I think the reported intended to differentiate "drinking in moderation" from "not smoking" because the drinking results are actually ambiguous at low doses. The

I think the reported intended to differentiate "drinking in moderation" from "not smoking" because the drinking results are actually ambiguous at low doses. There are a lot of possible confounders, but a few drinks a week actually correlate with better outcomes. As far these findings, and all others I've seen, smoking is strictly negative. But jumping from "never" to "1-14 per day" does seem like a pretty massive gap, and one that makes these results harder to use for a lot of people. I've known plenty of social smokers, as well as people who almost quit smoking but didn't kick a final 1/day habit. It certainly seems useful for a 10/day smoker to know what their risk profile would be like at 1/day or lower, which isn't available here. (I wonder if it's a data limitation here, though? Most of the social smokers are I know are 20-somethings who smoke when they're drinking with friends, or something similar. These cohort studies were on middle aged doctors and nurses, so there may not be many people falling between "nonsmoker" and "addicted"?)