On a reformulation of the theory of Lifshitz-van der Waals interactions in multilayered systems