NHS advice is don't drink at all (as per Government guidelines), but that if you are going to drink, stick to 1-2 units a week. They just recognise that people will have the odd drink, so it's best that they know that you should stick to 1-2 units a week.
DBennett - single studies aren't a good guide, especially when looking at abstracts.
In that first study, for example, there is no information given in the abstract about whether the drinkers were age & background matched with non-drinkers before comparing - important given that they do note that, compared with nondrinkers, drinkers were more likely to be smokers, 35 years of age or older, black, and multiparous (smoking and age being 2 definite risk factors and ethnicity (& possibly # of children) possibly being linked to lower income familes).
Also, in that first study, they found that there was a definite increase in risk with 5+ drinks a week. The small increase in risk with 1-2 drinks a week may not have been significant if a larger sample was used, or if better matching of cases was used.
The second abstract you linked to was a very small study and no information is given in the abstract about any other factors considered, e.g. whether the women who said yes to some alcohol during pregnancy were also more likely to be smokers.
Do you have any links to better quality studies or reviews on this subject?