I was browsing pubmed and came accross this study, which was conduced in California in the mid 90s. 40,000 adult diabetics were asked about their alcohol consumption, and their A1cs were drawn. The A1c results were sorted by self reported drinking habits. Guess who had the highest A1cs? The ones who didn't drink! Maybe the hypoglycemia pulls down A1c? Of course, the averages for all of the groups' A1c scores were over 8% (and below 9%), and the differences between the groups weren't that large.
I though that this was odd enough to deserve a blog post.
P.S. Still reading through pubmed, here's another study claiming that drinking at least weekly has a positive effect on diabetes outcomes. This one makes an even stronger argument, because instead of comparing A1c to drinking rates, it compares complication rates to drinking rates.