Among patients with chronic kidney disease (CKD), erythropoiesis-stimulating agents (ESAs) are approved for avoiding transfusions, a risk that increases when hemoglobin (Hgb) falls to <10 g/dl. Transfusions increase sensitization, prolong the waiting time to and the likelihood of transplantation, and when transplantation is performed worsen graft survival. Accordingly, the risk of transfusion among those expecting transplantation is comparable to that of cardiovascular events. Nonetheless, targeting Hgb to >13 g/dl is associated with increased cardiovascular events. Paradoxically, if this level is achieved mortality is lower. The anemia paradox--higher cardiovascular events when targeting higher Hgb but lower events when patients achieve these targets--appears to be at least partially attributable to a hyporesponse to ESAs. Whether it is the ESAs or conditions that lead to the increased ESA dose that provokes morbidity cannot be answered definitively at present. The lowest ESA dose to achieve the desired level of anemia correction appears to be a safer strategy. In acute illnesses, reducing the dose of ESAs or stopping it altogether may aggravate anemia, but this may be permissible. The rate of rise in Hgb>1 g/dl in any 2-week period is associated with an increase in blood pressure (BP) and poor outcomes. Accordingly, while initiating and maintaining ESA therapy, monitoring BP at home twice daily is warranted. The clinical decision-making process in managing anemia should consider the risks of transfusion; kidney transplant potential; presence of cancer; and the risks of stroke, heart failure, and possibly death. Above all, clinical decision-making should incorporate patient preference.