Local retrospective analysis of galactomannan cut-off values in bronchoalveolar lavage fluids for diagnosis of invasive aspergillosis.

Galactomannan antigen (GM) testing has been used for decades to screen immunocompromised patients for invasive aspergillosis (IA). Recent publications suggested that using a higher cut-off value than 0.5 in bronchoalveolar lavage fluid (BALF) could be more discriminant for hematology patients. We retrospectively analyzed the values of GM in BALF over 7 years (from 2010 to 2016). Performance indicators of the GM in BALF, according to three different cut-off values (0.5, 0.8, 1.5), were calculated using Stata 14.1. IA classification for hematology patients was based on European Organization for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) criteria, as defined in 2008. A number of 716 GM were performed on BALF from 2010 to 2016 (597 patients) and 66 were positive (> 0.5). Among these 597 patients, 27 IA were diagnosed, 13 with a positive GM in BALF, 9 with a negative GM in BALF, and 5 unclassified IA (ICU patients). The analysis of performance indicators, based on our local data, did not demonstrate any significant difference using a higher cut-off value of GM in BALF. This result may be explained by the local recruitment of patients and by pre-analytic variations during BALF realization.