Fungal biomarkers in HIV-associated disseminated histoplasmosis: a multicenter diagnostic accuracy study on the Guiana shield

Abstract

Objectives

Diagnosis of HIV-associated histoplasmosis remains challenging. Our objective was to compare the performances of (1→3)-β-D-Glucan (BDG) and aspergillus galactomannan (GM) antigen for the diagnosis of HIV-associated histoplasmosis.

Methods

We performed a diagnostic accuracy study using frozen primary serum specimens issued from consecutive hospitalized people living with HIV (PLWH) and blindly tested for BDG and GM using Fungitell^Ⓡ and Platelia^TM Aspergillus, respectively.

Results

We included 121 sera with 92 HIV-associated histoplasmosis cases and 29 negative controls. At thresholds of 150 pg/ml and 0.5 for BDG and GM, the sensitivity and specificity were 95% (85-100) vs 90% (77-100) and 52% (34-70) vs 83% (69-97), respectively. The receiver operating characteristics (ROC) curves showed area under the curves of 0.82 (0.68-0.91) vs 0.92 (0.80-0.98) for BDG and GM, respectively. Post-test probabilities showed best performances at lowest thresholds for a negative testing of BDG and GM and at the 0.7 threshold for a positive GM test.

Conclusions

If BDG alone may rule out histoplasmosis when negative, GM alone, either positive or negative, showed the best performances for the diagnosis of histoplasmosis. Given the poorer performances of BDG and GM than Histoplasma antigen detection assays commercially available, they should be considered as an alternative in settings where Histoplasma antigen detection assays remain unavailable. However, this study essentially provides insights in the performances of fungal biomarkers in disseminated histoplasmosis and does not represent recommendations for best practices.