In a recent paper, Klews and Schweizer [Phys. Rev. A 64, 053403 (2001)]presented a computational method for the three-dimensional propagation of wave packets in the hydrogen atom driven by a train of short unidirectional electric-field pulses. They argued that their computational scheme is valid for an arbitrary value of the pulse strength. We show, however, that such a scheme leads to both mathematical and physical inconsistences when strong pulses are considered, and, consequently, its validity is limited to the case of weak fields. The generalization to the case of an arbitrary pulse strength is also commented.