LAUSR

Introduction The diagnostic standard for Pneumocystis jirovecii pneumonia (PCP) is direct microscopic identification; however, in recent years, polymerase chain reaction (PCR) from bronchoalveolar lavage (BAL) samples to detect Pneumocystis nucleic acids has proven to be more sensitive and specific. Sputum samples have been presumed inferior to bronchoscopic samples secondary to variability and adequacy of sample collection. We observed several cases of positive sputum PCP-PCR results with negative PCP-PCR BAL results. The aim of the current study was to further characterize the clinical setting and outcomes in patients with positive sputum PCP-PCR samples and negative BAL PCP-PCR samples. Methods We identified all patients who underwent P. jirovecii-PCR testing at Mayo Clinic between 2011 and 2016. Patients with a positive sputum and negative BAL sample collected within a 14-day time frame were identified and underwent further chart review for demographics, immunocompromised state, and clinical outcome. Results From 2011 to 2016, 4431 respiratory samples from 3021 unique patients were tested for the presence of P. jirovecii by PCR. Fifty-five samples (1.2% of all samples collected) belonging to 24 unique patients (0.79% of patients tested) were identified as having a positive and negative sample collected within 14 days. Of these 24 patients, 10 (46%) patients had a positive sputum or tracheal secretion sample with negative BAL or bronchial washings. Out of these 10 patients, 8 were immunocompromised and 9 underwent treatment for PCP with 6 patients improving. Conclusion Our results suggest that discordant P. jirovecii-PCR testing results from sputum and bronchoscopic specimens are an infrequent occurrence. Patients with positive P. jirovecii-PCR sputum/tracheal secretion samples and negative bronchoscopic samples appear to be clinically infected and respond to PCP treatment. Sputum P. jirovecii-PCR testing may be a viable alternative to invasive testing.