Methods: We analysed a series of patients with proven or probable culture-documented IPA (EORTC/MSG criteria) in HM patients (1999-2015). All patients had BAL cultures that were positive for Aspergillus spp. and had concurrently obtained BAL cytology GMS available for analysis.
Results: We identified 67 such patients. BAL cytology based on GMS showed hyalohyphomycetes consistent with Aspergillus in 28/67 (41.8%) patients, whereas only in2/67 (3.6%) direct smear Calcofluor White stain was positive. Based on BAL GMS cytology, co-infections were identified in 6 patients: 2 Pneumocystis and 5 viral infections with cytopathic changes (one had both). The yield of cytology was not different in patients with IPA caused by non-fumigatus Aspergillus, although patients with IPA and >1 Aspergillus in BAL culture had more often positive cytology GMS (100% versus 0%, p 0.027). Cytology was also more often positive when obtained from a lesion-targeted BAL as compared to non-targeted bronchial washings (60.7% versus 7.1%, p 0.038). Patients with IPA and cavitary lesions (32.1% versus 5.1%, p 0.006), history of SCT (64.3% versus 33%, p 0.015) or prior exposure to itraconazole (75% versus 41%, p 0.007) had positive cytology GMS results more often than did patients without these characteristics. In the multivariate analysis, only cavitary lesions were significantly associated with positive BAL GMS cytology.
Conclusion: GMS stain in cytology of BAL in patients with HM and culture-documented IPA had a sensitivity of 41.8 % and was more often positive in patients with cavitary lesions. Although there were no differences in the proportion of GMS-positive cytology rates among differing Aspergillus spp causing IPA, mixed Aspergillus spp IPA was associated with an increase in positive cytology. BAL cytology was diagnostic for co-infections in more than 10% of patients. BAL cytology should be part of the diagnostic wok up in HM patients with suspected IPA.
S. T. Heo, None
S. Evans, None
J. Tarrand, None
D. P. Kontoyiannis, None