CLINICAL STUDY OF MINI-NUTRITIONAL ASSESSMENT FOR OLDER CHINESE INPATIENTS

LEI/c:04 LORD_c 8/10/09 11:48 Page 1 CLINICAL STUDY OF MINI-NUTRITIONAL ASSESSMENT FOR OLDER CHINESE INPATIENTS Z. LEI 1, D. QINGYI 2, G. FENG 1, W. CHEN 3, R. SHOSHANA HOCK 4, W. CHANGLI 1 1. Tianjin Lung Cancer Center, Department of Thoracic Surgery, Tianjin Medical University Cancer Institute and Hospital, Tianjin, P.R.China; 2. The Xiangya Medical School of Central-South University, Changsha, P.R.China; 3. Tianjin Medical University, Tianjin, P.R.China; 4. Wesleyan University, CT, U.S.A. Corresponding author for proof and reprints: Wang Changli, M.D. Tianjin Lung Cancer Center, Department of Thoracic Surgery, Tianjin Medical University Cancer Institute and Hospital, Tianjin, P.R.China 300060, Tel.: +86-22-23340123, Fax: +86-22-23359984, Email: raymd728@yahoo.com.cn Abstract: Background: Malnutrition is a common problem in older patients. Early detection of malnutrition is an important task in clinical practice. The MNA has become an extensively used tool to evaluate nutritional status in European countries and the United States. Objective: We evaluated the Mini-Nutritional Assessment (MNA) test and the short-form MNA (MNA-SF) test as screening tools for malnutrition in older Chinese inpatients and focused on finding an optimal cutoff point for MNA total score and MNA-SF score. Design: One hundred eighty-four older Chinese inpatients were enrolled in this study from July to August 2006. Nutritional assessment included MNA, anthropometric measurements, and biochemical markers. Results: According to the original cutoff point of the full MNA, 19.6% of those assessed were malnourished, 53.2% were at risk of malnutrition and 27.2% were well nourished. Correlations were found between MNA, MNA-SF and body mass index, triceps skinfold thickness, serum albumin, lymphocyte count, hemoglobin, lymphocyte ratio. With the most proper cutoff point lower than 19 indicating malnutrition, when using serum albumin (<35.0 g/l) as the indicator, the sensitivity and specificity of the MNA total score were 0.6286 and 0.7466; when using BMI (<18.5 kg/m 2 ) as the indicator, the sensitivity and specificity were 0.8636 and 0.7469. The incidence rate of malnutrition was 32.6%. The most proper cutoff point of MNA-SF was lower than 12. Conclusion: The MNA and MNA-SF were useful tools to identify older Chinese inpatients with malnutrition. However, the cutoff point of the MNA should be modulated for this population. Key words: Nutritional assessment, gerontology, screening. Introduction Malnutrition is a common problem in older patients as an important predictor of morbidity and mortality. In a study conducted in Dublin hospitals, it was reported that the incidence rate of malnutrition among all hospital patients was 11%-44% and among older patients, the incidence rate was 29%-61% (1). In a study in China, the prevalence of malnutrition was 41.6% among surgical patients (2). Therefore, early detection of malnutrition is essential. Especially concerning older patients, individuals that are undernourished or at risk of malnutrition should be identified immediately upon admittance to the hospital. Several nutritional screening tools have been developed to enable doctors to identify older adults who are undernourished or at risk of malnutrition (3). The most extensively evaluated tool to date is the Mini-Nutritional Assessment (MNA), which was designed by Yves Guigoz (4) and has been widely applied in European countries and the United States. Also, there are some reports recently suggesting MNA is a valid indicator of nutritional status in Africa and Asia population (5, 6). Rubensten and Guigoz further simplified the MNA test and devised the MNA short form (MNA-SF), which includes six items extracted from the original eighteen items of the MNA on the basis of high correlation with the MNA total score. The MNA-SF is used as a primary screening tool, and the full MNA is used if further evaluation with the remaining twelve items is deemed