Introduction
The availability of polymerase chain reaction (PCR)-based pneumonia (PN) panels in clinical practice is revolutionary to organism identification speed and has been shown to significantly reduce unnecessary empiric antibiotic use. Reference Niederman and Torres1 Buchan et al reviewed patients with PN panel results, identifying opportunities for antibiotic adjustment in 71% of patients, with antibiotic discontinuation or de-escalation occurring in 48% of patients. Reference Buchan2 Esplund et al examined the impact of a PN panel on an intensive care unit (ICU) population, finding opportunity for antibiotic changes in 75% of cases, an implementation rate of 54%, and shorter times to antibiotic changes by 29 hours with a 2-day reduction in median antibiotic days of therapy. Reference Esplund3 Markussen et al evaluated the utilization of a PN panel within an emergency department, finding participants with PN panel results were not only 3.5 times more likely to receive pathogen-directed treatment, but they also received it 9 hours faster than patients without a PN panel. Reference Markussen4
Indeed, there is a growing body of literature regarding the opportunity of a PN panel to enhance patient care; however, more data are needed to fully elucidate the best use of this emerging technology. The purpose of this study was to evaluate the impact of incorporating a PCR-based PN panel into the management of PN and to identify opportunities to optimize the interpretation and implementation of opportunities created by the panel results.
Methods
This retrospective, single-center study was conducted at a 400-bed community hospital. In December 2022, the BioFire® FilmArray® PN panel (33 targets, ∼1 hour turnaround) 5 was introduced as an orderable diagnostic test for acute care respiratory samples. The PN panel results published between December 2022 and September 2023 were extracted from the electronic medical record by laboratory informatics. Patients were screened and excluded if pregnant, incarcerated, found to have positive extrapulmonary cultures, or received systemic antibiotics more than 72 hours prior to the PN panel collection time.
The primary outcome was the incidence of actionable PN panel results. An actionable PN panel result was defined as a result that indicated an opportunity for antibiotic escalation and/or streamlining for PN. PN panel results that failed to identify atypical bacteria, methicillin-resistant S. aureus (MRSA), Pseudomonas aeruginosa (PSA), or multi-drug resistance gene markers when antibiotics targeting those pathogens were being utilized were designated as streamlining opportunities, as were PN panel results that did not identify any pathogens when antibiotics were being utilized. Secondary outcomes included the implementation of actionable PN panel results, time to implementation, hospital length of stay, all-cause in-hospital mortality, readmission within 30 days of discharge due to any cause, and intervention by the antimicrobial stewardship program pharmacist. Implemented actionable panel results were defined as any modification of the antibiotic regimen that was concordant with the opportunity created by the actionable PN panel. Durations of targeted therapy for MRSA and PSA were evaluated for patients with the respective streamlining opportunity.
Statistical analysis was performed using the Kruskal–Wallis test for continuous variables and either the chi-squared or Fisher’s Exact test for categorical variables. Interquartile ranges from the first quarter to the third quarter were reported for time-dependent variables in addition to the median value. A prespecified two-sided P value of <.05 identified statistical significance. The study was approved by the local institutional review board.
Results
Of 507 patients screened, 384 were included in the analysis. Reasons for exclusion were positive extrapulmonary culture (n = 65) and receipt of antibiotics 72 hours or more before PN panel (n = 58). Actionable results were identified in 333 cases (87%). Baseline characteristics for the entire cohort as well as patients with actionable versus non-actionable results are displayed in Table 1. Concordant PN panel and culture results were found in 361 cases (94%). Of the 23 discordant results, 22 (96%) were PN panels that identified pathogens that did not grow on culture (including 9 related to MRSA and 4 related to PSA). One discordant result was identified for a negative PN panel, but PSA grew on culture. Pharmacist intervention was found for 24 cases (6%).
Baseline characteristics including data on cases with actionable versus non-actionable results
Note. IQR, interquartile range; ICU, intensive care unit; PN, pneumonia; MRSA, methicillin-resistant S. aureus.
