PURPOSE: This study aimed to evaluate the agreement of Seoul National University Hospital-Nutrition Screening Index (SNUH-NSI) and Nutrition Risk Screening-2002 (NRS-2002) with patient generated-Subjective Global Assessment (PG-SGA) and the association between nutrition risk determined by these screening tools and operative morbidity after a gastrectomy for gastric cancer. METHODS: This study enrolled 174 patients who had undergone a gastrectomy for gastric cancer at Seoul National University Hospital from March to July 2009. We assessed a nutrition risk by two nutrition screening tools (SNUH-NSI, NRS-2002) and a nutrition assessment tool (PG-SGA) at hospital admission. We collected general patient information, serum albumin level, cholesterol amount, total lymphocyte count, hemoglobin, and body mass index, operative method, hospital stay, and operative morbidity. RESULTS: The mean age was 59.1+/-11.6 years, and 8.6% (n=15) of patients were assessed as having severe malnutrition by the PG-SGA. Agreement between the PG-SGA, SNUH-NSI (kappa=0.498, P<0.001), and NRS-2002 (kappa=0.439, P<0.001) was moderate. Patients with a high risk of malnutrition by PG-SGA, SNUH-NSI, or those with advanced gastric cancer showed more operative morbidity (P<0.05). There were no relationships between a high risk of malnutrition by NRS-2002 and operative morbidity. On multivariate analysis, malnutrition by PG-SGA (OR 2.159, 95% CI 0.693~6.721) or SNUH-NSI (OR 2.630, 95% CI 0.906~7.638) had a tendency to show higher operative morbidity, but it was not a significant independent risk factor. CONCLUSION: Both SNUH-NSI and NRS-2002 had moderate agreement with PG-SGA. Severe malnutrition risk as assessed by SNUH-NSI had an association with operative morbidity as PG-SGA did. SNUH-NSI was expected to be a valuable and efficient screening tool to detect malnutrition risk as much as PG-SGA.