That lancet paper always seemed to me a prime example of “statistically significant but not actually significant”. They massaged their data to just barely exclude 1 drink a day from their 95% confidence interval and still end up with an estimate of 1.01 relative all cause mortality at most from 1 drink a day. Realistically if you drink as much as the average American (around 500 drinks a year) the health effects will be extremely minimal unless you have some other health condition.