studies on how to actually do it, do it better do it cheaper, and yes, safer. Why on earth would a company want to harm, poison or kill their customers??

    On the other independent side, groups have been studying it all as well. From force-feeding lab animals to doing wide studies on any large scale effects and haven’t found anything significant. And then there is observable reality. Millions and Millions of cows, pigs, chickens have been fed on this stuff. Millions of people in this country and around the world have been eating this.. And I will add that if it wasn’t GMO, they would be starving.
    If there was something wrong at all, it would be completely obvious. 
    Instead, what can be seen is millions being fed, and living healthy longer better lives.

    1. I don’t agree with you, and think the evidence is pretty squarely against you, but I have to respect your ability to sum up one side of the argument concisely.  Well done.

    2. If there was something wrong at all, it would be completely obvious.

      To describe that statement as profoundly unscientific doesn’t even begin to cover it. You’ve mixed science and faith and gotten scienciness.

  16. in my mind, using tumor prone rats make sense, see if it effects tht rate of tumors.  otherwise you would be dealing in zero or 1 tumors, which would be less illustrative of the actual factor the GM product might be.  Other than that, yeah the study does seem to have a lot of holes.

  17. Hi all you well thinking and fact finding people. Sure you don’t want to jump to conclusions too fast. Wrong kind of rats? Great, the same were used in the studies that ‘proved’ the safety of this NK603 GM maize. So these were no good either? Statistics no good? Well, they seem to have followed international OECD guidelines. No dead animals? Not enough lifetime to develop these symptoms, the butcher is faster. No massive increase in cancer in US yet? Just wait and see what your human guineepig experiment will bring us. Ten years of human consumption of GM food is about 90 days in a rats life, the time when all the ‘thourough’ studies on GM food end. Wait for another 10 and you might be in for a very unpleasant surpise. In the meantime, try to complete your factfinding mission and don’t stop thinking. And have a look at this as well:

  18. On GMOs, I’ve been trying to get myself an opinion for years and I really can’t make up my mind, the benefits are quite obvious, so are the risks worth taking? It looks so, let’s hope so.

    On the study though, funny thing, it just looks so badly flawed one might even begin to, er… suspect… it was commissioned by Monsanto so it _could_ be discredited, or something :)

    1. The scientists say that the methodology used was exactly the same as that used to “prove” roundup-ready branded crops are “safe”.  Same rats, same conditions, they just ran the test for two years instead of 90 days.  I would not be so quick to discount this study, even though I agree it is flawed – you should check the links not_responsible posted above.

      I think GMOs should definitely be on the market – well labeled.  And farmers should have the right to seek damages if they can prove that GMO crops have contaminated their own crops without their consent.  That would be a science-based approach, in my opinion, whereas now we have a pseudo-science approach based on providing maximum profits for already wealthy corporations.

  19. Newell-McGloughlin  also works for the USDA and State Department. The head of the USDA is a former Monsanto director. Shouldn’t it be a little bit suspicious that she is the only one giving this study a harsh criticism?

    1. Kyle, She is not the only one. The study has some basic flaws that I would not have signed off as a reviewer. It is embarrassing for this journal’s reputation that they were not corrected in the peer review process. Sadly this is very common, but the results of most studies are not so controversial.

      Great work Maggie. Keep it up.

  20. You mentioned that the rats were allowed to consume an unlimited amount of food and that this is a flaw in the study.  We have a problem of obesity in this world, so don’t you think that over-eating of GMO food is a REAL problem vs. the control rat group?

  21. Useful article. Well done.

    I would quibble with the toxicology part of Sohn’s article though. Depending on the mechanism, increased concentration may not have a greater impact. A number of hormone response curves have a pretty sharp spike at a narrow range of concentration and low response everywhere else.


    Um. Sorry about the lack of signature at the top. I’m signed in via OpenID, but my username isn’t showing up for some reason.

  22. Both your and Discovery’s article are lacking merit.
    Better do some more reading. Here’s an accurate response to Monsanto’s non-science response.


    Response to Monsanto’s rebuttal of Seralini study (1)
    Monday, 24 September 2012 
     GMW: We’re happy to be able to alert GMW readers to what has become the commonest dodge employed by industry and regulators to fend off the inconvenient findings of independent science on risky products. In this dodge, industry and regulators claim that the study wasn’t done according to the protocols set by the Organisation for Economic Cooperation and Development (OECD) and that it wasn’t, or may not have been, conducted according to Good Laboratory Practice (GLP) rules. They conclude that they can ignore the findings of the study and the risky product can stay on the market.

    What are OECD protocols and GLP rules?

    OECD protocols were designed by industry and government representatives with the aim of creating an internationally harmonised system of tests that industry would do on its products in support of regulatory approval. The OECD protocols mean that industry only has to do one set of tests to gain approval in any OECD member country. 

