BACKGROUND: Assessment of patient factors is essential for selecting later-line chemotherapy in patients with metastatic colorectal cancer (mCRC). The efficacy, prognosis, and safety of each treatment regimen according to nutritional and inflammatory status still remain to be elucidated. PATIENTS AND METHODS: A total of 550 patients with mCRC who were registered in the REGOTAS study (Regorafenib versus TAS-102 as Salvage-line in patients with colorectal cancer refractory to standard chemotherapies: a multicenter observational study, UMIN 000020416) and treated with trifluridine/tipiracil (TFTD) or regorafenib as a later-line therapy were retrospectively stratified according to the modified Glasgow Prognostic Score (mGPS), which divided patients into mGPS 0 to 2 by serum albumin and C-reactive protein, and compared. RESULTS: The median overall survival (OS) of patients with mGPS 0, 1, and 2 was 10.0 months (95% confidence interval [CI], 9.2-11.6 months), 6.5 months (95% CI, 5.3-7.1 months), and 3.9 months (95% CI, 3.3-4.9 months), respectively. The median progression-free survival (PFS) with mGPS 0, 1, and 2 was 2.5 months (95% CI, 2.1-3.0 months), 2.0 months (95% CI, 1.9-2.3 months), and 1.7 months (95% CI, 1.4-1.9 months), respectively. There were significant differences by mGPS in both OS and PFS (all P < .001). No significant differences in OS and PFS were observed between the patient groups treated with TFTD and regorafenib in each mGPS group. In patients aged ≥ 65 years with mGPS 2, the OS and PFS were worse with regorafenib than with TFTD (OS: hazard ratio, 1.45; 95% CI, 0.93-2.25; P = .097; PFS: hazard ratio, 1.57, 95% CI, 1.01-2.44; P = .047), but there were no consistent trends observed as mGPS increased. The frequency of grade 3 and more adverse events was generally similar in each mGPS group. The multivariate analyses showed that mGPS was the strongest predictive factor for OS. CONCLUSIONS: The mGPS before later-line chemotherapy is strongly correlated with survival in patients with mCRC.