The 9-month all-oral regimen is not adequate for the treatment of patients with pre-XDR-TB or XDR-TB; it is also not adequate to treat MDR/RR-TB that has both inhA and katG mutations. It is recommended that samples be submitted for susceptibility testing to at least fluoroquinolones before the start of this regimen. In settings without access to the Xpert MTB-XDR cartridge, a line probe assay (LPA: MTBDRplus) can be used to detect the two most common mutations that confer resistance to isoniazid. These mutations are found in the inhA promoter and katG regions, and they confer resistance to isoniazid at different levels. Low-level isoniazid resistance is conferred when only inhA mutations are present, and high-level resistance is conferred when mutations in the katG gene are present. Mutations at the inhA promoter region are also associated with resistance to ethionamide and prothionamide. High doses of isoniazid (15–20 mg/kg) are generally considered to be effective in the presence of low-level isoniazid resistance when used as part of combination therapy, but the efficacy of high doses of isoniazid in the presence of katG mutations remains unclear. Nevertheless, high-dose isoniazid is always included in the 9-month all-oral regimen if either (but not both) of the mutations is present. The presence of mutations in both regions (i.e. inhA promoter and katG genes) suggests that neither isoniazid at a high dose nor thioamides may be effective and therefore the 9-month all-oral regimen is not appropriate in these cases. In the South African setting, the detection of both mutations in MDR-TB strains was considered a surrogate marker for more extensive drug resistance at the time that the 9-month linezolid-containing regimen was introduced. Patients in South Africa who had MDR-TB with both mutations were not considered eligible for the 9-month regimen, and so were not included in the routine dataset presented to the GDG for review. Thus, the efficacy of the 9-month regimen in such cases is largely unknown. In the absence of information on isoniazid resistance or mutation patterns in the case of an individual patient, knowledge of the prevalence of both mutations among locally circulating RR-TB strains (e.g. from DRS in the relevant epidemiological setting) may also inform decisions as to which treatment regimen would be most appropriate. DST for pyrazinamide and ethambutol is not carried out routinely in most settings and these results do not affect eligibility for the 9-month all-oral regimen. There are no rapid methods to detect M. tuberculosis resistance to clofazimine, linezolid and bedaquiline; however, the critical concentrations for mycobacterial growth indicator tube (MGIT) have been established, enabling NTPs to perform phenotypic DST. If resistance to these drugs is detected in isolates obtained from patients with MDR/RR-TB, the 9-month all-oral regimen should not be offered, or the patient must switch to a longer individualized treatment regimen. In the absence of DST for bedaquiline, linezolid and clofazimine, treatment decisions will rely on the likelihood of effectiveness of these medicines, based on an individual patient’s clinical history and surveillance data from the country or region. This should be considered a last resort and an interim measure until the capacity for DST for these drugs becomes available.