Comparison of a commercial real-time PCR panel to routine laboratory methods for the diagnosis of meningitis-encephalitis

Abstract

Meningitis-encephalitis can range from a mild, self limiting illness to a life threatening disease. Rapid microbial diagnosis allows for early targeted management. This study aimed to compare the BioFire FilmArray Meningitis/Encephalitis multiplex PCR panel (ME panel) to traditional testing algorithms for accuracy and turnaround time in the diagnosis of meningitis-encephalitis. From April to November 2018, cerebrospinal fluid (CSF) samples meeting existing laboratory testing criteria for suspected community acquired meningitis-encephalitis were tested on the ME panel and by routine laboratory methods. The methods were compared for accuracy of diagnosis and turnaround time. Where an organism was not identified, the study investigators came to a consensus on whether an infective aetiology was likely based on CSF parameters, clinical features, management and final discharge diagnosis. A total of 147 CSF samples met criteria for testing. Results were concordant in 143 (97%) of cases, including 27 samples where the same organism was identified by both methods. Of the four discordant samples, three organisms identified by the ME panel alone were considered clinically insignificant. One sample, which was culture and antigen positive for Cryptococcus neoformans, was not detected on the ME panel. The ME panel and routine methods identified an organism in 55% and 58% of clinically compatible cases of infection, respectively. The median turnaround time for the ME panel was 2.9 hours, compared to 21.1 hours for routine testing. The ME panel showed high concordance with traditional testing, simplified laboratory workflow, and significantly reduced turnaround time. The failure of the ME panel to detect Cryptococcus spp. is concerning. When cryptococcal meningitis is suspected, we would recommend using culture and cryptococcal antigen testing as the investigations of choice. Despite the availability of molecular assays targeting the common causes of CNS infection, the diagnostic yield remains suboptimal.