The study was conducted by researchers who have close ties to anti-GMO groups, who also provided funding for it to be done.
The paper in question was published in the Journal Of Organic Systems. This is a small journal that does not appear to have an "impact factor," a way that scientists analyze the importance of journals. Additionally, the journal is funded by a pro-organic farming federation, the journal does not specialize in genetically modified organisms but in organic farming and does not appear in an important repository for peer-reviewed research.
According to the paper [PDF], 168 pigs were fed either normal soy and corn feed or genetically modified soy and corn feed for five months, the average time a pig will spend on a farm before it is slaughtered. They studied the pigs over this time and autopsied them at the end when they were killed.
The study authors concluded that the pigs that ate GM feed were more likely to have severely inflamed stomachs and the female pigs fed GM feed had bigger uteri, and called for more long-term studies on GM foods.
Critics of the paper say that the study is flawed because the statistics used were abnormal. When Andrew Kniss, of the Control Freak blog performed more appropriate statistical tests, his results did not support the researchers' conclusions in the paper.
In reality, the study suggests that the GM pigs were no less healthy than those fed non-GM food when they were killed, though both groups had a high rate of illness, and almost 15% of them died before the end of the trial, possibly from bad practices.
Because the team measured so many variables, random chance meant the non-GM feed group showed abnormalities, too: 15% of non-GM feed pigs had heart abnormalities, while only 6% of GM feed pigs did. Similar data was presented in the liver problems category.
If the GMO feed really did cause stomach inflammation, there should have also been more cases of "moderate" inflammation and fewer cases of "nil inflammation," but in reality, the GMO fed group actually had twice as many pigs without stomach inflammation, says Mark Hoofnagle, of Denialism Blog.
Even the interpretation of what "inflammation" of the stomach looks like is subjective -- the vet performing the autopsy has to decide which group to put a given stomach in by how it looks. If two vets did the autopsies on the two groups, they could have made different judgment calls on the severity of the inflammation.
The "enlarged uterus" finding, while statistically significant, is not likely to be clinically relevant.
It could be related to the feed's soybean content, because soybeans are known to impact hormone activity in animals. The difference in uterine weight could come if on feed contained more soybeans than the other. There is also a link between swollen uteri and a chemical called Zearalenone, for which the researchers did not test the feed.
Analysis of the animals's feed also flagged warnings for some. David Tribe of GMO Pundit noted that the feed was moldy. While technically within allowed limits, levels of the fungal toxin fumonisin were at levels in the GM feed that when chronically fed to pigs could be toxic, and were more than double the levels in the non-GM feed.
Though the authors claim no conflicts of interest, the funding for the study itself was provided by Verity Farms, owned by one of the study's authors, which sells non-GMO grains. Funding also came from the Australian non-profit, Institute of Health and Environmental Research.