    While the OECD protocols make things easier and cheaper for industry, they have been criticised by independent scientists for being outdated, inflexible, and insensitive — in other words, they are likely to miss important toxic effects. They were never claimed to represent the best science; they were a compromise that industry could afford and regulators could accept. Independent scientists have no interest in following OECD protocols because they are not necessarily the best tool to detect effects.

    GLP is a set of laboratory management rules brought in by regulators in the 1970s and 1980s to combat widespread and serious industry fraud in the testing of chemicals for regulatory purposes (two high-profile cases of which involved Monsanto and glyphosate). GLP dictates how scientists commissioned to do testing for industry must carry out, record, and archive their experiments. While history shows that GLP doesn’t prevent fraud, it leaves a paper trail so that if fraud does later come to light, investigators are more likely to be able to identify the perpetrator. 

    Again, independent scientists usually do not follow GLP rules because it’s expensive in terms of labour hours and because they view them as irrelevant. Independent science has always had its own quality control system in the form of the peer-reviewed publication system, which subjects a study to the critical review of peers, the journal editor, and, after publication, fellow scientists who can repeat the experiment and test its findings.

    OECD protocols and GLP rules are not, and were never meant to be, a hallmark of good science. Yet industry and regulators misuse these two standards to dismiss studies done by independent researchers, which generally do not conform to OECD protocols or follow GLP rules. Thus OECD protocols and GLP rules have become a shield for industry to protect it from the findings of independent scientists. 

    Thus we’ve ended up in a situation where there are, for example, hundreds of independent studies showing harm from low doses of the food packaging chemical bisphenol A, and some limited industry studies that claim to show safety. Yet industry and regulators cling to the industry studies on the purported grounds that they are more “relevant to human risk assessment” (OECD-conformant) and “reliable” (GLP). 

    In fact, there are no good scientific reasons why industry studies should be considered more relevant or reliable than independent studies, and many reasons why they are less so. But if regulators were to abandon arbitrary definitions of relevance and reliability, they would have little alternative but to ban many chemicals, pesticides, and GMOs.

    For more, see: Antoniou, M., M. Habib, et al. (2011). Roundup and birth defects: Is the public being kept in the dark?, Earth Open Source. http://bit.ly/IP2FWH (section 4)

    How do OECD protocols apply to Seralini’s study?

    Seralini based his study on the chronic toxicity part of OECD protocol no. 453. It states that for a carcinogenesis trial you need a minimum of 50 animals of each sex per test group but for a toxicity trial a minimum of 10 per sex suffices. 

    Monsanto’s earlier 90-day feeding study on NK603, submitted to the EU in support of its approval, had been re-analysed by Seralini’s team. They found it reveealed signs of liver and kidney toxicity:

    de Vendomois, J. S., F. Roullier, et al. (2009). “A comparison of the effects of three GM corn varieties on mammalian health.” Int J Biol Sci 5(7): 706-726.

    So for this new experiment, Seralini’s team chose a chronic toxicity protocol to see if the signs of liver and kidney toxicity escalated into something serious, which they clearly did.

    We MUST remember that the study embarked on by Seralini’s team was NOT a carcinogenicity study but a chronic toxicity study. That is, they did not embark on a study to see if the GM maize or Roundup caused cancer. The rise in tumour incidence was unexpected and a surprise and thus not planned for. 

    So it is disingenuous to call this work a carcinogenicity study. The experimental design that Seralini used, compared with Monsanto’s 90-day study, was more extensive (longer; greater number of tests groups; greater range of parameters measured; diets better characterised including certainty that the control diet was non-GM, which Monsanto failed to provide data on in their 90-day feeding trial data; proper control diets as stipulated by EU GMO legislation instead of irrelevant control diets as used by Monsanto). 

    It is also worth remembering that Monsanto used 20 rats of each sex per group in its feeding trials but bizarrely, they only analysed 10, the same number as Seralini. So Monsanto does not have a leg to stand on on this point! We wonder why Monsanto only analysed 10 rats out of 20. Were these randomly chosen or were they selected because they were apparently healthy? Monsanto’s data, like most such industry feeding trial data on GMOs, is not published so we cannot check this.

    Monsanto is responsible for the fact that Seralini’s group did not design their study as a carcinogenicity study. Monsanto either did not know that the GM maize could have carcinogenic effects because it had failed to test it, or it did know and hid the fact. So it would not occur to any independent scientist to test for this. 

    Given Seralini’s results, it is now up to Monsanto to pay for a full carcinogenicity study on the NK603 maize, which, however, must be carried out by independent scientists with no conflicts of interest. 

    As EFSA has repeatedly said, it is industry’s responsibility to prove that its products are safe. Clearly it has not done that. So NK603 must be withdrawn from the market until it has been proven safe. 

Comments are closed.