Not at all. The most likely explanation to me is the causality is reversed. Poor health or limited social life causes both abstinence and mortality.

Nothing in the abstract suggests they did any analysis to determine causality which makes their statement at the end "showing that survival varies as a function of one’s own and one’s partner’s drinking" misleading, as it could easily be interpreted to imply causality. Of course they phrased it that way on purpose because implying causality makes their results sound way more important than they are.