necessary, which is called the two-step process (7, 8). In the case of China, the MNA was first introduced by Yu Kang. When it was used to evaluate the nutritional status of Chinese older surgical inpatients, the prevalence of malnutrition was close to that of European countries and the United States (2). Thus, the MNA is not only applicable to older patients in western countries, but also those patients in Asia. However, considering the differences in living style, dietetic culture and physical status between the East and the West, the original cutoff points do not accurately reflect the nutritional status of the older population in China. For this reason, the MNA has not been widely applied in China as a conventional assessment tool. This study focuses on determining the proper cutoff point of the MNA test that most precisely reflects the nutritional status of older Chinese inpatients. The study Aims The aim of this study is to evaluate the MNA and the MNA- SF tests as screening tools for malnutrition in older Chinese inpatients, focusing on finding proper cutoff points for MNA and MNA-SF. Participants One hundred and eighty-four older inpatients, comprised of 114 men and 70 women, were randomly selected from our 1

LEI/c:04 LORD_c 8/10/09 11:48 Page 2 CLINICAL STUDY OF MINI-NUTRITIONAL ASSESSMENT FOR OLDER CHINESE INPATIENTS hospital between July and August of 2006. All patients were investigated within five days following admission. All subjects were above 60 years old. Patients with malignant tumors, myelodysplastic syndromes, severe cognitive deficits (including dementia), and liver and renal diseases were excluded from the study. All participants provided written informed consent. Methods Medical doctors or nurses carried out nutritional assessments for the subjects. MNA characteristics Applying the MNA developed by Guigoz (4), the maximum score of the MNA-SF is 14. Scores of 12 and higher indicate satisfactory nutritional status. The maximum score of the full MNA is 30, with scores of 24 and above indicating satisfactory nutritional status. MNA scores between 17 and 23.5 suggest risk of malnutrition, and a total score below 17 indicates malnutrition (4). Considering the differences of population in height, weight, and dietary habits between the East and the West, and based on prior research conducted in China, the body mass index (BMI) item in the MNA was changed as follows: 18.5=0(from<19=0), 18.5-21.25=1(from 19-21=1), 21.25-24=2 (from 21-23=2), >24=3(from>23=3) (9). Since a BMI less than 20 kg/m 2 is widely accepted as underweight, particularly in well-developed countries, and 18.5 kg/m 2 is recommended as a practical lower limit for most populations, therefore, the diagnosis of malnutrition was made when BMI was less than 18.5 kg/m 2 and the diagnosis of risk of malnutrition was made when BMI was more than 18.5 kg/m 2 as well as less than 24 kg/m 2. In the dietary questionnaire section of the MNA, the cheese item, due to extremely infrequent consumption among Chinese, was replaced by soymilk and peanut milk while other items remained the unchanged. Anthropometric measurements The calculation of BMI is done by taking the weight in kilograms and dividing by height in meters squared. When patients were admitted to the hospital, their weight was measured by an RGZ-120-RT-weight scale. Their height was measured using measuring tape. Mid-arm circumference (MAC) was measured at the midway point between the acromion and the olecranon process with a tape. Calf circumference (CC) was measured by wrapping a tape around the calf at the widest part. Triceps skinfold (TSF) was measured with Harpenden calipers at the midway point between the acromion and the olecranon process. Skinfold thickness 1 cm apart from the thumb was measured by pinching the skin and subcutaneous tissue between the left thumb and other four fingers. The criteria of anthropometric measurements for nutritional status can be checked in the references, specifically (9, 10). Biochemical markers For each