Comparison of staining methods and a nested PCR assay to detect Histoplasma capsulatum in tissue sections.

To optimize diagnosis of histoplasmosis in tissue sections, 30 spleen specimens from mice, experimentally infected with Histoplasma capsulatum, were examined by H&E, Grocott stain, anti–bacille Calmette-Guerin antibody immunostain, Fungiqual A fluorochrome stain (Drs Reinehr and Rembold, Kandern, Germany), and a nested polymerase chain reaction (PCR) assay. Results were compared with the tissue burden determined by quantitative culture. By applying logistic regression, the nested PCR assay was the most sensitive method, but not significantly more sensitive than the Grocott stain. The 50% quantile to achieve a positive result was determined to be 3 colonyforming units per milligram of spleen tissue for the PCR assay, 11 for the Grocott stain, 27 for the fluorochrome stain, 190 for immunostaining, and 533 for the H&E stain. The Grocott and fluorochrome stains did not differ significantly in detecting fungal elements. The PCR assay unambiguously identified H capsulatum in tissue sections. The dimorphic fungus Histoplasma capsulatumis cosmopolitan, being endemic mainly in the Americas and on the African continent. 1 Outside endemic areas, histoplasmosis must be considered in diagnoses for patients with a travel history, but the diagnosis often is made accidentally by histologic examination and usually at a time when no specimen is left to grow the pathogen. In addition, culturing this hazardous fungus is restricted to biosafety level III laboratories in some European countries, and experience in isolating the pathogen is limited in nonendemic areas. Staining methods and, more recently, molecular tools have been described for the detection and identification of H capsulatum in tissue samples. 2-7