I'm thinking ... If meat-eating was a strong predictor for colon cancer, rates wouldn't vary so much geographically, unless Americans' meat consumption varies geographically.
It's not that meat-eating has no effect (some other factor(s) could be weakening or strengthening effect of meat), but there's clearly more to the story. What's going on along that strip from Maine to Louisiana? And why were the coastal states spared?
* The CDC source link also shows mortality rates.