The US death rate is about 2.5 million/year. Seems like somebody would have noticed a 20% jump due to a single drug. Can this 500,000 deaths/year number really be close to correct?
Edit: duh, not per year. Oh well, there may still be some little nugget of validity in there somewhere.