GM corn causes cancer in rats: a study in bad science

Last week a scientific paper appeared that reported that eating genetically modified (GM) corn causes cancer in rats.  Specifically, the scientists fed Roundup Ready® corn, or maize, to rats for two years, and reported that both females and males developed cancer and died at higher rates than controls.  

This is very surprising.  If GM corn causes cancer, why aren't Americans "dropping like flies," as one scientist asked?  We've been eating Monsanto's Roundup Ready® corn for over a decade, even if most of us aren't aware of it.  But our rates of cancer haven't increased more than Europeans, who eat far less GM corn.  Maybe the effect is limited to rats - in which case we should also have seen dramatic increases in cancer in lab rats.  But we haven't seen that either.

So what's wrong?  The best way to find out is to read the paper, which I did.  It turns out to be a very badly designed study, and the report itself omits many crucial details that may (and probably do) completely invalidate the findings.  The scientists leading the study have a strongly biases agenda and a conflict of interest, which they failed to reveal. I'll explain below, but meanwhile this study has already been taken up by politicians as proof (proof!) that GMO crops are harmful.  As Forbes blogger Tim Worstall explained, this paper is more politics than science.

Let's look at the study itself, which was led by Gilles-Eric Seralini (more on him below) and published last week in Food and Chemical Toxicology.  (A copy of the full paper is here.)

The authors studied 200 rats for 2 years, dividing them into 20 group of 10 rats each.  The test rats were fed a variety of diets:

1. Non-GM corn comprising 33% of the diet (this was the control group).
2. Roundup Ready corn comprising 11%, 22%, or 33% of the food.
3. Roundup Ready corn that had been treated with Roundup during cultivation.
4. Non-GM corn but with Roundup itself added to the rats' water.

So what happened?  Well, in some groups, the rats got more cancer than controls.  But not always.  In fact, the authors had to cherry-pick their own data to support their conclusions.

One major problem is that only 10% of the rats were controls - 10 male, 10 female. The study's main claim is that rates of cancer were significantly higher in the rats fed GM corn.  Martina Newell-McGloughlin from UC Davis, in an interview with Discovery News, said 

"The type of statistical analysis they used is really a type of fishing expedition.  One individual referred to it as 'fantasy statistics.' "

Another major problem is that there's no dosage effect.  In other words, if Seralini is right and GM food is bad for you, then more of it should be worse.  But the study's results actually contradict this hypothesis: rats fed the highest levels of GM corn lived longer than rats fed the lowest level.

A third problem, as Discovery News and other sources reported, is that the rats used in this study are a special laboratory strain that is highly prone to cancer.

Perhaps most damning, though, is the fact that rats fed Roundup directly had the longest survival times.  As Seralini's own Figure 1 shows, the longest-living rats in the entire study, out of all the conditions, where those that drank Roundup in their water.  These rats outlived the control rats.

Yum!  Maybe Perrier should start selling Roundup-enhanced spring water?

Seralini and colleagues struggle to explain the internal contradictions in their study.  They write, 

"Interestingly, in the groups of animals fed with the NK603 [Roundup Ready corn] without R[oundup] application, similar effects with respect to enhanced tumor incidence and mortality rates were observed."  

This tortured English is their way of admitting that rats did worse ("similar effects") when fed GM corn that was grown without Roundup.  They don't want to admit that this result contradicts their central hypothesis.

The study was designed to fail: the sample sizes (10 rats in each group) are so small that all the results are likely just due to chance, and none of the differences are meaningful.  It's exceedingly unlikely that the Roundup in the rats' water made them live longer, just as it's unlikely that Roundup Ready corn had any effect on the incidence of cancer.

I know that ad hominem attacks aren't valid, but I can't resist pointing out that Seralini's co-author, Joel de Vendomois, is a homeopath, with a "Homeopathy and Acupuncture Diploma", a double dose of quackery in a single degree.  Seralini has also published a book about the supposed dangers of GMOs, and he and de Vendomois are the lead scientists at CRIIGEN, an organization devoted to lobbying against GMOs. Of course, even if Seralini and de Vendomois are bad scientists, and even if they have a strong bias, their paper isn't necessarily wrong.  It's wrong simply because the science is wrong.

Not surprisingly, an anti-GMO group in California has gleefully embraced the claims of this dreadful paper to argue in favor of Proposition 37, a ballot initiative that will require labelling of genetically modified foods.  And Jose Bove of the European Parliament has used it to claim that all GM crops are harmful to human health.

Let's be clear about the science here.  Genetic modification of foods is a powerful technology that can be incredibly beneficial.  The recent development of salmon that can grow faster is an example: these salmon (developed by a company called AquaBounty) will make fish farming more efficient, and thereby help preserve the perilously endangered wild fish species in our oceans.  On the other hand, GM technology can be used, as Monsanto has done, simply to allow farmers to use more pesticides, which doesn't seem to benefit anyone other than the pesticide producers.  It's unfortunate that Monsanto's behavior has been used as an excuse to give all GMOs a bad name.

Now we have a bad study done by anti-GMO scientists who have allowed their political agenda to trump their scientific judgment.  What a mess.