Objectives: A recently developed bronchoalveolar lavage (BAL) galactomannan (GM) assay shows promising results. We evaluated the diagnostic performance of this assay and analyzed risk factors for false-positive results.
Methods: A prospective cohort study was performed in a tertiary hospital over a 9-month period. We reviewed all adult patients who underwent GM assays of BAL. Patients were categorized with proven, probable, or possible invasive pulmonary aspergillosis (IPA) according to revised EORTC/MSG definitions. Each patient with a false-positive BAL GM result was matched with three patients with true-negative BAL GM result, and the risk factors for false-positive BAL GM results were determined.
Results: Of 359 enrolled patients, 22 (6%) were diagnosed with IPA (1 proven, 17 probable, and 4 possible). Of the 22 patients with IPA, 17 (77%) had already received antifungal agents before the BAL GM assay was conducted. At an index cutoff value of ≥0.5, the BAL GM assay had a sensitivity of 64% (95% CI 41%-83%) and a specificity of 89% (95% CI 85%-92%). However, at an index cutoff value of ≥0.2, the BAL GM assay had a sensitivity of 86% (95% CI 65%-97%) and a specificity of 74% (95% CI 69%-79%). Of the 52 patients with positive BAL GM assay (≥0.5), 25 (7%) were false-positives. Univariate and multivariate analysis revealed that treatment with piperacillin-tazobactam or ampicillin-sulbactam was associated with false-positive BAL GM results.
Conclusions: The BAL GM assay appears promising for the diagnosis of IPA. However, treatment with certain antibiotics may interfere with the results of the BAL GM assay.
Copyright © 2010 The British Infection Association. Published by Elsevier Ltd. All rights reserved.