Fiberoptic bronchoscopy with bronchoalveolar lavage (BAL) is thought to be the procedure of choice to evaluate pulmonary infiltrates in hematopoietic stem cell transplant (HSCT) recipients. We retrospectively reviewed 91 bronchoscopies performed on 190 in-patient HSCT recipients admitted or treated for pneumonia from January 1994 to December 2004. These yielded a diagnosis 49% of the time with an overall survival of 35 days post-bronchoscopy. We were unable to detect any survival benefit from an addition to the treatment regimen after a positive result from analysis of the BAL fluid or transbronchial biopsy. The most common bacteria isolated was Pseudomonas that was often resistant to the patient's current antibiotics, suggesting that in lieu of this diagnostic procedure, changes to better cover resistant Gram-negative bacteria are reasonable. Although transbronchial biopsies provided an additional diagnosis in one out of 21 biopsies performed, six of the seven complications in our series were directly related to the transbronchial biopsy. With approximately a 50% yield from a bronchoscopy, additional treatment given after only 20% of all bronchoscopies, and no detectable survival benefit with a bronchoscopy that yielded a diagnosis, the utility of a bronchoscopy in this patient population is questioned by these data.