The regression line for predicting blood alcohol content from number of beers is y = 0.018x - 0.013.

The slope 0.018 of this line tells us that

the correlation between number of beers and BAC is 0.018
on the average, BAC increases by 0.018 for each (1) additional beer a student drinks
a student who drinks no beer will still have a BAC of 0.018
the average BAC of all the students in the study was 0.018