Evaluate which of two combinations of parameters based on International Society of Renal Nutrition and Metabolism recommendations could better identify patients with protein-energy wasting (PEW) and to compare the relationship of these two combinations with other clinical and body composition parameters. This was a multicentre longitudinal study with 24 months of follow-up. The PEW patients were characterized by: Group A (GA ) - normalized protein catabolic rate (nPCR) < 1.0 g/kg per day, albumin <3.8 g/dL and body cell mass index (BCMI) < 6.4 kg/m2 (n = 203); Group B (GB ) - nPCR <1.0 g/kg per day, albumin <3.8 g/dL and body mass index (BMI) <23 kg/m2 (n = 109). All the patients who did not meet these requirements were considered "well-nourished" (GA : n = 1818; GB : n = 3292). When compared to the well-nourished patients, PEW patients in the GA presented higher age, Kt/V, C-reactive protein, relative overhydration, fat tissue index (FTI); lower creatinine, albumin, nPCR, PTH, haemoglobin, phosphorus, calcium X phosphorus product, potassium, dry weight, BMI, BCMI, lean tissue index, %IDWG . In the GB , well-nourished patients FTI was significantly higher. In Cox analysis, the combination with BCMI was a strong independent predictor of mortality in these patients (hazard ratio: 1.48; confidence interval: 1.00-2.19; P = 0.048), even after adjustment. Although GB combination seemed to be also a predictor of death (hazard ratio: 2.67; confidence interval: 1.92-3.71; P < 0.001), when adjusted, the association remained no longer significant. A new combination of parameters including protein intake, albumin and BCMI demonstrated significant associations with other nutrition and inflammation parameters as well as with mortality.