The preoperative prognostic nutritional index (PNI) may forecast colorectal cancer (CRC) outcomes, but the evidence is not conclusive. Here, we retrospectively analyzed a cohort of patients from the Department of Surgical Oncology at the First Hospital of China Medical University (CMU-SO). We also conducted a meta-analysis of eleven cohort studies. Bayesian Information Criterion (BIC) was used to determine the optimal PNI cut-off values for classifying prognosis in the patients from the CMU-SO. The result from CMU-SO and meta-analysis both confirmed that low PNI was significantly associated with a poor prognosis and advanced TNM stages. Among the patients from the CMU-SO, the optimal cut-off values were "41-45-58" (PNI < 41, 41 ≤ PNI < 45, 45 ≤ PNI < 58, PNI ≥ 58), which divided patients into 4 stages. The BIC value for TNM staging combined with the PNI was smaller than that of TNM staging alone (-325.76 vs. -310.80). In conclusion, low PNI was predictive of a poor prognosis and was associated with clinicopathological features in patients with CRC, and the 41-45-58 four-stage division may be suitable for determining prognosis. PNI may thus provide an additional index for use along with the current TNM staging system to determine more accurate CRC prognoses.
Keywords: TNM staging; colorectal cancer; meta-analysis; prognosis; prognostic nutritional index.