When a German court ordered Monsanto to make public a controversial 90-day rat study on June 20, 2005, the data upheld claims by prominent scientists who said that animals fed the GM corn developed extensive health effects in the blood, kidneys and liver and that humans eating the corn might be at risk.
The 1,139 page research paper on Monsanto’s “Mon 863” variety also revealed that European regulators accepted the company’s assurances that their corn is safe, in spite of the unscientific and contradictory rationale that was used to dismiss significant problems. In addition, the study is so full of flaws and omissions, critics say it wouldn’t qualify for publication in most journals and yet it is the primary document used to evaluate the health impacts.
Mon 863 is genetically engineered to produce its own pesticide, a toxin called Bacillus thuringiensis or Bt, designed to attack a corn pest called the corn rootworm. Rats fed Mon 863 developed several reactions, including those typically found with allergies (increased basophils), in response to infections, toxins and various diseases including cancer (increased lymphocytes and white blood cells), and in the presence of anemia (decreased reticulocyte count) and blood pressure problems (decreased kidney weights). There were also increased blood sugar levels, kidney inflammation, liver and kidney lesions, and other changes.
According to top research biologist Arpad Pusztai, who was commissioned by the German government to evaluate the study in 2004, based on the evidence no one can say that Mon 863 will cause cancer or allergies or anything specific. The results are preliminary and must be followed-up to rule these out. He warns, however, “It is almost impossible to imagine that major lesions in important organs. . . . or changes in blood parameters. . . . that occurred in GM maize-fed rats, is incidental and due to simple biological variability.”
French Professor Gilles-Eric Seralini, a molecular endocrinologist at the University of Caen, agrees that the results indicate a toxic reaction. Seralini is a member of two French government commissions that evaluate GM food, one of which originally rejected a request for approval of the corn variety in October, 2003 due to the adverse findings of the study. Seralini won a French lawsuit allowing him to express his concerns in public, and now Greenpeace has won a German court battle that makes public the data that is the source of his concerns.
Pusztai and Seralini spoke about the Mon 863 study at a June 22 press conference in Berlin organized by Greenpeace. Both scientists are uniquely qualified to evaluate the study.
Seralini studies endocrine disruptors and the impact of pesticides on health. He was one of four experts appointed to respond to the WTO challenge filed by the US against the European Union’s policy on GM food and crops. He has read all of the industry’s GM-food submissions to Europe as well as all the commentaries on the submissions. Pusztai is the leading authority in his field of protein science (lectins) and had been commissioned by the UK government in the 1990s to develop the ideal testing protocol for all GM foods.
Although his protocol was supposed to be adopted by the UK government and eventually in Europe, Pusztai’s controversial finding that GM potatoes damaged the health of rats ultimately stopped the work. Pusztai has also been commissioned to evaluate all published studies on GM foods, and has analyzed most of the confidential submissions made by industry.
Both scientists have expressed alarm about the unsupported arguments that Monsanto and some European regulators use to force product approvals. Now that the Mon 863 study is available, other scientists and the public can evaluate the industry’s defense, which Pusztai and Seralini say contradict well established scientific principles. Chief among their concerns are the ways Monsanto explains away statistically significant effects.
Faulty Comparisons Hide Problems
In animal feeding studies, researchers attempt to minimize differences between the test animals and the control groups, so that only the impact of the item being analyzed will stand out.
In this study therefore, the test rats ate Mon 863 and the control group ate non-GM corn from the same parent line, i.e., corn whose genetics are the same except for the insertion of the genetic material and its impact. When comparing the results of these two appropriate groups, the health impacts were unambiguous and occurred at a rate that the scientific community accepts as not due to chance. But Monsanto and their supporters in the European Food Safety Authority (EFSA) appear to throw away the accepted methods of science that have been used for decades in order to rationalize the findings.
1. Researchers used six additional control groups, which were fed commercial corn varieties with entirely different genetics. While such comparisons are appropriate for commercial studies, it is entirely inappropriate for a safety assessment, according to Pusztai. Monsanto claimed that when the changes in the test rats were compared to this much larger, irrelevant control group, many changes were no longer significant.
2. In spite of the strained logic, many results were still statistically significant when compared to these six other controls and were reported as such by the laboratory that Monsanto used to conduct the study. Monsanto therefore ignored the study’s figures and claimed that since the changes in the rats were still within a wide range of reactions that are normal for the animals, they should be considered biologically irrelevant.
