I find that study hard to believe, as I know people who never drank, who had Alzheimer’s. I know a diabetes study said it was a risk for dementia, however, the meds taken were suspect more than disease. Do you have link for that alcohol study?
Statistical inference applied to groups does not preclude individual exceptions to statistical probability. I’m sure you do know teetotalers who had Alzheimers but what does that have to do with alcohol abuse correlating highly with dementia? Nothing, really.