It is clear from the trials described here that the number of different products being tested and the potential variation between batches of the same product present major problems in evaluating the safety and efficacy of hemoglobin-based oxygen carriers. The recent CBER "Points to Consider" document [42] makes clear that an understanding of the safety of oxygen carriers in humans is of paramount importance. In the event of phase II or indeed phase III trials being approved, the need may still remain for additional phase I or preclinical studies, particularly as unwanted or toxic properties of the solutions affect efficacy. It is likely that demonstrating safety and efficacy in acute hemorrhagic shock will be the most difficult task, as this is a complex clinical indication and is often accompanied by multisystem damage. The use of a hemoglobin-based oxygen carrier in this setting must have a distinct advantage over a plasma expander alone. In the application of perioperative transfusion, a decreased requirement for red cell transfusion has already been accepted as a basis for the efficacy for erythropoietin. However, in the case of a hemoglobin-based oxygen carrier, the reduction of red cell requirement in perioperative procedures would need to be balanced against any adverse drug reactions or unacceptable hemodynamic effects that may be caused by the product. It appears that there are still numerous hurdles to overcome in the development of hemoglobin-based red cell substitutes. Before these products can become established in medical practice, it is imperative that the potential mechanisms of toxicity of cell-free hemoglobin are clearly understood. Approval of hemoglobin-based oxygen carriers for clinical use will depend not only on clear demonstration of both safety and efficacy but also on risk-versus-benefit issues. Our understanding of the physiological effects of these products will evolve as progress is made in their clinical evaluation.