The relationship between alcohol use and cognitive function appears to be nonlinear, and indeed non-monotonic: light drinkers have better cognitive performance than nondrinkers. Reduction in cognitive performance for heavy drinkers is measured more in men than in women.
Source: Rodgers et al (2005), "Non-linear relationships between cognitive function and alcohol consumption in young, middle-aged and older adults: the PATH Through Life Project" — http://www.ncbi.nlm.nih.gov/pubmed/16128717
Chronic alcoholics do not have reduced numbers of neocortical neurons, but do have reductions in white matter volume.
Source: Jensen and Pakkenberg, "Do alcoholics drink their neurons away?" — http://www.sciencedirect.com/science/article/pii/014067369392185V
Neither of these studies speaks about the specific measurement you're asking for, IQ, but they do address the general topic.
(Chronic alcoholism is also associated with specific neurological conditions such as Wernicke-Korsakoff syndrome, which is caused by thiamine deficiency — someone who's getting most of their calories from booze is not getting enough nutrition.)
People usually abstain for reasons that might affect cognitive performance like depression or previous substance abuse for example.
Reduction in cognitive performance for heavy drinkers is measured more in men than in women.
They note that:
After adjustment for education and race, male hazardous/harmful drinkers no longer performed significantly less well than light drinkers, whereas male and female abstainers and occasional drinkers still did so.
-
...Chronic alcoholism is also associated with specific neurological conditions such as Wernicke-Korsakoff
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.