Every year, about 75 million units of blood are collected worldwide. Red blood cell (RBC) transfusion is one of the few treatments that adequately restore tissue oxygenation when oxygen demand exceeds supply. Although the respiratory function of blood has been studied intensively, the trigger for RBC transfusion remains controversial, and doctors rely primarily on clinical experience. Laboratory assays that indicate failing tissue oxygenation would be ideal to guide the need for transfusion, but none has proved easy, reproducible, and sensitive to regional tissue hypoxia. The clinical importance of the RBCs storage lesion (ie, the time-dependent metabolic, biochemical, and molecular changes that stored blood cells undergo) is poorly understood. RBCs can be filtered, washed, frozen, or irradiated for specific indications. Donor screening and testing have dramatically reduced infectious risks in the developed world, but infection remains a major hazard in developing countries, where 13 million units of blood are not tested for HIV or hepatitis viruses. Pathogen inactivation techniques are in clinical trials for RBCs, but none is available for use. Despite serious immunological and non-immunological complications, RBC transfusion holds a therapeutic index that exceeds that of many common medications.