Classification of actionable results and those that were implemented is provided in Table 2. For implemented actionable results, the median time from PN result to antimicrobial regimen modification was 22 hours overall, with a median of 23 hours for streamlining opportunities and 15 hours of escalation opportunities. With respect to the time of the culture result, the opportunities identified by the PN panel results were implemented by a median time of 24 hours before culture results were finalized. Implementation of streamlining and escalation opportunities were implemented in a median time of 22 hours and 45 hours, respectively, before culture results were finalized.
Classification of actionable pneumonia panel results and those that were implemented
Note. PSA, Pseudomonas aeruginosa; MRSA, methicillin-resistant S. aureus; MDR, multi-drug resistance.
1 Specific streamlining and escalation opportunities were counted per instance, not per patient.
Patients with actionable results that were implemented had a median hospital length of stay of 6 days versus 3 days for those with results not implemented (P < 0.01). All-cause mortality was 3% (n = 7) for patients with implemented results compared to 2% (n = 2) for those where it was not implemented (P = 0.72). Readmission within 30 days occurred in 29 patients (n = 13%) who had actionable results implemented versus 7 patients (6%) for those where it was not implemented (P = 0.19). Comparing results implemented within 24 hours (n = 118) to those implemented after 24 hours (n = 100), the median hospital length of stay was 4 days when implemented within 24 hours versus 6 days when implemented after 24 hours. Comparing patients with interventions implemented versus not implemented, no difference was found in all-cause mortality or readmission rates.
Discussion
This analysis found that a high percentage of PN panels result in actionable opportunities for antimicrobial stewardship. These actionable results mostly created opportunities for streamlining of atypical bacteria, PSA, and MRSA coverage, with occasional opportunities to escalate antibiotic therapy.
The patient population in this study was predominantly elderly and white, not requiring ICU-level care or mechanical ventilation at the time of PN panel collection. Internal medicine was the most frequent ordering provider type, and sputum specimens were the large majority of collected specimen types, often collected within 48 hours of hospital admission. Population characterization is important, as the lack of clinical benefit found within this study does not necessarily reflect a potential lack of benefit in a particular subgroup. For example, previous studies have demonstrated high sensitivity and specificity of the PN panel with bronchoalveolar lavage samples, while results from this study support the assertion that the PN panel maintains a high level of accuracy and reliability when utilized with sputum samples. Reference Buchan2
This study is limited by the single-center observational design, dependence on accurate specimen collection, and lack of focused education to prescribers. A key component to introducing new tests is ensuring that actionable opportunities are implemented. Although approximately 9 in 10 tests had an actionable result, only 65% were implemented, and only about half were implemented within 24 hours. Consistent with suggestions from existing literature, incorporation of PN panel results into routine antimicrobial stewardship activities completed by pharmacists (including provider education) will be one next step, while implementing a clinical decision support system will be a second step. Reference Timbrook6 A clinical benefit is more likely to be observed if opportunities are acted upon more frequently and expeditiously.
Conclusion
Implementation of a PN panel can produce a large amount of actionable results, which can improve antibiotic streamlining and escalation, but implemented opportunities may not globally improve clinical outcomes.
Acknowledgments
The authors thank Md Ashfaq Ahmed, PhD, for his involvement in the statistical analysis of this project.
Financial support
None reported.
Competing interests
Luke Manda – No conflicts of interest.
Anthony Wasielewski – No conflicts of interest.
Nishika Patel – No conflicts of interest.
Timothy P. Gauthier – Reports the following potential conflicts of interest: AbbVie Pharma (advisor/consultant), Antimicrobial Therapy, Inc (collaboration/advertising), DoseMe Pty Ltd (advisor/consultant, collaboration/advertising), Ferring Pharma (advisor/consultant), Firstline Mobile Health (advisor/consultant, collaboration/advertising), Gilead Pharma (advisor/consultant), GoodRx (advisor/consultant), GSK Pharma (advisor/consultant), LearnAntibiotics.com & IDstewardship.com (ownership interest), Melinta Pharma (advisor/consultant), Pattern Biosciences (advisor/consultant), Pfizer Pharma (advisor/consultant), and ProCE (honoraria). Other authors report no potential conflicts.
Open access
