Comparison of respiratory and Meningitis/Encephalitis viruses detected by FilmArray® multiplex PCR versus real-time PCR

Introduction: Fast and reliable pathogen detection is important for adequate management of infections. Although real-time PCR (rtPCR) is usually the most sensitive method for direct pathogen detection, it requires experienced technicians, includes several working steps and has a turnaround time of multiple hours. Therefore this method is not ideal for emergency diagnostics. The FDA cleared, fully automated sample to answer, FilmArray® (FA) multiplex PCR system (BioFire/bioMérieux) detects a broad spectrum of pathogens in ∼70 min. To optimize our diagnostic services during weekends and off-peak times, we compared the FA Respiratory Panel (RP) and FA Meningitis/Encephalitis (ME) Panel to our routinely used rtPCR assay. The FA panels detect 20 respiratory pathogens (17 viruses, 3 bacteria) in nasopharyngeal swabs (NPS) and 14 M/E pathogens (7 viruses, 6 bacteria, Cryptococcus neoformans/gattii) in cerebrospinal fluids (CSF). Materials and methods: With FA we tested 84 retrospective samples (23 NPS, 29 broncheoalveolar lavages [BALs], 32 CSF) and 60 prospectively collected NPS that required urgent testing during the 2015/2016 flu season by FA and rtPCR. FA sample input volume was 300 ml for RP and 200 ml for ME. Commercial RP and ME quality control panels (MMQC Inc., Scarborough, USA), containing samples positive and negative for each analyte detected by the FA panels, were tested multiple times. For rtPCR, nucleic acids were extracted from 220 ml of sample and eluted in 55 ml using NucliSENS easyMAG (bioMérieux). Respiratory viruses were analyzed by real-time PCR using a combination of 7 duplex Respiratory Multi Well System r-gene™ (RG) assays (influenza A/B, RSV/hMPV, HRV&EV/cell control, ADV/HBoV, HCoV/HPIV1-4) (Argene/bioMerieux), according to manufacturer's instructions. Additionally, we expanded FA RP testing to include (BALs), by implementing one additional sample preparation step. CSF was analyzed for virus using laboratory developed tests (LDTs) certified by the Swiss authorities. Results: RP and ME quality control panel results were 100% concordant with expected results. For all NPS, both tests, FA RP and RG, identified one or more viruses in 45/83 (54.2%) samples. FA RP and RG results correlated for 42/48 viruses detected (87.5%). FA RP detected an additional 3 HRV/EV and RG detected additionally 1 FluA, 1 ADV and 1 HRV/EV. Positive percent agreement (PPA) between RG (laboratory standard) and FA RP for NPS was 93.3% and negative percent agreement (NPA) was 92.7%. Overall correlation was 93.2%. Results from BALs yielded 92% PPA, 93.1% NPA and overall correlation of 92.4%. For FA ME testing, 31/33 CSF samples had identical FA ME and LDT results with an overall correlation of 94.4%. FA ME did not detect 2 parechovirus low level LDT positive samples (Ct 36.3 and 37.0). Using LDTs as the laboratory standard, FA ME PPA and NPA were 93.9% and 100%, respectively. Conclusion: Results obtained with the FilmArray® RP and ME panels were highly concordant with our currently used diagnostic methods, demonstrating excellent performance. The simplicity of the FilmArray® system, requiring less than 5 min of hands-on time, easy to read reports, and low sample volume allows for testing during off shifts and when urgent results are required. The comprehensiveness of the FilmArray® panels is ideal for diagnosing clinical syndromes where there are many potential causes