Grouping longer-term abstainers and recent quitters together as the reference group inflated the observed protective effect of alcohol, the authors note.