The biggest question always raised in these types of studies is whether the specific form of alcohol matters. In other words, does it matter whether someone drinks beer mostly, or wine mostly, or mixed drinks mostly. It turns out that regardless of the type of alcohol consumed, the risk of dementia is lower compared to teetotalers.
Do recall that this is an observational study, demonstrating only an association. It is not a randomized, double blind, placebo controlled study that demonstrates causation. More importantly, those who drank low-to-moderate amounts of alcohol also tended to be healthier, both physically & mentally. So who wants to volunteer for the study of causation as a follow up to this study of association?
In the meantime, it's always good to know that this study's findings are not a fluke but merely confirm similar conclusions published 4 years ago and again just 3 years ago in a systematic review of the literature .