As leukemia progresses and becomes refractory to treatment, the autopsy diagnosis is more often blood, edema, fibrosis, or the primary disease process rather than a specific infectious pathogen. Since OLB and autopsy do not provide a specific diagnosis in at least 20 percent of some subsets, including leukemia, we should not expect BAL and other less invasive procedures to have a high yield for a specific diagnosis. When BAL findings are individualized to each patient setting, they can be used to confirm diagnostic impressions earlier in the course of this subset. The authors hold BAL to the strictest criteria for its contribution to diagnosis: autopsy confirmation of histology and culture within three weeks of the procedure. Their report is meticulous in describing their subset and important factors to consider in discerning the diagnostic yield from any pulmonary procedure. The pathologic behavior of Candida, which awaits the stress of combination therapy on the immune system and invades tissue an average of 19 days before death in leukemia, is reconfirmed. Since this fungus is often identified in other tissues (funduscopic examination) and cultures at this stage, lavage showing abundant yeast and pseudohyphae is useful confirmatory evidence of invasive candidiasis. Colonization seems a remote issue in this particular setting.
On the other hand, the data of Saito et al cast doubt on previous reports of efficacy for BAL in diagnosis of invasive aspergillosis. It has been suggested that washings, lavage, or a single sheathed catheter isolate may be enough evidence to confirm pulmonary infection with this organism. Aspergillus, like Candida, infects lung and causes alveolar hemorrhage. For this reason, early reports of positive washings with normal roentgenograms most likely represented colonization rather than invasion. The study of Saito et al refutes the suggestion that BAL can stand alone to distinguish colonization from invasion in many situations. If BAL is accepted as adequate for diagnosis without other supportive cultures, a low risk opportunity for definitive histology and culture by transbronchial lung biopsy (TBBX) or thin needle aspirate (TNA) may be lost.
When the available pulmonary procedure options are the topic at bedside rounds, risk benefit reported for each procedure in the immunocompromised host literature is of paramount importance. In any given situation, sputum, BAL, TBBX, or TNA may provide at least one pathogen to treat. When these procedures are negative for pathogens, an aggressive use of them in combination does provide adequate support for continued empiricism or observation despite the acknowledged limitations of sample size.