I looked at the referenced study. This a meta-analysis and the number of studies contributing to the findings in each section appears to be rather low (2-8 stu

I looked at the referenced study. This a meta-analysis and the number of studies contributing to the findings in each section appears to be rather low (2-8 studies in the sections I looked at), although the sample sizes of some of the studies are large. The main thing that jumps out at me is that in most cases the moderate drinker comparison groups were, as they say, "at the lower end of moderate alcohol consumption". Their example (in the section on all-cause mortality): "five of the eight studies compared an average of about 0.7 U.S. drinks/day (8.4 g/d) with never consuming alcohol". So this is comparing the difference between teetotalers and people who drank very little. Although definitions of "moderate drinking" can allow up to 2 drinks a day, they had no findings comparing the effects of abstinence vs. consumption at the higher end of that range. The meta-analysis is pretty formulaic and a lot of text in the main chapters is boilerplate. This isn't a bad thing as it can indicate a systematic approach not tempted by fishing expeditions, but it does mean the ultimate findings are only a tiny portion of the size of the document. To my eye most of the effects appear rather moderate. This is an area I've been reading up on lately and it's rather fascinating to me that there seems to still be so much uncertainty about this question, given its obvious relevance to public health. I wonder if we'll ever see a durable conclusion on this matter.