Diagnostic accuracy of the 1,3-β-D-glucan test for pneumocystis pneumonia in a tertiary university hospital in Denmark: A retrospective study.

1,3-β-D-glucan (BG), a cell-wall component of most fungi including Pneumocystis (PC), is recommended by international guidelines for screening for pneumocystis pneumonia (PCP) in hematologic patients. We retrospectively validated the BG test in our tertiary university hospital. Forty-five patients (median age 53 years, 33% female) tested for PC by polymerase chain reaction (PCR) and/or immunoflourescence (IF)-microscopy with a stored blood sample within ±5 days of the PC test were tested by the Fungitell (cutoff <60 and >80 pg/ml). Cases had symptoms and radiology compatible with PCP and positive IF-microscopy (proven PCP, n = 8) or positive PCR (probable PCP, n = 10). Controls had no compatible symptoms/radiology and negative tests for PC on conventional testing (no PCP, n = 24), or positive PCR/IF-microscopy (colonized, n = 3). Median BG-levels were 1108 pg/ml (proven PCP), 612 pg/ml (probable PCP), 29 pg/ml (colonized), and 48 pg/ml (controls, P < 0.001). Compared to the PCP case/control classification, the BG test showed sensitivities of 83-89% and specificities of 64-74%, positive likelihood ratio (LR) of 3.2 and negative LR of 0.23 at recommended cutoff and moderate agreement between tests. Optimal cutoff was ≥73 pg/ml. In PCR-positive cases, the agreement between the BG test and IF-microscopy was 78-89% with fair/moderate agreement. Elevated BG levels were seen in controls with probable invasive fungal infections (n = 4), hemodialysis, bacterial infections and/or betalactams. To conclude, 11% of patients with PCP would be missed if the BG test had been used for diagnosing PCP. Specificity was moderate. Among PCR-positive patients, the BG test identified more cases than IF-microscopy. BG testing is potentially helpful but sensitivity is insufficient to exclude PCP.