inpatient, albumin, prealbumin, triglyceride (TG), total cholesterol (CHOL), lymphocyte count (Lc), hemoglobin (Hb), and lymphocyte ratio (Lc%) were all provided by the clinical laboratory of our hospital. The normal values of above reference values are as follows: albumin 35-55g/L, prealbumin 200-400mg/L, TG<1.17mmol/L, CHOL 2.90-5.2mmol/L, Lc 1.3-3.3( 109/L), Hb 110-160(g/L), Lc% 20-40. According to the standardized reference values, the nutritional status of a subject can be evaluated based on these combined results. Data analysis Considering biochemical markers and anthropometric measurements in accordance with normal distribution, all results are shown in the form of mean ± standard deviation. Data evaluation was carried out with SPSS 11.5 software (SPSS Inc., USA). Analysis of variance and Kruskal-Wallis test were utilized to determine the differences between groups of the MNA (MNA total scores <17,17 23.5, and 24) depending on the distribution of the analyzed variable. Partial rank correlation coefficients adjusted for age were performed to measure the relations between MNA, MNA-SF and nutritional parameters. The Chi-square test, correlation analysis, and t test were utilized to study correlations between the MNA and other indicators. The statistical significance level was set at 0.05. To identify optimal threshold values for predicting malnutrition, receiver operating characteristic (ROC) curve analysis was adopted. The area under the ROC curve was also evaluated. A value of close to 1.0 or a larger area under the ROC curve indicates a greater reliability and discrimination of the scoring system. Results The average age of the one hundred eighty-four older inpatients was 68.0 ± 5.7 y (60-86 y) in which 114 male patients averaged 68.1 ± 6.0 y (60-86 y) while the 70 female patients averaged 67.9 ± 5.3 y (60-83 y). For all inpatients, MNA total scores averaged 20.7 ± 4.4 with a range from 9.5 to 30 and a median at 21.5. MNA-SF scores averaged 9.7 ± 2.6 and ranged from 4.0 to 14.0, with a median at 10.0. Assessment of the nutritional status with the MNA and According to the original cutoff point of the full MNA, 27.2% of older inpatients had adequate nutrition, 53.2% were at risk of malnutrition, and 19.6% were malnourished (shown in Table 1). The detection rate of malnutrition using nutritional parameters varied from 4.9% to 36.1%. Except for CC, hemoglobin, albumin, and prealbumin, there were significant differences of malnutrition detection rate between the MNA and other nutritional markers (p<0.05). 2

LEI/c:04 LORD_c 8/10/09 11:48 Page 3 Table 1 Comparison of the nutritional assessment using MNA and Normal Adequate Risk of malnutrition p value nutrition(%) malnutrition(%) (%) MNA 27.2 (50/184) 53.2 (98/184) 19.6 (36/184) BMI 18.5~24 (kg/m 2 ) 65.8 121/184 22.3 41/184 11.9 22/184 * 0.045* MAC 21(cm) 53.8 99/184 16.8 31/184 29.4 54/184 * 0.029* CC 31(cm) 64.7 119/184 35.3 65/184 0.185 Hb 110~160(g/l) 79.4 143/180 20.6 (37/180 0.813 Lc 1.3~3.3( 109/L) 67.2 121/180 32.8 59/180 * 0.004* Lc% 20~40 63.9 115/180 36.1 65/180 * <0.001* Albumin 35~55(g/L) 80.7 146/181 19.3 35/181 0.956 Prealbumin 200~400(mg/L) 75.0 123/164 25.0 41/164 0.223 CHOL 2.90~5.20 95.1 78/82 4.9 4/82 * 0.002* (mmol/l) BMI, body mass index; CC, calf circumference; CHOL, cholesterol; Hb, hemoglobin; LC, lymphocytes; Lc%, lymphocyte ratio; MAC, midarm circumference; MNA, Mini- Nutritional Assessment. * Chi-square test (p<0.05) Comparisons of with the three groups of the MNA According to the original cutoff points of the MNA, the subjects were divided into three groups: adequate nutrition (score 24), risk of malnutrition (score = 17-23.5), and malnutrition (score < 17). In Table 2, the results of anthropometric measurements and biochemical markers were compared in the three groups. Nutritional parameters declined with the deterioration of nutrition status. Only BMI and CC differed significantly for each pair of groups as well as showed statistical significance between each group (p<0.05). Table 2 Comparison of in the three groups