BACKGROUND:Malnutrition, a predictor of increased mortality in dialysis patients, can be estimated using the subjective global assessment (SGA), a semiquantitative scale with three severity levels. This semiquantitative feature restricts the SGA's reliability and precision. METHODS:Using the components of the conventional SGA, we developed a fully quantitative scoring system (the dialysis malnutrition score) consisting of seven variables: weight change, dietary intake, gastrointestinal symptoms, functional capacity, comorbidity, subcutaneous fat and signs of muscle wasting. Each component was assigned a score from 1 (normal) to 5 (very severe). The sum of all seven components in this malnutrition score lies between 7 (normal) and 35 (severely malnourished). To evaluate nutritional status in chronic dialysis patients, anthropometric measurements including mid-arm circumference (MAC), triceps skin-fold thickness, calculated mid-arm muscle circumference (MAMC), body mass index (BMI, ratio of weight to square of height) and laboratory parameters were used. Forty-one patients (20 men and 21 women) were randomly selected from a pool of 120 haemodialysis patients. Patients were aged between 26 and 81 years (mean SD, 57 +/- 12 years) and had undergone haemodialysis for between 7 months and 12 years (mean +/- SD, 3.0 +/- 2.1 years). RESULTS:The malnutrition score of each patient was assessed by a dietitian within 5-20 min (12.0 +/- 3.5 min) with no knowledge of anthropometric findings. Pearson correlation coefficients between the malnutrition score and biceps skin-fold (r= -0.32) MAC (r= -0.55), MAMC (r= -0.66), BMI (r= -0.35), total iron-binding capacity (TIBC, r= -0.77), the serum albumin concentration (r= -0.36) and total protein (r= -0.33) were all significant, whereas the conventional SGA had significant correlation only with TIBC (r= -0.35) and MAMC (r= -0.37). Malnutrition score showed a significant correlation with age (r= +0.34) and years dialysed (r= +0.28). Multiple regression analysis showed a significant correlation between the malnutrition score and the combination of the MAMC, BMI, serum albumin concentration and TIBC (r= 0.81, P<0.001). There was no correlation between the malnutrition score and sex, urea reduction ratio, protein catabolic rate, and the absolute lymphocyte count. CONCLUSIONS:We conclude that our invented malnutrition score, which can be performed in minutes, reliably assesses the nutritional status of haemodialysis patients. We suggest that our malnutrition score may be superior to the SGA. More comparative and longitudinal studies are needed to confirm the validity of this scoring system in nutritional evaluation of dialysis patients.