We assessed the safety and efficacy of Fluosol-DA as a red-cell substitute in acute anemia. Twenty-three surgical patients with blood loss and religious objections to receiving blood transfusions were evaluated. Fifteen moderately anemic patients with a mean hemoglobin level (+/- SE) of 7.2 +/- 0.5 g per deciliter had no evidence of a physiologic need for increased arterial oxygen content and did not receive Fluosol-DA. Eight severely anemic patients with a mean hemoglobin level of 3.0 +/- 0.4 g per deciliter met the criteria of need and received the drug until the physiologic need disappeared or a maximal dose of 40 ml per kilogram of body weight was reached. We observed no adverse reactions to Fluosol-DA. The average peak increment in arterial oxygen content with the drug was only 0.7 +/- 0.1 ml per deciliter. There were no appreciable beneficial effects of Fluosol-DA, perhaps because of the small increase in arterial oxygen content, the brief half-life of the drug (24.3 +/- 4.3 hours), and the limited total dose. Six of the eight patients receiving Fluosol-DA died. One of the survivors received red-cell transfusions against his wishes, under a court order, after his total Fluosol-DA dose. Fourteen of the 15 moderately anemic patients survived. The data in this select group of patients refusing blood products suggest that, after blood loss, Fluosol-DA is unnecessary in moderate anemia and ineffective in severe anemia.