of MNA MNA total score F P <17 17-23.5 24 N 36 98 50 Male/Female 17/19 64/34 33/17 2.524 0.112 BMI(kg/m 2 ) 20.2±3.3* 22.1±2.8* 24.2±2.4* 21.750 <0.001 TSF mm 8.8±6.4 11.1±11.7 13.9±8.4 2.755 0.066 MAC(cm) 20.7±2.8 22.2±3.1 22.7±3.0 5.125 0.007 CC(cm) 30.0±2.6* 32.3±3.2* 34.7±2.5* 26.819 <0.001 albumin (g/l) 34.7±7.4 39.7±5.6 40.9±3.7 14.422 <0.001 TC (mmol/l) 4.1±0.6 4.3±1.2 4.4±0.8 0.502 0.607 Lc( 10 9 /L) 1.3±0.5 1.6±0.7 1.9±0.9 8.006 <0.001 prealbumin 204.9±60.0 232.0±62.4 253.4±117.8 3.301 0.039 (mg/l) TG(mmol/L) 1.0±0.4 1.3±0.8 2.1±1.4 7.563 0.001 Hb(g/l) 108.2±24.4 123.4±19.2 126.7±14.9 10.807 <0.001 Lc% 19.4±10.4 25.7±11.9 28.0±10.4 6.282 0.002 BMI, body mass index; CC, calf circumference; Hb, hemoglobin; Lc, lymphocytes; MAC, midarm circumference; TG, triglycerides; TSF, triceps skinfold thickness; #Kruskal-Wallis test; * BMI and CC have significant differences between each group (p<0.05). Correlation analysis between the MNA, the MNA-SF and In Table 3, correlations were found between the MNA and BMI, TSF, MAC, albumin, prealbumin, lymphocyte count, hemoglobin and lymphocyte ratio (p<0.05). Furthermore, the MNA had better correlations with BMI and albumin (p<0.05). There were also positive correlations between the MNA-SF and other indicators (p<0.05), excluding age, MAC, total cholesterol and prealbumin. The changes of nutritional parameters were well reflected by MNA and MNA-SF scores. The receiver operating characteristic (ROC) curves (Figures are not shown)plot the sensitivity versus (1-specificity) for MNA in older Chinese inpatients using albumin (<35.0 g/l) and BMI (<18.5 kg/m 2 ) as markers of malnutrition. The area under the ROC curves, which represents the overall accuracy of the MNA total score as a test for malnutrition, was found to be 0.747 ± 0.045 (95% CI: 0.660~0.835) for albumin and 0.885 ± 0.039 (95% CI: 0.779~0.932) for BMI. Table 3 Analysis of the correlation between MNA, MNA-SF and age, No.of Correlation with Correlation with subjects MNA total score* MNA-SF score* r P r P Age 184 0.019 0.797# 0.070 0.346# BMI 184 0.472 <0.001 0.475 <0.001 TSF 184 0.202 0.006 0.223 0.002 MAC 184 0.232 0.002 0.093 0.211 Albumin 181 0.347 <0.001 0.294 <0.001 CHOL 82 0.158 0.155 0.113 0.314 Lc 180 0.274 <0.001 0.210 0.005 Prealbumin 164 0.175 0.025 0.143 0.069 Hb 180 0.302 <0.001 0.202 0.006 Lc% 180 0.254 0.001 0.171 0.022 MNA total score 184 0.838 <0.001 BMI, body mass index; CHOL, cholesterol; Hb, hemoglobin; Lc, lymphocytes; MAC, midarm circumference; MNA, Mini-Nutritional Assessment; TSF, triceps skinfold; * Partial rank correlation coefficients adjusted for age were used to measure the association between MNA or MNA-SF score and except for age; # Spearman s rank correlation. Validity values of the MNA total score and MNA-SF for malnutrition Based on albumin and BMI as the indicators of nutritional status, the sensitivity specificity and Youden index for the cutoff points of different MNA total scores were presented in Table 4. According to the original cutoff point (<17) indicating malnutrition devised by Guigoz, when using albumin (<35.0 g/l) as the malnutrition indicator, the sensitivity and specificity of the MNA total score were 0.4286 and 0.8562, respectively; when using BMI (<18.5 kg/m 2 ) as the indicator, the sensitivity and specificity were 0.6364 and 0.8642 (4). Using a new cutoff point of below 19 indicating malnutrition, when using albumin (<35.0 g/l) as the indicator, the sensitivity and specificity of the MNA total score were 0.6286 and 0.7466, respectively; when using BMI (<18.5 kg/m 2 ) as the indicator, the sensitivity and specificity were 0.8636 and 0.7469. Although calculation of the MNA-SF score includes only BMI, we pointed out correlations between the MNA-SF score 3

