Plus another 50,000 deaths from second hand smoke. Plus another 85,000 deaths from alcohol and alcohol-related fatalities. Plus another 7.5-10K from NSAIDS. Plus another 20K or so from illicit drugs. Plus another 106K from prescription drugs.[1] That's about 29.4% of the 2.4 million annual U.S. deaths right there.
Plus IIRC most of these categories are undercounts because they are looking mostly at the proximate causes. (Smoking- and alcohol-related deaths are the obvious exceptions.) So they aren't necessarily considering someone who takes accutane and dies of liver failure 20 years later, someone who gets AIDS or Hep C from drug use, someone who gets MRSA after getting sent to prison for drug use, etc.
And that's completely ignoring people who die because they weren't taking drugs they should have been taking, which is the other side of the same issue-- making poor decisions about drug use. Now I think we would need to be really conservative in any calculations we made here in order to for the fact that individuals can make decisions that we wouldn't personally make but that are still acceptable, but there are still a lot of blatantly obvious cases like people refusing antibiotics for staph infections and such. Plus all of the above is only counting mortality and completely ignoring morbidity, which is obviously much greater.
[1] Lazarou J, Pomeranz B, Corey P. Incidence of adverse drug reactions in hospitalized patients. JAMA. 1998;279:1200-1205.
Suh DC, Woodall BS, Shin SK, Hermes-De Santis ER. Clinical and economic impact of adverse drug reactions in hospitalized patients. Ann Pharmacother. 2000 Dec;34(12):1373-9.