Invasive fungal infections (IFIs) are an increasing problem in intensive care units (ICUs), and conventional diagnostic methods are not always reliable or timely enough to deliver appropriate antimicrobial therapy. The dosage of fungal antigens in serum is a promising diagnostic technique, but several confounding factors, such as treatment with immunoglobulins (Ig), albumin, or antifungals, could interfere with the correct interpretation of the (1,3)-beta-D-glucan (BG) assay. This study assessed the reliability of the BG assay and the influence of timing and dosage of major confounding factors on circulating levels of IFI biomarkers. 267 ICU patients who underwent a BG assay were retrospectively studied. The timing and dosage of albumin, use of azole treatment, and infusions of intravenous IgG, red blood cells, concentrated platelets, and frozen plasma were analyzed to find possible correlations with the BG results. The sensitivity and specificity of the BG assay were calculated. The BG test in serum showed high sensitivity (82.9 %) but low specificity (56.7 %). The optimal cut-off for the test was 95.9 pg/mL. The mean BG level in proven invasive candidiasis was around 400 pg/mL. The only factor that was found to significantly confound (p < 0.05) the diagnostic performance of the BG assay was the administration of more than 30 g of albumin within 2 days prior to BG testing. The BG assay remains a useful diagnostic test in ICU patients and the levels of BG are useful in evaluating the positive predictive value of this biomarker. The only confounding factor in our study was the use of albumin.