There was some interesting discussion over at Discover magazine:

http://blogs.discovermagazine.com/80beats/2010/01/15/gm-corn...

They mainly channel Karl Haro von Mogel, who notes on his blog "the authors have a little history of making mistakes with statistics with this same data set" and points to further information:

http://www.biofortified.org/2010/01/organ-failure-organ-dama...