Performance Characteristics of Galactomannan and β-d-Glucan in High-Risk Liver Transplant Recipients.

BACKGROUND The utility of Aspergillus galactomannan (GM) and β-D-glucan (BG) in liver transplant recipients remains uncertain. METHODS As part of a randomized, double-blind trial of antifungal prophylaxis in liver transplant recipients at risk for invasive fungal infections (IFIs), GM and BG were assessed in 199 patients at baseline (enrollment) and weekly thereafter for the duration of study drug. Receiver operating characteristic (ROC) analysis was used to evaluate the accuracy of these for the diagnosis of IFIs. RESULTS Overall, 46.4% of the patients at baseline had positive GM (index ≥ 0.5) and 89.6% had BG of 80 pg/mL or greater with BG level of 500 pg/mL or greater in 31.8%. Patients with invasive aspergillosis (IA) (3/3) had positive GM at baseline as did 45.5% of those without IA (P = 0.098); the area under the ROC curve for the diagnosis of IA was 0.77 (fair test, ie, good sensitivity but poor specificity). Using BG cutoff of 80 pg/mL or higher, 100% (12/12) of the patients with IFI had positive baseline BG and as did 88.9% (160/180) of those without IFI (P = 0.618); the area under the ROC curve for predicting IFIs was 0.56 (poor test). In multivariate analyses, GM positivity was associated with study site (P = 0.041), and BG positivity with renal replacement therapy (P = 0.05) and study site (P = 0.01). The GM and BG levels declined over time; positivity at subsequent time points was lower in comparison with baseline (P < 0.001). CONCLUSIONS The GM and BG tests had significant center variability and limited accuracy for the diagnosis of IFIs in high-risk liver transplant recipients.