大腸直腸癌病患營養篩檢工具之比較

Hospitalized patients usually suffer from malnutrition, especially those of the departments of gastrointestinal surgery and oncology, with prevalences which are as high as 50%~60%. Thus, nutritional screening assessments of admitted patients should be performed as early as possible. However, most screening tools were developed based on the dietary habits and body constitution of Western Caucasian people, which might not be appropriate for East Asian populations. The aim of study was to investigate power differences among the following nutritional screening and assessments tools, including the Patient-Generated-Subjective Global Assessment (PG-SGA), Mini Nutritional Assessment-Long Form Taiwan revision (MNA-LF-T1), Malnutrition Universal Screening Tool Taiwan revision (MUST-T1), and Nutritional Risk Screening (NRS-2002). This study included 200 colorectal cancer patients, and their information on several aspects was collected after admission. We found that the respective kappa coefficients of agreement of the MUST-T1, MNA-LF-T1, and NRS-2002 were 0.545, 0.545, and 0.613, when using the PG-SGA as the reference method (p＜0.001 for all tools). All tools showed statistically significant relationships with biochemical data, anthropometric measurements, and length of hospital stay. In addition, when comparing the PG-SGA with other tools using the receiver operating characteristic (ROC) curve, the NRS-2002 was the best (0.908) in terms of the area under the ROC curve. The results of this study showed that the NRS-2002 had a better screening capability for malnutrition risk for colorectal patients, compared to the other tools on the basis of the PG-SGA.