A variety of viral, bacterial, and fungal pathogens cause infections of the central nervous system (CNS), which range in severity from mild and self-limiting to severe and life-threatening. Initial symptoms like headache, fever, photophobia, and neck stiffness aren’t organism-specific so can’t be used to guide therapy. Yet a 24-hour delay in specific therapy confers a 3- to 11-fold increased risk for death or long-term neurologic deficit. These circumstances underscore the need for rapid and accurate identification of the infecting organism.

Current microbiologic methods identify a specific organism in only 30%-50% of patients with presumed CNS infections, largely due to poor culture sensitivity stemming from the low concentration of organism in cerebrospinal fluid (CSF) and low volume of CSF collected for microbiologic analysis. Nucleic acid amplification tests (NAATs) demonstrate significantly greater sensitivity than bacterial or viral culture; however, few such tests are commercially available. These tests also detect only one or two pathogens associated with CNS infections, thereby limiting their applicability.

Multiplex Testing Review

In November 2015, the FilmArray ME panel (FA-ME) received Food and Drug Administration clearance for use in identifying six bacteria, seven viruses, and one yeast commonly associated with CNS infections. Initial evaluations of the FA-ME demonstrated overall sensitivity and specificity of 90%-100%; however, closer inspection of these studies reveal limitations that are important to consider when interpreting the data. The lone prospective clinical evaluation contained just 8/1,560 (0.5%) CSF specimens that were culture-positive for a bacterial pathogen, leaving two completely unevaluated (J Clin Microbiol 2016;54:2251-61). Among viral targets, nearly 50% of positive specimens contained enterovirus while other important pathogens such as herpes simplex virus type 1 (HSV-1), cytomegalovirus (CMV), and varicella-zoster virus were present in four or fewer specimens. This low prevalence of positive specimens resulted in wide 95% confidence intervals of 34%-100% for individual target sensitivities.

Additional evaluations have addressed this shortcoming by including large numbers of archived CSF specimens that have tested positive for each FA-ME target. These studies confirmed the overall high sensitivity of FA-ME for most targets though sensitivity for HSV-1 was only 50%-73% when compared to HSV-specific real time polymerase chain reaction tests (J Clin Microbiol 2018;56:e01927-17; Diagn Microbiol Infect Dis 2017;87:92-4; J Clin Microbiol 2016;54:785-7). In addition, inclusion of only preselected positive specimens in these studies prevents accurate calculations of sensitivity and specificity.

A multiplex testing approach simplifies ordering for clinicians and provides a relatively comprehensive result in as little as 60 minutes. Using multiplexed tests to analyze CSF resulted in a 44%-600% increase in specimens with an identified organism, mostly due to NAATs’ increased sensitivity compared to culture (J Clin Microbiol 2016;54:2251-61). These tests also detected viral targets that hadn’t been tested for initially, either because they weren’t in the differential diagnosis or because specific laboratory tests weren’t available.

While potentially beneficial, the significance of detecting these additional organisms needs to be considered in the context of other laboratory values and a patient’s clinical status. The FA-ME identified S. pneumoniae in an additional 12 CSF specimens when compared to culture; however, seven of these patients had no clinical or laboratory evidence of S. pneumoniae infection, suggesting a false positive result potentially due to external contamination of the specimens.

Consequently, despite having a specificity of >99% for S. pneumoniae the test’s positive predictive value was just 60%. Similarly, a definitive diagnosis of CNS infection was made in only 11% of human herpesvirus-6 and 33% of CMV-positive specimens, possibly due to latent virus present within leukocytes in CSF rather than an indication of active disease.

Combined, these data underscore the need to correlate FA-ME results with other laboratory values and host factors to validate a result, especially in cases with results unexpected or inconsistent with a patient’s risk factors and clinical course.

Establishing Usage Criteria

The FA-ME test identifies 14 microorganisms frequently associated with community acquired CNS infections. However, other patient populations including those with traumatic injury or surgery involving the CNS are susceptible to pathogens that are not part of the FA-ME panel. In these patients, FA-ME lacks broad utility, and a negative FA-ME result could be misleading. For these reasons, labs need to develop criteria to optimize the benefit from FA-ME testing including selecting appropriate patient populations and rejecting specimens.

If used judiciously as a component of a full diagnostic strategy, multiplexed meningitis testing provides rapid, actionable results that can shorten time to definitive diagnosis, reduce unnecessary antibiotic utilization, and ultimately improve the management of patients with signs and symptoms of CNS infection. To achieve these goals, users must be aware of the strengths and weaknesses of this test and take active steps to accentuate the benefits and mitigate potential pitfalls.

Blake W. Buchan, PhD, D(ABMM), is an assistant professor of pathology at the Medical College of Wisconsin and associate director of clinical microbiology at Wisconsin Diagnostic Laboratories in Milwaukee.+Email: bbuchan@mcw.edu