Abstract
Diagnosis of pulmonary aspergillosis (PA), a fungal disease caused by Aspergillus species, is challenging since symptoms are unspecific. The galactomannan (antigen secreted by Aspergillus species) test in bronchoalveolar lavage (BAL) fluid is a valuable diagnostic adjunct test in the diagnosis of PA. However, BAL collection is invasive and may not be suitable to severely ill patients. Sputum is non-invasive, easily collected, and lung specific and may be an alternative to BAL. The aim of this research was to retrospectively evaluate the utility of sputum galactomannan in the diagnosis of pulmonary aspergillosis in patients with chronic respiratory diseases and to estimate the sputum galactomannan cut-off value. We collected data from patients with clinical suspicion of pulmonary aspergillosis who had sputum galactomannan, culture, and Aspergillus IgG tests performed within four weeks. Sputum galactomannan was validated against the clinical diagnosis of aspergillosis, Aspergillus culture, and Aspergillus IgG tests. In total, 218 patients met inclusion criteria. Overall, sputum GM showed satisfactory agreement with clinical diagnosis of aspergillosis, Aspergillus culture, and Aspergillus IgG. When a receiver operating characteristic curve was constructed using Aspergillus culture/IgG and clinical diagnosis, the same cut-off (CO) of 0.71 (AUC: 0.83; CI: 0.69-0.86, p < 0.001) was determined. Against clinical diagnosis, sputum GM gave sensitivity and specificity of 70% and 71%, respectively. Sensitivity of 77% and specificity of 78% were found when sputum GM was evaluated against Aspergillus culture/IgG. In conclusion, this study showed that sputum galactomannan antigen testing has utility in the diagnosis of chronic forms of pulmonary aspergillosis and further prospective validation is indicated.