Can you use 1/4 of the total model, apply symmetry and finally run the FEA.
Maybe you tried this already, but how about using local mesh refinement for the smaller items and a relatively coarse grid for the overall model? That may get the model down to a reasonable size.
To answer your question, no I don't think there is a way to get a surface deformation plot result and then transfer that to the other bodies. It would be cool, though. Wouldn't it? Just too complicated for Simulation.
To join in with the chorus of replies, I can't imagine a reason why this can't be done in one simulation study. You can use H-adaptive meshing to automatically refine the mesh in the small areas for you. Maybe set the # of trials to 5 or so to make sure it gets small enough.
Also, in future, please fully describe what issues you ran into so that others can accurately give you advice. Finally, I also recommend symmetry. If 1/4 cannot be done, then at least use 1/2 symmetry. This might even allow your fully detailed model to run.