Diagnostic accuracy of serum (1-3)-β-D-glucan for Pneumocystis jirovecii pneumonia: a systematic review and meta-analysis

Background: Pneumocystis jirovecii pneumonia (PJP) can be a life-threatening opportunistic infection in immunocompromised hosts. The diagnosis can be challenging, often requiring semi-invasive respiratory sampling. The serum 1,3-beta-D-glucan (BDG) assay has been proposed as a minimally invasive test for the presumptive diagnosis of PJP. Method: We carried out a systematic review and meta-analysis using articles in the English language published between January 1960 and September 2019. We estimated the pooled sensitivity and specificity of BDG testing using a bivariate random effects approach and compared test performance in human immunodeficiency virus (HIV) and non-HIV subgroups with meta-regression. Data from the pooled sensitivity and specificity were transformed to generate pre- and post-test probability curves. Results: Twenty-three studies were included. The pooled sensitivity and specificity of serum BDG testing for PJP were 91% (95%CI 87-94%) and 79% (95%CI 72-84%) respectively. The sensitivity in patients with HIV was better than in patients without (94%, 95%CI 91-96%) versus 86% (95%CI 78-91%) (p 0.02), with comparable specificity (83%, 95%CI 69-92% versus 83%, 95%CI 72-90%) (p 0.10). A negative BDG was only associated with a low post-test probability of PJP (<= 5%) when the pre-test probability was low to intermediate (<= 20% in non-HIV and <= 50% in HIV). Conclusions: Among patients with a higher likelihood of PJP, the pooled sensitivity of BDG is insufficient to exclude infection. Similarly, for most cases, the pooled specificity is inadequate to diagnose PJP. Understanding the performance of BDG in the population being investigated is therefore essential to optimal clinical decision-making. (C) 2020 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.