Quantification of 1,3-β-d-glucan by Wako β-glucan assay for rapid exclusion of invasive fungal infections in critical patients : a diagnostic test accuracy study

Objectives: Rapid and reliable exclusion of invasive fungal infections (IFI) by markers able to avoid unnecessary empirical antifungal treatment is still a critical unmet clinical need. We investigated the diagnostic performance of a newly available β-d-Glucan (BDG) quantification assay, focusing on the optimisation of the BDG cut-off values for IFI exclusion. Methods: BDG results by Wako β-glucan assay (lower limit of detection [LLOD] = 2.16 pg/mL, positivity ≥ 11 pg/mL) on two consecutive serum samples were retrospectively analysed in 170 patients, admitted to haematological wards (N = 42), intensive care units (ICUs; N = 80), or other wards (N = 48), exhibiting clinical signs and/or symptoms suspected for IFI. Only patients with proven IFI (EORTC/MSG criteria) were considered as true positives in the assessment of BDG sensitivity, specificity and predictive values. Results: Patients were diagnosed with no IFI (69.4%), proven IFI (25.3%) or probable IFI (5.3%). Two consecutive BDG values < LLOD performed within a median of 1 (interquartile range: 1-3) day were able to exclude a proven IFI with 100% sensitivity and negative predictive value (primary study goal). Test's specificity improved by using two distinct positivity and negativity cut-offs (7.7 pg/mL and LLOD, respectively), but remained suboptimal in ICU patients (50%), as compared to haematological or other patients (93% and 90%, respectively). Conclusions: The classification of Wako's results as negative when < LLOD, and positive when > 7.7 pg/mL, could be a promising diagnostic approach to confidently rule out an IFI in both ICU and non-ICU patients. The poor specificity in the ICU setting remains a concern, due to the difficulty to interpret positive results in this fragile population.