Malnutrition is a major negative prognostic factor in dialysis patients. Simple and reliable estimations of nutritional status may therefore prove of particular value in the follow-up of these patients. To validate subjective global assessment (SGA) in dialysis patients we compared subjective global assessment with objective measurements (anthropometry, bioelectrical impedance, biochemical measurements) in 59 chronic uraemic patients treated by haemodialysis (n = 36) or CAPD (n = 23). Subjective global assessment was performed by an observer unaware of the results of objective measurements and was related to serum albumin (r = -0.51, P < 0.001) and bioelectric impedance phase angle (r = -0.58, P < 0.001) as well as with MAMC (r = -0.28 P = 0.028), %fat (r = -0.27, P = 0.042) and nPCR (r = -0.29 P = 0.027). Multiple regression analysis showed that the relationship of subjective global assessment (as a dependent variable) with objective measurements (covariates) was stronger (multiple r = 0.77) than the relationship found with univariate analysis. This finding indicates that subjective global assessment gives a well-based and balanced estimation of nutritional status. Our data show that subjective global assessment is a clinically adequate method for assessing nutritional status in dialysis patients. Being an inexpensive method of well-proven reliability, subjective global assessment can be recommended for a more frequent assessment of nutritional status in dialysis patients.