That statement/summary was published after another large scale study was published in the same journal. IIRC that Lancet study is also kind of complex in that the harms are different with age. So before 40 years or so, the risk of alcohol is due to things like car accidents, falls, fights, etc, and after age 40 it's things like cancer etc.
I think they even see the protective effect of light drinking on things like cardio outcomes in younger individuals, it's just outweighed by accidents etc.
Take what I'm saying with a grain of salt though because this is based on recent memory of from reading it before.
But if so, I think it changes the nature of the message a bit — although it's literally true there is no safe level of alcohol consumption in a population mortality risk sense, what that means is sort of complex and age-dependent, and might be something you can counteract, or might not.