744. Clinical Performance of Real-time PCR in the Diagnosis of Pneumocystis jirovecii Pneumonia compared with Immunofluorescence Assay

BACKGROUND: The laboratory diagnosis of Pneumocystis jirovecii pneumonia (PJP) has been traditionally based on microscopy techniques, which have suboptimal sensitivity and depends on the experience and skills of the microbiologist. Molecular detection assays based in PCR (Polymerase chain reaction) could improve sensitivity. Our aim was to evaluate the utility of real-time PCR in the diagnosis of PJP compared with IFA (Immunofluorescence assay) performed in different respiratory samples of patients with PJP suspicion for routine use in a clinical laboratory setting. METHODS: From September 2015 to April 2018, we studied by a real-time PCR targeting the large subunit of rRNA gene of P. jirovecii (PJ-PCR RealCycler PJIR kit Progenie Molecular) and Immunofluorescence assay (MONOFLUO P. carinii IFA BioRad) in all respiratory samples received for microbiological diagnosis of PJP. The definite clinical diagnosis of PJP was established by infectious disease physicians considering symptoms, radiological and laboratory findings. RESULTS: Overall, 302 samples were included (182 bronchoalveolar lavage, 67 sputum, 53 tracheal aspirates). PJ-PCR was positive in 51 (16.9%) and IFA in 11 (3.6%) of the patients with PJP. There were not IFA positive/PCR negative samples. Sensitivity, specificity, PPV and NPV for IFA were 26% (95%CI 15.9-39.6%), 100% (95%CI 98.5-100%), 100% (95% CI 77.2-100%) and 87.2% (95% CI 82.6-90.6%). Whereas, sensitivity, specificity, PPV and NPV for PCR was 92% (95%CI 81.2-96.8%), 98% (95% CI 95.4-99.2%), 90.2% (95% CI 79.0-95.7%) and 98.4% (95% CI 96.0-99.4%). PJ-PCR had sensitivity > 80% and specificity > 90% in all type of samples included. A definitive diagnosis of PJP was considered in 50 (16.6%) patients, including 4 (1.3%) cases with negative PJ-PCR. Five cases (9.8%) with positive PJ-PCR were considered as colonization. CONCLUSION: P. jirovecii PCR improves the sensitivity and NPV of PJP diagnosis respecting to IFA, regardless of respiratory sample type. Our results suggest that Microbiology laboratories should use PCR techniques to diagnose PJP better than IFA. DISCLOSURES: All Authors: No reported disclosures