This study had several limitations. The network meta-analysis identified study heterogeneity, which might be explained by the racial differences in the study populations. The CONCUR trial [11] and the TERRA trial [6] were conducted in patients with mCRC from Asian populations, while the FRESCO trial [7] was conducted only in patients in China. Although a subgroup pooled analysis was performed to compare the efficacy of regorafenib, TAS-102, and fruquintinib in Asian patients with refractory mCRC, future global, multicenter, prospective, randomized studies may still be required. In addition, the type of AEs was not detailed in the presentation of the outcome data. Further studies should be considered to detail the comparisons of toxicities for fruquintinib, regorafenib, and TAS-102. Detailed clinicopathological data that included aspects of past medical history, comorbidities, and other drug treatments were not analyzed in the present study. These demographic and clinicopathological factors may be of interest in future studies. Finally, although there was no significant publication bias shown by the funnel plot, the existence of potential publication bias may not have been excluded.