Background: The BIOFIRE® FILMARRAY® Blood Culture Identification 2 Panel is a rapid diagnostic test that provides results for 7 fungal and 26 bacterial bloodstream pathogens as well as 10 bacteria-associated antimicrobial resistance (AMR) genes from positive blood culture (PBC) samples. A multi-center prospective clinical evaluation using an investigational-use-only (IUO) version of the Panel yielded an overall sensitivity of 99.2% and specificity of 99.6% for pathogen identification. All BCID2 Panel organism assay results that were discordant with the reference method were investigated.
Methods: At 7 US and 2 EU sites, 1074 residual PBC were enrolled prospectively. The reference method for evaluation of the BCID2 Panel organism assay performance was microbial culture performed at clinical sites as part of routine patient care. All false positive (FP) and false negative (FN) results were investigated using alternate PCR assays and confirmed by sequencing.
Results: Overall, 74 discordant results involving 16 bacterial and 3 fungal assays were investigated. In all 54 FP cases,18 from polymicrobial samples, the presence of the detected analyte in the PBC was confirmed by sequencing. Of the 20 FN results, 7 were attributed to isolate misidentification by the clinical laboratory; these were resolved as true negative results by sequencing. Low analyte level (below the assay’s limit-of-detection) was determined to be the root cause of 9 FN results; 8 of which were encountered in PBC samples identified as polymicrobial by the reference method. Currently, 4 FN cases remain unresolved. Notably, 38/74 (51.3%) discrepancies needing investigation involved Staphylococcus epidermidis. In 26/29 FP cases, S. epidermidis was not recovered by culture but was sequenced from the PBC. Nine discrepancies (6 FN, 3 FP) were caused by incorrect identification of the Staphylococcus species recovered by the clinical laboratory. Low analyte titers in polymicrobial PBC was responsible for the remaining 3 FN results.
Conclusion: With 61/74 discordant results resolved favorably, the BCID2 Panel would have yielded an overall success rate of 99.8%. Therefore, the BCID2 Panel is expected to provide highly accurate identification of causative agents of bloodstream infections.
Data presented are from assays that have not been cleared or approved for diagnostic use.