The leucocytes present in red cell and platelet components have been implicated in several important immunological and infective complications of blood transfusion. Recent developments in blood filtration technology allow the production of leucodepleted blood products (residual leucocytes < 5 x 10(6) per transfused unit) in the laboratory or at the bedside with the potential to prevent these adverse effects. Quality assurance remains an important problem, particularly for bedside filtration. Prestorage filtration may have significant advantages for red cell and platelet production. There is strong clinical evidence that 3 log10 leucodepletion prevents or delays febrile reactions in patients receiving multiple red cell transfusions and can reduce cytomegalovirus transmission. Leucodepletion to prevent HLA alloimmunisation, platelet refractoriness and febrile reactions in patients receiving red cell and platelet support remains controversial. Transfused leucocytes induce 'immunosuppressive' changes in the recipient, but recent studies cast doubt on the association with cancer recurrence after surgery. However, leucodepleted blood may reduce the incidence of postoperative infection. Leucodepletion by filtration is expensive and there is a requirement for well-designed prospective clinical studies focusing on appropriate filtration technology (and alternatives), clinical outcome and cost-effectiveness.