Comparison of different methods for nutrition assessment in patients with tumors.

Nutrition screening to identify patients at risk of malnutrition is vital for cancer patients because of the high prevalence of malnutrition in this population. The aim of the present study was to compare different methods of nutrition assessment in patients with tumors. From June 2013 to June 2014, we conducted an observational multicenter study to compare the assessment of nutritional status in patients with tumors by anthropometry, biochemical indicators, nutritional risk screening (NRS-2002) and patient-generated subjective global assessment (PG-SGA). Mann-Whitney test and Kruskal-Wallis H non-parametric test were used for intergroup comparisons. Spearman's rank correlation coefficients were calculated to evaluate the association between different methods of nutritional assessment. The kappa statistic was used to evaluate the agreement between two assessment methods. A total of 927 oncology inpatients underwent full nutritional assessment and nutrition screening. The PG-SGA tool determined that 13.7% of patients were well-nourished (PG-SGA from 0-1) and the rest (86.3%) were malnourished. Among the malnourished patients, 57.8% were moderately malnourished (PG-SGA from 2-8) and 28.5% were severely malnourished (PG-SGA >= 9). According to NRS-2002, 30.7% of patients were at nutritional risk (NRS-2002 >= 3). There was a significant positive correlation between PG-SGA scores and NRS-2002 scores in both men and women. Compared to albumin, the PG-SGA had a sensitivity of 93.78% and specificity of 21.80%. In comparison, NRS-2002 had a low sensitivity of 43.13% and relatively higher specificity of 82.16%. In conclusion, the relationship between PG-SGA, NRS-2002 and nutritional status is statistically significant. Compared with NRS-2002, PG-SGA is a suitable screening tool for detecting the risk of malnutrition in patients with cancer.