P013

1:Department of Biomedical Sciences, Institute of Tropical Medicine, Antwerp; 2:University of Antwerp; 3:Medical Department, MSF, Paris, France; 4:Sorbonne University, Centre d'Immunologie et des Maladies Infectieuses (Cimi-Paris), Paris, France; 5:AP-HP, Bactériologie-Hygiène, Hôpital Pitié-Salpêtrière, Centre National de Référence des Mycobactéries, Paris, France; 6:Socios en Salud, Lima, Peru; 7:National Tuberculosis Reference Lab, Almaty, Kazakhstan; 8:Department of Microbiology, Sir JJ Hospital, Mumbai, India; 9:Regional Tuberculosis Reference Lab, Ho Chi Minh, Vietnam; 10:National Tuberculosis Reference Lab, Hanoi, Vietnam; 11:The Indus Hospital laboratory, Karachi, Pakistan; 12:Brigham and Women's Hospital, Boston, Massachusetts, USA; 13:Partners In Health, Boston, Massachusetts, USA; 14:Department of Global Health and Social Medicine, Harvard Medical School, Boston, Massachusetts, USA

Rifampicin-resistance in Mycobacterium tuberculosis is associated with mutations in the rpoB gene. ‘High-confident’ rpoB mutations, such as Ser450Leu, are likely to have concordant, resistant results on phenotypic drug susceptibility test (pDST) in liquid media MGIT. In contrast, ‘borderline’ rpoB mutations, recognized by the World Health Organization (WHO), result in in-vitro growth defects, leading to false-susceptible rifampicin pDST results in MGIT. The effect of these growth defects on liquid-based pDST for other anti-TB drugs is unknown.

We evaluated whether ‘borderline’ rpoB mutations affect minimum inhibitory concentrations (MICs) for bedaquiline, clofazimine, linezolid, and delamanid using 688 isolates (30-lineage-1, 301-lineage-2, 82-lineage-3, 275-lineage-4) from the endTB/Q trials, tested in the EUCAST-derived broth-microdilution method (BMD). MIC distributions of 39 ‘borderline’ rpoB mutants (13-His445Asn, 9-Asp435Tyr, 7-Leu452Pro, 5-His445Leu, 4-Leu430Pro, 1-His445Ser) were compared to 649 (508-Ser450Leu, 99-Asp435Val, 20-His445Tyr, 15-His445Asp, 7-Ser450Phe) ‘high-confident’ rpoB mutants, overall and by lineage. All isolates were from participants unexposed to the tested drugs and were classified as ‘likely-susceptible’ based on whole-genome sequencing.

Overall median turnaround time for MIC results was 14 days (IQR 10–14) in both groups. Overall, MIC distributions showed no differences between groups for any drug. However, in lineage-1, isolates with ’borderline‘ mutations had higher median linezolid MICs (1.0µg/ml) than those with “high-confidence” mutations (0.5µg/ml). Our findings suggest that ‘borderline’ rpoB mutations do not affect the validity of pDST results for the drugs tested using the BMD method. Higher linezolid MICs in lineage-1 ‘borderline’ rpoB mutants warrant further investigation, considering the small sample size of lineage-1 included in this study.