Comment on "Dense coding in entangled states"

In a recent Brief Report, Lee et al. [L. Lee, D. Ahn, and S.W. Hwang, Phys. Rev. A 66, 024304 (2002)] have claimed that using pairwise entangled qubits gives rise to an exponentially more efficient dense coding when two parties are involved than using maximally entangled qubits shared among N parties. Here, we show that their claim is not true.