Bedaquiline susceptibility testing of Mycobacterium tuberculosis in an automated liquid culture system

Objectives: The objective of this study was to evaluate the performance of the BACTEC MGIT960 system to test the susceptibility to bedaquiline for Mycobacterium tuberculosis complex. Methods: We determined the quality control (QC) range of bedaquiline using the M. tuberculosis H37Rv reference strain and the epidemiological cut-off (ECOFF) in MGIT960 and on Middlebrook 7H11 agar (M7H11) using 47 strains from bedaquiline treatment-naive patients. The accuracy of MGIT960 was evaluated versus M7H11 using 74 'probably susceptible to bedaquiline' and 18 'probably resistant to bedaquiline' strains. Repeatability and reproducibility of MGIT960 were assessed using five strains showing different resistance levels. Results: The QC range for the H37Rv strain was between 0.125 and 0.50 mg/L. The WT MIC distribution ranged from <= 0.03 to 1.00 mg/L in MGIT960 and from <= 0.008 to 0.25 mg/L on M7H11 with suggested ECOFFs of 1.00 and 0.25 mg/L, respectively. Applying these ECOFFs, the probably susceptible and probably resistant strains were distinguishable by both methods, albeit with only a 2-fold increased MIC for one of the resistant strains compared with the ECOFF. Intermethod agreement to classify the isolates was excellent (100%). All replicates in the repeatability and reproducibility experiments fell within the normal range. Conclusions: The MGIT960 system proved to be highly stable, reproducible and accurate relative to the M7H11 agar method for determining the bedaquiline MIC. The small margin between the suggested ECOFF and the lowest MIC for the mutant strains risks making both methods prone to discordant results. Further validation in clinical settings linked to treatment outcome data is needed.