In this article, we study the influence of long-range Coulomb interaction on three-dimensional second-order topological insulator (TI) by renormalization group theory. We find that both the analysis method and conclusion in the recent Letter Phys. Rev. Lett. 127, 176001 (2021) are unreliable. There are two problems in this Letter. Firstly, the characteristic described by the RG flows m → ∞ and D → 0 can not be used as the criterion for transition from second-order TI to TI, since this characteristic could be essentially not induced by Coulomb interaction but only results from the trivial power counting contribution of fermion action. Indeed, this characteristic is satisfied even for free second-order TI. Second, the flow of B is not paid attention, which is very important and should be seriously studied. In this article, we analyze carefully the corrections for the flows of the model parameters induced by Coulomb interaction. We find that the sign of m changes but the sign of B holds if the initial Coulomb strength is large enough, while the sign of m holds but the sign of B changes if the initial Coulomb strength takes small values. These results indicate that the second-order TI is unstable to trivial band insulator not only under strong Coulomb interaction but also under weak Coulomb interaction.