LEI/c:04 LORD_c 8/10/09 11:48 Page 4 CLINICAL STUDY OF MINI-NUTRITIONAL ASSESSMENT FOR OLDER CHINESE INPATIENTS and BMI, TSF, CC, albumin, lymphocyte count, triglyceride, hemoglobin, lymphocyte ratio or MNA total score in Table 3 (p<0.05). Furthermore, the highest correlation existed between the MNA-SF and the MNA total score (r=0.838, p<0.05). The optimal MNA-SF cutoff point was below 12 (sensitivity = 0.8955, specificity = 0.8800 and Youden index = 0.7755, in Table 5), which could also be tested in the ROC curve for the MNA-SF in older hospitalized Chinese patients. The area under the curve was 0.932 ± 0.018 (95%CI 0.897~0.966), which suggested a good correlation with MNA. Table 4 Validity values of the MNA total score for malnutrition The cutoff point of the MNA total score <16 <16.5 <17 <17.5 <18 <18.5 <19 <19.5 <20 Albumin (<35.0 g/l Sensitivity 0.3714 0.4286 0.4286 0.4571 0.4857 0.5714 0.6286 0.6286 0.6857 Specificity 0.8973 0.8699 0.8562 0.8493 0.8151 0.7671 0.7466 0.7192 0.6849 Youden index 0.2687 0.2984 0.2847 0.3065 0.3008 0.3386 0.3751 0.3477 0.3706 BMI (<18.5 kg/m 2 ) Sensitivity 0.5000 0.5909 0.6364 0.6818 0.7273 0.8182 0.8636 0.8636 0.8636 Specificity 0.8951 0.8704 0.8642 0.8580 0.8210 0.7716 0.7469 0.7222 0.6728 Youden index 0.3951 0.4613 0.5006 0.5398 0.5483 0.5898 0.6105 0.5859 0.5365 Table 5 Validity values of the MNA-SF The cutoff point of MNA-SF <9 <10 <11 <12 <13 Sensitivity 0.4552 0.6119 0.7687 0.8955 0.9776 Specificity 1.0000 1.0000 0.9600 0.8800 0.4400 Youden index 0.3134 0.6119 0.7287 0.7755 0.4172 Discussion The prevalence of malnutrition in older inpatients is high. Yin assessed the nutritional status of 280 inpatients and found a malnutrition rate of 22.5% for older inpatients (11). Research also showed that long-term older inpatients had a lower weight, lower body mass index, decreased total number of lymphocyte and transferrin, and creatinine-height index was significantly lower than control (12). In this study, according to the original cutoff point for the MNA test (<17), the prevalence of malnutrition for older inpatients was 19.6% and the rate of those at risk of malnutrition was 53.2%. According to the revised cutoff point (<19), the rate of malnutrition was 32.6%. If patients were in malnutrition, functional failure of organs and tissues was accelerated, immune activity declined, and the opportunity for infection and the risk-level of undergoing operation increased. Furthermore, malnutrition delays healing of operative incision and increases the incidence of postoperative complications (13). This results in a decrease in the quality of life for older inpatients and may lead to more severe consequences, such as prolonged hospital stays, increased hospital expenses and even a higher mortality rate 4 (14). Therefore, it is necessary to promptly and accurately assess the nutritional status of older inpatients in order to give nutritional intervention to those who are malnourished or at risk of malnutrition as soon as possible. At present, there still lacks a generally accepted assessment tool for determining malnutrition in China. The MNA, a convenient, cost-effective screening tool for malnutrition, is currently being introduced. We believe this is the first study to evaluate and set the threshold for the MNA in Chinese older population. According to the original scale devised by Guigoz (4) patients with MNA total scores under 17 are considered malnourished; as the MNA tool was designed and tested on older inpatients in Europe, considering the differences in living style, dietetic culture and physical status, such a cutoff point could lead to significant differences in assessment from for older inpatients in China. Compared with the classical assessment, all the differences with the MNA have statistical significance (p<0.05) except for CC, Hb, albumin and prealbumin (Table 1). This suggests that the assessment result of the MNA was not completely in accordance with that of the. Subjects in table 1 are grouped according to values of different markers, while in table 2, when comparing with the three groups of the MNA, only BMI and CC showed statistical significance between each group (p<0.05). Thus, older patient nutritional status was not distinguished well by the originally proposed grouping of the MNA, furthermore suggesting that