Using this argument, for example, they declared that a 52% decrease in reticulocytes (immature blood cells) was “attributable to normal biological variability.” According to Pusztai, an allowance of 5% variability is the norm in food experiments. Similarly, he says that the increase in blood sugar levels by 10% “cannot be written off as biologically insignificant, given the epidemic of diabetes.”
To put Monsanto’s claims into perspective, suppose that a large number of women who were fed a carefully controlled diet had a 25% increase in breast cancer compared to matched controls on another diet.
Using Monsanto’s logic, the findings can be dismissed because the increase was still within the normal variability of breast cancer for the whole population.
3. In spite of the statistical slight-of-hand, several results could still not be dismissed since they were well beyond the range Monsanto had defined as normal. So the company claimed that the potentially dangerous health effects were not considered significant because the reaction among the rats was not consistent between males and females. “This is really ridiculous,” says Seralini, because everyone studying cancer and endocrinology, for example, knows that there are differences between genders.
4. When even the gender defense could not be applied to a particular finding, Monsanto dismissed it since the reactions were not always dose specific. Specifically, the results observed in rats fed a diet that was 11% Mon 863 were sometimes more pronounced than results found in rats fed a 33% diet. Seralini notes that in endocrinology and toxicology research, differences are not always proportional to effects noted. A small dose of a hormone, for example, can cause a woman to ovulate, while a larger dose can make her infertile.
5. When all other excuses failed, Monsanto claimed that with such a large study, one would expect lots of results to fall in the statistically significant category purely by chance. Thus, no follow-up is required.
Seralini says, “It is dishonest not to do the tests again if you have statistical significance.” Pusztai similarly asks, “What is the point of doing a study if you dismiss the results you find?” He insists that you design a study specifically so that statistical significance indicates biological significance.
In spite of the fact that Monsanto’s explanations were at odds with time-honored principles of science, the European Food Standards Agency (EFSA) recommended that Mon 863 be approved.
In fact, the agency’s justification mimics that of Monsanto, point for point. In spite of EFSA’s recommendation to approve Mon 863, the majority of the countries in the EU Council of Ministers voted not to approve the corn on July 24, 2005. But EU law requires a “qualified majority” on such a vote, and so the pro-GM European Commission is now authorized to make the decision and is expected to approve Mon 863 within a few months.
Mon 863 will not be the first approved GM food in Europe to have shown significant health effects in rats. According to Seralini, an oilseed rape (GT 73), Roundup Ready corn (NK 603), and two Bt corn varieties (Bt11 and Mon 810) all showed statistically significant problems that regulators did not pursue with follow-up research. Seralini said that the effects of the GM crops were similar to that of pesticides. Some included inflammation disorders and problems in the livers and kidneys, the two major organs involved with detoxification. Seralini is part of a research group raising money to do independent research on a GM variety he says showed more than 50 significant rat anomalies.
GM Food is Prone to Unpredicted Effects
How can a GM crop create so many significant unpredicted side effects? There are several ways. The process of gene insertion, for example, typically results in hundreds or thousands of mutations throughout the genome. Insertion also changes the amount of protein that natural genes produce (5% of the genes in one study) and can destroy natural genes altogether.
The protein created by the inserted gene may also create allergies or toxins. Several studies indicate, for example, that the Bt toxin may cause allergic or immune system effects. Furthermore, according to Monsanto’s submission on Mon 863 to Australia and New Zealand, some of the foreign genetic material that was added into the corn was mutated during the insertion process. This means that the composition of the Bt protein that the corn creates is actually different than the one scientists intended.
With so many ways to create side effects, many scientists and consumer groups are demanding extensive evaluations and insist that a simple 90-day rat experiment is not competent to protect the public.
In the EU, pesticide approvals require research on three types of mammals, with feeding studies ranging from 90 days to two years. Seralini points out that Bt crops create new pesticides. Mon 863, for example, is unique; it differs from the natural version of Bt toxin in seven ways and should, according to Seralini, require at least the same level of evaluation as chemical pesticides. The same holds true for herbicide tolerant crops, which are engineered to survive large applications of weed killers such as Monsanto’s Roundup.
Seralini points out that these GM plants have far more herbicide residues in the edible portions and extensive toxicity tests must be performed. But the biotech industry claims that they could not afford to introduce GM crops if they had to pay for the tests normally required for pesticides in Europe. For GM crop approvals in the US, they spend even less. US authorities require only 30-day studies for the Bt plants and no safety tests whatsoever are required for herbicide tolerant varieties.
