Concordance of BioFire Pneumonia Panel with Traditional Respiratory Bacterial Culture Results

RATIONALE: Pneumonia causes significant morbidity and mortality and is a leading cause of death in the U.S. Treatment requires organism identification to minimize resistance and antimicrobial side effects. Culture techniques are slow, potentially prolonging antibiotic exposure, and can be negative even in clinically suspected infection. Molecular assays, such as the BioFire Pneumonia Panel, expedite organism identification by DNA detection and screens for 8 genetic markers of resistance. The BioFire Panel has >96% sensitivity and >97% specificity for BAL and sputum. We evaluated (1) concordance between bacterial cultures and the BioFire panel for organism identification and antimicrobial resistance; and (2) if utilization of the panel was clinically appropriate. METHODS: We performed a retrospective analysis of 135 BioFire Pneumonia Panel results during a 1-month period. Concordance of organism identification and resistance patterns was examined. Internal utilization of the pneumonia panel was assessed based on high pre-test probability of PNA from hypoxemia, severity of illness, and radiologic abnormalities. Appropriateness was assigned as indicated, weak indication or not indicated by clinical and radiological data. RESULTS: 131/135 pneumonia panels ordered were indicated (97.0%), 3 were weakly indicated (2.2%) and 1 was not indicated (0.7%). Both the pneumonia panel and culture results were positive in 26/135 (19%) patients. The pneumonia panel was positive and the culture results were negative in 32/135 (24%) patients. Both the pneumonia panel and culture results were negative in 57/135 (42.2%) patients. The pneumonia panel was negative and culture results were positive in 1/135 (0.7%) patients. No cultures were ordered in 19/135 (14%) patients. 14 patients had resistant bacteria detected by either the pneumonia panel or cultures. 4/14 (28.5%) demonstrated concordant resistance between panel and culture results. 6/14 (42.8%) patients demonstrated resistance detected only in the panel but not in culture. None of the patients had resistance patterns detected only by culture. 4/14 (28.5%) patients demonstrated resistance genes in the panel but no culture data was available for comparison. CONCLUSIONS: The majority of BioFire panels ordered during a 1month period were clinically indicated. Excluding patients with no culture data for comparison to the BioFire results, there was organism identification concordance in 70% of cases. The panel detected a potential pathogen in an additional 27% of cases, where culture was not informative. Excluding patients with no culture data for comparison, resistance patterns were concordant in 40% of patients. Combining the BioFire panel and culture results can maximize organism identification and patient benefit.