I'm guessing you're looking at a sheet/plate with various loadings. each will have it's own principal stress, and so it's own direction for maximum fatigue damage, yes? (fatigue damage being related to principal stress).
A simple, conservative, approach is to take the principal stress from each loading and determine fatigue from this. This is conservative as it is taking the worst (most damaging) stress from each loadcase, but these won't typically align.
So if you wanted to be super-precise then you'd work in global directions and use Mohr's circle to determine the normal stress, producing fatigue damage, for each load case. Then figure out the direction that maximises fatigue damage.
Fatigue and crack growth are treated the same.
another day in paradise, or is paradise one day closer ?