We examined the value of transferrin concentrations in estimating nutritional status as determined by the subjective global assessment (SGA) score. Fifty-nine hemodialysis patients (37 men and 22 women, aged 59+/-16 years, dialyzed for 3.6+/-3.9 years) were selected by predetermined criteria. All received erythropoietin (EPO) and oral iron therapy. SGA evaluation was conducted twice by both a dietitian and a physician. Serum iron, total iron-binding capacity (TIBC; which is linearly correlated with transferrin), transferrin saturation ratio, ferritin, albumin, total protein, and cholesterol were measured. Twenty-seven (46%) patients were well nourished (group A), 20 (34%) were moderately nourished (group B), and 12 (20%) were poorly nourished (group C) according to the SGA. TIBC values were 276+/-47 mg/dL, 217+/-54 mg/dL, and 176+/-41 mg/dL, respectively (P < 0.00001), and thus directly correlated with the state of nutrition. The relationship between TIBC and nutritional status was independent of age and number of years on hemodialysis. Serum ferritin values were 104+/-93 ng/mL, 161+/-154 ng/mL, and 363+/-305 ng/mL, respectively (P < 0.0003), and thus inversely correlated with the state of nutrition. Transferrin saturation ratios were slightly higher in the severely malnourished patients. The number of years on dialysis were a determinant of nutritional status. These values were 2.4+/-2.4 years for group A, 3.9+/-4.0 years for group B, and 5.7+/-3.9 years for group C (P < 0.05). The average age of the poorly nourished patients was 10 years older than the well-nourished patients. Serum iron values were lower but transferrin saturation ratios were higher in the severely malnourished patients. The required EPO doses were higher in the poorly nourished patients. We suggest that transferrin values are superior to other laboratory tests in assessing nutrition and will supplement SGA criteria. Serum ferritin may be useful as a predictor of illness. Older patients who have been on dialysis longer warrant special concern. Malnutrition may be an indicator of EPO resistance in dialysis patients. Finally, since a decreased TIBC level in poorly nourished patients may erroneously increase the transferrin saturation ratio, our findings may have implications in making the diagnosis and treatment of anemia and iron deficiency in malnourished dialysis patients.