Comment on ''A path integral leading to higher-order Lagrangians, arXiv:0708.4351 [hep-th]''

We show that the ``effective Lagrangian’’ constructed in [1] is inconsistent with the exact result for the complete Lagrangian presented in [2]. We trace the origin of the inconsistence to the peculiar way in which the path integral methods are used to eliminate variables that are not auxialiry.