Search

Bone and joint infections (BJIs) are complex infections that require precise microbiological documentation to optimize antibiotic therapy. Currently, diagnosis is based on microbiological culture, sometimes complemented by amplification and sequencing of the 16S rDNA gene. Clinical metagenomics (CMg), that is, the sequencing of the entire nucleic acids in a sample, was previously shown to identify bacteria not detected by conventional methods, but its actual contribution to the diagnosis remains to be assessed, especially with regard to 16S rDNA sequencing. In the present study, we tested the performance of CMg in 34 patients (94 samples) with suspected BJIs, as compared to culture and 16S rDNA sequencing. A total of 94 samples from 34 patients with suspicion of BJIs, recruited from two sites, were analyzed by (i) conventional culture, (ii) 16S rDNA sequencing (Sanger method), and (iii) CMg (Illumina Technology). Two negative controls were also sequenced by CMg for contamination assessment. Based on the sequencing results of negative controls, 414 out of 539 (76.7%) bacterial species detected by CMg were considered as contaminants and 125 (23.2%) as truly present. For monomicrobial infections (13 patients), the sensitivity of CMg was 83.3% as compared to culture, and 100% as compared to 16S rDNA. For polymicrobial infections (13 patients), the sensitivity of CMg was 50% compared to culture, and 100% compared to 16S rDNA. For samples negative in culture (8 patients, 21 samples), CMg detected 11 bacteria in 10 samples from 5 different patients. In 5/34 patients, CMg brought a microbiological diagnosis where conventional methods failed, and in 16/34 patients, CMg provided additional information. Finally, 99 antibiotic resistance genes were detected in 24 patients (56 samples). Provided sufficient genome coverage (87.5%), a correct inference of antibiotic susceptibility was achieved in 8/8 bacteria (100%). In conclusion, our study demonstrated that the CMg provides complementary and potentially valuable data to conventional methods of BJIs diagnosis., Competing Interests: GGu and GGe were employed by bioMérieux SA. SF was employed by Genoscreen. ER received consulting fees from Illumina and Pathoquest. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 d’Humières, Gaïa, Gueye, de Lastours, Leflon-Guibout, Maataoui, Duprilot, Lecronier, Rousseau, Gamany, Lescure, Senard, Deconinck, Dollat, Isernia, Le Hur, Petitjean, Nazimoudine, Le Gac, Chalal, Ferreira, Lazarevic, Guigon, Gervasi, Armand-Lefèvre, Schrenzel and Ruppé.)

The French National Reference Center for Staphylococci currently uses DNA arrays and spa typing for the initial epidemiological characterization of Staphylococcus aureus strains. We here describe the use of whole-genome sequencing (WGS) to investigate retrospectively four distinct and virulent S. aureus lineages [clonal complexes (CCs): CC1, CC5, CC8, CC30] involved in hospital and community outbreaks or sporadic infections in France. We used a WGS bioinformatics pipeline based on de novo assembly (reference-free approach), single nucleotide polymorphism analysis, and on the inclusion of epidemiological markers. We examined the phylogeographic diversity of the French dominant hospital-acquired CC8-MRSA (methicillin-resistant S. aureus ) Lyon clone through WGS analysis which did not demonstrate evidence of large-scale geographic clustering. We analyzed sporadic cases along with two outbreaks of a CC1-MSSA (methicillin-susceptible S. aureus ) clone containing the Panton-Valentine leukocidin (PVL) and results showed that two sporadic cases were closely related. We investigated an outbreak of PVL-positive CC30-MSSA in a school environment and were able to reconstruct the transmission history between eight families. We explored different outbreaks among newborns due to the CC5-MRSA Geraldine clone and we found evidence of an unsuspected link between two otherwise distinct outbreaks. Here, WGS provides the resolving power to disprove transmission events indicated by conventional methods (same sequence type, spa type, toxin profile, and antibiotic resistance profile) and, most importantly, WGS can reveal unsuspected transmission events. Therefore, WGS allows to better describe and understand outbreaks and (inter-)national dissemination of S. aureus lineages. Our findings underscore the importance of adding WGS for (inter-)national surveillance of infections caused by virulent clones of S. aureus but also substantiate the fact that technological optimization at the bioinformatics level is still urgently needed for routine use. However, the greatest limitation of WGS analysis is the completeness and the correctness of the reference database being used and the conversion of floods of data into actionable results. The WGS bioinformatics pipeline (EpiSeq TM ) we used here can easily generate a uniform database and associated metadata for epidemiological applications.