Flaws in the Mon 863 Study Should Have Caused It to be Rejected
According to Pusztai, the quality of Monsanto’s study was well below that normally required for a peer reviewed publication. He says, “It is odd, therefore, that it remains the central document considered by government regulatory authorities upon which to make a decision to protect the health of European citizens.”
Several features of the study appear to have been rigged to avoid finding problems.
Nutritional studies, for example, typically use young, fast-growing animals, which are sensitive to toxic and nutritional effects. By using a mix of young and old animals, Monsanto’s research design may have hidden serious problems.
Similarly, they used rats with a huge range of starting weights. According to Pusztai, the starting weights in a rat feeding study should not vary more than 2% from the average. By contrast, the male starting weights in Monsanto’s study ranged from 198.4 to 259.8 grams (or 143 to 186 grams according to the conflicting data in the study’s appendix). In either case, says Pusztai, the wide range “can make it impossible to find significant differences in animal weights at the end of the experiment.”
Monsanto tested the effects of two diets: in one Mon 863 constituted 33% of the rats’ diet, and in the other, it was 11%. Even in the 33% group, GM corn protein comprised only about 15% of the rats’ total protein. According to Pusztai, researchers should have started with the maximum amount of corn possible (while maintaining a balanced diet), and then used lower concentrations to evaluate any dose effect. (Since rats are stand-ins for humans, it is interesting to note that African aid recipients typically rely on corn for 90% of their total caloric intake.) Researchers also supplemented the corn with a commercial animal feed. Although its composition wasn’t reported, it may have contained GM soy, which could have skewed the results.
The study relied on analytical methods that are half a century old and ignored powerful new methods, such as profiling techniques, DNA chips, proteomics, and others.
They relied on just two observation times (week 5 and week 14), which will not give data about the intervening periods. And the short 90-day time period will miss chronic and reproductive problems, as well as problems in the next generation.
The analysis of the findings was obscured by using six irrelevant control groups fed commercial diets, as well as data from historical databases. Such comparisons are totally unacceptable in the field of nutrition. According to Pusztai, “The study should have included a control group fed the non-GM parent line, spiked with the Bt” toxin obtained from the Mon 863.
If GM-fed rats reacted worse than those fed this control diet, it would show that the genetic engineering process and its unpredicted side effects, and not the Bt toxin, were responsible. Pusztai says, “A second parental line spiked with a known toxin would also be useful as a positive control,” to make sure the measurements are sensitive enough to detect the expected impact of the toxin. Without this, it is difficult to know if the methods were working properly.
Monsanto also defended changes in kidney weights by comparing the values with a separate study, which used different corn genetics and a different lab. According to Pusztai, this absurd inter-experimental comparison is never done and should be disregarded. Some of the reported weight measurements were also bizarre, suggesting possible problems with animal management or faulty data.
One rat dropped 53 grams in one week and gained 102 grams in the next. Some that were heaviest at the beginning of the experiment were the lightest at the end. And the rats hardly grew at all during the last four weeks.
Overall, the research paper was confusing, conflicting, and poorly reported. It failed to disclose, for example, the methods used to measure changes in the animals and it did not provide sufficient chemical analysis to demonstrate that the nutritional composition of the feed remained stable for the duration of the 90-day experiment.
Since these most basic requirements for a nutritional study were not provided, the research cannot be repeated and the results remain suspect. Referring to the study as a whole, Pusztai says, “Nutritional scientists and leading journals would not accept these blatant inadequacies and misinterpretations.”
The Politics of Science Fails to Protect
When Seralini wanted to voice his concerns about the industry’s safety studies, he was told by French authorities that he was legally bound to keep even his opinions confidential. A lawsuit eventually granted him the right to speak, but until June 20, 2005, biotech companies were able to keep their feeding studies hidden by claiming that they contained confidential business information.
Seralini says that “No one can understand, even among EU regulators, why the composition of the blood of rats that have eaten the GM is secret.” The precedent established by the German court may open the door for more biotech studies to be made public. Without disclosure, says Seralini, just a few toxicologists can make the decision without public evaluation. And too often, the decision-making body is heavily influenced by the applying company.
Critics demand that regulators use independent studies, not industry studies, to prevent manipulation of data. But there are only a few independently funded researchers. Biology professor Bela Darvas of Hungary’s Debrecen University is one of them. After discovering that one of Monsanto’s Bt corn varieties, Mon 810, is lethal to two Hungarian protected species and one insect classified as a rare, he ran into an unexpected obstacle. Now Monsanto refuses to give him any more Mon 810 corn to use in his tests. They also refused his request for Mon 863.