As suggested by previous results, we show that the conventional analysis of the results of C-13 NMR line shifts used for metals do not apply in GICs : an orbital contribution to the shift must be taken into account, To evaluate quantitatively this contribution, we apply a recently developed formalism for orbital shifts in crystals to the case of first and second stage GICs (for B // c), in the independent subsystem approximation. It is shown chat the sign of the orbital contribution is different for donor and acceptor compounds but its magnitude is almost independent of the charge transfer amplitude. For second stages, the calculation predicts a splitting of the lines though dipolar contribution does not These results explain the experimental data