If you define alcoholics as the set of people who, for whatever reason* , tend to drink to excess, that is, who as a rule do not drink moderately, then these health benefits are irrelevant to them. For example, if they try to drink moderately, but only succeed half the time and otherwise seriously overdo it, then the net effect on health is likely to be very negative.

* The debate whether alcoholism is a disease, a personality trait, or something else entirely is irrelevant here.

The health benefits of alcohol are an unusual case, also -- many of its benefits are related to other chemicals that frequently accompany it, such as polyphenols in red wine. This does not apply to "LSD/Cocaine/Heroine" (sic), though there are comparable benefits for coffee and tea (particularly green tea), mostly antioxidants.

Many doctors are reluctant to recommend moderate drinking for its health benefits because it is also addictive and impairs judgment, and in cases where people drink excessively, the health costs are severe. This doesn't apply to most things recommended for health (exercise, eating fruits and vegetables, getting enough sleep, nethack, green tea, pomegranate juice (or whatever the health craze of the moment is, ac,ai berries? peach fuzz?)), because most of those are less likely to be taken to excess, and would probably be several orders of magnitude less harmful even if they were.

"I would think the real damage of teetotalism would be the stress involved in constantly thinking about and addressing it."

Then you would be wrong: http://query.nytimes.com/gst/fullpage.html?res=9E05EED9143FF...

This article shows benefits from moderate drinking for individuals at high risk of heart disease, not the population at large:

For many people moderate drinking has absolutely no benefit at all, Dr. Klatsky said. This applies to all adolescents: teenagers have a negligible risk of heart disease and for them the risks of heavy drinking vastly outweigh any benefits of moderate drinking.

The same generally applies to men under 40 and women under 50, except for those with known heart disease risks. Pregnant women, people with liver disease, known drinking problems or a family history of alcoholism should never be advised to begin to drink for their health, he said.

But for others with known coronary heart disease or a risk of heart disease because of obesity, cholesterol levels, or smoking, moderate alcohol use may be a way to reduce that risk.

It is important to look at context when reading about these sorts of studies.

The large population studies this article cites aren't exactly convincing. Anyone can perform similar studies showing other correlations. The "feeding studies" are more important to scientific understanding. Also, I would think an individuals personal health profile would outweigh blanket advice like "drink moderately".