Objectives: To investigate an outbreak of aspergillosis in a leukemia and bone marrow transplant (BMT) unit and to improve environmental assessment strategies to detect Aspergillus., Design: Epidemiological investigation and detailed environmental assessment., Setting: A tertiary-care university hospital with a 37-bed leukemia and BMT unit, Participants: Leukemic or BMT patients with invasive aspergillosis identified through prospective surveillance and confirmed by chart review., Interventions: We verified the diagnosis of invasive fungal infection by reviewing medical charts of at-risk patients, performing a case-control study to determine risk factors for infection, instituting wet mopping to clean all floors, providing N95 masks to protect patients outside high-efficiency particulate air (HEPA)-filtered areas, altering traffic patterns into the unit, and performing molecular typing of selected Aspergillus flavus isolates. To assess the environment, we verified pressure relationships between the rooms and hallway and between buildings, and we compared the ability of large-volume (1,200 L) and small-volume (160 L) air samplers to detect Aspergillus spores., Results: Of 29 potential invasive aspergillosis cases, 21 were confirmed by medical chart review. Risk factors for developing invasive aspergillosis included the length of time since malignancy was diagnosed (odds ratio [OR], 1.0; P=.05) and hospitalization in a patient room located near a stairwell door (OR, 3.7; P=.05). Two of five A. flavus patient isolates were identical to one of the environmental isolates. The pressure in most of the rooms was higher than in the corridors, but the pressure in the oncology unit was negative with respect to the physically adjacent hospital; consequently, the unit acted essentially as a vacuum that siphoned non-HEPA-filtered air from the main hospital. Of the 78 samples obtained with a small-volume air sampler, none grew an Aspergillus species, whereas 10 of 40 cultures obtained with a large-volume air sampler did., Conclusions: During active construction, Aspergillus spores may have entered the oncology unit from the physically adjacent hospital because the air pressure differed. Guidelines that establish the minimum acceptable pressures and specify which pressure relationships to test in healthcare settings are needed. Our data show that large-volume air samples are superior to small-volume samples to assess for Aspergillus in the healthcare environment.