Are GMOs safe to eat? Current, inadequate requirements for feeding studies and what happens when you exceed them

摘要

Most government food regulators only require a test of "substantial equivalence" between a genetically modified (GM) crop and a similar non-GM crop to determine safety. They do not require tests on animals or people to determine toxic, allergic or reproductive effects, or risk of cancer. Even less is needed for crops containing several "stacked" genes if all the genes in the stack have previously been individually approved for use in the same kind of plant. Animal feeding studies submitted are generallyanimal production studies, where farmed animals that are not physiologically comparable to humans are fed GM feed, and end-points that are mostly not relevant to human health are measured, e.g., breast meat yield. Long-term toxicology studies on animalsrelevant to human health are uncommon. Animals are usually fed for an insufficient time for most adverse effects to develop, and the number of animals per group is generally insufficient for clinically significant findings to reach statistical significance. Furthermore, often only body weights and death rates are measured, organs are rarely inspected internally, blood biochemistry results are rarely mentioned, and adverse findings are not investigated further. As a result of anecdotal reports of reproductive and gastro-intestinal problems in pigs fed GM crops in the USA, we conducted a long-term GM crop feeding study on pigs in the USA which exceeded these standards. We randomised and fed pigs (n=168), which are physiologically similar to humans, either a mixed GM soy and GM corn (maize) diet or an equivalent non-GM diet for 22.7 weeks. The GM corn contained stacked genes. GM-fed pigs had uteri that were 25% heavier than those of non-GM-fed pigs (p=0.025) and a higher rate of severe stomach inflammation than non-GM-fed pigs (2.6 times more likely, p=0.004). The severe stomach inflammation was worse for males (4.0 times more likely, p=0.041) than females (2.2 times more likely, p=0.034).