Since the introduction of recombinant human erythropoietin (rHuEPO) in 1989, the level of anemia correction has been debated. Information developed from clinical trials and observational studies have given conflicting results. The normal hematocrit (Hct) trial of Besarab et al showed no benefit, and a possible risk, of correcting the Hct from 30% to 42% in hemodialysis patients with NYHA class I to III cardiac disease. The study of Moreno et al in non-cardiac patients showed improved Sickness Impact Profile and Karnofsky scores in hemodialysis patients when their Hct was increased from 31% to 38.5%. Hospitalizations were significantly reduced. Observational studies from Madore et al, Xia et al, Ma et al, and Collins et al all show that increased morbidity and mortality are significantly associated with Hct <33%. Recent data on incident hemodialysis patients indicate the associated risk of death was not different for patients with Hct 33 to <36% versus Hct 36 to <39%. Hospitalization risks and associated costs were significantly less in the patients with Hct 36 to <39%, suggesting that higher Hct values may be less of a concern than previously considered. The current data suggest that patents with advanced cardiac disease should avoid Hct values in the normal range. In others, Hcts at least up to 39% appear to be safe and effective. Based on this review, a reasonable target Hct range may be 33 to 39%, which balances the risks and potential benefits.