Although it is common practice to use some form of isolation to protect allogeneic stem cell transplant patients from infection, the necessity for these practices in all environments has not been demonstrated. The current study evaluated patterns of infection and 100-day transplant-related mortality in 288 patients with myelodysplasia and leukemia transplanted without isolation. Patients were allowed out of hospital at any time within constraints of the medication schedule. Fever, foci of infection, and positive cultures within 28 days and death within 100 days because of the transplant procedure were recorded. Fever occurred in 57% of patients, and 10% had a clinical or radiographic focus of infection. Most infections were apparently endogenous; blood cultures from 24% of recipients grew organisms, 87% of which were gram-positive bacteria. Four patients (1%) died with aspergillus infection in circumstances indicating that isolation would not have been helpful. Twenty percent of patients remained without evidence of infection throughout. Transplant-related mortality at 100 days was 1% for 108 patients with early leukemia receiving transplants from matched siblings. For patients at higher risk, by virtue of donor and/or disease status, mortality was 21%. These figures compare favorably with those reported to the International Bone Marrow Transplant Registry, the majority of patients having been subjected to some form of isolation. We conclude that allogeneic stem cell transplantation can be safely performed in some environments without confining patients continuously to the hospital.