Infections are an almost inevitable complication of human bone marrow transplantation and account for the majority of deaths in transplant recipients. Even prior to the initiation of the transplantation procedure, patients may present with infections complicating previously unsuccessful chemotherapy for hematological malignancy or aplastic anemia. Nevertheless, these pre-transplantation infections should not exclude the possibility of bone marrow transplantation if they can be successfully controlled with specific antimicrobial therapy and necessary adjunctive measures. The immediate post-transplantation period prior to engraftment is characterized by severe marrow aplasia that results from high-dose chemotherapy and total-body irradiation. Infections are primarily septicemias and localized processes caused by bacteria and fungi and their incidence increases as the intensity of immunosuppression is escalated. The high mortality associated with bacterial septicemia makes early, empirical antibacterial therapy mandatory. However, the reduction in mortality from bacterial infection resulting from such an aggressive approach may be offset by a higher mortality from invasive fungal infection, especially in patients with prior fungal colonization and undergoing prolonged conditioning therapy. Thus, until more specific and sensitive tests for the diagnosis of invasive fungal infection become available, empirical intravenous amphotericin should be considered in patients who are persistently febrile and deteriorate clinically in the face of appropriate antibacterial therapy. Interstitial pneumonia associated with severe GVHD is the major infectious complication after successful marrow engraftment and is the most significant barrier to long-term survival. Trimethoprim-sulfamethoxazole is effective prophylaxis against interstitial pneumonia due to Pneumocystis carinii, but one half of the patients still develop a pneumonitis either associated with CMV or of unknown etiology. Mortality from interstitial pneumonia is related to prior radiation therapy while survival is associated with a four-fold rise in CMV CF antibody titer. The latter observation supports the need to investigate passive immunization with CMV antibody as a means of preventing some interstitial pneumonias. Despite the progress made in many areas of human bone marrow transplantation, the majority of graft recipients still die of infectious complications. Thus, new approaches to the management of infections in transplant recipients are urgently needed. Better-tolerated oral nonabsorbable antibiotics, laminar-air-flow rooms, granulocyte transfusions, and chemotherapy and immunotherapy for CMV are among the prophylactic and therapeutic measures that must be critically evaluated in well-controlled, prospective studies. Continued assessment of the infectious complications of bone marrow transplantation is a critical aspect of any ongoing transplant program, not just a research goal...