Objective Peritoneal dialysis (PD) patients are characterized by protein malnutrition and muscle wasting. Reliable, easy, and cheap methods for evaluating nutrition are desirable. Three methods are commonly available: dual-energy X-ray absorptiometry (DXA), bioimpedance (BI), and subjective global assessment (SGA).The objective of the study was to compare the previously mentioned methods for assessment of body composition and nutritional status in PD patients. Design The study is cross-sectional and consisted of 72 PD patients from a single center PD ambulatorium. Methods Participants were measured twice by DXA, twice by BI, and once by SGA. Measurements included lean tissue mass (LTM), fat tissue mass (FT) and, for BI, overhydration (OH), intracellular water (ICW), and extracellular water (ECW). LTM and FT were indexed to body area (Lean Tissue Index [LTI] and Fat Tissue Index [FTI], respectively), and ICW for height (ICW/ht). We assessed conventional biochemical and clinical variables, using values for normal individuals as a reference. Results There was good overall agreement between BI and DXA but considerable intra-individual variation (1 standard deviation: FT 5.7 kg; LTM 5.6 kg). Factors affecting the differences were FT, ICW, LTM, and ICW. Obesity (DXA 43%; BI 54%) and muscle wasting (BI 28%; SGA 53%) were common. Agreement between BI and SGA was poor. Thirty-eight percent of patients judged malnourished by SGA also had a low LTI; 23% with normal SGA had low LTI. SGA was closer related to LTI (BI) than LTI (DXA). Plasma albumin was correlated to LTI, FTI, and ICW/ht, and comorbidity to OH, clinical malnutrition, reduced FTI, but not LTI. Conclusion Agreement between DXA and BI was high on a population basis but not at an individual level. Obesity and muscle wasting were common in this population. OH might reduce DXA accuracy in PD patients. LTI and ICW may be useful measures to supplement SGA in assessing nutrition.