1. IntroductionCommunity acquired meningitis and encephalitis are severe, potentially fatal infections of the central nervous system (CNS) that require a timely diagnosis and rapid administration of antimicrobials [1,2,3]. Patients who contract viral or bacterial meningitis have a mortality rate of 0.5–11.7% and 10–15%, respectively [4,5]. Cerebrospinal fluid (CSF) cultures remain a critical tool in pathogen identification, but take days to finalize and are affected by receipt of empiric antimicrobials prior to the lumbar puncture (LP) [2]. Viruses such as herpes simplex virus (HSV) 1 and 2, varicella/zoster virus (VZV), and enterovirus are common pathogens in encephalitis, and can present similarly to meningitis [3,4]. These viral pathogens are not easily identified through culture and often require specialty tests performed at reference laboratories, which can take several days to return.The BioFire® FilmArray® Meningitis/Encephalitis (ME) Panel is a multiplex polymerase chain reaction (PCR) test of the CSF fluid that can detect the 14 most common pathogens among all age groups in 1 h [6,7]. The sensitivity and specificity of each individual pathogen detected in this rapid diagnostic test (RDT) varies. Leber and colleagues demonstrated a sensitivity/positive percentage of agreement (PPA) of 100% for the following pathogens: Escherichia coli K1, Haemophilus influenzae, Streptococcus pneumoniae, human cytomegalovirus, HSV-1, HSV-2, Human Parechovirus, VZV, and Crytococcus gattii [8]. Two pathogens had lower sensitivities: 95.7% for Enterovirus and 85.7% for Human Herpesvirus 6 (HHV-6). The specificity or negative percentage of agreement (NPA) was 99.2% or greater for all organisms other than Streptococcus agalactiae [8]. In a meta-analysis of the sensitivity and specificity of the ME Panel in both adults and pediatrics, authors found similar results with a sensitivity and specificity of >90.8%. The main difference was among HSV, as there were concerns of false negative results in HSV-1/2 [9]. There is limited evidence evaluating PCR positivity after antibiotic initiation. Prior to the creation of the ME Panel, a study in 2006 by Bronska et al. found that when CSF PCR was performed prior to antibiotics versus in the presence of antibiotics, the yield only decreased from 100% to 81%, which was higher than rates reported for standard CSF culture that declined from 52% to 7% [10].Implementation of the ME Panel into clinical practice has yielded mixed results. Some institutions have found a decreased length of stay associated with the ME Panel [11,12]. Implementation of the ME Panel has also been shown to reduce acyclovir days of therapy in several clinical pre-post analyses [12,13,14,15]. Another pre-post study performed by Walker and colleagues showed significantly fewer patients receiving overall antimicrobials such as acyclovir, vancomycin, and ampicillin, but there were not significant decreases in duration except for in the ampicillin days of therapy [16].

The purpose of this study is to evaluate how implementation of the ME panel has affected antimicrobial duration in patients with suspected meningitis/encephalitis in the absence of formal guidance for interpretation.

4. DiscussionSeveral studies have established high sensitivity and specificity for bacterial pathogens for the BioFire® FilmArray® ME panel [8,9]. Despite this, implementation of the ME panel at our institution did not significantly impact antimicrobial duration of therapy, except for acyclovir. Total duration of therapy in our study was 3 and 2 days in the pre- and post-groups, respectively. This suggests after implementation of the ME panel healthcare providers continued to rely on CSF culture results before making antibiotic changes, decreasing the benefits of an RDT. These results are similar to what other studies have found [15,16]. Previous studies have demonstrated mixed results regarding impact of the ME panel on clinical outcomes [11,12,13,14,15,16]. In our study, clinical outcomes including length of stay, 30-day readmission, and all-cause mortality duFring index admission were not found to be statistically different between groups.At our institution, the ME panel was implemented with no formal guidance. While rapid diagnostics often yield rapid and actionable information, implementation of RDTs does not always lead to rapid changes in therapy. Antimicrobial stewardship programs (ASP) are optimally positioned to improve response and outcomes associated with RDTs. The best evidence for impact of ASP with RDTs is with multiplex PCR-based blood culture identification (BCID) panels [20,21]. Macvane and colleagues showed that compared with no intervention and ASP alone, ASP plus BCID decreased time to effective therapy in patients and more rapid de-escalation in patients with positive blood cultures [20]. Institutions looking to implement rapid diagnostic tests should carefully consider guidance and requirements around ordering and who will respond to results.

Diagnostic stewardship is becoming increasingly important, and our study demonstrates several opportunities to implement interventions to improve ordering of and reaction to the ME panel. First, there was a low rate of ME panel positivity and the majority of patients (70%) had a normal CSF WBC. This may indicate over-utilization of the test. There was also a large difference noted between the adult and pediatric positivity rate of the ME panel, which could imply over-utilization of the ME panel particularly in immunocompetent, adult patients. Institutions could consider only performing the ME panel in patients with pleocytosis, but certain exceptions should be allowed. Some populations in which the ME panel should be performed regardless of CSF WBC count are neonates, children, patients who are neutropenic or those who received a prolonged duration of antibiotics prior to the LP. For example, in our study 9 of the 18 patients with a positive ME panel had a CSF WBC < 10 cells/ mm3, all of which were children with viral pathogens identified.

Although there was no significant difference in days of therapy overall, there was a reduction of acyclovir days of therapy in this study. This mirrors findings from other studies, as previously discussed [12,13,14,15]. Interestingly, this reduction was seen despite the concern regarding sensitivity of the ME panel for HSV found in the study by Tansarli and colleagues. Given similar clinical outcomes between groups, this may be an opportunity for an additional antimicrobial stewardship intervention at centers implementing the ME panel.An example of “low-hanging fruit” for antimicrobial stewardship that could be targeted with the ME panel is the discontinuation of vancomycin. Vancomycin is specifically added for coverage for drug-resistant S. pneumoniae. In one of the largest studies evaluating accuracy of the ME panel, sensitivity/positive percentage of agreement (PPA) and specificity/negative percentage of agreement (NPA) of the ME panel for S. pneumoniae were 100% and 99.2%, respectively [8]. Sixteen patients had S. pneumoniae detected on the ME panel with no false negatives identified, but seven patients were determined to have false positives. Therefore, if the LP is obtained in a timely fashion and the ME panel is negative for S. pneumoniae, providers should feel confident in discontinuing vancomycin immediately rather than waiting for CSF culture results.

These results demonstrate the challenge of implementing RDT. They possess a great potential with the rapid, accurate results, but translating that to clinical change continues to be difficult. Our results highlight the importance of pairing institutional policy and ASP intervention with RDT. Based on these results, it may be beneficial to limit use of the ME panel through restriction of ordering to a specific team or by use criteria. In addition, oversight by or consultation with an infectious diseases or ASP prescriber or pharmacist should be considered with the use of this RDT. Finally, ASP interventions using clinical decision support systems, included alerts in the electronic medical record, stop dates for empiric antimicrobial therapy, and/or antimicrobial timeouts could be considered to enhance the ME panel as a RDT.

Our study is not without limitations. First, 25% of patients in this study received no antimicrobial therapy despite performance of CSF diagnostic tests. A majority of these patients were being evaluated for ocular or neurological diseases, such as multiple sclerosis. Institutions should evaluate strategies to limit ME panel utilization in patients without clear evidence of CNS infection. While power was met for this study, it did include a small sample of patients from a single center. A larger sample size over a wider geographical region would enhance the generalizability of this study. Finally, the interpretation of these results is limited by the overall low rate of positivity, particularly in the adult population.