RESUMO
BACKGROUND: Galactomannan (GM) detection in serum samples has been used to diagnose invasive aspergillosis (IA). Limited sensitivity has been observed in lung transplant recipients, for whom bronchoalveolar lavage (BAL) testing has been advocated. Because airway colonization with Aspergillus species occurs frequently in these patients, false-positive GM results have been reported if the cutoff validated for sera is used (i.e., 0.5). METHODS: Herein, we prospectively studied BAL fluid samples from 60 lung transplant patients to determine the optimal cutoff for BAL GM testing. Only one sample per patient was studied. BAL samples were vortexed and processed according to the manufacturer's instructions for serum samples. Sensitivity, specificity, and likelihood ratios were calculated in reference to proven or probable IA cases using receiver operating characteristic analysis. RESULTS: Eight patients had IA during the study (incidence 13.3%), including four patients with proven IA. Aspergillosis increased 5-fold the risk of death in lung transplant recipients. The positive predictive value of a positive BAL GM test at the 0.5 cutoff was low (24.2%). Raising the cutoff improved test specificity without compromising sensitivity. The best cutoff was defined at 1.5 (sensitivity 100% and specificity 90.4%). CONCLUSIONS: This study reinforces the importance of BAL GM testing in lung transplant recipients, particularly to exclude the diagnosis of IA. To minimize the frequency of false-positive results, a higher test cutoff should be applied to BAL samples, in comparison with serum samples.