Do those stats account for different populations of users? I imagine there are far more people using alcohol and tobacco.

No, because it's completely true. Repeated use of either tobacco or alcohol cause definite harm - cancer for tobacco (note, not nicotine...), brain and liver damage for alcohol. Heroin doesn't. Long term opioid use can cause hyperalgesia for _some_ people, but by no means all, and that's really if you take them every day at high doses (and heroin is actually not so bad for this - it differs from opioid to opioid), which is a little weird since morphine is quite the opposite - I suspect this is because the active metabolite responsible for the primary effects (heroin is really a prodrug) is morphine-6-glucoronide (M6G), not actually morphine as many would expect.