1701. Differences in Diagnostic Performance of β-d-Glucan Testing in Patients with Varying Degrees of Susceptibility to Invasive Fungal Infections

BACKGROUND: Usefulness of β-d-glucan (BDG) testing in high-risk patients for invasive fungal infection (IFI) diagnosis has been well demonstrated. However, data on its usefulness in patients without risk factors are limited. We evaluated differences in the diagnostic performance of BDG testing in patients with varying degrees of susceptibility to IFI. METHODS: From April 2017 to May 2018, all consecutive patients (≥18year-old) who were performed BDG testing (Beijing Gold Mountainriver Tech) were enrolled. Patients were classified into three groups: Group A for patients with host factors defined by 2008 European Organization for Research and Treatment of Cancer-Mycoses Study Group diagnostic (EORTC-MSG) criteria, Group B for patients with malignancy receiving recent chemotherapy within 1 month without host factors, and Group C for others. Cases of proven and probable IFI defined by EORTC-MSG criteria, Pneumocystis pneumonia and all fungemia were considered as true IFIs. Sensitivity, specificity, positive and negative predictive value (PPV and NPV) were calculated with a cut-off value for positivity ≥80 pg/mL. RESULTS: Among 473 eligible patients, 190, 142, and 141 patients were classified into group A, B, and C, respectively. Rates of true IFI were significantly different in each group (57/190, 19/142, and 10/141 in each group, P < 0.001). Sensitivities were 0.83, 0.68, and 0.70 and specificities were 0.62, 0.59, and 0.63 in group A, B, and C, respectively. PPVs were considerably different among three groups (PPV for 0.48, 0.20, and 0.12; NPV for 0.89, 0.92 and 0.97 in each group, respectively). CONCLUSION: The BDG test is a useful assay for IFI diagnosis; however, the clinical interpretation should be different by patient risks. Whereas BDG testing could be considered as a tool for predicting IFI in high-risk patients, it only could be a tool for excluding IFI in patient without risk factors. DISCLOSURES: All authors: No reported disclosures.