Résumé
The FilmArray Blood Culture Identification 2 panel (BCID2; bioMérieux) is a fully automated PCR-based assay for identifying bacteria, fungi, and bacterial resistance markers in positive blood cultures (BC) in about 1 h. In this multicenter study, we evaluated the performance of the BCID2 panel for pathogen detection in positive BC. Conventional culture and BCID2 were performed in parallel at four tertiary-care hospitals. We included 152 positive BC-130 monomicrobial and 22 polymicrobial cultures-in this analysis. The BCID2 assay correctly identified 90% (88/98) of Gram-negative and 89% (70/79) of Gram-positive bacteria. Five bacterial isolates targeted by the BCID2 panel and recovered from five positive BC, including three polymicrobial cultures, were missed by the BCID2 assay. Fifteen isolates were off-panel organisms, accounting for 8% (15/182) of the isolates obtained from BC. The mean positive percent agreement between the BCID2 assay and standard culture was 97% (95% confidence interval, 95 to 99%), with agreement ranging from 67% for Candida albicans to 100% for 17 targets included in the BCID2 panel. BCID2 also identified the bla CTX-M gene in seven BC, including one for which no extended-spectrum b-lactamase (ESBL)-producing isolate was obtained in culture. However, it failed to detect ESBL-encoding genes in three BC. Two of the 18 mecA/ C genes detected by the BCID2 were not confirmed. No carbapenemase, mecA/C, or MREJ targets were detected. The median turnaround time was significantly shorter for BCID2 than for culture. The BCID2 panel may facilitate faster pathogen identification in bloodstream infections. IMPORTANCE Rapid molecular diagnosis combining the identification of pathogens and the detection of antibiotic resistance genes from positive blood cultures (BC) can improve the outcome for patients with bloodstream infections. The FilmArray BCID2 panel, an updated version of the original BCID, can detect 11 Gram-positive bacteria, 15 Gram-negative bacteria, 7 fungal pathogens, and 10 antimicrobial resistance genes directly from a positive BC. Here, we evaluated the real-life microbiological performance of the BCID2 Editor