Retrospective evaluation of routine whole genome sequencing of Mycobacterium tuberculosis at the Belgian National Reference Center, 2019. Issue 5 (3rd September 2022)

• Record Type: Journal Article
• Authors: Soetaert, Karine; Ceyssens, Pieter-Jan; Boarbi, Samira; Bogaerts, Bert; Delcourt, Thomas; Vanneste, Kevin; De Keersmaecker, Sigrid C.J.; Roosens, Nancy H.C.; Vodolazkaia, Alexandra; Mukovnikova, Marina; Mathys, Vanessa
• Is Part Of: Acta clinica belgica. Volume 77:Issue 5(2022)
• Page Start: 853
• Page End: 860
• Publication Date: 2022-09-03
• Subjects: Whole-genome sequencing -- tuberculosis -- resistance -- diagnostic -- bioinformatic; Clinical medicine -- Periodicals; Diagnosis, Laboratory -- Periodicals; 616.005
• DOI: 10.1080/17843286.2021.1999588
• Languages: English
• ISSNs: 1784-3286

Abstract

ABSTRACT: Objectives: Since January 2019, the Belgian National Reference Center for Mycobacteria (NRC) has switched from conventional typing to prospective whole-genome sequencing (WGS) of all submitted Mycobacterium tuberculosis complex (MTB) isolates. The ISO17025 validated procedure starts with semi-automated extraction and purification of gDNA directly from the submitted MGIT tubes, without preceding subculturing. All samples are then sequenced on an Illumina MiSeq sequencer and analyzed using an in-house developed and validated bioinformatics workflow to determine the species and antimicrobial resistance. In this study, we retrospectively compare results obtained via WGS to conventional phenotypic and genotypic testing, for all Belgian MTB strains analyzed in 2019 (n = 306). Results: In all cases, the WGS-based procedure was able to identify correctly the MTB species. Compared to MGIT drug susceptibility testing (DST), the sensitivity and specificity of genetic prediction of resistance to first-line antibiotics were respectively 100 and 99% (rifampicin, RIF), 90.5 and 100% (isoniazid, INH), 100 and 98% (ethambutol, EMB) and 61.1 and 100% (pyrazinamide, PZA). The negative predictive value was above 95% for these four first-line drugs. A positive predictive value of 100% was calculated for INH and PZA, 80% for RIF and 45% for EMB. Conclusions: Our study confirms the effectiveness of WGS for the rapid detection of M. tuberculosis complex and its drug resistance profiles for first-line drugs even when working directly on MGIT tubes, and supports the introduction of this test into the routine workflow of laboratories performing tuberculosis diagnosis.