Purpose: To define the role of lower-respiratory-tract cultures in the diagnosis of invasive pulmonary aspergillosis (IPA) in immunocompromised hosts.
Methods: Immunocompromised patients with a positive, nonbiopsy, lower-respiratory-tract culture for Aspergillus species were classified as having definite, probable, indeterminate, or no IPA. Culture data, positive predictive values (PPVs), correlation with clinical and radiographic findings, and the relationship between the number of specimens submitted and the likelihood of recovering Aspergillus were assessed.
Results: Definite or probable IPA was diagnosed in 72% of episodes from patients with hematologic malignancy, granulocytopenia, or bone-marrow transplant; in 58% of those with solid-organ transplant or using corticosteroids; and in 14% of those with human immunodeficiency virus infection. The PPV of cultures ranged from 14% in the latter group to 72% in the first group (bone-marrow-transplantation subgroup, 82%). Fungal cultures were more often positive than were routine cultures (P < 0.001). Clinical and radiographic findings suggestive of IPA were present more frequently in infected than uninfected patients (59% versus 24%, P < 0.025); and 73% versus 6%, (P < 0.0001, respectively). Infected patients with > or = 1 positive node had more cultures submitted than a control group of patients with no positive cultures (5.8 +/- 4.7 versus 2.1 +/- 2.2 cultures, P < 0.001).
Conclusion: Recovery of Aspergillus species from high-risk patients is associated with invasive infection. Clinical and radiographic correlations help to separate true- from false-positive cultures. At least 3 sputum specimens should be submitted for fungal culture whenever fungal infection is suspected.