If you apply perturbation theory without making a multipole expansion first, there are 2 main complications:
– What you consider to be H0 will be different, since you cannot include the monopole term of the interaction term in H0 anymore. This means that the calculation of the eigenstates and eigenvalues E0 will be different and can benefit less from the well-known electron problem.
– The calculation of the correction to the energies will also run into difficulties, because of the complexity of the full interaction term H1 for a nucleus of general shape.