Diagnostic Yield of Bronchoalveolar Lavage and Bronchoscopic Lung Biopsy for Detection of Pneumocystis carinii

Abstract

To assess the need to perform a bronchoscopic lung biopsy (BLB) in addition to bronchoalveolar lavage (BAL) to obtain a definitive diagnosis of Pneumocystis carinii pneumonia. We retrospectively reviewed the results of concurrently collected paired BAL and BLB specimens to determine the diagnostic yield of both methods for the detection of P. carinii organisms. During a 3-year period, the BAL fluid specimens stained with a commercially available direct immunofluorescence monoclonal antibody (DFA) reagent and the BLB specimens stained with Grocott methenamine-silver nitrate (GMS) were assessed for the presence of P. carinii. BAL fluid was routinely collected from multiple sites and combined into a single specimen for testing. During the 3-year period of study, 119 patients were identified who had paired BAL fluid and BLB specimens tested for the presence of P. carinii. Of the 119 patients, 16 had either BAL fluid that could not be interpreted or BLB tissue that was inadequate. Of the other 103 patients, 21 had P. carinii pneumonia. The sensitivity of the DFA method on BAL fluid and of the GMS method on BLB was 95% and 43%, respectively. For detection of P. carinii, the diagnostic yield is significantly higher for DFA-stained BAL specimens than for GMS-stained BLB specimens.