Purpose: To assess the performance of (1→3)-β-D-glucan (BDG) and Candida albicans germ tube antibody (CAGTA) for the diagnosis of invasive candidiasis (IC) in a prospective cohort of 107 unselected, non-neutropenic ICU patients.
Methods: BDG (cutoff positivity ≥80 pg/mL) and CAGTA (cutoff positivity ≥1/160) assays were performed twice a week. Confounding factors included amoxicillin-clavulanate and piperacillin-tazobactam treatments, recent surgery, Gram-positive bloodstream infection, renal replacement therapy, and enteral nutrition. Patients were classified as neither colonized nor infected (n = 29), Candida spp. colonization (n = 63) (low grade, n = 32; high grade, n = 31), and invasive candidiasis (IC) (n = 15).
Results: BDG levels were higher in patients with IC and high-grade colonization than in the remaining groups (p = 0.012), and two consecutive measurements ≥80 pg/mL discriminated IC from the remaining groups (sensitivity 80%, specificity 75.7%). For the discrimination between IC and Candida spp. colonization, the AUC for the maximum value of BDG was 0.667 (95% CI 0.544-0.790) and for the maximum value of CAGTA 0.545 (95% CI 0.395-0.694). Significant changes of BDG and CAGTA kinetics in IC patients treated with antifungals were not observed. In patients neither colonized nor infected or with low-grade Candida spp. colonization, none of the confounding factors was associated with a significant increase in BDG positivity.
Conclusions: Two consecutive BDG levels ≥80 pg/mL allowed discrimination among IC and high-grade colonization. Systemic antifungal therapy could not be monitored with biomarker kinetics, and BDG levels were not subject to interference by confounding factors in either colonized or infected patients or with low-grade colonization.