P008
Evaluating the performance of broth microdilution for the detection of rifampicin-resistant tuberculosis caused by borderline rpoB mutants
N S Hidayatallah(1,2) P Rupasinghe(1) J Vereecken(1) N Barreda(9) L Chingisova(7) S V Omar(8) A Ashraf(6) L Guglielmetti(3,4,5) L Rigouts(1,2) B C de Jong(1)
1:Institute of Tropical Medicine Antwerp; 2:University of Antwerp; 3:Medical Department, Médecins Sans Frontières, Paris; 4:Sorbonne University, Centre d'Immunologie et des Maladies Infectieuses (Cimi-Paris), Paris; 5:AP-HP, Bactériologie-Hygiène, Hôpital PiIé-Salpêtrière, Centre National de Référence des Mycobactéries, Paris; 6:The Indus Hospital laboratory, Karachi; 7:National Tuberculosis Reference Lab, Almaty; 8:National Institute of Communicable Diseases, Johannesburg; 9:Socios en Salud, Lima
The clinical significance of borderline rpoB mutants was reflected in World Health Organization (WHO) guidance in 2021, when seven such mutations were recognised as causing rifampicin-resistant tuberculosis (RR-TB). Borderline rpoBmutations carry fitness costs that impair in vitro growth and result in poor sensitivity of rapid phenotypic drug-susceptibility tests (pDST) such as the liquid culture MGIT 960 system.
The WHO recently recommended broth microdilution (BMD) for pDST in M. tuberculosis, providing minimum inhibitory concentration (MIC) values with relatively short turnaround times, and the possibility to extend incubation in case of fastidious growth We assessed the performance of BMD with borderline rpoB mutants, testing 52 strains representing 6 borderline rpoB mutations (9-Leu430Pro, 3-His445Leu, 10-His445Asn, 10-Asp435Tyr, 10-Ile491Phe); 26 strains representing 5 of the most common high confidence rpoB mutations (5-Asp435Val, 5-His445Asp, 4-His445Tyr, 8-Ser450Leu, 4-Ser450Trp); and 47 wildtype rpoB strains.
MIC distributions for rifampicin showed high MICs for high confidence rpoB mutations (2 to >16 µg/ml), while borderline rpoB mutants showed a wide range of MICs (0.06 to >16 µg/ml), with significant overlap with wildtype MICs (0.06 to 1 µg/ml). The MIC distributions for bedaquiline, clofazimine and linezolid were similar for borderline and common rpoB mutations. Extended incubation led to a non-specific increase in MIC values and therefore does not improve borderline detection in BMD.
Our data indicates that although BMD performs well with high confidence rpoB mutants, half of the borderline rpoBmutants tested gave false rifampicin-susceptible results, despite yielding valid results for other antibiotics.