Subjective global assessment (SGA) is a method to score nutritional status in a standardized way. The original 3-point scale has been replaced by a 7-point scale. The reliability of the latter scale has never been tested. We therefore assessed inter-observer and intra-observer reliability. Furthermore, we examined the relationship of SGA with other objective nutritional parameters. In 13 hemodialysis and 9 peritoneal dialysis patients, two nurses assessed SGA. They re-examined the same patients two weeks later. Anthropometric measurements and blood samples were taken at the first assessment. According to SGA, 2 patients (9%) were classified as severely malnourished, 6 (27%) as mildly malnourished, and 14 (64%) as well nourished. The 7-point SGA scale showed fair inter-observer reliability [intraclass correlation (ICC) = 0.72] and good intra-observer reliability (ICC = 0.88). A strong correlation was present between the 7-point SGA scale and body mass index (BMI) (r = 0.79, p < 0.001), % fat (r = 0.77, p < 0.001), and mid arm circumference (r = 0.71, p < 0.001). Lower correlations were found with mid arm muscle circumference and serum albumin. With respect to biochemical markers, the strongest relationship was found with prealbumin (r = 0.60, p = 0.004). We conclude that the 7-point SGA scale is a valid and reliable tool to assess nutritional status among end-stage renal disease patients. We suggest that one observer or a select group of observers perform the assessments to gain maximum benefit from the reliability of the SGA instrument.