I see your point about the mathematical numbers of patient in different subgroups assuming equal distribution - too small to have an impact, won't bother to poke with a "but what if?" scenario :>)
I think this one from your list is the biggie we're after, because it's the Holy Grail - preventing risk factors from progressing to established CVD:
P.S. Is anyone besides me feeling pretty giddy right now over FDA finally awarding us NCE!?!