Gomori Methenamine Silver Stain on Bronchoalveolar Lavage Fluid Is Poorly Sensitive for Diagnosis of Pneumocystis jiroveci Pneumonia in HIV-Negative Immunocompromised Patients and May Lead to Missed or Delayed Diagnoses

Arch Pathol Lab Med. 2020 Jan 6. doi: 10.5858/arpa.2019-0394-OA. Online ahead of print.

Abstract

Context.—: Direct visualization of Pneumocystis jiroveci organisms, using Gomori methenamine silver (GMS) staining in bronchoalveolar lavage fluid (BAL), is a historical gold standard that has been widely used for the diagnosis of P jiroveci pneumonia (PJP). However, the stain may be less sensitive in human immunodeficiency virus (HIV)-negative immunocompromised patients owing to a lower burden of organisms.

Objectives.—: To assess the sensitivity of the GMS stain on BAL fluid for the diagnosis of PJP in HIV-negative immunocompromised patients as compared to HIV-positive patients.

Design.—: We conducted a retrospective review from 2012 to 2018 to identify immunocompromised patients (≥18 years old) who underwent bronchoscopy with BAL GMS staining for the diagnosis of PJP. To assess for sensitivity, we sought to identify BAL GMS-positive cases and BAL GMS-negative cases of PJP. The BAL GMS-negative cases were categorized into proven and probable PJP.

Results.—: We identified 45 adult immunocompromised patients with proven and probable PJP, including 24 HIV-negative (11 BAL GMS-positive and 13 BAL GMS-negative) and 21 HIV-positive cases (all were BAL GMS-positive). The sensitivity of BAL GMS for the diagnosis of PJP in HIV-negative immunocompromised patients was 11 of 24 (46%) versus 21 of 21 (100%) in HIV-positive patients (CD4: median, 10 cells/mL; range, 3-300 cells/mL). Delayed or missed diagnoses were seen in 3 cases of BAL GMS-negative PJP. Re-examination of BAL GMS slides showed rare P jiroveci cysts in 1 case.

Conclusions.—: BAL GMS has poor sensitivity for PJP in HIV-negative immunocompromised patients. Using BAL GMS as a sole method for PJP may result in missed or delayed diagnoses in this population.