Abstract
We point out that the effective U due to vibrational couplings is a factor of 2 larger than calculated by Lannoo et al., due to off-diagonal interactions neglected in their work.