| 摘要: | 住院病患營養不良發生率約20％~60％，尤其是腸胃外科及腫瘤科的病患甚至高達50％~60％。因此建議推薦住院病患應及早入院後作營養篩檢。但相關篩檢工具是依照西方白人的飲食習慣及體型結構發展而來，並不適用於東方亞裔族群。本研究目的為探討不同營養篩選工具，包括：長式簡易營養評估量表（Mini Nutritional Assessment-Long Form, MNA-LF）及其台灣修訂版（MNA-LF Taiwan revision, MNA-LF-T1）、萬用營養不良篩檢工具（Malnutrition Universal Screening Tool, MUST）及其台灣修訂版（MUST Taiwan revision, MUST-T1）與營養危險因子篩檢方法（Nutritional Risk Screening, NRS-2002）等五種，比較其運用於評估大腸直腸癌病患之營養不良篩選能力之差異性，以及分析哪些是預測癌症病患營養不良之重要預測因子。本研究對象為民國101年11月至102年3月於某南部大型區域教學醫院收案的200名大腸直腸癌病患。受試者於入院後進行多項營養篩檢問卷收集，包含：病患主觀性整體評估表（Patient-Generated-Subjective Global Assessment, PG-SGA）、MNA-LF 、MNA-LF-T1、MUST、MUST-T1、及NRS-2002，並且收集體位測量與血液生化值的資料。研究結果發現：PG-SGA判定有55.5%的個案有營養不良或具營養風險；MUST及MUST-T1則分別判定37.5%及34.0%；MNA-LF、MNA-T1-LF則為45.8%及41.8%；NRS則為44.2%。當以PG-SGA為營養篩選結果之參考值時，MNA-LF及其T1版與PG-SGA評估結果之一致性，分別為0.52及0.545；MUST及其T1版分為0.552及0.527；NRS-2002則為0.613（以上皆p<0.001）。各工具與白蛋白、前白蛋白、血紅素、白血球、急性反應蛋白、血清總蛋白以及身體質量指數、體重流失率、臂中圍、小腿圍的一致性測量結果良好且具有顯著意義，惟與淋巴球總數評估結果之一致性皆不高。此外，以PG-SGA為標準計算ROC曲線面積（Receiver Operating Characteristic curve），MNA-LF及其T1版、MUST及其T1版，以及NRS-2002分別為0.810、0.799、0.696、0.693及0.908。在預測住院天數及生活品質方面，MNA-LF-T1及MUST-T1則顯著性則高於NRS-2002。綜言之，本研究發現當以PG-SGA為標準時，NRS-2002比MUST、MUST-T1及MNA-LF、MNA-LF-T1更適合用於評估大腸直腸癌病人的營養不良風險；而預測住院天數及生活品質方面，則以MUST-T1較佳。
It was widely acknowledged that approximately 20%-60% of hospitalized patients suffered from malnutrition at the time of admission, especially for gastrointestinal surgery and oncology patients whose prevalence were even as high as 50% to 60%. Therefore, it was suggested that the inpatients should be examined on their nutritional status early soon after admission by nutritional screening assessments. However, all screening tools were developed based up on the eating habits and body structure of Western white people, which might not be proper to be employed in the East Asian populations. To determine whether the adoption of Taiwanese-modified versions of nutritional screening tools with adjusted anthropometric cut-off points could improve the nutritional status grading capability of these tools in the target population, this study aimed to compare the differences among the nutritional screening tools, containing the Mini Nutritional Assessment-Long Formï¼MNA-LF) and itâs revisionï¼MNA-LF Taiwan revision, MNA-LF-T1ï¼, Malnutrition Universal Screening Toolï¼MUSTï¼and itâs revisionï¼MUST Taiwan revision, MUST-T1) , and the Nutritional Risk Screeningï¼NRS-2002), in terms of the malnutrition screening functions of patients with colorectal cancer. It also attemped to identify what were the most influential predictive factors of malnutrition among the colorectal cancer patients. A total of 200 subjests were recruited, who were the inpatients with colorectal cancer admitted to the large, regional, teaching hospital located in the Southern Taiwan during the time period of November, 2012 to March, 2013. They were examined on their nutritional status by several screening tools, containing the Patient-Generated-Subjective Global Assessmentï¼PG-SGAï¼, MNA-LF-T1, MUST-T1 and NRS-2002. Anthropometric data on percentage of weight loss, Mid-Arm Circumferenceï¼MACï¼, Calf Circumferenceï¼CCï¼and Body Mass Indexï¼BMIï¼were gathered, and biochemical data including serum albumin, prealbumin and Total Lymphocyte Countï¼TLCï¼were also collected. The results show that prevalences of malnutrition were 55.5%, 37.5%, 34.0%, 45.8%, 41.8%, and 44.2% by the measurements of PG-SGA, MUST, MUST-T1, MNA-LF, MNA-LF-T1 and NRS-2002, respectively. While use of PG-SGA as a reference to compare grading capability with other screening tools in terms of the level of agreement, the kappa coefficients for MNA-LF, MNA-LF-T1, MUST, MUST-T1, & NRS-2002 were 0.52, 0.545, 0.552, 0.527, & 0.613ï¼p < 0.001 for all toolsï¼. All tools showed a statistically significant relationships with the biochemical data and anthropometric measurements including serum albumin, prealbumin, hemoglobin, white cell count, C-reative protein, total protein, BMI, percent of weight loss, MAC and CC, except TLC. Besides, while comparing PG-SGA with other screening tools in terms of ROC, the calculations of areas under ROC curves for MUST, MUST-T1, MNA-LF, MNA-LF-T1, NRS-2002 were 0.696, 0.693, 0.810, 0.799& 0.908, respectively. However, in the perspectives of comparison with LOSï¼Length of Hospital Stayï¼and QoLï¼Quality of Lifeï¼, both MUST-T1 and MNA-LF-T1 showed better correlation with PG-SGA than NRS-2002 did. In conclusion, NRS-2002 showed a better screening capability for malnutrition risk among all tools, and MUST-T1 was the better predicting indicator in terms of LOS and QoL in this study. |
