We defined erythropoietin (EPO) resistance by the ratio of the weekly EPO dose to hematocrit (Hct), yielding a continuously distributed variable (EPO/Hct). EPO resistance is usually attributed to iron or vitamin deficiency, hyperparathyroidism, aluminum toxicity, or inflammation. Activation of the acute-phase response, assessed by the level of the acute-phase C-reactive protein (CRP), correlates strongly with hypoalbuminemia and mortality in both hemodialysis (HD) and peritoneal dialysis (PD) patients. In this cross-sectional study of 92 HD and 36 PD patients, we examined the contribution of parathyroid hormone (PTH) levels, iron indices, aluminum levels, nutritional parameters (normalized protein catabolic rate [PCRn]), dialysis adequacy (Kt/V), and CRP to EPO/Hct. Albumin level serves as a measure of both nutrition and inflammation and was used as another independent variable. Serum albumin level (deltaR2 = 0.129; P < 0.001) and age (deltaR2 = 0.040; P = 0.040) were the best predictors of EPO/Hct in HD patients, and serum albumin (deltaR2 = 0.205; P = 0.002) and ferritin levels (deltaR2 = 0.132; P = 0.015) in PD patients. When albumin was excluded from the analysis, the best predictors of EPO/Hct were CRP (deltaR2 = 0.105; P = 0.003) and ferritin levels (deltaR2 = 0.051; P = 0.023) in HD patients and CRP level (deltaR2 = 0.141; P = 0.024) in PD patients. When both albumin and CRP were excluded from analysis in HD patients, low transferrin levels predicted high EPO/Hct (deltaR2 = 0.070; P = 0.011). EPO/Hct was independent of PTH and aluminum levels, PCRn, and Kt/V. High EPO/Hct occurred in the context of high ferritin and low transferrin levels, the pattern expected in the acute-phase response, not in iron deficiency. In well-dialyzed patients who were iron replete, the acute-phase response was the most important predictor of EPO resistance.