Look at the error bars in Figs 4 and 5 a-d - they're so large they overlap in some cases! 4a is also odd, drug group starts out with EPA levels 5 points below control group for low dose case. I wouldn't want to have depend on this tiny study for anything important.