A randomized clinical trial of granulocyte transfusions for infection in acute leukemia.

In a prospective, controlled, randomized study to evaluate the efficacy of filtration-leukapheresis granulocytes in granulocytopenic, febrile patients with leukemia, 19 patients received antibiotics alone, and 12 received antibiotics plus daily granulocyte transfusions from ABO-matched donors. In skin-chamber studies the granulocytes appeared at sites of inflammation for at least six hours after transfusion. Infected subjects survived longer if they received granulocytes. Differences between control and transfused patients were greatest in patients with persistent bone-marrow failure, the 21-day survival being 20 per cent in controls, and 75 per cent in transfused patients. Granulocytes appeared to have no effect on the outcome of febrile episodes in which infection was not documented, the 21-day survival being 79 per cent for controls and 88 per cent for transfused patients. The transfusion of granulocytes thus appears to offer a survival advantage to infected, persistently granulocytopenic patients.