Background: With standard diagnostic methods, the etiologic pathogen of community acquired pneumonia (CAP) is detected in ≤ 50% hospitalized CAP patients. In our previous studies using a diagnostic “bundle”, we were able to detect etiologic pathogens in ≥ 70 % of the patients. Our bundle consisted of a nasopharyngeal swab for Biofire film array, that detects 9 viral genera, nucleic acid amplification testing (NAAT) for Streptococcus pneumoniae (S. pneumo) and Staph. Aureus, urine antigens for Legionella, and S. pneumo, and cultures of sputum and blood. The current study is designed to determine if a next generation PCR Biofire Multiplex Investigational Sputum Pneumonia Panel (BIPP) will improve detection of candidate pathogens.
Methods: Starting in January 2017, we compared the standard bundle to BIPP. Herein we report results in the first 120 patients admitted with CAP. The BIPP requires sputum or a sputum equivalent. The panel includes probes for 9 viral Genera, 18 bacterial pathogens, and 7 antibiotic resistance genes. Serum procalcitonin (PCT) levels were used to separate bacterial colonization from bacterial invasion.
Results: Of the first 120 patients 64 were evaluable. Patients with inadequate data collection or an alternative diagnosis were non-evaluable. Influenza was detected in 19 of the 64 (29.7%) patients with the standard bundle and 29 of 64 (45.3%) patients with the BIPP, p < 0.07. Similarly, 38 potential bacterial pathogens were detected with the standard bundle vs 52 with the addition of BIPP. Detection of both potential bacterial and viral pathogens occurred in 19 of 64 (29.7%) patients with the standard bundle as compared to 41 of 64 (64.1%) patients with the BIPP, p < .0.0001. The median PCT level in 8 virus only patients was 0.07 ng/ml. In comparison, there were 14 patients wherein BIPP detected a mixture of virus plus bacteria (V+B) but the median PCT was 0.09, similar to the virus only patients, p = 0.70. In contrast, 23 patients had V + B detection with a PCT median value of 3.4 ng/ml, which was significantly higher than virus only patients, p = 0.001, and similar to bacterial only patients, p = 0.6. Hence, if a concomitant PCT was < 0.25 ng/ml, the detected bacteria were interpreted as colonizing and not invading. Overall the standard bundle detected a pathogen in 84% of evaluated patients as compared to 98% when BIPP results were included.
Conclusions: With a rapid diagnostic bundle, a candidate etiologic pathogen can be detected in 80% of the patients with CAP and using the BIPP platform, the detection rate increases to 98%. Moreover, PCT levels successfully adjudicated whether detected bacteria were colonizing or invading. Thus, the increase in rapid pathogen detection coupled with PCT levels offers great potential for improvements in antibiotic stewardship, including reduction in overuse of antibiotics and allowing for targeted treatment in patients with CAP.