the original cutoff point of MNA needed revision in the Chinese population. Our results showed that, according to correlation analysis between the MNA, the MNA-SF, and, the MNA had a good correlation with anthropometric measurements (BMI, TSF, MAC) and biochemical markers (albumin, lymphocyte count, prealbumin, hemoglobin, lymphocyte ratio); this was especially true for correlations with BMI and albumin, with results similar to those of He et al (15). Consequentially, albumin and BMI were chosen as the best parameters for revising the cutoff point of the MNA total score (Table 3). Based on albumin and BMI as the indicators of nutritional status, the validity values of the MNA and MNA-SF total score were revised. Youden index (sensitivity + [specificity - 1]) was used to compare the proportion of the cases correctly classified. The higher the Youden index, the more accurate the prediction at the cutoff point. The cutoff point of MNA was revised to 19, having the highest Youden index (Table 4). Compared with the original cutoff point (<17), its specificity slightly decreased while the sensitivity greatly increased. Since a high sensitivity and a high negative predictive value are required for clinical screening purposes in establishing a cutoff value, a cutoff point below 19 was determined best for detecting malnutrition in Chinese older inpatients. Based on this cutoff point, our research showed the incidence rate of malnutrition is 32.6%. In table 3, we show correlations between the MNA-SF and BMI, TSF, CC, albumin, lymphocyte count, triglyceride,

LEI/c:04 LORD_c 8/10/09 11:48 Page 5 hemoglobin, lymphocyte ratio or MNA total score (p<0.05). Results suggest that the MNA-SF, as the first step of the nutrition screening procedure, could fairly accurately reflect the nutritional status of the patients. Furthermore, according to the correlation coefficient and the area under the ROC curve between the MNA and MNA-SF, the highest correlation existed between the MNA-SF and the MNA total score (r=0.838, p<0.05). This can also be observed in the ROC curve. The area under the curve was 0.932 ± 0.018 (95%CI 0.897~0.966), which indicates that the MNA-SF appropriately reflects the MNA. Based on the MNA as the indicator of nutritional status, we determined the optimal MNA-SF cutoff point was below 12 based on high sensitivity (0.8955) and the highest Youden index (0.7755). Therefore, a cutoff point of 12 for MNA-SF is most suitable for older hospitalized patients in China (Table 5). Due to our limited scope and focus solely on older patients in China, one obvious limitation of the study is that our results may only apply to older inpatients in China; naturally, more extensive research is needed with a much broader subject pool if it is applied to the entire older population. Also, if the patients have nutritional related disease, including cancer or liver failure, more specific assessment tools need to be adopted to determine nutritional status. Since the MNA is only a primary screening tool for malnutrition and risk of malnutrition, the final nutritional status must be confirmed by other anthropometric, biochemical, and dietary parameters to have a full evaluation. Conclusion Such a high prevalence of malnutrition in older inpatients in China necessitates the adoption of a conventional screening tool for malnutrition risk. The MNA is presently the best and most attractive screening tool for nationwide use in China. Setting the MNA score cutoff point below 19 would be the most accurate and sensitive method for screening out malnourished older patients, thereby enabling early nutritional intervention and prevention of malnutrition and further complications during and following hospital stay. While there are limitations on our MNA studies and methods for use, specifically our application and MNA cutoff scores being almost exclusively applicable to older Chinese patients, it raises the question of whether or not a universal nutritional assessment tool, being the MNA test or otherwise, can ever be