UC San Diego

Invasive pulmonary aspergillosis (IPA) is a life-threatening disease that affects mainly immunocompromised hosts. Galactomannan testing from serum and bronchoalveolar lavage fluid (BALF) represents a cornerstone in diagnosing the disease. Here, we evaluated the diagnostic performance of the novel Aspergillus-specific galactomannoprotein (GP) enzyme-linked immunosorbent assay (ELISA; Euroimmun Medizinische Labordiagnostika) compared with the established Platelia Aspergillus GM ELISA (GM; Bio-Rad Laboratories) for the detection of Aspergillus antigen in BALF. Using the GP ELISA, we retrospectively tested 115 BALF samples from 115 patients with clinical suspicion of IPA and GM analysis ordered in clinical routine. Spearman's correlation statistics and receiver operating characteristics (ROC) curve analysis were performed. Optimal cutoff values were determined using Youden's index. Of 115 patients, 1 patient fulfilled criteria for proven IPA, 42 for probable IPA, 15 for putative IPA, 10 for possible IPA, and 47 did not meet criteria for IPA. Sensitivities and specificities for differentiating proven/probable/putative versus no IPA (possible excluded) were 74% and 96% for BALF GP and 90% and 96% for BALF GM at the manufacturer-recommended cutoffs. Using the calculated optimal cutoff value of 12 pg/mL, sensitivity and specificity of the BALF GP were 90% and 96%, respectively. ROC curve analysis showed an area under the curve (AUC) of 0.959 (95% confidence interval [CI] of 0.923 to 0.995) for the GP ELISA and an AUC of 0.960 (95% CI of 0.921 to 0.999) for the GM ELISA for differentiating proven/probable/putative IPA versus no IPA. Spearman's correlation analysis showed a strong correlation between the ELISAs (rho = 0.809, P < 0.0001). The GP ELISA demonstrated strong correlation and test performance similar to that of the GM ELISA and could serve as an alternative test for BALF from patients at risk for IPA.