Clinical nutrition scores are superior for the prognosis of haemodialysis patients compared to lab markers and bioelectrical impedance.

Background. Malnutrition is closely related to inflammation and atherosclerosis in uraemic patients. There is still debate on how to quantify nutritional status in order to achieve the best prediction of mortality and hospitalization. Methods. Different methods to detect malnutrition were prospectively investigated for their prognostic impact on mortality and hospitalization of haemodialysis (HD) patients. We compared clinical nutrition scores (body mass index, BMI; subjective global assessment, SGA; malnutrition inflammation score, MIS; and nutritional risk screening, NRS) to lab parameters of protein and lipid metabolism, or bioelectrical impedance analysis (BIA) in 90 HD patients. Over a 3-year follow-up, all-cause mortality and hospitalization were evaluated using a Cox regression model. Results. The scores SGA, NRS, MIS, serum albumin, prealbumin, transferrin and BIA were predictive of both mortality and hospitalization. Elevated CRP predicted only a significantly higher mortality. After adjustment for age, gender, dialysis vintage and diabetes status, the best prognostic parameters for mortality were the clinical nutrition scores, MIS-Index >= 10 [HR 6.25 (2.82-13.87), P < 0.001], NRS [HR 4.24 (1.92-9.38), P < 0.001] and SGA B/C [HR 2.70 (1.14-6.41), P < 0.05]. Conclusions. In HD patients, serum markers of protein metabolism and BIA can be used for evaluation of the nutritional status. However, with regard to mortality and hospitalization risk, the individual clinical nutrition scores are superior compared to lab markers and BIA. To confirm malnutrition, we propose using clinical nutrition score generally or at least in the case of two malnutrition-positive parameters (lab, BIA, BMI).