D. BANDEIRA¹, L. BERTAIOLA-MONNERAT², M.-L. TRITTEN¹, H. H. SIEGRIST¹, R. LIENHARD<sup>1

1</sup>ADMed Microbiologie, 2300 la Chaux-de-Fonds, Suisse, ²HFR Hôpital cantonal, 1708 Fribourg, Suisse

Two genomic tests for the rapid identification of the agent of whooping cough were compared. Eighty-nine nasopharyngeal aspirates or swabs from patients and 14 external quality controls from QCMD were analysed. Native samples were tested on the Biofire FilmArray Respiratory Panel (FA) (bioMérieux, Marcy-l'Etoile, F) to detect Bordetella pertussis and 21 other pathogens. The same samples were then extracted by Nuclisens®easyMag and tested with the RIDA®GENE Bordetella real-time multiplex PCR (RG) (r-biopharm, Darmstadt, D) using the RotorGene Q to detect B. pertussis, B. parapertussis and B. holmesii. The comparison of results from patient swabs showed several discrepancies. Ten B. pertussis were found by FA versus 17 with RG. None B. parapertussis or B. holmesii were detected by RG. In the 10 samples positive for Bordetella with FA co-infections were detected in 3 samples with rhino/enterovirus (RV/EV) and in 2 with RV/EV and adenovirus (ADV), 5 were positive for Bordetella only. From 79 negative samples only 22 (27.8%) were totally negative. All the others were positive for at least one pathogen; 36 (45.6%) for RV/EV, 9 (11.4%) for respiratory syncytial virus, 7 (8.9%) for parainfluenza virus, 6 (7.6%) for human metapneumovirus, 5 (6.3%) for influenza virus, 5 (6.3%) for C. pneumoniae, 4 (5.1%) for M. pneumoniae, 3 (3.8%) for adenovirus, 2 (2.5%) for coronavirus. Testing EQC samples (QCMD) out of 6 positive B. pertussis 5 were found by RG versus only 3 by FA. Furthermore, RG correctly identified 1 B. parapertussis, 2 B. holmesii and the 4 negatives (2 B. bronchiseptica and 2 Bordetella spp. negatives). FA gave two false positive results for B. bronchiseptica revealing a specificity problem with the primers. Due to its lack of sensitivity for samples with low pathogen concentration, FA appears less effective than RG to diagnose the aetiology of pertussis cough. However, despite its lack of sensitivity and specificity; the broad spectrum approach of FA gives useful information about the aetiology of respiratory problems and remains a rapid and easy to use test to diagnose respiratory infections without specific clinical symptoms of pertussis.