established. In our study, the test s accuracy was based on and designed for Chinese older inpatients. Earlier studies, including the original study and development of the MNA by Guigoz (4) were also targeted to a specific country, culture, or group, and were limited by a largely homogeneous subject pool. We have already pointed out the need to differentiate certain dietary items for the Chinese patient screening, including changing the cheese item to soymilk and peanut milk; nevertheless, further cultural differences, not limited only to extremely different dietary practices, but also consideration of cultural or regional differences, should be further researched. In addition, the test should not be universally applied in an arbitrary fashion. For example, if a patient is vegetarian or vegan, additional questions or methods may be necessary. A multinational and multicultural subject pool study, taking note of race and other factors, would be an efficient way to further expound on this issue. Taking evaluation of the MNA a step even further, it appears feasible to establish scores that apply to certain groups of individuals rather than including everyone over a certain age. A healthy BMI range that applies to men and women without distinction is still an issue of debate among researchers; likewise, nutritional assessment may require such distinctions to even more accurately evaluate individuals. Such distinctions and considerations, while it will make tests like the MNA more complex, can ultimately serve both a patient s and health provider s interests in the best manner. Financial disclosure: None of the authors had any financial interest or support for this paper. References 1. Corish CA & Kennedy NP. (2000) Protein-energy undernutrition in hospital inpatients. Br J Nutr 83, 575-591. 2. Yu K & Chen W. (1999) The Nutritional Assessment of Surgical Elderly Inpatients. Acta Nutrimenta Sinica 21, 212-215. 3. Green SM & Watson R. (2006) Nutritional screening and assessment tools for older adults: literature review. J Adv Nurs 54, 477-490. 4. Guigoz Y, Vallas BJ & Garry PJ. (1996) Assessing the nutritional status of the elderly: the Mini-Nutritional Assessment as part of the geriatric evaluation. Nutr Rev 54, 59-65. 5. Charlton KE, Kolbe-Alexander TL & Nel JH. (2007) The MNA, but not the DETERMINE, screening tool is a valid indicator of nutritional status in elderly Africans. Nutrition 23,533-542 6. Kabir ZN, Ferdous T, Cederholm T, Khanam MA, Streatfied K, Wahlin A. (2006) Mini Nutritional Assessment of rural elderly people in Bangladesh: the impact of demographic, socio-economic and health factors. Public Health Nutr 9, 968-974. 7. Rubenstein LZ, Harber JO & Salva A. (2001) Screening for undernutrition in geriatric practice: developing the short-form mini-nutritional assessment (MNA-SF). Gerontology: Medical Sciences 56, 366-372. 8. Vellas B, Guigoz Y & Garry PJ. (1999) The mini nutritional assessment (MNA) and its use in grading the nutritional state of elderly patients. Nutrition 15, 116. 9. Liao EY & Chao CS. (2001) Endocrinology. The People's Health Publishing House, Beijing, 1767-1973. 10. Chen CM & Shao ZM. (1994) The status of food nutrition and health in seven provinces in China. China Statistical Publishing House, Beijing. 11. Yin QF, Chen WK & Mo ZF. (2005) Nutrition status evaluation of 280 in-patients. Modern Hospital 5, 26-27. 12. Wang L, Li YY & Jiang W. (2005) A study of nutritional assessment for the long term hospitalized elderly patients. Chinese Journal of Geriatrics 24, 589-591. 13. Sungurtekin H, Sungurtekin U, Balci C, Zencir M, Erdem E.(2004) The influence of nutritional status on complications after major intra-abdominal surgery. J Am Coll Nutr 23,227-232. 14. Chen CH, Schilling LS & Lyder CH. (2001) A concept analysis of malnutrition in the elderly. J Adv Nurs 36, 131-142. 15. He YL, Jian ZJ & Ouyang M. (2004) Using Mini-Nutritional Assessment to Investigate the Nutritional Status of the Aged Hospitalized Patients. Chinese Journal of Clinical Nutrition 12, 93-96. 5